Urban Habitat Constructions Under Catastrophic Events: COST C26 Action Final Report

COST ACTION C26 URBAN HABITAT CONSTRUCTIONS UNDER CATASTROPHIC EVENTS FINAL REPORT

COST ACTION C26 Urban Habitat Constructions under Catastrophic Events Final Report Chair Federico M. Mazzolani Department of Structural Engineering, University of Naples “Federico II”, Naples, Italy

Chairman Federico M. Mazzolani Editorial Board Mike Byfield Gianfranco De Matteis Dan Dubina Beatrice Faggiano Maurizio Indirli Alberto Mandara Federico M. Mazzolani Jean-Pierre Muzeau Frantisek Wald

CRC Press/Balkema is an imprint of the Taylor & Francis Group, an informa business

Typeset by MPS Ltd. (A Macmillan Company), Chennai, India Printed and bound in Great Britain by Antony Rowe (A CPI Group Company), Chippenham, Wiltshire All rights reserved. No part of this publication or the information contained herein may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, by photocopying, recording or otherwise, without written prior permission from the publisher. Although all care is taken to ensure integrity and the quality of this publication and the information herein, no responsibility is assumed by the publishers nor the author for any damage to the property or persons as a result of operation or use of this publication and/or the information contained herein. Published by:

CRC Press/Balkema P.O. Box 447, 2300 AK Leiden, The Netherlands e-mail: [email protected] www.crcpress.com – www.taylorandfrancis.co.uk – www.balkema.nl

ISBN: 978-0-415-60686-8 (Hbk + CD-Rom) ISBN: 978-0-203-83362-9 (eBook)

Urban Habitat Constructions under Catastrophic Events (Final Report) – Mazzolani (Ed). © 2010 Taylor & Francis Group, London, ISBN 978-0-415-60686-8

Table of Contents

Introduction

1

Chapter 1: Characterization of catastrophic actions on constructions 1.1 Design fires for structural engineering M. Gillie, F. Wald & K. Horová

7

1.2 Characterization and modelling of seismic action D. Lungu, A. Stratan & R. Vacareanu

19

1.3 Characterization of catastrophic actions on constructions: Explosive loads A. Tyas

43

1.4 Actions due to natural catastrophes, except volcanic eruptions T. Stathopoulos, I. Zisis, A. Talon, J.-P. Muzeau, C. Coelho, J.-P. Carlier & S. Wolinski

53

1.5 Actions due to volcanic eruptions J.-P. Muzeau, A. Talon, J.-C. Thouret, T. Rossetto, B. Faggiano, D. De Gregorio, G. Zuccaro & M. Indirli

73

1.6 Tsunami hazard and risk evaluation in the Gulf of Naples: State of the art and perspectives S. Tinti, F. Zaniboni, A. Armigliato & G. Pagnoni

99

Chapter 2: Analysis of behaviour of constructions under catastrophic events 2.1 Analyses of structures under fire D. Bacinskas, E. Geda, V. Gribniak, G. Kaklauskas, G. Cefarelli, B. Faggiano, A. Ferraro, F.M. Mazzolani, E. Nigro, C. Couto, N. Lopes, P. Vila Real, M. Hajpál, Á. Török, M. Kaliske, L. Kwasniewski, D. Pintea & R. Zaharia

111

2.2 Evaluation of structural response under exceptional seismic actions M. Fischinger & G. Della Corte

139

2.3 Analysis of behaviour of constructions under impact and explosions: Approaches for structural analysis, from material modeling to structural response G. De Matteis, E. Cadoni & D. Asprone

161

2.4 Consequences of natural disasters on constructions C. Coelho, R.P. Borg, V. Sesov & M. Indirli

179

2.5 Consequences of volcanic eruptions on constructions R.P. Borg & M. Indirli

201

Chapter 3: Evaluation of vulnerability of constructions 3.1 Vulnerability of existing buildings under fire E. Nigro, G. Cefarelli, F. Wald, M. Hajpál, R. Zaharia, N. Lopes, P. Vila Real, L. Kwasniewski, Z. Drabowicz, D. Pantousa, E. Geda, D. Bacinskas, V. Gribniak & M. Heinisuo V

219

3.2 Performance based evaluation and risk analysis E. Mistakidis, R. Vacareanu & A.J. Kappos

227

3.3 Vulnerability and damageability of constructions under impact and explosion F. Dinu

247

3.4 Performance assessment under multiple hazards D. Vamvatsikos, E. Nigro, L.A. Kouris, G. Panagopoulos, A.J. Kappos, T. Rossetto, T.O. Lloyd & T. Stathopoulos

271

Chapter 4: Protecting, strengthening and repairing 4.1 Fire damaged structures Y.C. Wang, F. Wald, J. Vácha & M. Hajpál

293

4.2 Innovative seismic protection technologies and case studies M. Kaliske & A. Mandara

303

4.3 Blast loading assessment and mitigation in the context of the protection of constructions in an urban environment P.D. Smith

327

4.4 Mitigation options for natural hazards, with a special focus on volcanic eruptions M. Indirli, E. Nigro, L. Kouris, F. Romanelli & G. Zuccaro

343

4.5 Avalanche risk assessment in populated areas A. Talon & J.-P. Muzeau

365

Chapter 5: Strategy and guidelines for damage prevention 5.1 Fire design in Europe M. Heinisuo, M. Laasonen & J. Outinen

375

5.2 Demands and recommendations for assessment and mitigation of risk under exceptional earthquakes A. Plumier, R. Landolfo & D. Dubina

403

5.3 Impact and explosion M.P. Byfield & P.P. Smith

423

5.4 Multi-hazard risk assessment methodology M.H. Faber & H. Narasimhan

435

5.5 A framework and guidelines for volcanic risk assessment H. Narasimhan, R.P. Borg, G. Zuccaro, M.H. Faber, D. De Gregorio, B. Faggiano, A. Formisano, F.M. Mazzolani & M. Indirli

443

List of COST papers

459

Author index

471

VI

Urban Habitat Constructions under Catastrophic Events (Final Report) – Mazzolani (Ed). © 2010 Taylor & Francis Group, London, ISBN 978-0-415-60686-8

Introduction

According to the Memorandum of Understanding, the main objective of the COST Action C26 on “Urban Habitat Constructions under Catastrophic Events” was to increase the knowledge on the behaviour of constructions located in urban habitats and subjected to exceptional loading conditions producing catastrophic events. With this perspective, the main aim of the Action was to define suitable tools to predict the ultimate response of such constructions under extreme conditions, occurring when both loading and structural resistance are combined in such a way to reduce the safety level below acceptable values. Such extreme situations can be produced by both natural (i.e. earthquakes, fire, wind storms, heavy snow, avalanches, floods, volcanic eruptions, …) and man-made (i.e. gas explosions, accidental impact from projectiles or vehicles out of control and occasionally due to bomb blasts during terrorist attacks) events. A big effort has been done to characterise the performance of structures under such exceptional loading conditions, as well as the consequences of a catastrophic event occurring in a given region, with regard to life safety, economic losses due to direct damage and indirect social costs, related to loss of use of a facility or a class of facilities. In addition, ad hoc guidelines for the damage prevention as well as for the repairing of constructions hit by the above actions have been properly set out. The participation of 23 European eligible Countries (Austria, Belgium, Cyprus, Czech Republic, Finland, Fy Republic of Macedonia, France, Germany, Greece, Hungary, Italy, Lithuania, Malta, Netherlands, Poland, Portugal, Romania, Slovenia, Spain, Sweden, Switzerland, Turkey, U.K.) confirmed the international interest of this Action. The activity of COST Action C26 was developed by means of four Working Groups (WGs), dealing with the main issues related to catastrophic events: WG 1 “Fire resistance” (chairman Frantisek Wald, Czech Republic; vice-chairman Yong Wang, U.K.) WG 2 “Earthquake resistance” (chairman Dan Dubina, Romania; vice-chairman Alberto Mandara, Italy) WG 3 “Impact and explosion resistance” (chairman Mike Byfield, U.K.; vice-chairman Gianfranco De Matteis, Italy) WG 4 “Risk assessment for catastrophic scenarios in urban areas” (chairman Michele Faber, Switzerland; co-chairman Maurizio Indirli, Italy), In addition, an “ad hoc” Working Group was created WG “Lexicon” (chairman Jean-Pierre Muzeau, France) Since the beginning, a very fruitful cooperation among the Management Committee (MC) members and the WG experts has been created, also thank to the effectiveness of the WG chairman-ships. One positive issue consisted in the already consolidated experience to work together of a large number of participants in many previous international research projects (i.e. TEMPUS, COST C1, COST C12, RECOS, PROHITECH). The planning of the whole activity during the period (2006–2010) has been done according to the Technical Annex of the MoU. From the evaluation of the obtained results, the main aspects, which deserve special consideration, can be identified as follow. • The main challenge of the COST Action C26 consists in the willing to create a common methodology for the assessment of the effects of extreme loading conditions produced by catastrophic events on constructions belonging to urban habitats. • At the same time, from the results of the structural analysis of constructions under different extreme loading conditions, it was possible to set-up specific recommendations for the mitigation of the damage state, by means of appropriate repairing and/or protecting systems. 1

• The methodological approach is based on the consolidated experience of out-standing experts in the field of Structural Engineering. • It has to be recognized that this kind of activity was never developed in the past, following such homogeneous and unitary way, so now this approach must be considered as innovative. During the whole period (2006–2010), the Action developed a lot of relevant activities, among them: organization of the Workshop in Prague the Symposium in Malta (23,24,25 October 2008), the Seminar in Southampton (27 March 2009), the C25&C26 Training School for Early Stage researchers inThessaloniki (18 to 23 May 2009), the Final Conference in Naples (16 to 18 September 2010), where this Final Report Volume is distributed to all participants. Important in situ scientific missions have been organized: many surveys in the Vesuvius area, in the Abruzzo earthquake damaged zones, in the damaged zones of the Chile earthquake. Many short term scientific missions (STSMs) have been finalized to develop this survey activity, giving the possibility to young early stage researchers to live unrepeatable experiences. The contents of this Final Report represent the out-put of the COST Action C26. The main results of the research work have been summarized in five Chapters, where each WG contributed for its specific subject. Chapter 1: Characterization of catastrophic actions on constructions – – – – – –

Design fires for structural engineering Characterization and modelling of seismic action Characterization of catastrophic actions on constructions: Explosive loads Actions due to natural catastrophes, except volcanic eruption Actions due to volcanic eruptions Tsunami hazard and risk evaluation in the Gulf of Naples: State of the art and perspectives

Chapter 2: Analysis of behaviour of constructions under catastrophic events – Analyses of structures under fire – Evaluation of structural response under exceptional seismic actions – Analysis of behaviour of constructions under impact and explosions: Approaches for structural analysis, from material modeling to structural response – Consequences of natural disasters on constructions – Consequences of volcanic eruptions on constructions Chapter 3: Evaluation of vulnerability of constructions – – – –

Vulnerability of existing buildings under fire Performance based evaluation and risk analysis Vulnerability and damageability of constructions under impact and explosion Performance assessment under multiple hazards

Chapter 4: Protecting, strengthening and repairing – Fire damaged structures – Innovative seismic protection technologies and case studies – Blast loading assessment and mitigation in the context of the protection of constructions in an urban environment – Mitigation options for natural hazards, with a special focus on volcanic eruptions – Avalanches risk assessment in populated areas Chapter 5: Strategy and guidelines for damage prevention – Fire design in Europe – Demands and recommendations for assessment and mitigation of risk under exceptional earthquakes – Impact and Explosion – Multi-hazard risk assessment methodology – A framework and guidelines for volcanic risk assessment 2

At the end of the Volume, the list of papers prepared by the COST C26 experts and presented during the three main events, which have been organized for the dissemination of project results: – the Workshop in Prague (30–31 March 2007) – the Symposium in Malta (23,24,25 October 2008) – the Final Conference in Naples (16,17,18 September 2010) The text of this Volume is also included in the attached DVD. A second DVD, also attached, contains the out-put of the WG “Lexicon”, where about 400 technical words related to the Action subject are translated into 18 languages (English, Czech, Dutch, Finnish, French, German, Greek, Hungarian, Italian, Lithuanian, Macedonian, Maltese, Polish, Portuguese, Romanian, Slovenian, Swedish, Turkish). In order to show the successfulness of this Action, it is ease to compare what we have initially promised with what we have actually produced. The objectives identified in the Technical Annex of the Memorandum of Understanding are listed below. All the objectives are related to the Chapters of this Final Report. The fulfillment of the objectives is clearly demonstrated by this strict relation. It is also important to notice that these Chapters are coincident with the main Topics of the Final Conference in Naples. Objective N.1 “To increase the knowledge on the behaviour of constructions located in urban habitats and subjected to exceptional and catastrophic events”. The description of the main recurrent cases of extreme loading conditions due to catastrophic events has been given (fire, fire after earthquake, exceptional earthquakes, explosions, impacts, landsides, avalanches, tsunami, volcanic eruptions,….), as well as the definition of the corresponding actions to be used in structural analysis. The fulfillment of this objective is given in: Chapter 1: “Characterization of catastrophic actions on constructions” Object N.2 “To define suitable tools for predicting the ultimate response of such constructions under extreme conditions, occurring when both loading and structural resistance are combined in such a way to reduce the safety level below acceptable values”. The structural models for analysing the behaviour of constructions under extreme conditions have been set-up and the behaviour of the constructions has been investigated under these conditions. Objective N.3 “To characterise the performance of structures under extreme loading conditions”. The structural behaviour of urban habitat constructions has been examined in order to evaluate their vulnerability and damageability with respect to the considered extreme loading conditions. Suitable systems for protecting them from the effects of catastrophic events have been also proposed. The fulfillment of both objectives N.2 and 3 is given in: Chapter 2: “Analysis of behaviour of constructions under catastrophic events” Objective N.4 “To analyse the consequences of catastrophic events occurring in a given region, with regard to life safety and economic losses due to direct damage, including the indirect social costs related to loss of use of facilities. This analysis has been performed by evaluating the vulnerability of constructions under various catastrophic conditions (fire, earthquake, impacts, explosions). The risk assessment methodology has been applied also in complex scenarios, like the one of the Vesuvius eruption, which has been assumed as a reference case study for exemplifying the adopted approach. The fulfillment of the objective N.4 is given in: Chapter 3: “Evaluation of vulnerability of constructions” Objective N.5 “To prepare ad-hoc guidelines for the damage prevention as well as for the repairing of constructions hit by extreme actions during catastrophic events.” The collection of data for the most typical situations has given the possibility to set-up general recommendations, which could be useful for the damage prevention against catastrophic events. 3

Consolidation systems for both repairing and upgrading the damaged constructions have been also provided. Objective N.4 is covered in two Chapters: Chapter 4: “Protecting, strengthening and repairing” and Chapter 5: “Strategy and guidelines for damage prevention”. As it can be emphasized by the contents of this volume, during the four years of the activity of COST Action C26, very important results have been produced by the research group, not only because they are in accordance with the MoU of the Action, but also because they are perfectly integrated in the environmental reality. In fact, this period (2006–2010) has been dramatically characterized by typical catastrophic events (four high intensity earthquakes: Suchuan 2008, Abruzzo 2009, Haiti 2009, Chile 2010, Yushu 2010, together many landslides and floods in Europe, last but not least the volcanic eruption in Island in 2010), whose consequences, in term of effect and damage, demonstrated the importance and the actuality to set-up a methodology for preventing damage to the built heritage due to such events. The activity of COST C26, as summarized in this Volume, represents a milestone in this complex subject involving several aspects. This Volume will represent in the near future a background document, as a reference point for new research projects, which have been already formulated as a natural development of this Action. One on “Integrated Fire Engineering and Response” (TU 904, chair Frantisek Wald) has been already approved and other two, proposed by Dan Dubina and Maurizio Indirli, are now under examination. Good luck to these new proposals. Concluding this introduction, I would like to express my sincere thanks to the members of the Management Committee representing 23 European Countries and to the experts of the WGs with their very effective leaders. The success of this Action is particularly merit of the friendly atmosphere of co-operation which has been created within the group, making the common work as an unrepeatable experience. Federico M. Mazzolani Chair of COST C26 Naples, June 2010

4

Chapter 1: Characterization of catastrophic actions on constructions

Urban Habitat Constructions under Catastrophic Events (Final Report) – Mazzolani (Ed). © 2010 Taylor & Francis Group, London, ISBN 978-0-415-60686-8

1.1 Design fires for structural engineering M. Gillie School of Engineering, University of Edinburgh, Edinburgh, UK

F. Wald & K. Horová Czech Technical University in Prague, Czech Republic

1.1.1 THE NATURE OF BUILDING FIRES 1.1.1.1 General To design structures to resist fire loading, some assessment of the likely gas temperatures to which they may be subjected is required. The first part of this paper summarizes the nature of fires that occur in buildings with reference to relevant tests. The study of fire dynamics is a subject in its own right so only an overview of the available knowledge as applicable to the design of building structures can be given here. For a complete survey of fire dynamics see Drysdale (1998). Since fire is a highly complex phenomenon and the various factors affecting its behaviour are unlikely to be known with any certainty for most structures, it is necessary to develop models of building fires that simultaneously account for the likely variation in the possible fires that could occur, and are sufficiently simple to apply in design. The second part of the paper discusses the various ways in which building fires may be modelled. In the final section of the paper three compartment fire tests are reviewed and the results from these tests compared with the predictions of gas temperatures that the various models of compartment fires make. 1.1.1.2 Fires in small compartments Most building structures are divided into a number of compartments such as offices, meeting rooms etc. From a fire safety perspective these divisions are significant for two reasons. Firstly they provide a means of preventing fire spread and hence often allow a structure to be designed on the assumption that a fire will only occur in a single compartment. Secondly a fire occurring within a compartment will develop in a different manner to a fire in an unrestricted space. Historically most research into the behaviour of compartment fires has focussed on small compartments and assumed a reasonably uniform distribution of fuel. This work has informed most of the design fires currently used in structural design and will be discussed in this section. The following section will consider the behaviour of fires in the larger compartments that are increasingly present in modern structures. Immediately after ignition the behaviour of a fire in a small compartment will not be affected by the compartment boundaries. However, hot gases will not be free to escape from the compartment but, due to buoyancy effects, will begin rapidly to accumulate in a layer under the ceiling. As the fire grows in size and the layer of gases develops, a point will be reached when the downward radiation from the smoke layer becomes sufficiently intense to ignite objects distant from the seat of the fire. Assuming a sufficient supply of air, this will result in full involvement of all combustible materials in the fire. The transition from localised to fully involved burning tends to be rapid and is known as “flashover”. The above process is indicated in Fig. 1. After flashover, the rate of heat release within the compartment is controlled by the rate of supply of air through openings such as doors and windows. This is a ventilation controlled fire and in sufficiently small compartments will result in fairly uniform temperatures at any level within the compartment. 1.1.1.3 Fully-developed fires The rate of heat release rate in fully-developed (flashed-over) fire in buildings is normally controlled by the amount of oxygen available. The most widely quoted relationship between the rate at which 7

Figure 1. The growth of a compartment fire. Table 1. Heat of combustion, hc , of common fuels. Obtained from DiNenno (2002), Drysdale (1998) and EN 1991-1-2, 2004. Fuel MJ/kg

Heat of combustion

Wood (Beech) Wood (Pine) Leather Wool Nylon Polyester Alcohols Petroleum Diesel Coal

19.5 17.5 20 20.5 30 24–30 30 45 45 30

fuel in consumed in ventilation controlled fires and degree of ventilation is that of Kawagoe (1958), reported, for example, in Rasbash et al. (2004) as

where A is the area of ventilation opening and H its height. Kawagoe derived this result from fire tests in compartments containing wooden cribs. The value of constant the K is somewhat uncertain but is often quoted between 0.5 and 0.9 kgs−1 kg5/2 , depending on the size and shape of the ventilation opening. Various more refined versions of relationships similar to Kawagoe’s are available. By multiplying the rate of mass loss by the heat of combustion of the fuel, hc , an estimate of the rate of heat release within a compartment fire can be obtained as

Typical value of hc for various fuels are shown in Table 1. 1.1.1.4 Fire in large compartments and travelling fires Experimental and analytical evidence increasingly suggests that in larger fire compartments, such as typical office spaces, the assumption of uniform temperatures at any level within the compartment is not valid. Instead, fires in larger compartments will tend to travel within the compartment as fuel is consumed at a rate governed by the available ventilation. This causes variations in gas temperatures within such compartments that are not present in older descriptions of compartment fires. The first clear evidence of this was presented by Cooke (1998) who undertook a number of fire tests with uniform fire loads of wooden cribs in a long, thin (4.5 × 8.75 × 2.75 m high) compartment in which ventilation was provided at one end. The results showed a clear progression of temperature within the compartment. Peak values occurred near the source of ventilation early in the fire and then progressed away from the opening as fuel was consumed. This progression occurred even when ignition was distant from the source of ventilation. The progression of the peak temperature 8

Figure 2. The nominal standard fire curve as defined in EN 1991-1-2.

from the front to the rear of compartment took 20-30min, with higher levels of ventilation resulting in more rapid fire spread. Similarly Welch et al. (2007) report on a series of fire tests undertaken in a 12 × 12 × 3 m high compartment in which a combination of wood and plastic fuel was burnt. Ventilation was provided either along one or two walls of the compartment. The tests were heavily instrumented which allowed both temperature and heat-flux maps to produced after various periods of burning. Although the differences in temperature were modest, when converted to incident heat-fluxes on the compartment ceiling, very significant differences arose as a result of the T 4 dependence of radiation. Correspondingly, the effect of non-uniform gas temperatures on structural temperatures has been shown to be significant in recent work by Gillie and Stratford (2007) who reported on temperatures in a concrete slab above a fire compartment. Lower surface concrete temperatures varied by as much as 400◦ C despite a compartment size of only 3.5 m by 4.5 m. The above results have all been recorded in compartments that are still small in comparison the many modern office spaces and so highlight the need for a new approach to defining design fire for structural fire engineering. 1.1.2 DESIGN FIRES 1.1.2.1 The nominal standard fire The nominal standard fire curve is the best known and most widely used method of estimating temperatures in compartment fires. It assumes that the temperature in a fire compartment is uniform and that it increases indefinitely according to a logarithmic relationship with time, see Figure 2. The nominal standard fire curve has been incorporated, with minor differences, into a number of design standards worldwide. In EN 1991-1-2, 2002 the gas temperature, θ in ◦ C, at time t in minutes, is given as

This form of temperature-time relationship was originally derived from measurements of tests taken early in the 20th century and has been shown to have only a very limited similarity to the temperatures in real compartment fires. Notable shortcomings in the Standard Fire curve include the lack of a cooling branch and no dependence on either fuel load or the available ventilation. Thus it is often referred to as a “nominal” temperature-time curve. Until recently almost all structural fire design was based on prescriptive methods where there is no requirement to make an explicit assessment of the likely response of a structure to fire. Instead, fire protection is specified based on tests of single structural elements subject to Standard Fires. The relationship between such single element tests and real, global structural behaviour in fire is limited at best so the crude nature of the Standard Fire curve is not seen as problematical for prescriptive design. Over the last ten to fifteen years, performance-based methods have been introduced to structural fire engineering. These do require the designer to make assessments of structural behaviour in 9

Figure 3. Predicted temperatures within a compartment for various ventilation conditions using the parametric approach. The Standard Fire curve is shown for comparison.

fire, as assessments are made for other types of loading. Adopting the Standard Fire curve in performance-based design is difficult to justify on scientific grounds due to its lack of similarity with real fires and cannot be advised. Despite this, its use remains widespread, partly due to its ubiquity in other branches of fire engineering and partly due to the high and sustained temperatures that it predicts being seen as conservative. 1.1.2.2 Heat balance and Zone models Recognising that the Standard Fire curve was not physically reasonable, researchers in Sweden in the 1970s (Pettersson et al., 1976) developed a method of predicting fire temperatures by considering the heat balance in a fire compartment. By assuming • the temperature within a fire compartment is uniform, • all available fuel is burnt within the compartment, • and that the thermal properties of the compartment walls are uniform, the heat balance can be expressed as

where q˙ c is the rate of heat release due to combustion, q˙ L is the rate of heat loss due to the replacement of hot gases by cold, q˙ W the rate of heat loss through the compartment boundaries and q˙ R the rate of heat loss due to radiation through the compartment openings. By evaluating these terms it is possible to arrive at a differential equation that relates the temperature within the compartment to the fuel load, the available ventilation and thermal properties of the compartment walls. The solution to this equation cannot be expressed explicitly so compartment temperatures based on Pettersson’s approach are normally presented graphically for various fire loads and ventilation conditions, e.g. Drysdale (1998) or Buchanan (2000). The lack of an explicit solution to Petterson’s model was addressed in EN 1991-1-2 by instead using a parametric approach that fits a curve to the experimental data originally used to validate Petterson’s approach. The input variables are very similar and take account of fire load, ventilation conditions and the thermal properties of the compartment. Additionally, a fire growth rate is included and different behaviour is predicted for ventilation and fuel controlled fires. Full details of how to use this approach are given in EN 1991-1-2 and are also presented by Franssen and Zaharia (2005) in a rather clearer manner. A comparison of predicted gas temperatures for a typical compartment with various ventilation conditions is shown in Figure 3. Petterson’s model is the simplest example of a class of compartment fire models known collectively as “zone models”. These all represent compartment temperatures by considering energy, mass and momentum conservation with various levels of sophistication. One-zone models, such as Pettersson’s assume that all the gases within a compartment are at an equal temperature, whereas “two-zone” models divide the gases into an upper, hot zone and a lower, cooler zone. The more sophisticated zone models allow for factors such as compartment boundaries with varying thermal properties and multiple compartment openings to be included in analyses. Models that account the 10

interaction between fires in more than one compartment are also available. All but the simplest zone models require numerical solutions Approaches such as these remove some of the shortcomings of the Standard Fire test. Notably cooling behaviour is included and the effects of compartment geometry and ventilation conditions are accounted for. Thus, their use in performance-based design can be justified on the basis that the key factors affecting fire behaviour are included in the predictions of gas temperatures. Since the models are also reasonably straightforward to use they are attractive for routine designs where more onerous calculations would not be economic. Despite their benefits, heat-balance and zone models do have restrictions on their applicability and fail to capture some aspects of fire behaviour. The most significant simplification is the assumption that at any level within a compartment gas temperatures are uniform. This has always been regarded as a reasonable assumption for small compartments – the EN 1991-1-2 parametric equation is applicable for compartments up to 500 m2 floor area – although the recent work discussed above suggests significant variations in temperature will occur even in small compartments. For compartments larger than around 500 m2 uniform burning, and therefore temperatures, are unlikely because of restrictions on air supply to the fire. If simple heat balance estimates are applied to large compartments, a more severe fire than is in fact likely will be predicted because peak temperatures will not be reached simultaneously across the whole area of the compartment (Cooke, 1998). 1.1.3 COMPUTATIONAL FLUID DYNAMICS Computational Fluid Dynamics (CFD) models of fire growth and behaviour have been available for some years. CFD modelling is a numerical approach to representing fluids that divides a fluid domain into small volumes and considers conservation of mass, energy etc. within each volume. Software exists that can represent the very wide range of physical phenomena known to affect fire behaviour including compartment geometry, heat release rates of burning fuel, complex ventilation conditions, turbulent gas flow, soot production and many others. Using such software is complex and time-consuming and for this reason CFD models are currently little used in design work. If the greater resolution of fire behaviour available from CFD models is considered to worth the additional effort, great care must be taken when obtaining and interpreting predictions as the output can be influenced hugely by even minor differences in input data. Obtaining the full range of input data needed in a sufficiently accurate and precise manner will be impractical for most structural engineering problems. The implications of not having the correct input data was highlighted in a recent study where Rein et al. (2007a) compared the blind predictions of a fire in a very well defined compartment by nine different analysts using CFD methods. The predictions varied very widely. Rein et al., concluded that at present CFD predictions of fire growth are not sufficiently reliable to be used in engineering design unless directly supported by experimental validation. However, they also noted that if the use of CFD models is restricted to predicting gas temperatures for a given fire heat release rate, good predictions can be made. The use of CFD in this way is likely to be advantageous for structures with complex compartment geometries for which the use of zone models can not be justified. 1.1.4 COMPARISON TO FIRE TESTS 1.1.4.1 Cardington fire test 2003 The structural integrity fire test, large scale test No. 7, was carried out in a centrally located compartment of the building, enclosing a plan area of 11 m by 7 m on the 4th floor. The preparatory works took four months. The fire compartment was bounded with walls made of three layers of plasterboard (15 mm + 12,5 mm + 15 mm) with a thermal conductivity of between 0,19–0,24 Wm−1 K−1 . In the external wall the plasterboard is fixed to a 0,9 m high brick wall. The opening of 1,27 m high and 9 m length simulated an open window to ventilate the compartment and allow for observation of the element behaviour. The ventilation condition was chosen to produce a fire of the required severity in terms of maximum temperature and overall duration. The columns, external joints and connected beam, about 1,0 m from the joints, were fire protected to prevent global structural instability. The fire protection used was 18–22 mm of Cafco300 vermiculite-cement spray, with a thermal conductivity of 0,078 Wm−1 K−1 . 11

Figure 4.

Location of thermocouples in the compartment below the ceiling and on steel structure.

Figure 5. Comparison of the prediction of the gas temperature to the measured temperatures.

The geometry and measured material properties of the flooring system are summarised by Wald et al. (Wald et al., 2003). Wooden cribs with moisture content 14% were used to provide a fire load of 40 kg/m2 . The instrumentation used included thermocouples, strain gauges and displacement transducers. A total of 133 thermocouples were used to monitor the temperature of the connections, the steel beams within the compartment, the temperature distribution through the slab and the atmosphere temperature within the compartment, see Figure 4. The quantity of fuel and the dimensions of the opening in the facade wall were designed to achieve a representative fire in an office building. Figure 5 shows the measured time-temperature curve within the compartment. In the initial stages of the fire the temperature within the compartment grows rapidly to reach a maximum temperature of 1107,8 C after about 54 min. The maximum recorded compartment temperature occurred near the wall of the compartment. Figure 5 also compares the temperatures predicted by the parametric curve given in EN 1991-1-2: 2003 with the test results. The parametric curve predicts a maximum temperature of 1078◦ C after 53 min and this compares well with the test results, see (Wald et al., 2004). The influence of the secondary beam with 0,5 m height placed at the ceiling in the draining and in the distribution of temperatures of the involving fluid was simulated by CFD code, see Figure 6 from Numerical simulation of the 7th Cardington Compartment Fire Test using a full conjugate heat transfer approach. The study confirming the difference till 10% lead to the conclusion that the beam affected the results quite significantly. This was due to the manner in which the beam changed 12

Figure 6. The influence of the secondary beam in the ceiling simulated by CFD code Almeida et al. (2007).

Figure 7.

Geometry of the fire compartment.

the flow pattern. Figure 7 shows the velocity field at the symmetry plane, for both simulations (with and without beam). Analysis of this results revealed that the vortex from the simulation with the beam was reduced in the middle back of the compartment and was directed towards the opening as time proceeded. The presence of the beam showed to have a significant role in the general vortex development, thus affecting the velocity and temperature field. 1.1.4.2 Ostrava fire test 2006 The structure of Ostrava fire test was composed of tree storey steel structure with the composite slabs, the beam-to-beam and beam-to-column header plate connections, and the diagonal wind bracings. Internal size of fire compartment was designed 3,80 × 5,95 m with height of 2,78 m. The structure of enclosure was made from the light silicate and ceramic bricks. Opening of 2400 × 1400 mm ventilated the room during the fire. The doors of fire compartment 1400 × 1970 mm and columns were equipped by the fire isolation by boards, see Figure 7. Fire load was represented by the unwrought timber bars 50 × 50 mm of length 1 m from softwood with moisture till 13% For the compartment fire were the bars placed into eight piles. The gas temperature in the fire compartment was measured by four thermocouples 300 mm below ceiling, marked at Figure 8 as TGi. Two thermocouples were placed in front of the fire compartment 0,5 m and 1 m from front wall. Figure 9 shoes the comparison of the predicted gas temperature in the fire compartment by parametric fire curve according to EN 1993-1-2:2005 Annex A, and of the predicted primary beam temperature to the measured values. 1.1.4.3 Mokrsko fire test 2008 The structure in Mokrsko fire test represents one floor of the administrative building of size 18 × 12 m, see Figure 10. The composite slab on the castellated beams was designed with a span 9 to 12 m and on beams with corrugated webs with a span 9 to 6 m. The deck was a simple trapezoidal composite slab of thickness 60 mm with the height over the rib 120 mm with sheeting. Two walls were composed from cladding, linear trays, mineral wool and external corrugated sheets. In two 6 m spans were compared the system with the internal grid and horizontal sheeting and with vertical sheeting without the internal grid. Two other walls are made of sandwich panels of thickness 13

Figure 8. The position of the thermocouples for recording of gas and beams temperatures.

Figure 9. Comparison of the predicted temperature by the parametric fire curve to the measured average gas temperature and primary beam temperature.

Figure 10.

Fire compartment with description of the major floor and wall structures.

14

Figure 11.

Location of thermocouples below the ceiling and in the window openings.

Figure 12. Comparison of prediction of the gas temperature by nominal and parametric fire curve to the measured average temperatures from part and whole fire compartment measure 500 mm under the ceiling.

150 mm filled with mineral wool. In front of the concrete wall was brick over by plaster blocks. The fire protection of columns, primary and edge beams as well as bracings was designed for R60 by board protection 2 × 15 mm Promatect H. The fire load created 15 m3 unwrought wooden cribs 50 × 50 mm of length 1 m of softwood. The cribs were placed into 50 piles. The openings of height 2,54 m and total length 8,00 m with parapet 1 m ventilated the compartment. To allow a smooth development of fire no glassing was installed. The gas temperature in the fire compartment was measured by 14 jacketed 3 mm thermocouples located 0,5 m below the ceiling in the level of the beams lower flanges. Two thermocouples were placed in the openings. The temperature profile along the compartment height was measured between the window and in the back of the fire compartment below the backward secondary beam. The location of the thermocouples is shown in Figure 11. Figure 12 shows the comparison of prediction of the gas temperature by nominal and parametric fire curve to the measured average temperatures from part and whole fire compartment. The sensitivity of the prediction by zone model to the combustion factor m for fire growth rate coefficient tα = 300 s is documented on Figure 13. 1.1.5 FURTHER DEVELOPMENTS The recently acknowledged variability of temperatures in large fire compartments, have led researchers to begin developing new approaches to modelling compartment fire behaviour that are more sophisticated than the simple heat-balance approach but avoid the complexities and uncertainties of CFD models. Rein et al. (2007b) proposed a model consisting of “near-field” and “far-field” temperatures where the far-field temperatures result from hot gases and near-field temperatures from direct impingement of a flame. They proposed that the duration of exposure to 15

Figure 13. Prediction by zone model with different combustion factors m for fire growth rate coefficient tα = 300 s.

Figure 14. Predictions of far-field temperatures for different sizes of travelling fire in a 1250 m2 compartment using a near-field/far-field approach. The nominal standard fire curve and parametric curve predictions for a fire load 420 MJ/m2 are shown for comparison (Rein, 2007b).

near-field temperatures in well a ventilated fire is governed by the available fuel load and that for office fires this will be of the order of 15 minutes. Exposure to far-field temperatures was found to be around 10 times longer and dependent on the size of the fire (itself governed by the ventilation conditions) and geometry of the compartment. This method produces predictions of fire size, temperature and rate of travel based on fuel load and ventilation and thus appears to offer a useful method for predicting gas temperatures in large compartments that avoids the assumption on uniform temperatures present in other methods. However, it is still under active development and will probably be refined further. Figure 14 shows predictions of far-field temperatures for an example case. If should be noted that in addition to the temperatures shown in this figure, structural elements may experience short exposure to near-field temperatures. This will be of the order 1250◦ C. REFERENCES Buchanan A. (2001) Structural Design for Fire Safety, Wiley, ISBN-10: 047189060X. Cooke, G.M.E. (1998) “Tests to determine the behaviour of fully developed natural fires in a large compartment”, Fire Note 4, Fire Research Station, Building Research Establishment. Drysdale D. (1998) An Introduction to Fire Dynamics, 2nd Edition Wiley. ISBN 0-471-97291-6. EN 1991-1-2 (2002) Eurocode 1: Actions of Structures Part 1-2 Actions of Structures Exposed to Fire, CEN, Brussels.

16

Franssen J-M., Zaharia R. (2005) Design of Steel Structures Subject to Fire. Les Éditions de l’Universite de Liege. ISBN 2-930322-99-3. Gillie M., Stratford T. (2007) The Dalmarnock Fire Tests: Experiments and Modelling: Chapter 8 Behaviour of the Structure During the Fire. School of Engineering and Electronics, University of Edinburgh ISBN 978-0-9557497-0-4. Pettersson O., Magnuson S.E. and Thor J. (1976) Fire engineering design of structures Swedish Institute of Steel Construction, Publication 50. Rein G. et al. (2007a) The Dalmarnock Fire Tests: Experiments and Modelling: Chapter 10 A Priori Modelling of Fire Test One. School of Engineering and Electronics, University of Edinburgh ISBN 978-0-9557497-0-4. Rein G. et al. (2007b), “Multi-story Fire Analysis for High-Rise Buildings”, Proceedings of the 11th International Interflam Conference, London, Sept. 2007. Wald F., Chladná M., Moore D., Santiago A. (2006) The temperature distribution in a full-scale steel framed builing subject to a natural fire, Steel and Composite Structures, Vol. 6, No. 2. Wald F., Chlouba J., Kallerová P. (2006) Temperature distribution in the header plate connection subject to a natural fire, Czech Technical University in Prague. Wald F., Kallerová P. (2009) Results from Fire test in Mokrsko 2008, Prague, 2010, ISBN 978-80-01-04267-0. Almeida N., Lopes A., Vaz G., Santiago A., Silva L. (2007) Steel and Composite Structures: Numerical simulation of the 7th Cardington Compartment Fire Test using a full conjugate heat transfer appoach, London, ISBN 978-0-415-45141-3. Cadorin J.-F., Franssen J.-M., Zone model OZone, University of Liège.

17

Urban Habitat Constructions under Catastrophic Events (Final Report) – Mazzolani (Ed). © 2010 Taylor & Francis Group, London, ISBN 978-0-415-60686-8

1.2 Characterization and modelling of seismic action D. Lungu Technical University of Civil Engineering of Bucharest, Romania

A. Stratan Politehnica University of Timisoara, Romania

R. Vacareanu Technical University of Civil Engineering of Bucharest, Romania

1.2.1 STATE OF THE ART 1.2.1.1 General The two orthogonal horizontal components and the vertical component of the ground motion produced at a site by an earthquake provide the characterisation of the shaking. The ground motion representation for structural design may be one of the following: (i) Time histories (ii) Power spectral density (iii) Response spectra. Any of these representations can be either site-dependent or site-independent. A site-dependent representation of the seismic input must be appropriate for the soil characteristics where the structure is located, as well as for the source mechanism and the geologic path from the source to the site. The time histories can be recorded or simulated. The limitations of number, spectral content and especially of the acceleration level of recorded accelerograms are the main reasons for the synthesising earthquakes. The power spectral density is the most representative function to describe the frequency content of the ground motion. The response spectra for a specified ground motion represents the maximum response of the SDOF structure to that motion. The response can be elastic (linear) or inelastic (nonlinear). The response ordinates can be normalised with respect to the peak value of the corresponding ground motion parameter. The normalised response is called the dynamic amplification. 1.2.1.2 Ground motion modelling 1.2.1.2.1 Peak ground acceleration and peak ground velocity Recorded ground motions should be digitised at a time interval of 0.005 s. Afterwards the correction procedure apply zero base-line, instrument correction and filters. The corrected accelerograms are integrated to find velocity and displacement time-histories. The main characteristics of the ground motion time history are: (i) Peak ground acceleration (PGA) and corresponding peak ground velocity (PGV) and peak ground displacement (PGD); (ii) Frequency content; (iii) Motion duration. 19

The peak ground acceleration is the amplitude of ground acceleration. The peak ground velocity is the amplitude of ground velocity. Both parameters are very simple measures of severity of the ground shaking. PGA and PGV recorded at a site during an earthquake are random variables and they should be defined with a specified probability to be exceeded on a certain area during a specified time interval. The prediction of the peak ground parameters at a site is the target of the probabilistic seismic hazard analysis (PSHA). 1.2.1.2.2 Strong motion duration The strong motion duration is defined as the time interval between two specified fractions of the total cumulative energy of accelerogram:

The cumulative energy of accelerogram monotonically increases from zero toward the total energy accumulated during the total duration of the motion, td . The specified fractions can be: 0.05 − 0.95, 0.15 ÷ 0.85, 0.05 ÷ 0.75, 0.10 ÷ 0.90, etc. The strong motion duration represents the time interval over which the motion power is almost constant and near its maximum (the power is the slope of the cumulative energy plot). Hence different strong motion duration definitions can be used for different seismic records. For simulated accelerograms, the durations of the strong phase of the motion and of the transient motions can be described by linear duration enveloping functions, I(t) multiplying the stationary acceleration: t ≤ tr I(t) = t/tr Rise time, tr Strong motion duration, tm tr < t < tr + tm I(t) = 1.0 I(t) = −t/tr + td /(tr + tm ). Decay time t ≥ tr + tm The above time intervals depend on the intensity of the ground shaking. For rock conditions and magnitudes of order of 5.5 ÷ 7.5, one may assume tr = 1–3 s, tm = 5–15 s and decay time 4–10 s. 1.2.1.2.3 Frequency content The frequency content of ground motion is the crucial concept for the understanding of the mechanism of ground motion to damage structures. The maximum values of structure response are when the structure frequency and the major part of the frequency content of ground motion fall in the same frequency band. The frequency content can be described: 1. Directly, by the power spectral density function (PSD), obtained from stochastic modelling of the acceleration process; 2. Indirectly, by the response spectra obtained from numerical integration of the motion equation for the SDOF structure. 1. The stochastic measures of frequency content are related to the power spectral density function of stationary segment of the ground motion. They are: (i) The dimensionless indicators ε (Cartwright & Longuet – Higgins) and q (Vanmarcke); (ii) The f10 , f50 and f90 fractile frequencies below which 10%, 50% and 90% of the total cumulative power of PSD occur and the frequencies f1 , f2 and f3 corresponding to the highest 1,2,3 peaks of the PSD. The ε and q frequency bandwidth measures are defined as functions of the spectral moments of the one-sided spectral density of the stationary process of ground acceleration, G(ω). To define the frequency content indicators, one has to introduce first the spectral moment of order “i”:

20

It follows that:

The Vanmarcke frequency content indicator is:

The Cartwright & Longuet-Higgins indicator is:

Wide frequency band processes have ε values close to 2/3 and smaller than 0.85. Narrow frequency band seismic-processes of long predominant period (i.e. superposition of a single harmonic process at a short predominant frequency, fp and a wide band process) are characterised by ε values greater than 0.95. The Kennedy-Shinozuka indicators are f10 , f50 and f90 and can be defined as:

The physical meaning of the above indicators is that the area enclosed by the normalized spectral density situated to the left of f10 , f50 and f90 is equal to 10%, 50% and 90% respectively, of the total area. The RMS value of the ground acceleration process is the square root of two-order spectral 1/2 moment λ0 and is very sensitive to the strong motion duration definition. Cumulative power of the PSD is defined by:

where the cut-off frequency is the reverse of the double of time interval used in process digitization. The duration for computing PSD of the stationary segment of the acceleration process should be selected as D = T0.9 − T0.1 , where T0.9 and T0.1 are the times at which 90% and 10% of the total cumulative energy of the accelerogram are reached. Alternative duration definitions are: D = T0.95 − T0.05 or D = T0.75 − T0.05 . For simulating time – histories, a minimum power spectral density compatible with a specified response spectra functions must be recommended. 21

Figure 1. Variation of q values with respect to γ.

– Note on the values of frequency content indicators If one considers a random signal with constant power spectral density G0 within the range [ω1 , ω2 ], the spectral moment of order “i” is:

where: γ = ωω12 . Using relation (12) in the definition of q indicator, one gets:

One can notice that: – for γ → 0 (ideal white noise), q → 1/2; – for γ = 1 (pure sinusoid – harmonic), q = 0; – since the q indicator is a measure of the scattering of the PSD values with respect to the central frequency, the q value increases as the frequency band of the PSD increases; – for narrow band random signals, the q values are in the range [0, 0.25]; – for wide band random signals, the q values are in the range (0.25, 0.50]. The variation of q values with respect to γ is represented in Figure 1. Using relation (12) in the definition of ε indicator, one gets:

One can notice that: – for γ → 0 (ideal white noise), ε → 2/3; – for γ = 1 (pure sinusoid – harmonic), ε = 0; – for wide band random signals, the ε values are in the vicinity of 2/3. The variation of ε values with respect to γ is represented in Figure 2. Let us consider a random signal expressed as a sum of two random signals, one with a PSD characterized by a constant value GA within a narrow frequency band in the range [ω1 , ω2 ] and 22

Figure 2. Variation of ε values with respect to γ.

Figure 3.

PSD of a random signal – situation 1.

one with a PSD characterized by a constant value GB within a wide frequency band in the range [ω2 , ω3 ], Figure 3. The spectral moment of order “i” is:

where: γA = ωω12 and γB = ωω23 . Using relation (15) in the definition of q indicator, one gets:

23

Figure 4.

PSD of a random signal – situation 2.

Figure 5. Values of q for situation 1.

Using relation (15) in the definition of ε indicator, one gets:

In the following developments, two situations are considered: – Situation 1, of Figure 3, where the narrow band random signal is of high frequency; γA = 0.05 and γB = 0.95; – Situation 2, of Figure 4, where the narrow band random signal is of low frequency; γA = 0.95 and γB = 0.05. The values of the frequency content indicators are presented in the following. The results are represented in Figures 5–8, as a function of the ratio GA to GB . From Figures 5 and 6 one can notice that: – for situation 1 • for small GA /GB values, the wide band random signal has little influence, the value of q approaching 0, as for a pure sinusoid; 24

Figure 6. Values of q for situation 2.

Figure 7. Values of ε for situation 1.

• for large GA /GB values, the value of q approach 0.50, the value corresponding to a white noise random signal. The value is only approaching 0.50, but it never reaches 0.50 since it is a band limited white noise random signal; • a superposition of a wide band random signal with a high frequency narrow band random signal with strong contrast in terms of PSD ordinates (GA /GB << 1) will produce a random signal very close to a pure sinusoid. In other words, only the narrow band random signal will influence the value of q; – for situation 2 • for small GA /GB values, the value of q approach 0.50, the value corresponding to a white noise random signal. The value is only approaching 0.50, but it never reaches 0.50 since it is a band limited white noise random signal; • for large GA /GB values, the value of q is increasing above 0.50. The maxim q value is 0.85; • a superposition of a wide band random signal with a low frequency narrow band random signal with strong contrast in terms of PSD ordinates (GA /GB >> 1) will produce a random signal with q values larger than 0.50. The increase in q values is steadier. From Figures 7 and 8 one can notice that: – for situation 1 • for small GA /GB values, the wide band random signal has little influence, the value of ε approaching 0, as for a pure sinusoid; • for large GA /GB values, the value of ε approach 2/3, the value corresponding to a white noise random signal. The value is only approaching 2/3, but it never reaches 2/3 since it is a band limited white noise random signal; • a superposition of a wide band random signal with a high frequency narrow band random signal with strong contrast in terms of PSD ordinates (GA /GB << 1) will produce a random 25

Figure 8. Values of ε for situation 2.

signal very close to a pure sinusoid. In other words, only the narrow band random signal will influence the value of ε. – for situation 2 • for small GA /GB values, the value of ε approach 2/3, the value corresponding to a white noise random signal. The value is only approaching 2/3, but it never reaches 2/3 since it is a band limited white noise random signal; • for large GA /GB values, the value of ε is steadily increasing. The maxim ε value is 0.99; • a superposition of a wide band random signal with a low frequency narrow band random signal with strong contrast in terms of PSD ordinates (GA /GB >> 1) will produce a random signal with ε values larger than 2/3. The increase in ε values is steadier and larger as the PSD of the narrow band random signal is narrower. The deterministic measures of frequency content are related to the structure maximum response to the ground motion. They are the control frequencies and corresponding control periods:

fc is the border between the maximum acceleration branch and the maximum velocity branch of response spectra and fD is the border between the maximum velocity branch and maximum displacement branch of response spectra. SD, SV and SA are respectively the relative displacement, relative velocity and absolute acceleration spectra of the SDOF structure response. For classifying the frequency content of the accelerograms the most significant measures are: (i) (ii) (iii) (iv)

the control frequency fC the indicator ε and the frequency bandwidth f10 − f50 − f90 the first peak of PSD f1 .

The correlation between median frequency f50 and the control frequency fC was found very strong but the stability of control frequency is better than that of the median frequency. 1.2.1.3 Modelling of response spectra 1.2.1.3.1 Accurate response from time history Response spectra from the time history must be completed at sufficient frequency (period) intervals to have a good resolution of spectral ordinates. A set of 100 frequencies (periods) should be selected to produce accurate response spectra. 26

1.2.1.3.2 Effective peak ground acceleration and effective peak ground velocity Since the PGA and PGV are not sufficiently consistent scaling factors for the spectral values of both acceleration and velocity, the effective peak acceleration (EPA) and effective peak velocity (EPV) are recommended to be considered as normalizing factors for response spectra. The EPA is defined as the average of the maximum ordinates of acceleration spectra in the period range of 0.1 to 0.5 seconds, divided by a standard (mean) value of 2.5 (for 5% damping):

The EPV is defined as the average of the maximum ordinates of velocity spectra in the period range of 0.8 to 1.2 seconds, divided by the same standard value of 2.5 (for 5% damping):

The somewhat arbitrary definitions are intended for the broad frequency band ground motions having a relatively small or medium control period (say: TC < 0.4 ÷ 0.5 s). For other frequency content and especially for the narrow frequency band motions of long predominant period the above definitions must be definitely changed. In that case, the 0.4 s averaging period interval must be centered on the period range corresponding to the dominant peak of each of the spectra SV and SA respectively. (The period corresponding to the peak of the SV is close to the period corresponding to the predominant peak of the PSD. The period corresponding to the peak of the SA is slightly less than the period corresponding to the dominant peak of the SV). The new parameters EPA and EPV characterize the intensity of a ground motion by averaging the effects of shaking on the most exposed – to that spectral content – structures. Individual EPA and EPV may be either lower or higher than corresponding PGA and PGV, respectively. Regression between EPA and PGA shows that EPA = 0.65 ÷ 0.85 PGA (for peak ground accelerations > 0.05 g). Regression curves relating EPV and PGV show similar results. 1.2.1.3.3 Elastic response spectrum The elastic response spectrum is given as smoothed acceleration spectrum (for 5% damping) having a specified probability to be exceeded; 0.5 median, 0.1, etc. Generally, the spectra are computed for free field ground motions. In certain cases, the output motions at the top of a rock underlying relatively soft soil layers can be used. Seismic input motions must be compatible with local soil condition, intensity of shaking, seismic source mechanism and hypocentral distance. The recommended response spectrum in this code is the Model spectrum No 1 in Table 1. Coefficient of variation of the normalised (with respect to PGA or EPA) spectral shapes, β(T) is a parabolically increasing function of structure period: from about COV = 0.1 for T = 0.1 s to a constant value COV = 0.4 ÷ 0.7 for T > TC . The higher values correspond to a larger diversity (in the spectral content) of the input motions. For computing the normalised spectral shapes having a specified probability to be exceeded the lognormal distribution is recommended. Observations. The Model 1 is calibrated by EC8 with respect to the control period as follows: (i) TC = 0.4 s for rock underlaying ≤5 m of weak material and for gravel and overconsolidated clays at least of several tens of meters and having vs ≥ 400 m/s at a depth of 10 m; (ii) TC = 0.6 s for soils having shear wave velocity Vs > 200 m/s at a depth of 10 m increasing to Vs > 350 m/s at a depth of 50 m; (iii) TC = 0.8 s for Vs ≤ 200 m in the uppermost 20 m. The Model 2 is calibrated by the NEHRP Seismic Provisions (1991) and the ASCE 7-95 draft standard (1995) as follows:

27

Table 1. Smoothed acceleration elastic response spectra for horizontal free-field ground motions (damping 0.05).

1)

2.5 indicates the mean response spectrum EPA, EPV or PGA values should be calibrated for 10% probability to be exceeded during 50yr (475 yr mean recurrence interval) 2)

where: Vref is the shear wave velocity of the reference soil condition (rock or even other soil condition). Vs – the mean shear wave velocity in the uppermost 30 m at the site. The exponents ma and mb are given as functions of the shaking intensity (Borcherdt, 1994): EPA

0.1 g

0.2 g

0.3 g

0.4 g

ma for mean Fa mb for mean plus 1st dev Fb

0.35 0.65

0.25 0.60

0.1 0.53

−0.05 0.45

The mean shear wave velocity in the uppermost 30 m, Vs is classified as follows: (i) 750–1500 m/s for rock; (ii) 370–750 m/s for gravelly soils, soft rock or soils with >20% gravel having more than 10 m thickness; (iii) 180–370 m/s for stiff clay and sandy soils (including silt loams, sandy clays, silty clays) having ≥5 m thickness; (iv) <180 m/s for soft soils (very soft clays of standard penetration resistance Nblows < 16 and having ≥3 m thickness and silty clays <37 m). Both models are intended to be applied for wide and intermediate frequency band ground motions with TC ≤ 1.0 s. 1.2.1.3.4 Design spectrum and inelastic response See Table 2. 1.2.1.3.5 Acceleration-displacement response spectra The acceleration-displacement response spectra format (Mahaney et al., 1993), termed ADRS, is an alternative representation of response spectra. Periods in these ADRS are represented by a series of radial lines extending from the origin of the plot. The elastic and inelastic response spectra of seismic motions recorded in free-field in Bucharest during 1977 and 1986 Vrancea subcrustal earthquakes were computed using NONSPEC computer program. Some results are presented in ADRS format in Figure 9. 1.2.1.4 Probabilistic seismic hazard assessment Probabilistic seismic hazard assessment (PSHA) has a cornerstone position for the prediction of the strong ground motion likely to occur at a particular site. For most seismic regions, the basic information for the hazard analysis (source properties, catalogue of events, records at relevant intensity, soil geological data, etc.) is very limited. 28

Table 2. Response modification (behaviour) factor for the response of nonlinear structure

1) 2) 3)

1 R

=

1 1 Rµ Rov

ξ – structure initial damping Rµ – mean Rµ σln Rµ – standard deviation of ln Rµ

Figure 9.

Elastic and inelastic ADRS format for Bucharest ‘86.

Moreover, the adequacy of earthquake catalogues, the model of earthquake occurrence, the structure of attenuation relation, the nonhomogenity of recording conditions, etc. artificially contribute to the randomness inherent in the hazard analysis. One may observe that the quality of deterministic information has an important influence on the accuracy of probabilistic results. The general PSHA is based on the following methodology: (i) Identification of the independent sources of seismic activity and determination of the Gutenberg-Richter relationship from contribution of each source; 29

Table 3. Hazard levels of the factors involved in the assessment of seismic design force. EPA, EPV (or PGA, PGV) hazard induced by:

Probability of non-exceedance of structural response Response modification factor 1/R

Source magnitude

Attenuation law

Normalised elastic response spectra β(T)

1/Rµ

1/Rov

T = 100 yr. 0.1 probability of exceedance in 50 yr. (T ∼ = 475 yr.)

Mean Mean plus one standard deviation

0.5 0.9

0.5 0.9

Deterministic

Soil factor, F (to EPA and EPV) Mean (for EPA) Mean plus one standard deviation (for EPV)

(ii) Fitting the attenuation relationship on peak ground motion (or structural response) parameter, properly classified according to the soil category; (iii) Calculating the peak ground motion (or structural response) parameter at the site with a specified probability of non-exceedance during structure lifetime (recurrence intervals may alternatively be used) (iv) Delineation of isoseismal maps in one of the following formats: – Peak ground acceleration (PGA); – Effective peak acceleration EPA and effective peak velocity EPV; – Spectral acceleration SA and spectral velocity SV at specified frequencies (periods) developed for a damping ratio of 0.05; (v) Construction of uniform hazard response spectra for design. The global hazard level involved in the assessment of seismic design force in building codes is assembled in Table 3. Details on PSHA is out of the scope of this chapter. However, some information on the PSHA related to Vrancea subcrustal seismic source are given hereinafter. The Vrancea region, located when the Carpathians Mountains Arch bents, is the source of subcrustal (60–170 km) seismic activity, which affects more than 2/3 of the territory of Romania and an important part of the territories of Republic of Moldova, Bulgaria and Ukraine. According to the 20th century seismicity, the epicentral Vrancea area is confined to a rectangle of 40 × 80 km2 having the long axis oriented N45E and being centered at about 45.6◦ Lat.N 26.6◦ and Long. E. The average number per year of Vrancea subcrustal earthquakes with magnitude equal to and greater than Mw is (Lungu et al, 1999):

In Eq.(23), the threshold lower magnitude is Mw0 = 6.3, the maximum credible magnitude of the source is Mw,max = 8.1, and α = 3.76 ln10 = 8.654, β = 0.73 ln10 = 1.687. The following model was selected for the analysis of attenuation (Mollas & Yamazaki, 1995):

where: PGA is peak ground acceleration at the site, Mw – moment magnitude, R – hypocentral distance to the site, h – focal depth, c0 , c1 , c2 , c3 , c4 – data dependent coefficients and ε – random variable with zero mean and standard deviation σε = σln PGA . The values of the coefficients, based on the strong motion accelerographic data from all regions of Romania subjected to Vrancea subrustal 30

Figure 10.

Seismic hazard curve for Iasi City.

source generated earthquakes are given as: c0 = 3.098, c1 = 1.053, c2 = −1.000, c3 = −0.0005, c4 = −0.006, σln PGA = 0.502. Details are given elsewhere (Lungu et al., 1999). The mean annual rate of exceedance of PGA – the hazard curve – for Iasi City due to Vrancea subcrustal seismic source is represented in Figure 10. The results in Figure 10 are based on the research done within CERSIS Project, Contract #31-023/2007, finaced by the National Authority for Scientific Research of Romania.

1.2.2 CONTRIBUTION TO THE RESEARCH DEVELOPMENT There is a high probability that seismic action will at some time exceed the one considered in design of a structure. This fact is related to the inherently uncertain character of the seismic action. Though knowledge on characterisation and modelling of seismic action is continuously improving as more data from past earthquakes becomes available, the design seismic action can still be specified in probabilistic terms only. An example of a catastrophic earthquake is the recent event from May 12, 2008 in Sichuan province of China. From observations on the failure and collapse pattern, it seems that the earthquake intensity was much higher than what has been designed for in that area (Wibowo et al., 2008). 1.2.2.1 Seismic motion leading to exceptional actions on structures Several phenomena that can lead to exceptional effects of seismic action on structures are summarised by Mistakidis et al. (2007) in a paper developed by members of WG2 “Earthquake resistance” of COST Action C26. There are two aspects that can lead to exceptional characteristics of ground motion on structures often neglected in design: near-fault ground motions and local site conditions. In the case of near-field ground motions, with the distance to the fault up to 20–60 km, the azimuth of the site with respect to the hypocenter may affect considerably the characteristics of the seismic motion. The effect of forward directivity is produced when the rupture propagates towards a site and the slip takes place also towards the site (Stewart et al., 2001). Due to the fact that velocity of fault rupture is close to the shear wave velocity, an accumulation of energy is observed at the rupture front. Ground motion in a site affected by forward directivity has the form of a long duration pulse. Vertical component of the ground motion is generally smaller than the horizontal one, and its effect on structural response is generally ignored. However, in the near-fault regions, vertical component of the ground motion may be important (Gioncu and Mazzolani, 2002) and its influence on seismic performance of structures deserves attention. Vertical component is believed to have contributed to some brittle failure modes in steel structures during the Northridge (1994) and Kobe (1995) earthquakes (Gioncu and Mazzolani, 2002). Local site conditions have been recognized for a long time as important parameters affecting ground motion characteristics. Recordings of strong-motion vary significantly with respect to: local 31

Figure 11.

City-site interaction (Bard et al., 2007 in Gioncu and Mazzolani, 2010).

geotechnical conditions, possible basin effects, and surface topography (Stewart et al., 2001). From the above factors, local geotechnical conditions were studied in most detail. Studies performed by Idriss et al. (in BSSC, 2001) show a dependence of the amplification of peak ground acceleration (PGA) by the soil layers on the intensity of the ground motion. Amplification is maximum (between 1.5 and 4.0) for small values of PGA at the base rock (0.05–0.1 g), and tends to decrease for ground motions of larger intensities (factors close to 1.0 for values PGA at the base rock about 0.4 g). Reduced amplification of at large intensities is attributed to nonlinear soil response. Another effect of soft soil conditions is a significant amplification of spectral accelerations in the medium and long period range (periods larger than 1 second). Surface topography can affect ground motion characteristics in several ways. If the seismic wave enters the basin through its edge, it may be “trapped” inside the basin. The effects of multiple reflections are the amplification and increase of duration of the seismic motions (Graves, 1993, in Stewart et al., 2001). Amplification of seismic motion may be observed also for irregular topographies, such as crest, canyon, and slope. A description of typical topographic amplifications was presented by Castellani et al., 1982 (in Athanasopoulos et al., 1998). 1.2.2.2 Seismic action in urban habitats City-site interaction and near-source effects were addressed by Gioncu and Mazzolani (2010) in a paper prepared for the final COST C26 Conference in Naples. City-site interaction effects refer to the influence on the ground motions of densely urbanized cities. The main effects of this secondary seismic source are produced in the case of a dense constructed area situated in a soft soil (see Figure 11), where the presence of buildings radically changes the characteristics of ground motions. The superposition of the vibrations produced by buildings over the soil vibrations coming from the source gives rise to a modification of the actual ground motions, which differ from the free-site ones. In addition, each point on the surface can have different movements, explaining the strange and highly variable damage within identical building sets (see the damage in urbanized areas during the very destructive Kocaeli and Taiwan earthquakes). Due to the limited number of investigated cases and the differences between regular and irregular cities, it is difficult to derive general rules to characterize these effects, but the increase of seismic design loads is expected in case of densely constructed areas. The near-source effects are characterized by pulse with very high velocity, important vertical components, short duration, and reduced number of pulses. Due to very high velocity of pulse seismic actions, the effects of strain-rate increase, the brittle fracture being the real danger for structure collapse (examples are Northridge and Kobe earthquakes). Therefore, the design approach for structures in near-source areas must mainly consider the strength problems. Due to the reduced number of pulses and short duration, the amount of dissipated seismic energy is reduced and the reduction of design seismic loads due to ductility, as it is considered in code provisions, must be revaluated (Gioncu. and Mazzolani, 2010). 1.2.2.3 Seismicity of Vrancea seismic source and soil conditions in Bucharest A specific study on the seismicity of Vrancea seismic subcrustal source and soil conditions in Bucharest, Romania was addressed by Lungu et al. (2008a & 2008b) in two datasheets prepared for 32

the COST C26 workshop in Prague. The major information related to the seismicity of Romania, the existing seismic instrumentation, the available strong motion records and the new seismic zonation map from the P100-1/2006 Romanian seismic design code are presented. Vrancea source dominates seismic hazard not only in Romania but also in Republic of Moldova and also affects large areas in Bulgaria and Ukraine. Strong Vrancea earthquakes have been felt on areas of about 2 millions km2 . Based on the available data obtained from more than 400 boreholes and using the GIS techniques, significant soil parameters were mapped for the territory of Bucharest. It is shown that the geological results, correlated with (i) shear wave velocity measurements in several locations having depth between 30 m and 200 m and (ii) analysis of recorded strong earthquakes permit seismic microzonation of Bucharest to be used as a tool for urban planning and earthquake risk reduction.

1.2.2.4 Selection of time-history records for dynamic analysis of structures Time-history analysis is increasingly used in design of new structures and evaluation of existing ones. In the case of time-history analysis, seismic action is described by a suite of ground acceleration records, which in the current practice are selected on a case by case basis. A summary of code requirements and an overview of several references related to selection of earthquake records for time-history analysis of structures were collected by Stratan and Dubina (2008) as part of the activities within WG of the COST Action C26. 1.2.2.4.1 Code provisions Provisions related to selection of earthquake records for time-history analysis from the following three codes are summarised in the following: EN 1998-1 (2004), NEHRP (FEMA 450, 2003) and FEMA 356 (2000). Eurocode 8 (EN 1998-1, 2004) stipulates that when a spatial model of the structure is required, the seismic action should be represented by three time-history records (accelerograms) applied along the two horizontal and one vertical directions. The obvious fact hat the two horizontal components should be different is explicitly stated by the code. Considering that vertical component of the ground motion may be neglected for most of the structures, analysis of a spatial model would require two (different) horizontal records, while the analysis of a plane model would require one horizontal record. Accelerograms obtained through three different procedures are allowed by Eurocode 8: artificial, recorded and simulated accelerograms. Artificial accelerograms should be generated so as to match the codified elastic response spectra at the building site for 5% viscous damping. The duration of accelerograms should be consistent with the magnitude and other relevant features of the seismic event used in establishment of the design seismic action. When site-specific data are not available, the minimum duration of the stationary part of the accelerogram should be equal to 10 s. Recorded accelerograms may be used, provided that the samples used are adequately qualified with regard to the seismogenetic features of the sources and to the soil conditions appropriate to the site and their values are scaled to the value of peak ground acceleration atop of soil layers (ag · S) for the zone under consideration. Simulated accelerograms should be generated through a physical simulation of source and travel path mechanisms, complying with the requirements for recorded accelerograms stated above. Irrespective of the procedure used to obtain the suite of accelerograms, they should observe the following rules: – The mean of the zero-period spectral acceleration (peak ground acceleration) values of individual time-histories should not be smaller than the codified peak ground acceleration atop of soil layers (ag · S). – In the range of periods between 0.2T1 and 2T1 (where T1 is the fundamental period of the structure in the direction where the accelerogram will be applied) no value of the mean 5% damping elastic spectrum, calculated from all time histories, should be less than 90% of the corresponding value of the 5% damping elastic response spectrum. 33

The following requirements are stipulated in Eurocode 8 regarding the number of accelerograms: – A minimum of 3 accelerograms should be used for time-history analysis. – If at least 7 accelerograms are used for time-history analysis, the behaviour of the structure should be verified using the average values of response quantities (forces, displacements, deformations). – If less than 7 accelerograms are used, the most unfavourable value of the response quantity should be used. NEHRP (FEMA 450, 2003) requires that for analysis of plane models of structures, a horizontal component of an accelerogram obtained from actual recorded events should be used. Accelerograms should be obtained from events having magnitudes, fault distance, and source mechanisms that are consistent with those that control the maximum considered earthquake. Where the required number of appropriate recorded accelerograms is not available, simulated accelerograms should be used to make up the total number required. The ground motions should be scaled such that for each period between 0.2T1 and 1.5T1 (where T1 is the fundamental period of vibration of the structure for the direction of response being analyzed) the average of the 5% damped response spectra for the suite of motions is not less than the corresponding ordinate of the target elastic response spectrum. In the case of analysis of spatial models of structures, two horizontal components are required. Additionally to the criteria specified for single-component records, each pair of motions shall be scaled such that for each period between 0.2T1 and 1.5T1 (where T1 is the fundamental period of the structure) the average of the SRSS (Square Root of Sum of Squares) spectra from all horizontal component pairs should not be less than 1.3 times the corresponding ordinate of the target elastic response spectrum. The following requirements are stipulated in NEHRP (FEMA 450, 2003) regarding the number of accelerograms: – If at least 7 accelerograms are used for time-history analysis, the behaviour of the structure should be verified using the average values of response quantities (forces, displacements, deformations). – If less than 7 accelerograms are used, the most unfavourable value of the response quantity should be used. FEMA 356 (2000) provisions are similar to the ones from NEHRP (FEMA 450, 2003). The differences between the two concern the explicit requirement of the minimum number of records allowed for in analysis (3 accelerograms), and the scaling of two-component records, whose average SRSS spectrum should not be less than 1.4 times the corresponding ordinate of the target elastic response spectrum. 1.2.2.4.2 Discussion of code requirements Code provisions are generally concerned with four aspects of selection of earthquake records: (1) how are the records obtained (through artificial generation, from existing recordings of past earthquakes, or through simulation); (2) the compatibility between earthquake records and the seismic source, travel path and site characteristics that control seismicity at the building location; (3) matching between the target response spectrum and the response spectra of earthquake records, accounting for properties of the analysed structure and (4) the number of records used and implications of result interpretation. Eurocode 8 (EN 1998-1, 2004) seems to favour artificial accelerograms, though allowing all three types of earthquake records. NEHRP (FEMA 450, 2003) and FEMA 356 (2000) give preference to recorded accelerograms, but acknowledge the fact that enough data may not be available, and some of the records could be simulated ones. It seems that the “simulated” accelerograms referred in US codes include both “artificial” and “simulated” accelerograms as defined in Eurocode 8. Selection or generation of a suite of earthquake records need to account for the seismic source (magnitude, type of faulting mechanism), travel path (site to source distance) and local site effects (soil type, topography, etc.). A seismologist would be required to analyse and process such kind of data. Though earthquake records obtained as above would be “appropriate” from the seismological point of view, codes impose a supplemental requirement that response spectra of accelerograms fit in some way the code (“target”) elastic response spectrum. The average of the suites of earthquake records should fit the code spectrum in a range of periods specified with respect to the fundamental 34

Figure 12. Artificial accelerograms, generated for different site-source distances, from near field (W1) to far field (W5), Chang and Kawakami, 2006.

period of vibration of the analysed structure. The lower limit (e.g. 0.2T1 ) accounts for higher modes of vibration of the structure, while the upper limit (e.g. 2T1 in Eurocode 8) accounts for effective period increase due to “softening” of the structure in the case of plastic response. This spectrum matching has little seismological reason, and seems to assure the structural engineer that the time-history representation of seismic action would be compatible with the response spectrum representation for the particular case of the analysed structure. Due to the probabilistic nature of seismic action, it is not relevant to characterise seismic response of a structure under a single ground motion record. Therefore, most codes require a minimum of three earthquake records to be used in analysis. As it is a relatively small number of records from a statistical point of view, in this case the structural assessment is to be based on the most unfavourable response. When more earthquake records are used (a minimum of 7), structural assessment can be based on the mean response, due to a more representative number of records.

1.2.2.4.3 Time-history records There are a lot of procedures available in literature regarding the development of earthquake records for time-history analysis of structures. Their classification is quite varied and therefore difficult to synthesise. According to Eurocode 8 (EN 1998-1, 2003) earthquake records for time-history analysis of structures can be of three types: artificial, simulated or recorded accelerograms. Artificial accelerograms are generated using stochastic algorithms so that the response spectrum of the generated signal with a prescribed duration matches the “target” (usually code) response spectrum. Ground acceleration is modelled as a filtered Gaussian white noise modulated by a deterministic envelope function (Safak, 1988 in Erdik and Durukal, 2003). One of the well-known procedures for generation of artificial accelerograms is that of Gasparini and Vanmarcke, 1976. This approach is often criticised for generating accelerograms that do not reflect the real phasing of seismic waves and cycles of motion (Iervolino et al., 2008). An improved class of methods for generating artificial accelerograms modifies the procedure by accounting for some seismogenetic such as magnitude and distance (Sabetta and Pugliese 1996, Boore, 2002 in Dall’Ara et al., 2006; Hanks and McGuire, 1981, Boore, 1983 in Erdik and Durukal, 2003; Chang and Kawakami, 2006). Figure 12 shows an ensemble of five artificial accelerograms generated for different site-source distances, from near field (W1) to far field (W5), Chang and Kawakami (2006). 35

Figure 13. Pulse types used to represent fault-normal components of near-fault ground motion by Sasani and Bertero, 2000 (a) and Alavi and Krawinkler, 2000 (b).

Another option is to use semi-artificial accelerograms, where existing recorded accelerograms are modified to obtain full spectrum matching with the target spectrum. The modification can be realised in the frequency-domain (e.g. changing the Fourier amplitude spectra) or in the timedomain (e.g. wavelet transform, Iervolino et al., 2008). Records obtained using this procedure have the advantage over pure artificial records that the essential character of the original record is preserved. Thus, records that conform to the type of source characteristics expected (e.g. strike/slip, subduction, near-field forward directivity etc.) can be selected (Priestley, 2006). Simple pulses can be used in order to model the ground motion. A review of existing research in this field is available in Gioncu and Mazzolani, 2002. Synthetic pulses were used often in studies concerning seismic response of structures under near-fault ground motions (Sasani and Bertero, 2000; Alavi and Krawinkler, 2000, see Figure 13). Simulated records are obtained through physical simulation of source and travel path mechanisms, and may account for site effects. A number of different methods have been used for the deterministic simulation of strong ground motion, including the three-dimensional finite difference method, discrete wave-number method, indirect boundary element method, modal summation method, ray theory, 2.5-D discrete wave number-boundary integral equation method, 2.5-D pseudospectral method, and the 2.5-D finite difference method (Erdik and Durukal, 2003). Essentially, all of these deterministic methods convolve the source function with synthetic Green’s functions to produce the motion at ground surface (Erdik and Durukal, 2003). Recorded accelerograms are ground-motion time-histories obtained from real seismic events in the past. The rapid development of digital seismic networks worldwide and the availability of strong-motion databases have increased the accessibility to recorded accelerograms. Strong-motion recordings and a summary of recorded strong-motion parameters can be obtained from a variety of sources, a list of these databases being available in Bozorgnia and Campbell, 2004. Most engineers feel more comfortable using recorded time histories than artificial/simulated time histories because recorded time histories represent real events, and because they do not feel familiar enough with the methods used to generate artificial/simulated time histories to have confidence in their reliability (Stewart et al., 2001). Typically, the time series are selected from databases with recorded ground motions with similar magnitudes and similar distances to the design earthquake. Other factors such as the site condition, fault type, and spectral content may also be considered (Watson-Lamprey and Abrahamson, 2006; Stewart et al., 2001). It is generally not possible to obtain the magnitude-distance pairs directly from 36

code specifications, as the constant hazard response spectra in seismic design codes are obtained from a probabilistic seismic hazard analysis, representing the aggregated contribution of a range of earthquake magnitudes occurring at various rates on each of several seismic sources located at various distances from the site (Stewart et al., 2001). Therefore, before selecting earthquake records, the magnitude, distance, site conditions, and other parameters that control the ground motion characteristics are obtained by de-aggregation of the seismic hazard (Stewart et al., 2001). As the accelerograms selected from databases based on specific criteria (magnitude, distance, site conditions, fault type) will not fit exactly the target spectrum, the records are scaled so that average response spectrum of the suite of ground motion would fit the target response spectrum within a period range representative for the analysed structure. Though the idea of scaling a ground motion is not theoretically correct from the seismological point of view, it is used as a standard practice in order to reduce the scatter in the seismic response of the analysed structure, by having a rational fit of the average response spectrum of records and the target spectrum. There is a wealth of different scaling procedures available in the literature. The simplest procedure represents scaling individual accelerograms so that the area under their response spectra is equal to the one under the target response spectrum, within a prescribed period range. The most “intrusive” are the modification of frequency content of the ground motion, by generating semi-artificial accelerograms. Iervolino et al. (2008) investigated the possibility to find suites of unscaled recorded accelerograms from the European Strong-Motion Database complying with the Eurocode 8 spectra, while accounting for additional constraints believed to matter in the seismic assessment of buildings (e.g. magnitude). It was possible to find suites of 7 unscaled recorded accelerograms, for site categories A, B and C only, and for low to moderate peak ground accelerations (ag = 0.15 g and 0.25 g). For soft soil conditions (site categories D and E) and for strong ground motion (ag = 0.35 g) matching to the Eurocode 8 spectra required scaling. Naiem et al. (2004) investigated an approach for selection of a suite of recorded earthquake ground motions that in combination match a given site-specific design spectrum with minimum alteration. The proposed method is capable of searching a set consisting of thousands of earthquake records and recommending a desired subset of records that match the target design spectrum, by using a genetic algorithm, which treats the union of 7 records and corresponding scaling factors as a single “individual”. Watson-Lamprey and Abrahamson (2006) suggested a procedure to select time series for use in non-linear analyses that are intended to result in an average response of the non-linear system. A simple model of a yielding system is used as a proxy for the non-linear behaviour of a more complicated yielding system. Beyer and Bommer (2007) reviewed code provisions regarding selection and scaling of ground motions for bi-directional analysis (of spatial structures) and showed that the guidelines provisions are frequently inconsistent or are lacking transparency regarding the underlying assumptions. Several issues involved when selecting and scaling records for bi-directional analysis and post-processing results of such analyses were discussed. 1.2.2.5 Simulation of accelerograms for fuzzy analysis of structures Fuzzy stochastic tools for structural analysis and reliability assessment were investigated by Sickert et al. (2008) and Mistakidis et al. (2007) within the activities of WG2 of COST Action C26. The uncertain character of both earthquake recording and structural system behaviour was considered within a response history procedure. 1.2.2.5.1 Uncertainty of seismic action The earthquake excitation can be described with the aid of artificially generated, site specific acceleration-time-dependencies. These dependencies may be generated in a way that they are compatible to a prescribed target spectrum. Additional information can be obtained considering all phases of the earthquake process, from origin and propagation until the transmission from underground to the structure. However, relevant seismic centres in the environment and the distance dependent decrement of the spectral amplitude are uncertain. Geological conditions, (e.g. stratigrahic sequence, intensification, and damping effects) and the registration of cyclic characters (e.g. strong earthquake duration) contain further uncertainty. 37

Figure 14.

Flowchart of generation algorithm.

The uncertainty of the earthquake excitation results from aleatory and epistemic sources. Because of the different characteristic of aleatory and epistemic uncertainty adequate set theoretical uncertainty models have to be utilized for quantification. On one hand, random fluctuations may be described with probabilistic methods which represent the traditional uncertainty model. On the other hand non-traditional uncertainty models are more suitable for the modelling of the imprecision because they reflect the available rare information adequately. The developed generalized uncertainty model should provide the possibility to distinguish the sources of uncertainty within the assessed structural responses. 1.2.2.5.2 Fuzzy random processes The uncertainty model fuzzy randomness fulfils the above suggested requirements. It represents a marriage between fuzzy methods describing the imprecision and traditional probabilistic methods and contains fuzziness and randomness as special cases. Thus, it may be understood as a generalized uncertainty model and describes imprecise probabilities as fuzzy sets of probability measures. The associated fuzzy probabilities represent weighted bounds of probability. Utilizing the model fuzzy randomness, earthquake excitations are described as fuzzy random processes which represent a fuzzy set of real valued random processes. Then, structures are loaded by fuzzy random acceleration ˜ processes a˜ (τ), fuzzy random velocity processes v˜ (τ), or fuzzy random displacement processes d(τ). 1.2.2.5.3 Generation of fuzzy accelerogams The algorithm presented in Figure 14 starts with a white noise process which is modified in the frequency domain by means of fuzzy filter functions. As example, two filters are formulated in dependency of fuzzy parameters. The high-pass filter according to:

can be used for eliminating unwanted long-period components. A meaningful choice of the fuzzy corner frequency ω˜ H is the fuzzy triangular number ‹0.9, 1.0, 1.1›rad/s. The more important step is the filtering by the Kanai-Tajimi-low-pass filter according to

The fuzzy parameters ω˜ 0 and ξ˜0 can be regarded as the natural circular frequency and the damping of the soil, respectively. After back transformation to the time domain a fuzzy intensity function is applied as shown in Figure 15. The approach results in a fuzzy random acceleration process according to Figure 16. 1.2.2.6 Scenarios based earthquake hazard assessment In a study performed within the COST Action C26, Romanelli et al., (2010) questioned the applicability of standard probabilistic estimates of seismic hazard (PSHA) for historical and strategic buildings, when considering time intervals of about a million year. An alternative capable of 38

Figure 15.

Fuzzy intensity function.

Figure 16.

Example of a fuzzy random acceleration processes.

minimizing the drawbacks of PSHA is represented by scenario earthquakes, also named neodeterministic seismic hazard assessment (NDSHA), characterized at least in terms of magnitude, distance and faulting style, and by the treatment of complex source processes. NDSHA approach is a hybrid method consisting of modal summation and finite difference methods (Panza et al., 2001). Modal summation is applied along the bedrock (1D) model that represents the average path between the assumed source and the local, laterally heterogeneous (2D) structure beneath the area of interest. These signals are numerically propagated through the laterally varying local structure by the finite difference method. Synthetic seismograms of the vertical, transverse and radial components of ground motion are computed at a predefined set of points at the surface. The NDSHA naturally supplies realistic time series of ground motion, which represent also reliable estimates of ground displacement readily applicable to seismic isolation techniques, useful to preserve historical monuments and relevant man made structures. An integrated neo-deterministic approach to seismic hazard assessment has been developed that combines different pattern recognition techniques, designed for the space-time identification of strong earthquakes, with the procedure for the NDSHA (Romanelli et al., 2010). The integrated approach allows for a time dependent definition of the seismic input (realistic synthetic seismograms), through the routine updating of earthquake predictions. The time information given by the intermediate-term medium-range earthquake prediction is very useful to plan preparedness and rescue actions and to define priority criteria for the investigations required by the seismic microzonation. Even if strong motion records in near-fault, soft soil, or basin conditions have been recently obtained, their number is still very limited to be statistically significant for seismic engineering applications: the realistic computation of the seismic input, taking in account the source and site effects, combined with the evaluation of the seismic response of buildings provides an effective approach to the assessment of seismic risk. The NDSHA was recently used to compute ground motions for the 6 April 2009 earthquake (Mw = 6.3) in L’Aquila, Italy (Nunziata et al., 2010). 1.2.3 RECOMMENDATIONS FOR FURTHER DEVELOPMENT Characterization of seismic action is a challenging task due to its inherent high uncertainty. A combination of source characteristics, travel path effects, local site effects and soil-structure interaction contributes to the complex nature of the seismic action at the building site. 39

There seems to be a gap between the existing knowledge on characterization of seismic action and the code provisions. Consider for example directivity effects in near-fault regions and soft soil conditions. Both can generate ground motions with long period pulse-type form. The acceleration response spectrum of this type of motions is characterized by a large value of the control period TC (limiting value between the constant acceleration and constant velocity region of the spectrum). While modern design codes generally recognize this effect in the case of soft soil conditions, it is not considered in the case of near-fault ground motions. A further issue that requires attention and further research is influence of near-fault and soft-soil ground motions on seismic performance of structures with fundamental period of vibration lower than the TC control period of the ground motion. Earthquake force reduction factors valid for standard ground motions may be inappropriate in this cases. As nonlinear dynamic analysis of structures is increasingly used in design of new structures and evaluation of existing ones, time-history representation of seismic action becomes more important. Current code provisions for selection of earthquake records for time-history analysis of structures are of little help in obtaining a set of ground motions for design. Sometimes codes use contradictory criteria for selection of acceleration time-histories. On one hand, time-histories should be compatible with the characteristics of the seismic source, travel path and site effects. Whether records matching these criteria are selected from historical earthquakes, obtained using artificial generation or through simulation, expertise in seismology is required. Few structural engineers would have it. On the other hand, codes require that response spectra of the earthquake records match the target elastic response spectrum, in order to have compatibility between the two alternative representations of the seismic action (response spectrum and time-history), and to reduce the scatter in analysis results. However, elastic code spectra are “envelopes” of response spectra from real seismic records. Therefore, in many practical applications it is quite difficult, or even impossible to find a set of recorded time-histories that would match the elastic code spectra over a wide range of periods. Often the matching requires scaling of records which alter the “seismological” compatibility. Seismic engineering is multidisciplinary in its nature. Often there is a lack of cooperation between the seismological and engineering communities. A variety of information on characterisation, selection, generation and simulation of seismic action is available in literature, but most of it requires specialised knowledge, being developed and understood by seismologists only. A close collaboration between seismologists and structural engineers is needed to advance the current state of practice in structural analysis under seismic action. As a practical aid for seismic performance evaluation of structures using time-history analysis, ground motion suites can be developed for different seismic zones and site conditions within a country, and made available to structural engineers.

REFERENCES Alavi, B., and Krawinkler, H., 2000. Consideration of Near-Fault Ground Motion Effects in Seismic Design. Proc. of the 12th World Conference on Earthquake Engineering (12WCEE), Auckland, New Zealand, Jan. 30 – Feb. 4, paper 2665. Athanasopoulos, G.A., Pelekis, P.C., Leonidou, E.A., 1998. Effects of surface topography and soil conditions on the seismic ground response – including liquefaction – in the Egion (Greece) 15/6/1995 earthquake. 11th European Conference on Earthquake Engineering, Balkema, Rotterdam. Beyer, K. and Bommer, J.J., 2007. Selection and Scaling of Real Accelerograms for Bi-Directional Loading: A Review of Current Practice and Code Provisions. Journal of Earthquake Engineering, 11:1, 13–45. Bozorgnia, Y. and Campbell, K.W., 2004. Engineering Characterization of Ground Motion. Earthquake Engineering. From Engineering Seismology to Performance-Based Engineering, Yousef Bozorgnia and Vitelmo V. Bertero (ed.), CRC Press. Building Seismic Safety Council, BSSC, 2001. NEHRP Recommended Provisions for Seismic Regulations for New Buildings and Other Structures, Part 1 — Provisions and Part 2 — Commentary. Federal Emergency Management Agency, Washington D.C. Chang, H-Y. and Kawakami, H., 2006. Effects of ground motion parameters and cyclic degradation behaviour on collapse response of steel moment-resisting frames. Journal of Structural Engineering, Vol. 132, no. 10, pages 2553–1562.

40

Dall’Ara, A., Lai, C.G. and Strobbia C., 2006. Selection of spectrum-compatible real accelerograms for seismic response analyses of soil deposits. First European Conference on Earthquake Engineering and Seismology, Geneva, Switzerland, 3–8 September 2006. Paper number: 1240. EN 1998-1, 2004. Eurocode 8: Design of structures for earthquake resistance – Part 1: General rules, seismic actions and rules for buildings. CEN – European Committee for Standardization. Erdik, M. and Durukal, E., 2003. Simulation Modeling of Strong Ground Motion. Earthquake Engineering Handbook, W.F. Chen and Charles Scawthorn (ed.), CRC Press. FEMA 356, 2000. Prestandard and commentary for the seismic rehabilitation of buildings. American Society of Civil Engineers, Reston, Virginia for Federal Emergency Management Agency, Washington, D.C. FEMA 450, 2003. NEHRP recommended provisions for seismic regulations for new buildings and other structures (FEMA 450). Building Seismic Safety Council, National Institute of Building Sciences, Washington, D.C. for Federal Emergency Management Agency, Washington, D.C. Gioncu, V. and Mazzolani, F.M., 2002. Ductility of Seismic Resistant Steel Structures. Spon Press, London and New York. Gioncu, V., and Mazzolani, F.M. 2010. Problems of seismic behaviour of buildings situated in urban habitats. Proc. of the Int. Conference “Urban Habitat Constructions under Catastrophic Events”, Naples, 16-18 September 2010, COST Action C26. Iervolino, I., Maddaloni, G. and Cosenza, E., 2008. Eurocode 8 Compliant Real Record Sets for Seismic Analysis of Structures, Journal of Earthquake Engineering, 12:1, 54–90 Lungu, D., Arion, C., Calarasu, E. 2008a. Bucharest soil conditions and input ground motion for the structural performance analysis, Proc. of the Int. Symposium “Urban Habitat Constructions under Catastrophic Events”, Malta, 22–23 Oct., COST Action C26, Editors: Mazzolani, Mistakidis, Borg, Byfield, De Matteis, Dubina, Indirli, Mandara, Muzeau, Wald, Wang, p. 135–140. Lungu, D., Vacareanu, R., Aldea, A. 2008b. Seismicity of Vrancea subcrustal source and corresponding seismic instrumentation. Proc. of the Int. Symposium “Urban Habitat Constructions under Catastrophic Events”, Malta, 22–23 Oct., COST Action C26, Editors: Mazzolani, Mistakidis, Borg, Byfield, De Matteis, Dubina, Indirli, Mandara, Muzeau, Wald, Wang, p. 129–134. Lungu, D., Arion, C., Aldea, A., Demetriu, S., 1999, “Assessment of seismic hazard in Romania based on 25 years of strong ground motion instrumentation”, NATO ARW Conference on strong motion instrumentation for civil engineering structures, Istanbul, Turkey, June 2–5, 1999. Mistakidis, E., Apostolska, R., Dubina, D., Graf, W., Ne-cevska-Cvetanovska, G., Nogueiro, P., Pannier, S., Sickert, J.-U., Simões da Silva, L., Stratan, A., Terzic, U. 2007. Typology of seismic motion and seismic engineering design. Proc. of Workshop in Prague 30-31/3/2007 “Urban habitat constructions under catastrophic events”. COST Action C26. Ed. Wald, F., Mazzolani, F., Byfield, M., Dubina, D., Faber, M., p. 130–157. Naeim, F., Alimoradi A. and Pezeshk, S., 2004. Selection and Scaling of Ground Motion Time Histories for Structural Design Using Genetic Algorithms. Earthquake Spectra, Volume 20, No. 2, pages 413–426. Nunziata, C., Vaccari, F., Panza, G.F. (2010). The Mw 6.3, 2009 L’Aquila earthquake: linear and nonlinear site effects. Proc. of the Int. Conference “Urban Habitat Constructions under Catastrophic Events”, Naples, 16–18 September 2010, COST Action C26. Panza, G.F., Romanelli, F., Vaccari, F. 2001. Seismic wave propagation in laterally heterogeneus anelastic media: theory and applications to seismic zonation. Advances in Geophysics, vol. 43. Academic Press, 1–95. Priestley, N., 2006. Seismological information for displacement-based design – a structural engineer’s wish list. First European Conference on Earthquake Engineering and Seismology, Geneva, Switzerland, 3–8 September 2006. Paper number: Keynote Address K1a Romanelli, F., Peresan, A., Vaccari F. & Panza, G.F. 2010. Scenarios based earthquake hazard assessment. Proc. of the Int. Conference “Urban Habitat Constructions under Catastrophic Events”, Naples, 16–18 September 2010, COST Action C26. Sasani, M. and Bertero, V.V., 2000. Importance of severe pulse-type ground motions in performance-based engineering: historical and critical review. Proc. 12th World Conference on Earthquake Engineering (12WCEE), Auckland, New Zealand, Jan. 30 – Feb. 4, paper 1302. Sickert, J.-U., Kaliske, M., Graf, W. 2008. Fuzzy stochastic earthquake analysis of structures. Proc. of the Int. Symposium “Urban Habitat Constructions under Catastrophic Events”, Malta, 22-23 Oct., COST Action C26, Editors: Mazzolani, Mistakidis, Borg, Byfield, De Matteis, Dubina, Indirli, Mandara, Muzeau, Wald, Wang, p. 141–145. Stewart, J.P., Chiou, S-J., Bray, J.D., Graves, R.W., Somerville, P.G., Abrahamson, N.A. 2001. Ground Motion Evaluation Procedures for Performance-Based Design. PEER Report 2001/09, Pacific Earthquake Engineering Research Center, College of Engineering, University of California, Berkeley.

41

Stratan, A. and Dubina, D. 2008. Selection of time-history records for dynamic analysis of structures, Proc. of the Int. Symposium “Urban Habitat Constructions under Catastrophic Events”, Malta, 22-23 Oct., COST Action C26, Editors: Mazzolani, Mistakidis, Borg, Byfield, De Matteis, Dubina, Indirli, Mandara, Muzeau, Wald, Wang, p. 123–128. Watson-Lamprey, J. and Abrahamson, N., 2006. Selection of ground motion time series and limits on scaling. Soil Dynamics and Earthquake Engineering 26 (2006) 477–482. Wibowo, A., Kafle, B., Kermani, A.M., Lam, N.T.K., Wilson, J.L., Gad, E.F., 2008. Damage in the 2008 China Earthquake, http://www.aees.org.au/

42

Urban Habitat Constructions under Catastrophic Events (Final Report) – Mazzolani (Ed). © 2010 Taylor & Francis Group, London, ISBN 978-0-415-60686-8

1.3 Characterization of catastrophic actions on constructions: Explosive loads A. Tyas Department of Civil & Structural Engineering, University of Sheffield, Sheffield, UK

1.3.1 HIGH-EXPLOSIVE BLAST IN AIR The willingness of terrorists to attack “soft” civilian targets in urban locations has meant that, increasingly, blast loading from the detonation of high explosive charges must be considered by the designers of civilian structures. The prediction and characterization of such loads was, until relatively recently, the domain of specialist military designers of weapons and defensive structures. This chapter considers the approaches that civilian analysts can take to help predict and quantify loading from blast events. 1.3.2 HOPKINSON CRANZ SCALING In the early part of the 20th century, Hopkinson (1915) and Cranz (1926) independently determined that the parameters of blast waves produced by detonations of different masses of explosive at different distances from a target could be related by the concept of scaled distance (Z),

where S = distance from detonation point to a target and W = the mass of the explosive charge. In general, the peak pressure experienced by targets exposed blast waves generated by detonations of different masses of explosive will be the same at the same scaled distance, provided that the chemical composition of the explosive and the geometrical form of the two events are identical. Time-based parameters of the loading (duration of blast wave, delay between detonation and arrival at the target, impulse under the load-time curve) have a similar relation, but vary with the linear scale factor of the event. Since the mass of the explosive charge is the cube of the linear scale factor, it is convenient to use W 1/3 as the scale factor. Hence, for example, the relative duration or impulse associated with the detonation of charges with mass W1 and W2 will be in the ratio (W1 /W2 )1/3 . The form of a high explosive blast wave loading pulse on a target, in the absence of obstacles or reflections takes the typical “Friedlander” form (Friedlander 1946) shown in Figure 1, in which pmax is the peak overpressure (pressure above atmospheric), ta is the arrival time of the wave at a target, td is the positive duration and i is the positive impulse per unit area. The loading comprises a near-instantaneous rise in pressure (effectively a shock front) followed by an exponential decay. The form of this positive phase of the blast wave for the period ta < t < ta + td is given by the Friedlander equation:

where α is a decay coefficient. The propagation of the blast wave through air results in air particles being displaced radially away from the detonation point; consequently, at the end of the positive 43

Figure 1.

Friedlander blast wave.

loading phase, a partial vacuum exists, resulting in a period of negative pressure until equilibrium is re-established. This negative phase is frequently ignored in analysis of blast effects, although it can result in a large reduction in the overall applied impulse. Methods of determining the loading parameters associated with the interaction of a blast wave on a structural target can be broadly grouped into four areas: • Closed form, theoretical predictions • Empirical or semi-empirical predictions • Physically based numerical models 1.3.3 THEORETICAL PREDICTIONS OF BLAST LOAD PARAMETERS Closed form methods essentially follow the Rankine-Hugoniot shock analysis method, solving the equations of conservation of mass, momentum and energy across the shock front. Given an equation of state for air (usually from the assumption of ideal gas behavior, with), the peak blast pressure can be stated as a function of the velocity of the shock front, or “Mach” number, M and ambient atmospheric pressure pa . Taking the specific heat ratio of air to be 1.4, the shock analysis gives:

The pressure given in (3) is the peak pressure associated with the compression of the air particles by the shock front. This is the pressure that would be experienced by a target parallel to the direction of propagation of the blast wave, and hence is often called the incident, or “side-on” pressure. However, if the target impedes the flow of the blast wave, a reflection of the shock occurs. This produces a significant increase in the pressure experienced by the target, due to the shock itself being reflected, and a proportion of the kinetic energy of the displaced air particles being converted to pressure energy as they impact upon the target. If the target is normal to the direction of propagation of the blast wave, the ratio of this reflected pressure, pr to the side-on pressure, ps is given by:

44

Figure 2.

Friedlander blast wave parameters for a spherical free-field explosion.

Equation (4) implies that the normally reflected pressures is always at least double the magnitude of the side-on pressure, and the ratio can be as high as eight. In fact, deviation from ideal gas behavior for very intense shocks means that the ratio can be significantly higher than eight.

1.3.4 EMPIRICAL AND SEMI-EMPIRICAL PREDICTIONS OF BLAST LOAD PARAMETERS Whilst this type of analysis is relatively simple and can work well for ideal scenarios (a similar approach was famously used by G.I Taylor to estimate the yield of the first atomic bomb detonations), it has serious limitations in practice. For example, the Mach number at any distance from a given detonation cannot easily be determined from theoretical considerations. An alternative approach is therefore to collate experimental records from numerous blast trials and develop empirical relationships between the blast wave parameters and scaled distance. Since the energetic output of explosives vary with their chemical composition, the mass of the explosive charge used in determining the scaled distance in (1) is given in terms of “equivalent mass” of a standard explosive, typically TNT, by multiplying the actual mass by an empirically derived equivalence factor. Thus, data from a large number of disparate experimental trials can be combined into a single load prediction relation. Such data are given in the US Tri-Forces design guide TM5-1300(1991), and the computer code ConWep (1991) for a limited number of geometrical arrangements of explosive charge (spherical free-field blast and hemispherical ground-burst explosion). An example of the type of prediction curve given by these approaches is shown in Figure 2, for a spherical blast wave from a free-field explosion striking normally onto a rigid target. In Figure 2, the scaled distance is in terms of kgTNT and the actual values of time and impulse can be found by factoring the scaled values from the chart by W 1/3 . TM5-1300 also gives similar relations for negative phase parameters and coefficients to factor the reflected pressure in cases of oblique shock-structure interaction (examples are given in Gebekken & Doge 2010). These approaches work very well for simple geometrical layouts. For example, Figure 3 shows a comparison between an laboratory experimental record from the University of Sheffield Blast & Impact laboratory, and ConWep prediction for a normally reflected blast pressure loading on a rigid target 1.010 m from a 70 g C4 explosive hemispherical ground-burst detonation. Predictive relationships are also given in TM5-1300 & ConWep for the development of the gas pressure, or “quasi-static pressure” (QSP) generated when a detonation occurs inside a vented structure. In such a case, a series of reflections of the initial blast wave from the walls, floor and roof produces a train of shock loads at a particular location within the building. The precise magnitude and timing of the reflections will vary according to on the size and geometrical complexity of the internal volume as well as the size, type and position of the explosive. However, the general trend will be for the shocks to lose their identity and merge to an increased general background 45

Figure 3. Experimental record and ConWep prediction for 70 g C4 hemispherical ground burst charge reflecting normally from a rigid target 1.010 m away.

Figure 4.

Quasi-static pressure in a vented internal explosion.

pressure, which dissipates through the vent opening(s). Figure 4 shows data from a University of Sheffield test on a 50 g C4 charge inside a vented 1.5 m cubical structural; the general trend of the experimental QSP agrees reasonably well with a ConWep prediction for the same event. Whilst these well-established empirical approaches work well for simple geometrical problems, their predictions must be treated with caution when they are applied to more realistic scenarios. Two examples have been highlighted by bespoke experimental work in recent years. Firstly, the TM5-1300 & ConWep predictions of reflected pressure parameters apply only to the case where the reflecting surface has large lateral dimensions. If the target is relatively small 46

Figure 5.

Example of diffraction wave or “clearing” effect on reflected pressure on a finite sized target.

compared to the physical length of the incoming pressure pulse, diffraction waves from the free edges will propagate along the target face and lead to premature reduction of the reflected pressure. Figure 5 shows an example of this “clearing” effect at the centre of a small scale test target, 700 mm square, exposed to a blast wave from a 250 g C4 hemispherical groundburst at 8m distance in a test conducted at University of Sheffield. Whilst both ConWep and TM5-1300 purport to allow for this effect, recent experimental evidence by Rickmann & Murrell (2007) casts strong doubt on the accuracy of the predictions. This article, suggests that the predictions may over-predict the impulse on a target by up to 60%. Rickman & Murrell and Smith et al. (1999) have proposed alternative predictive relationships for the clearing effect. The second major drawback of simple empirical predictive methods is their inability to allow for effects such as shielding of blast loading by obstacles between the detonation and the target, and the converse effect of reflections from adjacent structures preventing geometrical dispersion of the blast wave’s energy, and instead channeling or focusing the blast onto a target. This is a particular potential problem in predicting blast loading in complex urban environments. This theme has been studied extensively at Cranfield University, UK (e.g. Rose & Smith 2002, and Rose et al. 2006). Their experimental and numerical results have shown that the shielding and focusing effects can lead to magnitudes of blast loading parameters that are significantly higher or lower than those given by simple predictive methods. Furthermore, in scaled model studies of blast loads along city streets, the Cranfield researchers have shown that the most intense blast loading may not be experienced by the nearest targets, as reflections can focus additional loads onto more distant targets (Smith et al. 2001). Clearly the wide extensive range of different possible shielding and channeling scenarios makes the development of comprehensive empirical predictive tools difficult if not impossible. Nevertheless, there is scope for general indicative guidance, of the sort provided by Rose & Smith (2001) who used numerical modeling to determine the variations of blast pressure parameters along straight city streets.

1.3.5 NUMERICAL MODEL PREDICTIONS OF BLAST LOAD PARAMETERS This brings us neatly to the third and final type of predictive method; physically-based numerical modeling. Whilst computational numerical solutions of the shock equations were first conducted 47

Figure 6. Comparison of numerical model predictions and experimental pressure readings for complex internal detonations (Tyas 2007).

by the mathematician John von Neumann as part of the Manhattan Project in World War II, it is only in the last 10–15 years that improvements in computational power and the robustness and stability of numerical modeling codes have resulted in the migration of this method of blast load prediction from the academic or specialist military field to general use. Typical numerical modeling of detonation and blast wave propagation before interaction with a target uses an explicit time-stepping numerical integration scheme to solve the compressible Euler equations relating the energy, density, pressure and flow velocity of the air. Explosives can be modeled using an equation of state to model the energy release at detonation. Models for the air typically assume ideal gas behavior, which may limit the accuracy of the predictions very close to the detonation, but generally give good results at scaled distances >0.5–1 m/kg1/3 (see, for example, Pope & Tyas 2002). Many commercial and academic research tools use this approach; among the more common ones are Ansys Autodyn (2010), LS-Dyna (2007), AIR3D (Rose et al. 2006) and EUROPLEXUS (2010). The Eulerian discretization scheme means that the computational grid remains fixed and the air material flows through it. To determine the loading parameters when the blast wave strikes an effective rigid target (or at least a target whose inertia prevents it from deforming significantly during the loading time, but responds more slowly), the target may be modeled by a rigid surface. Spatial and temporal variations of load may then be determined from the model, and these used for a subsequent analysis of the structural response. Tyas (2007 & 2008) Presented examples of numerical predictions and experimental validation of blast wave loading from a detonation inside a partially confined structure with complex internal geometry (Figure 6), and of the prediction of the interaction between a blast wave from a hemispherical groundburst detonation with a rigid target (Figures 7 & 8). If the target is less massive, significant deformations may occur during the loading phase, and these may affect the load experienced by the structure. Gebbeken and Toge (2010) and Teich et al. (2009) and Teich et al. (2010) show that this lack of rigidity may significantly reduce the intensity of the reflected pressure pulse; they suggest that designers may deliberately use this effect to their benefit in the design of, for example, large span, lightweight glazed façades. In computational numerical analysis of blast-structure interaction, the effect of non-rigidity of a target means that the reflecting surface is continually changing its geometry. The problem then essentially becomes one of coupled Fluid-Structure interaction. Since structural deformations are most effectively and efficiently modeled using a Lagrange discretization approach, in which 48

Figure 7. Numerical model pressure contours and experimental high-speed video shadography of blast wave interaction with barrier (Tyas 2008).

Figure 8. Comparison of numerical predictions and experimental readings for blast pressure on face of target in Figure 7 (Tyas 2008).

the structural material is fixed to the numerical grid, the interface between the (non-deforming) Euler grid and the (deforming) Lagrange grid must be continually updated. In blast modeling, this is achieved by Arbitrary Lagrange-Euler (ALE) interaction, where the Euler grid is continually re-mapped onto the deforming Lagrange surface. This approach is used, for example, by Shi et al. (2007) in their study of blast loading on unclad columns and develop simplified prediction methods for the blast loads imposed on the columns. Work presented as part of the COST Action C26, includes studies by researchers at the Joint Reseacrh Centre Ispra, on predicting the effect of an internal detonation in a railway station (Solomos 2010). Figure 9 shows an example of the a blast load sufficiently intense to cause structural elements to deform so highly that they become detached from adjacent members. In this case, a robust ALE formulation can still, in principle, re-map the Euler grid onto the failed elements, and thus allow the blast load to continue to accelerate the elements. Correct modeling of the fragment break-up and the momentum transfer from the blast to the fragments is important in scenarios where secondary fragmentation from such failed elements may present a projectile hazard to personnel or other structural elements, particularly since such high momentum fragments typically present a danger over a much greater radius than the blast wave itself. Teich et al. (2010) used and ALE formulation to predict the response to blast loading of large curtain wall glazed façades with rigid and flexible supports. Their results give convincing evidence that support flexibility can play an important role in reducing the overall glazing damage, by “softening” the reflection of the blast wave (Figure 10), and hence reduce injuries inside a building from glazing fragments, a major cause of serious injury in urban explosion events. In addition to adapting the grid to follow the Lagrange surface, ALE methods can also be used to automatically track the propagation of the air shock, and compress the grid in areas of high spatial 49

Figure 9. EUROPLEXUS ALE model of blast inside a railway terminal building, showing gross failure of cladding panels (Solomos 2010).

Figure 10. ALE model of blast load on double glazed façade, showing the influence of support flexibility on loading, and hence, damage (Teich et al. 2010). (a) – stiff support. (b) – linear elastic soft support.

or temporal transients, in order to better capture the shock. This approach is computationally more efficient than simply making the grid size very small across the whole analysis domain, although, as Kwasniewski et al. (2010) show, care must be taken in choosing a suitable initial grid size. 1.3.6 VALIDATION OF PREDICTIVE METHODS AND STANDARDIZATION OF DESIGN BLAST LOADS The numerical modeling approaches described in the previous section have rapidly moved from the field of specialist research to mainstream practice over the last decade or so. Powerful and easy-tofollow user interfaces mean that even untrained users can quickly set up numerical models, without 50

Table 1. Blast classification scenarios and associated blast pressure loading parameters for testing of security glazing. From ISO 16934 (2007). Minimum values

Approx. equivalent threat scenario

Threat Classification

pr(max) (kPa)

ir (kPa.ms)

W (kgTNT )

S (m)

ER30 ER50 ER70 ER100 ER150 ER200

30 50 70 100 150 200

170 370 550 900 1500 2200

30 100 160 500 1000 2000 2500

33 34 33 39 41 46 or 49

necessarily understanding the effect of their choices of parameters. As with any numerical tool, the output can be highly sensitive to input parameters and choices of analysis approach; experienced users realize that the field of numerical blast analysis is as much an art as a rigorous science. For the inexperienced user, there is the ever-present danger of confusing computing power and precision with accuracy. Kwasniewski et al. 2010 show how the development of every stage of a modeling strategy should, ideally, be validated against high quality experimental data. This approach is routinely followed by defense engineers, who typically demand experimental validation of numerical model results of both blast loads and structural deformation, no doubt because of the high probability of their predictions being put to the test in reality. However, as Tyas (2008) notes, well controlled validation testing, particularly at large scale, is highly expensive. It is entirely unrealistic to expect civilian analysts routinely to commission bespoke experimental validation tests. This leaves the question of how civilian designers are to be able to access validation data. Whilst numerous experimental studies of blast load measurement have been published, there is no commonly agreed body of standard test methodologies or results. A step towards this is given in ISO 16933 (2007) & ISO 16934 (2007), recently published protocols for blast testing of security glazing. Both these standards present lists of typical explosion hazards and the associated peak reflected pressures, positive durations and positive impulses. These values are shown in Table 1. The correlation between the blast load parameters and the equivalent threat scenarios in Table 1 are based were determined from a comprehensive series of blast trials conducted in the UK, France and Germany, numerical models using Air 3D and predictions from ConWep. The threat scenarios are based on the assumption of hemi-spherical pressure waves from a surface detonation impinging normally on a target, with no shielding or focusing effects. They are thus simplistic to some extent, but provide a simple set of test criteria for the assessment of the effectiveness of security glazing. In the absence of other, more detailed prediction methods, these data also represent a useful set of approximate design load values for blast loading at moderately large distance from a range of likely terrorist bombing events. This approach suggests a direction for future work in this area. Instead of bespoke experimental testing to validate particular numerical models, there is a requirement for a series of benchmark experimental trials that capture the significant physical processes of different blast loading events. Such benchmark trials should be designed in collaboration with numerical modeling experts and conducted independently at a number of different test sites and across a range of test scales. Data from these trials could then be used both for validation of new numerical modeling techniques and for training of numerical modelers. REFERENCES Ansys Autodyn 2D/3D. 2010. (http://www.ansys.com/products/explicit-dynamics/autodyn/default.asp ConWep: Conventional weapons effects program 1991. Prepared by DW Hyde, ERDC Vicksburg MS.

51

Cranz C. 1926. Lehrbuch der Ballistik, Springer-Verlag, Berlin. EUROPLEXUS. 2010. A Computer Program for the Finite Element Simulation of Fluid-Structure Systems under Transient Dynamic Loading – User Manual. Commissariat a l’Energie Atomique/Joint Research Centre. (http://europlexus.jrc.ec.europa.eu/public/manual_pdf/manual.pdf) Gebbeken, N. and Döge, T. 2010. Determination of Design Loads from Explosion and Impact Scenarios. In. Proc Urban Habitat Constructions under Catastrophic Events Cost C26 Conference, Naples, Italy. Hopkinson B. 1915. British Ordnance Board Minutes, 13565. ISO 16933. 2007. Glass in Building – Explosion Resistant Security Glazing, Test and Classification for Arena Air Blast Loading. ISO 16934. 2007. Glass in Building – Explosion Resistant Security Glazing, Test and Classification by Shock Tube Loading. Kwasniewski, L., Balcerzak, M. & Wojciechowski, J. 2010. A feasibility study on modelling blast loading using ALE formulation. In. Proc Urban Habitat Constructions under Catastrophic Events Cost C26 Conference, Naples, Italy. Teich, M., Gebbeken, N., Lööf, A. and van Doormaal, A. and 2010. Windows and Glazing Systems Exposed to Explosion Loads: Part 1 – Lethality and Hazard Assessment. In. Proc Urban Habitat Constructions under Catastrophic Events Cost C26 Conference, Naples, Italy. LS-DYNA. 2007. Keyword User’s Manual v.971. Livermore Software Technology Corporation (LSTC). (http://www.lstc.com/pdf/ls-dyna_971_manual_k.pdf) Pope D.J. and Tyas A. 2002 Use of hydrocode modelling techniques to predict loading parameters from free air hemispherical explosive charges. In Proc. 1st Asia-Pacific Conference on Protection of Structures against Hazards, Singapore. Rickman, D.D. & Murrell, D.W. 2007. Development of an Improved Methodology for Predicting Airblast Pressure Relief on a Directly Loaded Wall, ASME Journal of Pressure Vessel Technology 129: 195–204. Rose, T.A. and Smith, P.D. 2002. Influence of the principal geometrical parameters of straight city streets on positive and negative phase blast wave impulses. Int. J. Imp. Engng. 27: 359–376. Rose T.A., Smith P.D. and May J.H. 2006. The interaction of oblique blast waves with buildings. Shock Waves 16(1): 35–44. Shi, Y., Hao, H. and Zhong-Xian, L. 2007. Numerical simulation of blast wave interaction with structure columns, Shock Waves 17: 113–133. Smith, P.D., Rose, T.A. and Saotonglang, E. 1999. Clearing of blast waves from building facades”. Proc. Instn Civ. Engrs Structs & Bldgs 134: 193–199. Smith, P.D., Whelan, G.P., Feng, L.J. and Rose, T.A. 2001. Blast loading on buildings from explosions in city streets, Proc. ICE, Structures & Buildings 146(1): 47–55. Solomos, G. 2010. Physical vulnerability assessment of critical structures. Presentation at Cost C26 WG3 meeting, Nicosia, Cyprus. Teich, M., Gebbeken, N., Warnstedt, P. and Nehring, G. 2009. Interaction of Blast Waves with Flexible Structures. Presentation at Cost C26 WG3 meeting, Aveiro, Portugal. (http://www.civ.uth.gr/cost-c26/documents/ 15th%20meeting_Aveiro/WG3/Teich%202009%20-%20Interaction%20of%20Blast%20Waves%20with% 20Flexible%20Systems%20(online%20COST).pdf) TM5-1300: Design of Structures to Resist the Effects of Accidental Explosions 1991 US Department of Defense. Tyas, A. 2007. Blast and Impact Research at University of Sheffield. Presentation at Cost C26 WG3 meeting, Timisoara, Romania. (http://www.civ.uth.gr/cost-c26/documents/6th%20meeting_Timisoara/presentations/ WG3/2.Tyas.pdf) Tyas, A. 2008. Blast and Impact Research at University of Sheffield. Presentation at Cost C26 WG3 meeting, Valetta, Malta. (http://www.civ.uth.gr/cost-c26/documents/10th%20meeting_Malta/WG3/Cost%20C26% 20-%20Malta%20-%20AT.pdf)

52

Urban Habitat Constructions under Catastrophic Events (Final Report) – Mazzolani (Ed). © 2010 Taylor & Francis Group, London, ISBN 978-0-415-60686-8

1.4 Actions due to natural catastrophes, except volcanic eruptions T. Stathopoulos & I. Zisis Department of Building, Civil and Environmental Engineering, Concordia University, Montreal, QC, Canada

A. Talon & J.-P. Muzeau LaMI – Polytech’Clermont-Ferrand (CUST), Blaise Pascal University – Clermont-Ferrand, France

C. Coelho Department of Civil Engineering, University of Aveiro, Portugal

J.-P. Carlier Laboratoire de Mécanique de Lille, Université de Lille 1, France

S. Wolinski Rzeszow University of Technology, Rzeszow, Poland

1.4.1 INTRODUCTION In addition to the catastrophic actions imposed by fires, earthquakes and volcanoes, there are actions due to other causes, such as extreme winds, floods, snow and avalanches with tremendous consequences. For instance, Figure 1 shows clearly that the catastrophes due to floods and extreme winds are much more significant, particularly in the recent years, in comparison with those due to earthquakes and volcanic eruptions. This chapter will provide some insight into these additional actions originating from extreme winds, floods, snow and avalanches. 1.4.2 EXTREME WINDS 1.4.2.1 Wind-induced catastrophes Some of the most fatal and costly natural catastrophes that occurred in the past few decades were the result of extreme wind incidents. Hurricanes, typhoons and tornadoes were reported and tracked around the globe. The social and economic impact associated to these natural disasters initiated a drastic response from a number of political and academic institutions. Both structural and wind engineers invested a significant amount of effort to better understand the impact of such events on structures. Several studies were initiated after major hurricanes and had as main objective to assess the induced damage to both low-rise and high rise structures. The assessment revealed in most cases local damages for tall buildings (façade and non-structural elements failure – see Figure 2), whereas low-rise buildings have in many cases collapsed or torn apart even during moderate intensity cyclone events (Figure 3). 1.4.2.2 Damage assessment In most of the damage assessment studies there was an effort to relate recorded wind speeds to observed damage intensity. Mehta et al. (1983) examined the aftermath of Hurricane Frederic that impacted the coastal areas of Alabama, Mississippi and Florida. In addition to the wind speed – damage intensity correlation, this study also evaluated the performance of various buildings for cases where the measured wind speed was reaching the code suggested design speeds. The observations revealed that the damage was closely related to the type of construction. Fully engineered 53

Figure 1. Comparison of the number of major natural catastrophes from 1950 to 2005 classified by type of event (Münchener Rück, 2006).

Figure 2.

Cladding damage in Houston, Texas (Minor, 2005).

buildings had a superior performance compare to pre-engineered or marginally engineered buildings both in the structural (main frame) and non-structural (facade and roof sheathing) components. Similar observations were reported by Perry et al. (1990), using this time damage information reported after three major hurricanes, i.e. Camille, Frederic and Elena. The focus of this study was on metal buildings and the observed damages were grouped into nine main categories. In addition, special attention was given to the concept of the continuous load path and its importance during extreme wind events. The mitigation strategies proposed in this study were mainly focused towards the proper use, and in some cases improvement, of wind standards. Indeed, Davenport (1990) pointed out the requirements that could improve the wind resistance of structures which among others include the appropriate application of wind standards in practice. In this particular study, the damage assessment after hurricane Gilbert revealed that the main building component suffered from excessive damage was the roof of the structure. It was also reported that 25% of the houses examined had significant damage. The poor performance of gable roofs during strong wind events reported by several post-disaster studies was examined by Meecham et al. (1991). This was an experimental approach which indicated the improved performance of hip roofs compared to gable roofs. The aftermath of hurricane Andrew initiated several studies mainly due to the highest estimated damage cost that exceeded the amount of $40 billion. A very detailed damage assessment was 54

Figure 3.

Low-rise house damage during hurricane Charley (Munchener Ruck, 2006a-b).

performed by a team consisting of engineers, scientists, building-code officials and manufacturer representatives. Keith & Rose (1994) discussed in detail the findings of this effort and reported the various damage forms observed mainly in wood-based structures. In agreement to previous studies, the major failure reason was the improper connections and the loss of the gable-end walls. Moreover, the authors indicated that the increased economic losses were mainly attributed to the failure of the wall and roof sheathing, which allowed rain and other debris to enter the house. This type of failure was also reported by Sparks et al. (1994) and Crandell (1998). An important observation was related to tie-down straps, which, when used, reduced significantly the intensity of the damage. In agreement to Meecham’s et al, (1991) experimental findings, Crandell (1998) also reported the improved performance of hip roofs over gable roofs. Moreover, the damage assessment revealed that one-storey houses suffered from increased damage compared to two-storey houses. Cochran (2000) summarized the most common failure mechanisms during extreme wind events in both low-rise and high rise buildings. Once again, the roof failure of low-rise structures along with the wall glazing and doors are identified as the most favorable to initiate a major damage and eventually total failure. In addition, certain mitigation techniques and actions are discussed and proposed, mainly towards the proper use and application of wind standards and inspection during the construction by local authorities. One of the most devastating and costliest hurricanes of all times was Hurricane Katrina which had a tragic impact on the southern coast of the United States. The multi-billion damage was assessed by several studies reporting the observed types of damage on various structures. Eamon et al. (2007) surveyed three main types of structures i.e. residential buildings, commercial buildings and some major types of civil infrastructure. In addition, the reported damage was identified based on the type of construction for each of these three types of structures. As the authors reported, the most vulnerable structures were the low-rise wood buildings. The damage was excessive and followed the common failure mechanisms discussed by previous studies such as roof, sheathing and gable-end wall failures. Metal and R/C structures performed adequately as far as the main wind force resisting system is concerned but suffered of equal intensity facade damage. Damage findings related mainly to wood structures were also discussed by Van de Lindt et al. (2007). This study considered 27 case studies and examined structural and non-structural damages. The authors observed in several cases the lack of a continuous uplift load path which, as expected, resulted in significant structural damage. Moreover, loss of roof sheathing and gable end walls were for one more time the most common type of damage for low-rise buildings. 55

1.4.2.3 Hurricane Risk assessment Hurricane events have been closely examined not only from an engineering but from an economic point of view as well. Several studies focused on the risk assessment of such events trying to develop some form of relation between the observed wind speeds and the damage intensity. In a study conducted by Sparks et al. (1994) the damage observed during hurricanes Hugo and Andrew was correlated to certain wind speed levels. For instance, major damage was initiated for wind speeds over 40 m/s whereas loss of the building envelope was observed for wind speeds over 80 m/s. Other influence factors have been incorporated in the risk assessment process. Friedman (1984) discusses in detail a four-step approach of assessing the potential damage due to an extreme wind event. In addition to the structural vulnerability, factors such as local conditions, topography and geographical patterns of intensity of wind can influence the degree of damage. Construction practices are also identified as an important criterion for hurricane risk assessment by Berz & Smolka (1988). Khanduri & Morrow (2003) present a more detailed approach related to the generation of specific from generic vulnerability models, taking into consideration specific building characteristics (e.g. occupancy, height etc). 1.4.2.4 Field studies A very important and detailed study carried out on a full-scale low-rise structure was the Aylesbury project. This study started in 1970’s, following a series of full-scale wind pressure measurements on tall buildings, and the data collected were used and compared with wind tunnel studies for nearly three decades. Details of the full-scale experiments are presented by Eaton & Mayne (1975). It appears that the Aylesbury project was the forerunner for the wind pressure monitoring in full-scale low rise buildings. The Texas Tech University Project (TTU) initiated in the 1980’s and had as main objective to extract accurate field data that describe the wind distribution on the surface of a low rise building and define the appropriate terrain characteristics for proper simulation in boundary layer wind tunnels (Ng & Mehta, 1990; Levitan et al., 1990). The very interesting and unique characteristic of the facility was the unrestrained base which allowed the rotation of the house to any desired orientation. Several wind tunnel studies were conducted and verified their results to the available field data. Another important project, connecting full-scale with model scale studies, started in the Silsoe Research Institute (formerly AFRC Institute of Engineering Research) in the late 1980’s (Robertson & Glass, 1988). Although there were more than one instrumented buildings, the main test structure was the Silsoe experimental building located at the Silsoe Institute, Wrest Park, Silsoe, UK on a relatively open field site. Field data were used extensively for wind velocity and upstream terrain analysis as well as for comparison to several wind tunnel experimental studies. One more interesting study, started a decade ago, is the Southern Shores project. An existing house located in the town of South Shores (North Carolina) was selected for full-scale monitoring. The two-storey house has a relatively complex geometry and is instrumented with meteorological, pressure, strain and deformation transducers. The focus of this project was mainly to examine extreme wind effects, therefore the pressure equipment was selected based on a large dynamic range criterion. Details of the house and the instrumentation used can be found in Porterfield & Jones (2001). The field monitoring started in 1997 and during the first three years more than 8000 records were available, including three hurricane events (Bonnie, Dennis and Floyd). Caracoglia et al. (2008) presented results from the analysis of the above records and made comparisons with current wind provisions. In addition to the pressure coefficient detailed comparisons, the authors performed a simplified model analysis and compared stress and deformation findings with simulation results. The discrepancies of the later comparison were justified by the modeling simplifications, the definition of tributary areas and the averaging method of local pressures. The Florida Coastal Monitoring Program is a collaborative project, which started in 1998. The participants include Clemson University, University of Florida, Florida International University, Florida Institute of Technology and Institute for Business and Home Safety. The objectives of this project are to collect field wind speed and pressure data during extreme wind phenomena and compare them to scale model test results. The project makes use of six mobile tower systems capable to monitor detailed weather data (wind speed, direction, temperature, atmospheric pressure etc). In addition, a large number of occupied residences are instrumented with pressure monitoring 56

equipment. There thirty-two houses located in Florida and six in North and South Carolina. The location of the pressure taps was chosen based on expected higher suction regions on the roof. Details of both weather and pressure instrumentation can be found in Datin et al. (2006). The ongoing Load Paths Project (Doudak 2005; Zisis & Stathopoulos, 2009) was also a joint effort among several Canadian Universities. The construction of a unique full-scale test house and two weather towers provided important information related to wind characteristics, windinduced envelope pressures and foundation loads. In addition, the field data were compared to wind tunnel experimental findings and simplified numerical modeling results. The concepts of structural attenuation and wind load transfer mechanisms were discussed and results were presented (Zisis & Stathopoulos, 2010). Last but not least, a recent project was initiated at the University of Western Ontario first under the name “Three Little Pigs” and currently as “Insurance Research Lab for Better Housing”. This multi-million project is intended to simulate wind pressure time histories and apply these loads directly on full-scale test structures. The generation of the fluctuating pressures will be provided by 100 pressure actuators. The primary objective of this study is the investigation of the performance of low-rise structures subjected to extreme wind loads. Information regarding this project was presented by Bartlett et al. (2007).

1.4.3 EXTREME FLOODS 1.4.3.1 General Floods are usually due to a volume of water exceeding the total storage capacity of a basin. As a result, part of water flows spread outside from its normal area. The worst floods are water overflows associated to high-water stages where water flows out from its natural or artificial banks, submerging and inundating more or less large low-lying areas of normally dry lands. They represent common and mostly natural disasters. Floods happen when soil and vegetation cannot absorb all the water. Consequently, the water runs off the land in quantities that cannot be carried in stream channels or kept in natural ponds, lakes or man-made reservoirs. Floods are usually local, short-lived events that can happen suddenly and sometimes with little or no warning. They are usually caused by intense storms that produce more water volume than an area can absorb or a stream can carry within its normal channel. Rivers can also flood their surroundings when a dam fails, when ice or a landslide temporarily blocks the course of the river channel, or when snow melts rapidly. Normally dry lands can also be flooded by high lake levels, by high tides, or by waves driven ashore by strong winds. Small streams are subjected to floods which may last from a few minutes to a few hours. On larger streams, floods usually last from several hours to a few days. A series of storms might keep a river above flood stage for several weeks. 1.4.3.2 Classification Floods can be divided into different categories depending on different parameters such as their location or their duration. Classification according to the location: – River floods: A river flood is the most common type of flood. When the actual quantity of a river flow is larger than the amount that the channel can contain, the river overflows its banks and floods the surrounding lands whose surface is at a lower level. Flooding along rivers is natural and inevitable. The Vltava flood in Prague on August 2002 is an example of such a phenomenon. Some river floods occur seasonally (due to spring rains or monsoon for instance). Flooding can also occur when the snow melts at a fast rate. Floating ice can accumulate at a natural or man-made obstruction and stop the flow of water creating a river flood (ice jam). – Coastal floods: Coastal floods occur along coastal areas. They can be due to severe sea storms (intense offshore low-pressure), hurricanes or tropical storms producing torrential rains. In addition, the wind 57

– –

–

–

generated by tropical storms or hurricanes can also drive sea water inland. The Katrina hurricane in New Orleans on August 2005 is an example of such a flood. Coastal floods can also be produced by giant tidal waves created by volcanoes or earthquakes (tsunamis for instance). The Sumatra-Andaman earthquake on 26 December 2004 is of that kind. Estuary or delta floods: Estuary or delta floods are mainly caused by the combination of sea tidal surges caused by storm-force winds and river floods. Arroyos floods: An arroyo is a water-carved gully or a normally dry creek found in arid or desert regions. When storms approach these areas, a fast-moving river can form along the gully which can cause severe but local damages. Urban floods: As lands are transformed from fields or forests to paved roads, parking lots, buildings, large industrial areas etc., they lose their ability to absorb rainfall. Urbanization decreases the ability to absorb water 2 to 6 times over what would occur on natural lands. During periods of urban flooding, streets are transformed into river channels and opened building basements become storage reservoirs. Accidental floods: The origin of an accidental flood can be very different. It can be caused by a dam failure increasing suddenly the quantity of water, by a landslide cutting or decreasing the river flow or by an earthquake diverting a river flow for instance.

Classification according to their duration – Slow floods: Slow floods usually last for a relatively long period – from one week to several months. They can lead to lose stock and agricultural products and to damages to roads and rail links. – Fast floods: Fast floods last for a much shorter period – one or two days. Although fast floods act shortly, they can cause severe damages and create important risks to life and properties as people usually have a very limited time for preventative actions. – Flash floods: Flash floods are floods that occur suddenly with little or no advance warnings. They are usually the result of intense rainfalls over relatively small areas. Flash floods may arrive within minutes or a few hours after heavy rainfalls, intense thunderstorms and tropical storms, dam or levee failures or releases of ice jams. Generally, flash floods create the greatest damage.

1.4.3.3 Significance of floods According to the Swedish International Water Institute (SIWI), about 80% of the natural disasters come from a hydraulic or a meteorological problem and floods concern more than 65 million people per year in the world. From the natural disasters, floods are those which cause the major economical damages and are responsible for about 60% of the fatalities due to natural catastrophes in the world (see Figure 1). Flood risk is generally highly localized and as a result, difficult to quantify. Maps exist which show the relative flood risk modeled for river flood, flash flood, dam burst, storm surge, and tsunami (Figure 4). River floods cause damage to lives, property, infrastructure and crops. Coastal floods cause overflow or overtopping of flood-defences like levees or any kind of natural barriers. Land behind the coastal defence may be inundated or experience damages. As many urban communities are located near coasts, this is a major problem. Periodic floods occur naturally on many rivers, forming an area known as the flood plain. Monsoon rainfalls can cause disastrous flooding in some equatorial countries. In Europe floods from sea may occur as a result from heavy Atlantic storms, pushing the water to the coast. Especially in combination with high tide, this can be damaging. Floods are the most frequent type of disaster worldwide. Thus, it is often difficult or impossible to obtain insurance policies which cover destruction of property due to flooding, since floods are a relatively predictable risk. 58

Figure 4.

European flood risk map.

1.4.3.4 Physical configuration of floods Generally, urbanization increases the size and frequency of floods and may expose communities to an increasing of flood hazards. Damages can be created on bridges by erosion on their foundations but also by dynamic horizontal action on their deck depending on the maximum water level, at the base of buildings leading to a possible brittle behavior due to the lack of support, on sewer systems, on roadways, on canals and, in fact, on most types of engineering structures including levees which can lead to an increasing of the flooding. Weak building and house foundations as well as weak wall basements can be strongly damaged. Due to the water velocity, strong bending effects combined with concentrated actions may occur on submerged structural parts of the constructions which act as a shield face to the water flow. Depending on the area, landslides can amplify the damage when important debris is included into the stream. Without considering casualties and public health problems, the physical damages due to floods are structural damages or collapses caused by washing waters, landslide triggered on account of water getting saturated. Depending on the flow speed, the carried added elements (floating trees or stones for instance), can act as battering rams whose impacts amplify the damages and create cracks. Then, the water becomes more powerful to erode the weakened structure. Due to the flow, the erosion around and under the foundations is one of the most important factors of damages. The particles carried by the flow amplify the abrasive effects. In the case of slow floods, the problem comes from the lower resistance of wet or soluble material (houses built of earth or weak foundations for instance). This is often crucial in the poor countries. In most of the industrialized countries, “vigilance maps” are drawn and daily updated in co-operation with different meteorological services to inform about the possible risks. These are generally based on four scales: – A “Green zone” with no real risk. – A “Yellow zone” meaning a probability of flood but with no risk of significant damages. Nevertheless, vigilance is required. – An “Orange zone” associated to a probability of an important flooding with significant risks to people and properties. – A “Red zone” corresponding to a probability of a major flood with real threat to people and properties. 1.4.3.5 Flood mitigation strategies 1.4.3.5.1 Vulnerability and risk assessment Erosion and floods are a common problem in Europe which is observable on a medium-term scale. Several causes of varying degree are contributing to it. On the one hand are the dynamic nature of coastal and river zones and climate change, and on the other are the anthropogenic influences 59

(Eurosion, 2004). To deal with the problem in an effective manner an adequate understanding of these causes is needed. However, actionable knowledge of these zones dynamics is limited, the effects of climate change are identified in a context of great uncertainty, and the effects of coastal and river interventions have not been systematically monitored. Nevertheless, although the causes are not clearly identified and understood, it is increasingly clear that people and assets in some littoral and fluvial urban fronts are endangered and that serious damage and high costs should be expected. The situation calls for measures to be taken, and it is increasingly important to make scientific and technical information available to decision makers to support their resolutions. To effect sustainable actions, plans should be conceived based on a medium to long-term evolution assessment. Due to inherent uncertainty the assessment of future conditions can only be done on the basis of scenario evaluations, for which numerical models comprising the present state of knowledge may be used as a tool. To help in ranking action priorities, vulnerability and risk must also be assessed. The risk can be accomplished by crossing vulnerability and degree of exposure information in risk maps depicting spatial analysis (as a recommendation from the European Union – EU, Directive 2007/60/CE). Parameters, such as wave energy, tides, bathymetry and topography, shoreline morphology, sediment budget and meteorological conditions are important for a vulnerability analysis. However, zones which are very vulnerable to floods may not necessarily be considered as at risk. An approach similar to the vulnerability analysis needed to be established for evaluating the degree of exposure. Storm surges in combination with river floods can be highly damaging to infrastructure or property, and cause substantial human and economic losses. In terms of exposure levels the parameters considered important are population density, economic activities potentially affected by floods, and ecological, cultural and historical values exposed to devastation by sea actions or floods. Spatial classification facilitates mapping the degree of exposure. Risk maps consist of a classification obtained from crossing vulnerability with degree of exposure. Analysis based on the spatial classification and weighting of the parameters is important for the evaluation of the vulnerability and the risk of floods (see also Coelho et al., 2009). 1.4.3.5.2 Modeling i. Fluvial processes While the damages and casualties of extreme flood events is a major concern for almost everyone, the scientific study of such events is only at its very beginning. Several reasons can justify this state of fact. – Data collection: Since by definition these events are of very high return period, a classical statistical analysis of existing data is difficult and error-prone, especially as far as poorly monitored watersheds are concerned. One has then to adopt modified approaches (e.g. Klemes, 1988). Large scale post-flood survey strategies are often necessary to collect the data and to identify the phenomena involved in the flood generation and propagation (Marchi et al., 2009). – Mechanism comprehension: The genesis and the propagation of extreme floods are the result of the conjunction of many factors and mechanisms which may be difficult to identify and to quantify. Altogether with the lack of data, this renders the relation between rainfall and generated flow difficult to predict. As an example, the behavior of small protection structures (levees and dams) is of critical importance. Since these are designed to be efficient for a relatively small return period, one can expect them to be over-flown in case of an extreme event, but the exact moment of their failure and its consequences on the flood is crucial. – Modeling strategies: For all these reasons, modeling extreme flood remains a challenge; the hydrologist has to select, or rather develop, a model that is able to predict accurately such events with a restricted number of available data (Moussa & Bocquillon, 2009). ii. Coastal processes Understanding of coastal processes to have the capability of prediction and improving the modeling of shore-line evolution and coastal protections behavior is still a challenge to the scientific community. Modeling capability is required to protect coastal areas of erosion and floods. 60

Some one-line numerical models of shoreline configuration (Silva et al., 2007) were especially designed for sandy beaches, where the main cause of medium-term shoreline evolution is the alongshore sediment transport, which is dependent on the wave climate, water levels, sediment sources/sinks, sediment characteristics and boundary conditions. Usually, the models input are the changing water level and the bottom elevation of the modeled area, which may change during calculations. The sediment volumes transported are estimated through the application of formulae (CERC, 1984; Kamphuis, 1991) dependent on the shoreline to wave breaking angle, the wave breaking height, the beach slope and the sediment grain size (Hanson & Kraus, 1989; Kamphuis, 1991). The models assume that each wave acts during a certain period of time (the computational time-step). The wave transformation characteristics depends on refraction, diffraction, reflection, shoaling and wave breaking phenomenon’s, and wave conditions may be imported from more complex wave models such as SWAN (Simulating Waves NearShore). Due to the importance of the boundary conditions, the definition of different conditions should be tested and a sensitivity analysis of the results is usually necessary. Moreover, different coastal protection work combinations should be considered (groins, breakwaters and seawalls, sediment sources/sinks sites or artificial nourishments). The model allows for assessment of some mediumterm scenarios of combined anthropogenic and natural actions, permitting also the modeling of climate change scenarios. Generally speaking, numerical models of shoreline evolution (Hanson & Kraus, 1989; Vicente & Clímaco, 2003) are based on the calculation of coastal sediment transport through the application of the continuity equation to the volumes of sand moved. The variation of the shoreline position is calculated in each cell of the domain by the differences between the volumes of sediments that enter and exit over a certain time interval. Thus, the shoreline position changes due to the spatial and temporal variations of the longitudinal solid transport. For this reason, situations of systematic tendency in position changing are better represented (e.g. erosion down-drift of a groin). The effects of the cross-shore transport are usually not considered in medium to long-term model simulations, since its evolution representation is still not realistic on temporal scales of years to decades. Over a time interval, the analysis along the shoreline for different cells allows establishing the relationship between the volume variation of the cell and the variation of the volume of sediments in transport. The variation of the volume of sand along the beach represents a variation in the depth level at points of the same profile. Erosion/accretion is distributed along the active cross-shore profile, between the closure depth and wave run-up limit. The closure depth can be estimated according to the inshore depth limit of Hallermeier (1978). According to Nicholls et al. (1998), Hallermeier’s approach allows robust estimates of the closure depth both for individual erosive events and for the time-dependent case. This conclusion was based on an analysis of a series of cross-shore profiles measured bi-weekly and with high precision on a sandy ocean beach. All the present parameters and variables are complex to estimate. Thus, numerical modeling of shoreline behavior and coastal floods probability is still poorly understood. 1.4.4 EXTREME SNOW 1.4.4.1 General Extreme snow is a natural action resulting from heavy snowfalls in areas where snow is usually low or an action resulting from any snow loads in regions not normally exposed to snowfalls. Repeated snow events that do not have time to melt and the rain that saturates the snow, which greatly increase its weight, can accumulate and significantly surpass the roof design’s live load and can cause a roof structure to fail. Snow covers on roofs are susceptible to drift action, which leads to removal of snow from some areas and an accumulation in others, and can bring to the extreme design states of snow loads. The collapse of roofs due to heavy snow accumulation may be considered as a catastrophic event and the risk-based approach may be utilized for safeguarding constructions against extreme snow actions. 1.4.4.2 Significance of snow loads Urban habitat constructions subjected to extreme snow loading are undergone through safety assessment. Extreme snow loading accounts for several roof collapses each year. Lightweight 61

roof structures, especially long span flat roofs of shopping centers, sport and concert halls, stadia, railway and bus stations, metal dome roofs of tanks and the like are the most frequent types of structures collapsed during recent years (Zuranski, 2007; Pavlov & Vostrov, 2005). Snow loads on roofs depend on different climatic variables (the amount and type of snowfall, the specific gravity and other snow properties, wind, air temperature, amount of sunshine, etc.), on roof variables (shape, thermal properties, etc.), on site exposure and surrounding environment variables. Calculation of extreme snow loads is largely based on statistics of extreme events which are conventionally utilized by the theory of extremes. Because this method may give very misleading results, purely empirical fit to the observed extremes has been recommended (Makkonen, 2005; Wolinski, 2007a). 1.4.4.3 Snow load configuration Imbalance of load on roofs due to redistribution of snow as result of wind action, sliding off snow cover, heat loss, solar radiation etc., is usually the main reason of a partial collapse or serious damage of roof structures. Complete roof collapse can occur due to heavy snow accumulation, i.e. exceptional winter storms or repeated snow events followed by rain. In regions not normally exposed to snow falls any snow loads can account for total roof collapse. Response characteristics of constructions subjected to extreme snow loads may be regarded in terms of load effects or damage/destruction measures. Structural risk analysis which involves the identification of catastrophic snow events, probabilities and consequences of these events is the recommended method of safety assessment of urban habitat constructions (Steward & Melchers, 1997; Faber et al., 2006; EN 1991-1-7; EN 1991-1-3). 1.4.4.4 Structural aspects There are two main strategies for designing roof structures exposed to extreme snow loads: – design a structure to sustain the action taking into consideration the probability of occurrence of the extreme snow load, – control snow load considering three main methods: snow removal (mechanical or with manpower), snow sliding and snow melting or combination of these methods The extreme value of snow resulting from heavy snow falls SAd may be calculated as: – the product of characteristic snow load on the ground Sk and the coefficient Cesl for exceptional snow: SAd = Cesl Sk , usually Cesl = 2 (EN 1991-1-3), – the maximum value of snow load assessed for an assumed return period, usually T = 100, 200, 500 years, using the asymptotic maximum values distribution of the seasonal extremes; conventionally the Gumbel distribution (Ellingwood & O’Rourke, 1985), – the maximum value for an assumed return period T, using the Monte Carlo simulation technique by Markov chain modelling and historical data converted to a time series (Dukes & Palutiko, 1995), – the maximum value for an assumed return period T, using the empirical method, i.e. a curve fitted directly to the historical seasonal extremes (Wolinski, 2007a). Repeated snow events that do not provide sufficient time for the snow to melt, occurrence of rain that saturates the snow, melting due to heat loss through the roof and subsequent refreezing can lead to additional loads in some areas of the roof. On flat roofs melted snow may accumulate in areas of greatest deflection causing “ponding”. Therefore, the flat roof may be loaded much more than it would have been anticipated according to standards of snow loads and may lead to roof collapse, e.g. the collapse of the ice rink roof in Bad Reichenhall, the disaster of the exhibition hall in Chorzow, Poland in 2006 (Zuranski, 2007). Redistribution of snow load can occur as a result of snow accumulation and drift due to wind action as well as due to sliding processes. Maximum value of the observed coefficient for exceptional snow load due to these actions was about Cesl ∼ = 5 ÷ 6 (Gooch, 2002; Mihashi et al., 1989). Imbalance of snow load on dome roofs, especially on lightweight aluminium and steel round in plan dome roofs of tanks and buildings, caused recently failures of several tank roofs in USA and 62

Figure 5. problem.

Scheme of response identification

Figure 6. Illustration of structure (system) representation.

Russia. Sliding off snow cover may bring two extreme design states: balanced snow load which is determined with account for sliding of whole snow ring, and unbalanced snow load when a part of snow cover slides while the other part of it remains (Pavlov & Vostrov, 2005).

1.4.4.5 Vulnerability analysis Building structures may be considered as structural systems, i.e. bounded groups of interrelated interdependent or interacting elements forming an entity that achieves defined objectives. Therefore, the approach based on the generic system characteristics such as exposure, robustness and vulnerability may be utilized for design and assessment of concrete structures (Steward & Melchers, 1997; Faber et al. 2006). The general scheme of response identification problem is shown in Figure 5 (Wolinski, 2007b). An exposure is related to any event with the potential to cause damage to the structure (loads, corrosion, errors or other disturbances). Robustness is considered to be a measure of the degree to which the specified or unpredictable perturbations influence performance of a structure and is characterized by means of the risk associated with all indirect consequences of its failure or collapse. Vulnerability of a structure is defined as the measure of extend to which changes would harm a structure and characterizes the direct risk associated with its damage or failure (Figure 6). The total risk R of the structure can be assessed by:

where the structure is subjected to nH different hazards that may damage the structure in nD different ways and the performance of the damage structure can be discretised into nK adverse states Kk with corresponding consequences C(Kk ), and p(Hi ) is the probability of occurrence of the i-th hazard, p(Dj |Hi ) is the conditional probability of the j-th damage state of the structure given in the i-th hazard and p(Kk |Dj ) is the conditional probability of the k-th adverse overall structural performance K given in the i-th damage state. The robustness index takes values 0 ≤ IR ≤ 1 depending on the source of risk. If all risk is due to direct consequences, the structural system is completely robust and IR = 1. If all risk is due to indirect consequences, the structural system has no robustness and IR = 0 (Faber, 2007). 63

Figure 7. Comparison of the Gumbel and empirical analysis for the extreme snow loads in two towns “BP” and “K” (Poland).

The concept of the empirical method to assess the maximum value of snow load assessed for an assumed return period is that sample ordering is applied directly to the observed historical data and not to a postulated cumulative distribution function (CDF) of the probability distribution. The basic steps in the empirical method procedure are as follows (Makkonen, 2005; Wolinski, 2007a): – select a number of the biggest seasonal extremes of S, – plot the extremes according to the order ranks by Pi = i/(n + 1) on S – P plot, – fit a curve that best models the data using a standard algorithm (the third degree polynomial may be used and weights to the points may be assigned according to their statistical confidence), – extrapolate or interpolate along the fitted curve to evaluate the value of S that correspond to an assumed return period (or to evaluate a return period corresponding to a given S-value).

1.4.4.6 Example of application Comparison of the Gumbel probability analysis and the empirical method for the extreme snow loads resulting from heavy snowfalls S for two regions of Poland is shown in Figure 7. The values in the figure represent annual extremes from T=50 years. The straight lines represent the Gumbel analysis and the curved lines represent the empirical method. The dashed lines correspond to extrapolation to periods T >50 years. The purely empirical fit to obtained data using the simple or weighted approximation or interpolation procedure can be then used for extrapolation to estimate the maximum value corresponding to a given return period. The extremes resulting from the Gumbel analysis are quite different from those of the presented empirical method as well as from those of the Markov chain modeling.

1.4.4.7 Further developments New methods such as the purely empirical method of determining extreme snow loads for structural design are necessary to compare with the extreme value theory via the Gumbel model, which is widely used in building codes to estimate snow loads. The approach to design and assessment of the lightweight roof structures subjected to infrequent loading conditions (heavy snow storms, unusual patterns of snow cover, combined snow, wind and ice actions, etc.) based on assessment of risk characteristics should be introduced as a helpful supplement to design and assessment procedures. 64

1.4.5 AVALANCHE ACTIONS ON CONSTRUCTION 1.4.5.1 General Extreme avalanches are, each year, the cause of very important injuries, deaths, environmental damages and material damages. These damages are due to inadequate knowledge of initial phenomena of these avalanches, of their process and their potential impact on human life and construction (Boissier & Muzeau, 2008). This section is focused on the presentation of the current knowledge on avalanche phenomena and their consequences (actions) on constructions in a context of risk analysis. The risk is the combination of hazard occurrence and issue damages. In the avalanche risk context, hazard is the avalanche and issues can be persons, structures, infrastructures, communications, environment, economy etc. Avalanches are phenomena very similar to landslides: analogies exist between soil grains and snow grains, soil layers and snow layers, landslides and avalanches, the effects of compaction and mode of deposit. It is possible to identify two main types of avalanches (Givry & Perfettini, 2004) depending on the snow state: on the one hand, aerosols, dry snow or powder avalanche and, on the other hand, wet snow avalanche. The difference between these kinds of phenomena defines their mode of failure, their way of displacement down the slope and their relative energy. An avalanche occurs when the loading conditions (internal weight, external forces due to a skier, an explosion or another avalanche for instance) become greater than the resistance of the snow cover. The snow deposit phase conditions and the snow characteristics are critical factors to the avalanche departure. Therefore, an avalanche classification is proposed that depends on these snow characteristics and the avalanche departure modes. Finally, the avalanche progress over constructed areas generates impacts or actions that have to be foreseen to reduce consequences on issues. 1.4.5.2 Avalanche risk analysis There are various important interests in avalanche risk analysis: – interests for people: in France, since 2000, an average of 14 people (skiers, hikers, ski resort users and workers, motorists, etc.) have died in avalanche accidents (55 for the winter 2005–2006); – interests for structures and infrastructures: buildings, ski lifts, roads, bridges, etc. can be damaged or destroyed by avalanches; – economic concerns: power cuts or road closings for instance may induce production falls. – social and political stakes: the occurrence of an avalanche may modify both risk perception and risk acceptance; the crisis management after the impact of an avalanche is under the responsibility of politicians. 1.4.5.3 Snow deposit The snow deposit depends on the snowfall frequency and intensity and on the snow base. The stratification of the snowy coat (Fig. 8) is made according to a wind-driven deposit mode. The order of magnitude is one centimeter by ten minutes. In comparison, the wind-driven deposit of loess average rate is one centimeter by century (Burlet, 2002). Several variability scales of snow stratification may be considered: – the temporal variability: hourly, daily (daytime, night-time or morning and afternoon), weekly, annual, secularly, millenary; – the spatial variability: vertical or lateral on one part, at a centimetre scale, and at metre to decimetre scale on other parts. Several snow bases may be considered: – snow coat, rock, soils or grass, – North, South, East, West mountain face, – high, moderate or low slope declivity. The variability of the snow deposit and of the snow base conditions, the snow characteristics and the avalanche departure. 65

Figure 8.

Illustration of snow stratification (LEGOS, 2005).

Table 1. Main snow characteristics (Naim-Bouvet et al., 2000). Unit weight ρ Cohesion Compression strength Elastic modulus

20 kg/m3 for new snow 500 kg/m3 for old snow 8 to 35 kPa for ρ contained between [300 to 460 lg/m3 ] 0 to 20 kPa for ρ < 300 kg/m3 σt = 58.3 (ρ/ρice )2.65 for faceted grains and cups σt = 79.7 (ρ/ρice )2.39 for all others type of grain 2462ρ02.826 kPa

1.4.5.4 Snow characteristics The mechanical properties of the snowy coat are based on the theory of soil mechanics. It is a multi-phases coat with the average values of the main characteristics presented Table 1. Several grain types can be encountered: – crystal of fresh snow that present a weak sintering cohesion and are stable on steep slope (Figure 9); – fine grains that have a good sintering cohesion (Figure 10); – faceted grains (Figure 11); – cups that have no frottage sintering cohesion and generate sliding on superior slopes (Figure 12); – round grains that have a capillary cohesion and frost cohesion. In consequence, the snowy coat has a more resistant cohesion (Figure 13). The sintering phenomenon (Figure 14) corresponds to the realization of a glass bridge between two grains. The bridge number is higher when the grains are smaller. This phenomenon has a stabilization effect on snowy coat due to the increasing of the cohesion. However, this phenomenon also encourages the propagation of failures due to the rigidifying of the snowy coat. The avalanche departure depends on the snow deposit phase and on the snowy coat characteristics. The mechanical system of antagonistic forces includes: the traction strength (FT) and the resistance strength (FR). The avalanche departure corresponds to an equilibrium break into the snowy coat: FT > FR. The mechanism of equilibrium break is the same for natural and accidental break (Figure 15). 1.4.5.5 Avalanche classification Depending on the considered spatial scale (local slope scale or massive scale), different types of avalanches can be encountered. Meteorological data are generally precise and reliable. Consequently, it is possible to provide a risk index but it is only global information. 66

Figure 9. 1998).

Crystal of fresh snow (Météo France,

Figure 11.

Faceted grains (Météo France, 1998).

Figure 13.

Round grains (Météo France, 1998).

Figure 15.

Mechanism of avalanche departure.

Figure 10.

Figure 12.

Figure 14.

Fine grains (Météo France, 1998).

Cups (Météo France, 1998).

Sintering phenomenon (Anena, 2008).

At the slope scale, three kinds of avalanches may be distinguished: – powder avalanche, – plate avalanche, – wet snow avalanche. 67

Figure 16.

Schemes of an avalanche progress (Avalanche, 2008).

At the massive scale, three kinds of avalanche may be distinguished: – bottom avalanche, – streaming avalanche, – valley avalanche. When considering the slope scale, an avalanche can be composed of the following generic items (Figure 16): – a departure zone (accumulation basin), – a flow zone (throat) – a stationary zone (dejection cone). 1.4.5.6 Uncertainties associated with avalanches One of the important required research activities is to evaluate the uncertainties leading the behavior and the consequences of avalanches. Uncertainties are involved in the three processes: the starting phase, the developing phase and the runout phase. Starting phase The uncertainties concerning the starting phase are, for instance, the quality of the snow layers and the applied loads. Each snowfall is affected by mechanical and thermodynamical evolutions. This implies a spatial variability of strengths and thermodynamic properties. – Vertical and plan variability The vertical variability is due to the structure of the snow cover; penetrograms into the snow covers do present a great variability in the vertical direction. The variability comes from the mode of deposit of the snow layers under various conditions and from the thermal and mechanical metamorphosis of their grains. The plan variability is explained by several phenomena: • the topography, the roughness, the nature and the slope angle influence the setting-up of the snow cover; • the slope determines the orientation of the temperature gradient, which induces different transformations. Wind modifies the spatial repartition of snow and this leads to local gradients and overloads. – Time variability It comes from the kinetic of metamorphosis within the snow cover. Two scales for the time variability can be observed: • daily variations close to the surface induced by the alternation day/night, • transformations due to the gradient intensity. Inside the snow cover, transformations are continuous but significant property changes can be noticed after a few days. 68

Moreover, time variability comes from wind, rain, new snowfalls and human or animals actions. Developing phase The uncertainties concerning the developing phase depend on the topography, the roughness, the nature and the slope angle of the snow cover or of the ground. So, the path of the avalanche can be considered as random. Runout phase In the runout phase, only avalanches that impact stakes are taken into account. The uncertainties concern the energy, the speed, the orientation and the mass of the avalanche. Another important point of uncertainty is that an avalanche often contains rock or ice blocks, trees and materials from skiers and ski lifts, that have been pulled away from the soil. All of them may damage buildings or structures and hurt people. The resistance of the structures or infrastructures, their behavior under dynamic loads, their damaging is also concerned with uncertainties. 1.4.5.7 Possible protection action Two main suggestions for possible protection actions as input for structural analysis and strengthening design can be made: – The first one concerns the building of avalanche walls or avalanche equipments. Paths of the avalanche and its maximum volume have to be precisely assessed; the wall has to be calculated under dynamic loads taking into account the consequences of its damaging. – The second one concerns the design of the building in front of the possible avalanche path. As for hurricanes, the orientation, the dimensions of openings and other configuration elements are to be taken into account. 1.4.5.8 Further development Avalanches may generate very important mechanical, social, environmental and economical impacts. These avalanches may be characterized at the snowy coat scale, at the slope scale and the massif scale. The parameters and properties that distinguished the different kinds of avalanche depend on the study scale considered. The knowledge of the avalanche actions on constructions in populated areas is of main interest in order to reduce the occurrence of avalanche and to minimize the consequences of avalanches, in a context of risk analysis. A major possible development is relevant to the prediction of the occurrence of the avalanche phenomenon. In this goal, a study has been carried out by Burlet et al. (1999) at the Ski resort scale. It is based on the idea that mechanical models can help with avalanche predictions as mechanical models do for land slopes. The development of this approach needs the knowledge of the snow layers (type, thickness and mechanical properties) each time the risk assessment is needed. Another possible development concerns the avalanche dynamics and more precisely its impact on works. In that way, research is carried out by Ma (2008) at the slope scale. It aims to estimate the avalanche action on an avalanche wall used to protect a road. It is based on the idea that an avalanche can be modeled as a granular flowing.

1.4.6 CLOSING REMARKS This chapter has reviewed the major characteristics of natural catastrophic actions other than fires, volcanoes and earthquakes causing however comparable, if not more critical, effects on man and his environment, particularly buildings and other structures. In particular, extreme winds, floods, snow and avalanches have been addressed, along with their physical characteristics, research activities related to their investigations and design risk-based methodologies to treat them in engineering terms. Clearly, there is a lot of research action and collaboration required in order to advance the current-state-of-the-art in these areas. 69

REFERENCES Anena, 2008. Association Nationale pour l’Etude de la Neige et des Avalanches. URL: http://www.anena.org (On line). Avalanche, 2008. Programmes d’études des avalanches. (URL: http://www.avalanches.fr). Bartlett, F. M., Galsworthy, J. K., Henderson, D., Hong, H. P., Iizumi, E. and Inculet, D. R. 2007. The three little pigs project: A new test facility for full-scale small buildings. 12th International Conference on Wind Engineering, Cairns, Australia. 1623–1630. Berz, G. & Smolka, A. 1988. Windstorm hazard and insurance. Journal of Wind Engineering and Industrial Aerodynamics, 30, 191–204. Boissier, D. & Muzeau, J.P. 2008. Avalanches, Int. Symp. COST C26: Urban Habitat Constructions under Catastrophic Events, Malte. Burlet, J.L. 2002. Mécanique de la neige et variabilité – Application à la prévision du risque d’avalanche. Civil Engineering Thesis. Université Blaise Pascal Clermont-Ferrand II. Burlet, J.L., Bacconnet, C., Boissier, D., and Gourvès, R. 1999. Modelling of the snow cover variability with a Gaussian stochastic field. 8th ICASP, Sydney, Australia, 869–875. Caracoglia, L., Sangree, R. H., Jones, N. P., & Schafer, B. W. 2008. Interpretation of full-scale strain data from wind pressures on a low-rise structure. Journal of Wind Engineering and Industrial Aerodynamics, 96(12), 2363–2382. CERC – Coastal Engineering Research Center 1984. Shore Protection Manual, U.S. Government Printing Office, Washington DC. Cochran, L. 2000. Wind engineering as related to tropical cyclones. Part of Storms, Vol. I, Routledge Press. Coelho, C., Granjo, M. and Segurado-Silva, C. 2009. Map of Costal Zone Vulnerabilities to Wave Actions – Application to Aveiro District (Portugal), International Symposium on Integrated Coastal Zone Management, Arendal, Norway, 318–330. Crandell, J.H. 1998. Statistical assessment of construction characteristics and performance of homes in HurricanesAndrew and Opal. Journal ofWind Engineering and IndustrialAerodynamics, 77 & 78, 695–701. Datin, P. L., Liu, Z., Prevatt, D. O., Masters, F. J., Gurley, K., & Reinhold, T. A. 2006. Wind loads on singlefamily dwellings in suburban terrain: Comparing field data and wind tunnel simulation. Structures Congress 2006, St. Lois, Missouri, USA., 201(40889) 13. Davenport, A.G. 1990. The impact of structural damage in Jamaica due to Hurricane Gilbert and the prospects for disaster reduction. Journal of Wind Engineering and Industrial Aerodynamics, 36, 53–62. Doudak, G. 2005. Field determination and modeling of load paths in wood light-frame structures. PhD Thesis, McGill University, Montreal, QC, Canada. Dukes, M.D.G. & Palutikof, J.P. 1995. Estimation of extreme wind speed with very long return periods, Journal of Applied Meteorology, Vol.34, pp. 1050–1961. Eamon, C.D., Fitzpatrick, P. and Truax, D.D. 2007. Observations of structural damage caused by Hurricane Katrina on the Mississippi Gulf Coast. Journal of Performance of Constructed Facilities, Vol. 21, No. 2, 117–127. Eaton, K. J., & Mayne, J. R. 1975. The measurement of wind pressures on two-storey houses at Aylesbury. Journal of Wind Engineering and Industrial Aerodynamics, 1, 67–109. Ellingwood, B. & O’Rurke M. 1985. Probabilistic Models of Snow Loads on Structures, Structural Safety, 2, pp. 291–299. EN 1991-1-3 :2003 2010. Eorocode 1 – Actions on structures Part 1–3: General actions – Snow loads, European Committee for Standardization EN 1991-1-7 :2006 2010. Eorocode 1 – Actions on structures – Part 1–7: General actions – Accidental actions, European Committee for Standardization. Eurosion 2004. Living with Coastal Erosion in Europe: Sediment and Space for Sustainability. www.eurosion.org. Faber M., et al. 2006. On the qualification of robustness of structures. Proceedings of OMAE2006, 25th Offshore Mechanics and Arctic Engineering Conference, Hamburg, Germany. Faber H.M. 2007. Framework for risk assessment of structural systems. Proceedings of the COST C26 Workshop, Prague, Czech Technical University in Prague, pp. 359–367. Friedman, D.G. 1984. Natural hazard risk assessment for an insurance program. The Geneva Papers on Risk and Insurance, 9, No. 30, 57–128. Givry, M., and Perfettini, P. 2004. Construire en montagne – la prise en compte du risque d’avalanche. Ministère de l’écologie et du développement durable et Ministère de l’équipement, des transports, du logement, du tourisme et de la mer. Gooch, C. 2002. Winter snow loads. Cornell University.

70

Hallermeier, R.J. 1978. Uses for a Calculated Limit Depth to Beach Erosion. Proceedings of 16th Coastal Engineering Conference, ASCE, Hamburg, Germany, pp. 1493–1512. Hanson, H., Kraus, N. C. 1989. GENESIS: Generalized Model for Simulating Shoreline Change; Report 1, Technical Reference, U.S. Army Corps of Engineers. Kamphuis, J. W. 1991. Alongshore Sediment Transport Rate, Journal of Waterways, Port, Coastal and Ocean Engineering, ASCE, Vol. 117(6), pp. 624–641. Keith, E.L. and Rose, J.D. 1994. Hurricane Andrew – Structural performance of buildings in South Florida. Journal of Performance of Constructed Facilities, Vol. 8, No. 3. Khanduri, A.C. and Morrow, G.C. 2003. Vulnerability of buildings to windstorms and insurance loss estimation. Journal of Wind Engineering and Industrial Aerodynamics, 91, 455–467. Klemes V. 1988. The improbable probabilities of extreme previous floods and droughts. In: O. Starosolszky and O.M. Melder, Editors, Hydrology of Disasters, James and James, London, UK, pp. 43–51. LEGOS, 2005. Cryosphère satellitaire, (URL: http://www.legos.obs-mip.fr). Levitan, M. L., Mehta, K. C., Chok, C.V., & Millsaps, D. L. 1990. An overview ofTexasTech’s wind engineering field research laboratory. Journal ofWind Engineering and IndustrialAerodynamics, 36(Part 2), 1037–1046. Ma,Y. 2008. Analyse de l’effet d’une avalanche de neige sur un ouvrage de protection type galerie paravalanche. Expérimentation, modélisation. PhD, University of Savoie, September 2008. Marchi L., Borga M., Preciso E., Sangati M., Gaume E., Bain V., Delrieu G., Bonnifait L. and Pogacnik N. 2009. Comprehensive post-event survey of a flash flood in Western Slovenia: observation strategy and lessons learned. Hydrological Processes, vol. 23, no26, pp. 3761–3770. Makkonen, L. 2005. Models for growth of rime, glaze icicles and wet snow on structures. Philosophical Transaction A, vol. 34, 2005. Meecham, D., Surry, D. and Davenport, A.G. 1991. The magnitude and distribution of wind-induced pressures on hip and gable roofs. Journal of Wind Engineering and Industrial Aerodynamics, 38, 257–272. Mehta, K.C., Minor J.E., and Reinhold T.A.1983. Speed-damage correlation in Hurricane Frederic. Journal of the Structural Division, Vol. 109, ASCE, No. 1. Metéo France,1998. Panneaux Série “Neige”. URL: http://www.cnrm.meteo.fr (On line). Mihashi, H., Takahashi, T., Izumi, M. 1989. Wind effects on snow loads on roofs. Proceedings of 1st. Int. Conference on Snow Engineering, CRREL Special report 86–6, pp. 168–167. Minor, J. 2005. Lessons Learned from Failures of the Building Envelope in Windstorms, Journal of Architectural Engineering, ASCE, Volume 11, No. 1, pp. 10–13. Moussa, R. and Bocquillon, C. 2009. On the use of the diffusive wave for modelling extreme flood events with over-bank flow in the floodplain. Journal of Hydrology, Volume 374, 1–2, pp. 116–135. Munchener Ruck, 2006a. Knowledge Series, Hurricanes-More intensive, more frequent, more expensive: Insurance in a time of changing risks. Munich Re Group. Münchener Rück, 2006b. Topics Geo, Rassegna delle catastrofi naturali 2005, Munich Re Group, http://www.munichre.com. Naim-Bouvet F., Prat M., Jacob J., Calgaro J.A., Raoul J. 2000. La neige – recherche et réglementation. Editions Presses École Nationale des Ponts et Chaussées & Cemagref. Nicholls, R. J., Birkemeier, W. A., and Lee, G. 1998. Evaluation of depth of closure using data from Duck, NC. Marine Geology, 148: 179–201. Ng, H. H. T., & Mehta, K. C. 1990. Pressure measuring system for wind-induced pressures on building surfaces. Journal of Wind Engineering and Industrial Aerodynamics, 36(1–3), 351–360. Pavlov, A., & Vostrov, V. 2005. Extreme snow loads on dome roofs of buildings and structures. IABSE Symposium “Structures and Extreme Events”, Lisbon, IABSE Reports, Volume 90, pp. 198–199. Perry, D.C., McDonald, J.R. and Saffir, H.S. 1990. Performance of metal buildings in high winds. Journal of Wind Engineering and Industrial Aerodynamics, 36, 985–999. Porterfield, M., & Jones, N. 2001. The development of a field measurement instrumentation system for low-rise construction. Wind and Structures, 4(3), 247–14. Robertson, A. P., & Glass, A. G. 1988. The Silsoe structures building – its design, instrumentation and research facilities. AFRC Inst. Engng. Res., Silsoe, Div. Note DN 1482. Silva, R., Coelho, C., Taveira-Pinto, F., and Veloso-Gomes, F. 2007. Dynamical Numerical Simulation of Medium Term Coastal Evolution of the West Coast of Portugal. 9th International Coastal Symposium, Gold Coast, Queensland, Australia. Sparks, P.R., Schiff, S.D. and Reinhold, T.A. 1994.Wind damage to envelopes of houses and consequent insurance losses. Journal of Wind Engineering and Industrial Aerodynamics, 53, 145–155. Steward, M.G., & Melchers, R.E. 1997. Probabilistic risk assessment of engineering systems. London, Chapman Hall.

71

Van de Lindt, J.W., Graettinger, A., Gupta, R., Skaggs, T., Pryor, S. and Fridley, K.J. 2007. Performance of wood-frame structures during Hurricane Katrina. Journal of Performance of Constructed Facilities, Vol. 21, No. 2, 108–116. Vicente, C. M., Clímaco, M. 2003. Evolução de Linhas de Costa. Desenvolvimento e Aplicação de um Modelo Numérico, Informação Técnica Hidráulica ITH 42, Laboratório Nacional de Engenharia Civil, Lisboa. Wolinski, Sz. 2007a. Framework for assessment of actions on structures due to catastrophic events. In monograph Research problems of building structures. Vol. III, Polish Academy of Sciences, Bialystok, pp. 51–58 (in Polish). Wolinski, Sz. 2007b. Identification and classification of response of urban habitat constructions subjected to extraordinary loading conditions. Proceedings of the COST C26 Workshop in Prague, Czech Technical University in Prague, pp. 368–371. Zisis, I. & Stathopoulos, T. 2009. Wind-induced cladding and structural loads on a low wooden building, Journal of Structural Engineering, ASCE, Vol. 135, No. 4, pp. 437–447. Zisis, I. & Stathopoulos, T. 2010. Wind-induced loads on the foundation of a low-rise building: Full-scale and wind tunnel experimentation, 2010 Structures Congress, ASCE, Orlando, FL, USA. Zuranski, J. 2007. Failures of roofs under snow loads in Poland. Proceedings of the XXIII Conference“Structural Failures 2007”, Szczecin, Poland, pp. 357–364 (in Polish).

72

Urban Habitat Constructions under Catastrophic Events (Final Report) – Mazzolani (Ed). © 2010 Taylor & Francis Group, London, ISBN 978-0-415-60686-8

1.5 Actions due to volcanic eruptions J.-P. Muzeau & A. Talon LaMI – Polytech’Clermont-Ferrand (CUST), Blaise Pascal University – Clermont-Ferrand, France

J.-C. Thouret Magmas and Volcanoes Laboratory, Blaise Pascal University – Clermont-Ferrand, France

T. Rossetto Department of Civil, Environmental & Geomatic Engineering, University College London, UK

B. Faggiano & D. De Gregorio Department of Structural Engineering, University of Naples ‘Federico II’, Naples, Italy

G. Zuccaro PLINIVS Centre. Hydrological, Volcanic and Seismic Engineering Centre. University of Naples ‘Federico II’, Naples, Italy

M. Indirli ENEA (Italian National Agency for New Technologies, Energy and Sustainable Economic Development), Bologna, Italy

1.5.1 INTRODUCTION Today, about 500 million people are at risk from volcanic hazards. In the past 500 years, over 200,000 people have lost their lives due to volcanic eruptions. An average of 845 people died each year between 1900 and 1986 from volcanic hazards and for the next years, these numbers are predicted rising (Tilling, 1991 & 2005). The reason is not due to increased volcanism, but to an increase in the amount of people populating the area surrounding the active volcanoes. In Europe, this is the case of the Neapolitan area (Italy), where Vesuvius and Phlegrean Fields volcanic complexes threaten the safety of about one million of people. Instead, in the rest of the World, metropolitan areas, like Tokyo (Mt. Fuji), Mexico City (Popocatépetl) and Auckland (Auckland Field) are affected by eruptive risk. The peculiar importance of this aspect has induced the COST Action C26 “Urban habitat constructions under catastrophic events” (2006–2010) to introduce the analysis of the volcanic risk in the urban areas in its research activities. The objectives are substantially two: to provide a methodology to evaluate the volcanic vulnerability of the urban environment towards an eruption and to propose simple and economical mitigation interventions. The first developed activity is relating to the definition of the actions produced by an eruption on the constructions (Mazzolani et al., 2008 and 2009a), with reference to physical phenomenon and the consequences on the structures. In particular, they are hereafter presented distinguishing between the effects due to the specific products of an effusive (lava flows) and an explosive (air fall deposits, pyroclastic flows and surges and flying fragments) eruption and that one produced by the secondary effects of an eruption, like lahar, tsunami and volcanic earthquake. The main goal of this paper is to describe the main actions due to volcanic eruptions and their consequences on the local environment and particularly on constructions. 1.5.2 ANALYSIS OF VOLCANIC HAZARD 1.5.2.1 Definition A volcanic hazard (Dumaisnil et al., 2008) is defined as an event that can occur in a given area or location, such as a lava flow or a volcanic earthquake, along with the probability of the event’s 73

Table 1. Indirect volcanic hazards with their physical characteristics (after Blong, 2000). Indirect hazards

Characteristics pertinent to risk

Example

Earthquakes and ground deformation Tsunami

Limited damage; subsidence may affect hundreds of km2 . Can travel great distances; exceptionally, waves to 30+ m. Can continue for years. Can affect extensive areas for years after eruption. Limited effects.

Sakurajima, 1914, Usu, 2000 Krakatoa, 1883

Secondary debris flows Post-eruption erosion and sedimentation Atmospheric effects Air shocks, lightning Post-eruption famine and disease

Limited effects at present.

Santa Maria, 1902–1920 Irazu, 1963–64, Pinatubo, 1991–2000 Mayon, 1814, Agung, 1960 Lakagígar (Laki), 1783

occurrence. It is important to understand the hazard, but it is difficult to reduce the hazard itself because most effects of the eruptions are beyond our control. Hazard refers to the annualized probability of the specific volcanic event (tephra fall, pyroclastic flows, lahars, etc.) occurring in the area under consideration. Various types of volcanic activity can affect people and property (death, injury, destruction, etc.) both in close proximity to and far from a volcano. Some hazards are more severe than others depending on the size and extent of the event taking place and whether people or property is within the path of the volcanic hazardous event. Although most volcanic hazards are triggered directly by an eruption, some occur when a volcano is in a state of quiescence. The range of volcanic hazards that an area is exposed to needs to be recognized and understood (see Tables 1–2, Figure 1).

1.5.2.2 Process to identify volcanic hazards Identifying and assessing the range of volcanic hazards that might be relevant at any given volcano, or site near a volcano, is a complex process that encompasses at least six tasks: 1) identifying a threatening volcano by using a set of physical, geological, and socio-economic criteria; 2) foreseeing an eruption at a given time and site by installing an adequate network of equipment for surveillance; 3) defining and ranking possible eruption scenarios based on the past and recent behaviour of the edifice; 4) delineating areas likely to be affected by the effects of the eruptive phenomena based on the recognized scenarios; 5) identifying and weighing people and communities at risk, vulnerable elements, social, economic, and cultural elements at stake; 6) undertaking a land-use policy and emergency procedures, including: civil defence, relief organisation, mitigation, preparedness and consciousness, and risk management aiming at sustainable development. 1.5.3 WHY TO STUDY VOLCANIC HAZARDS? 1.5.3.1 Main reasons to study volcanic hazards In 2000, an estimated 500 million people were living within a distance of 100 km from an active volcano. Twice during the twentieth century two large towns were destroyed by eruptions: St Pierre, Martinique, in 1902 and Armero, Colombia, in 1985. Major population centres lie within ten to twenty kilometres from several large volcanoes with a likelihood of eruption during this century, e.g., Naples near Vesuvius and Arequipa near El Misti (see Figure 2). 74

Table 2. Direct volcanic hazards with their physical characteristics (after Blong, 2000). Direct hazards

Characteristics pertinent to risk

Example

Fall processes: Tephra falls

Downwind transport velocity <10 to <100 km/h, can extend 1000+ km downwind, and can produce impenetrable darkness; surface crusting from tephra fall encourages runoff.

Vesuvius, 1631, 1906 Rabaul, 1994

Ballistic projectiles

Can affect a 10+ km radius from the vent; projectiles have high-impact energies; fresh bombs above ignition temperatures of many materials.

Soufrière St Vincent, 1812

Lava flow: Lava flows Domes

Bury or crush objects in their path; follow topographic depressions; can be tens of kilometres long; and produce a noxious haze from sustained eruptions.

Kilauea, 1960, 1983-present Merapi, Soufrière Hills Montserrat, 1995-present

Pyroclastic flow: Pyroclastic flows Pyroclastic surges

Concentrated gas-solid dispersion; small flows can travel to a distance of 5–10 km within topographic lows, whereas large flows can travel a distance of 50–100 km; large flows can mount topographic obstructions.

Pinatubo, 1991, Unzen, 1991–93, Mount Pelée, 1902, Taal, 1960

Laterally directed blast

Destroy all constructions.

Bezymianny, 1956, Mount St Helens, 1980

Debris flow: Primary (eruption-triggered) debris flows (lahars)

Velocities may exceed 10 m/s; rapid aggradations, incision or lateral migration may occur; the hazard may continue for months or years after eruption.

Nevado del Ruiz, 1985, Kelud, 1919

Jökulhlaulps

These high discharge flows are triggered by ice-dammed lake breakouts; can flood extensive pieces of land and can occur with little or no warning.

Katla, 1918 Grímsvötn, 1996

Sector collapse and flank failure: Debris avalanche Magmatic origin Phreatic origin No eruption, seismogenic

Emplacement velocities of up to 100 m/s; can create topography, pond lakes; and produce tsunamis in coastal areas.

Mount St Helens, 1980 Bezmianny, 1956 Bandaï-san, 1883, Ontake, 1984 Shimabara, 1792

Other eruptive processes: Phreatic explosions Volcanic gases and acid rains

Damage limited to proximal areas but can be lethal; corrosive, reactive; low pH in water; CO2 in areas of low ground.

Soufrière de Guadeloupe, 1976 Dieng plateau, 1979

It is therefore needed: 1) 2) 3) 4) 5)

to minimize the risk of loss of life from structure collapse or damage in the event of an eruption; to facilitate appropriate warning and evacuation systems; to protect food-producing areas and other areas of significant economic activity; to improve the expected performance of structures and lifelines; to improve the functional capability of structures and lifelines that are essential to post-eruption recovery during and after an eruption, and to minimize the risk of damage to hazardous facilities.

1.5.3.2 Field of action We wish to apply this hazard, vulnerability, and risk assessment to urban environments or to densely populated areas of the volcanic islands of the West Indies (such as Basse Terre, Guadeloupe or 75

Figure 1a. Smoke and ash from Mount St. Helens. Credit: Norman G. Banks, United States Geological Survey, courtesy NSF.

Figure 1b. Principal types of direct volcanic hazards (after USGS web site).

Figure 2. World map showing the main seismic and volcanic zones combined with the densely populated areas, well-studied volcanoes and some dates of the characteristic eruptions (Chester, 2000; Thouret, 2010).

Martinique), of Java, Indonesia (Semeru and Merapi volcanoes), and of Peru (e.g., El Misti) or Ecuador, or Colombia. In particular, we wish to apply the hazard and risk method to large cities at risk in the developing world such as Arequipa near El Misti in southern Peru and Lumajang near Semeru in eastern Java or Yogyakarta near Merapi in central Java. The flow chart of Figure 3 shows the field of application of the method, the principal aims, steps and tasks to be carried out, and the potential effects on a densely populated area or a city located on the ring plain of the edifice. A volcanic eruption (Figure 1) can be characterized by several actions, as explosions, projections of magma or pre-existing solid rock, lava flows, more or less dense clouds of ash-laden gas, pyroclastic flows, dust and lahars. Furthermore, earthquakes (and sometimes tsunami) accompany the event. 76

Figure 3.

Flow chart (Thouret, 2010).

1.5.4 EFFUSIVE ERUPTION 1.5.4.1 Lava flows During an effusive eruption, the lavas, constituted by totally or partially fused magma, emerge on the surface, flowing in a viscous mass from the crater itself, or from fissures or fractures. Their speeds are generally of few kilometers per hour and they, as well as the temperature, decreases with the distance from the vent. Lava can also be blown away in fragments to create kinds of avalanches moving down slopes at speeds as high as 150 km/h. The most abundant chemical component of lava is the silica. Depending on the SiO2 weight content, they are classifiable in: acid, intermediate, basic and ultrabasic. In general, more basic magmas present higher eruptive temperatures (1000–1200◦ C), than more acid magmas (700–900◦ C). Lava and magma can be considered as not Newtonian Bingham type fluids and so, the viscosity is their central peculiarity which influences eruptive behavior, mobility and shape. A construction invested by a lava front is subject to an unavoidable destiny of destruction. Potentially, the risk interests the settlements in areas around the volcano and it grows near the eruptive vents and parasitical cones. The lava flow produces a lateral horizontal pressure which can cause the collapse of the invested buildings. The crisis is also caused by the degradation of the materials produced by high temperatures of the magma. For example, during the Etna eruption of 2001, the temperature of lava flow, measured with the infrared radiometer, was 1075◦ C. However fortunately, the advancing speed of the lava flows is sufficiently low to allow the evacuation and the safeguard of human lives. So, the mitigation of lava risk can not require the reduction of structural vulnerability, but it must be pursued by means of two types of interventions: passive and active protection. The passive protection consists on a proper planning of the territory and of the emergencies management: the buildings must not be built near the areas with probable opening of eruptive vents and fractures or morphologically depressed areas exposed to potential invasion of lava flows. The active protection consists on the containment and/or the deviation of lava flows through barriers, able to withstand lateral thrust of flow and high temperatures. Today, the barriers of earth (generally of unconnected material) have proved their reliability (Colombrita, 1984). However useful contribution will be certainly the study of innovative barriers which are more efficient and more easily erectable than those made of earth (Marsella et al., 2008). 77

Figure 4.

Representation of the air fall phenomenon.

1.5.5 EXPLOSIVE ERUPTION 1.5.5.1 Air fall deposits During an explosive eruption, the air fall deposits are formed by the accretion of clasts which fall by gravity from the eruptive column or which are thrown directly in area from crater, according to ballistic trajectories (Figure 4). They fall down to a distance which depends on their speed and the initial ejection angle. The largest pyroclastic fall in the environs of the emission point, the most fragmented ones at greater distance and the smallest ones can be transported by stratospheric winds. Generally, air fall deposits cover the topography with uniform thickness, but, because of their poor consistency, they are removed from the most steep slopes (>20–30◦ ) and accumulated in the valleys. During violent explosive eruptions (Plinian and sub-Plinian), large deposits of pumice cover an area of elliptical shape around the crater, which is elongated in the direction of wind. The ashes deposit after very long time reaching large distances, above it deposits of pyroclastic flows often follow. Contrary, moderately explosive eruptions, on the contrary, produce deposit of clasts fall, whose distribution is symmetrical around the crater, because the launches are not sufficiently high to be influenced by the wind. Generally, the thickness of air fall deposits decreases with the distance from the eruptive centre. The air fall deposits action on the ground level can be considered as a gravitational distributed load, which can be estimated as qG = ρ gh, where g is the gravity acceleration (9.81 ms−2 ), h is the deposit thickness (m), ρ is the deposit density (kg.m−3 ). The last one depends on the following factors: the composition of pyroclasts, their compactness, the deposit moisture and the subsequent rains. So, the deposit density is weather dependent: in dry conditions it ranges from 400 kg.m−3 to 1600 kg.m−3 , according to its compactness; in damp conditions it ranges from 800 kg.m−3 to 2000 kg.m−3 (Spence et al., 2005). The air fall deposits action on the roofs is similar to the snow load, so, with reference to the Italian technical code (M.D. 2008) the air fall deposits action on the roofs can be determine through the following relation:

where µ is the shape coefficient, function of the angle pitch (α), qG is the air fall load on the ground level and CE is the exposition coefficient which take into account the effect of the topography of the construction site (De Gregorio et al., 2010). In addition, to the relationship (1), for completing the model of the air fall deposits action, it is necessary to consider the high temperatures (200–400◦ C) of the clasts, which are able to produce important thermal degradation of the mechanical properties of the materials (Mazzolani et al., 2008 and 2009a). Some eruptions may send ashes into the stratosphere to heights of 10–30 km above the earth’s surface. Combined with the wind, they can spread more or less heavy materials relatively far from the volcano itself. Most of building damages due to ash falls occur when the ash load exceeds the strength of either the roof-supporting structures or material used to cover the structure (sheet metal, plywood, etc.). According to an American study of the U.S. Department of the Interior, dry ash presents a weight ranking from 4 to 7 kN/m3 , and rainwater can amplify it by 50 to 100%. If the ash becomes saturated by rain, it can reach more than 20 kN/m3 . So, ash loading may be considered as similar to a specific snow load but with some major differences: – being heavier, it is a much more severe loading case (Table 3); 78

Table 3. Density & load comparison, 10 cm of snow and 10 cm volcanic ash. Load type

Unit weight (kg/m3 )

Load (kPa)

New snow Damp new snow Settled snow Dry uncompacted ash Wet compacted ash

50–70 100–200 200–300 500–1,300 1 000–2 000

0.05–0.07 0.1–0.2 0.2–0.3 0.5–1.3 1.0–2.0

– ash doesn’t melt; – ash can fill gutters and draining pipes leading to collapse, especially after rainfalls. For a dry layer of ash about 10 cm thick, the extra load on a building can range from 0.4 to 0.7 kN/m2 ; a wet layer might reach 1.0 to 1.25 kN/m2 . In areas where snow load cases exist, a relative protection against ash falls may be expected but it depends highly on the location of the considered structure because snow loads vary with altitude and geographical position. Ash is dense, abrasive and chemically corrosive. Volcanic ash is a frequent volcanic hazard which can have wide reaching affects on populations due to its distribution in the atmosphere. Most impacts are disruptive rather than destructive; however it is the hazard which most frequently affects the most people. The size of the eruption and the wind speed and direction affect the extent of the distribution. As a result, populations are vulnerable to the impacts caused by volcanic ash. Several recent eruptions have illustrated the vulnerability of urban areas which receive only a few millimetres or centimetres of volcanic ash. This has been sufficient to cause disruption of transportation, electricity, water, sewage and storm water systems. Ash fall is one of the eruptive phenomena with greater risk for existing buildings and infrastructure, as the expected impact involves (with different levels of intensity) a very large area, which definition is strictly linked to the direction and intensity of the wind, as well as to the type of eruption. In the case of Vesuvius and Campi Flegrei, the scenarios show an increase of roof loads due to ash fall between 1000–3000 kg/sqm inside the red zone and between to 300–400 kg/sqm for distances up to 30 km from the vent. Different types of damage may also occur in distal areas (more than 100 km from the vent), where the ash deposits are not likely to cause structural problems to buildings, but still could affect transportation networks and HVAC systems (ashes infiltration in filters and ducts). In case of eruption of Campi Flegrei, the direction with higher risk is the whole urban center of Naples, where the population is more than twice the area of the villages around Vesuvius. Ash deposit on roads and transport networks can cause considerable damages especially in proximal areas, causing localized or extended interruptions with direct effects on emergency management. In accordance with the holistic approach to impact studies, this review encompasses the main sectors where studies have been undertaken concerning the impacts from volcanic ash fall and hence their vulnerability (Sword-Daniels, 2010). The work has been brought together and summarised with the intention to inform future studies on ash fall impacts research, and to provide an insight into some of the collective knowledge in this field. Although the review of studies is not exhaustive, the purpose is to review the main sectors that have been investigated and to use a whole-systems perspective to unearth the gaps in our knowledge and understanding of ash fall impacts and vulnerability. The engineering and volcanological literature has been searched for references to volcanic ash and its effects, impacts and management. Studies are multi multiple, and vary in quality and depth. The most informative studies have been summarised in brief, in the following sector-by-sector impacts. It can be seen that despite multiple impact-related studies, only a handful of sectors have thus-far been considered. Sectors are broad and include many subtopics within them; however there are many potential areas of research that remain little-explored. The following sectors have been studied for ash fall impacts and the main findings are summarised below from the following references: Baxter, 2006; Blong, 1981, 1984 & 2003; Casadevall, 1996; Cook et al., 1981; Dobran, 2006 & 2007; Frameworks Architects et al., 1996; Gordon et al., 2005; Horwell & Cowie et al., 2003; Inbar et al., 1995; Johnston, 1997a, b; Johnston et al., 2000 & 2004; Newnham et al., 2010; Shriever & Hansen, 1964; Spence et al., 1996; Spence et al., 2005; Stewart et al. 2006 & 2009; Wilson et al., 2008; Wilson et al., 2009a, b. 79

Figure 5.

Representation of the pyroclastic flows and surges phenomena.

1.5.5.2 Pyroclastic flows The pyroclastic flows and surges are the most dangerous phenomena produced during an explosive eruption. They are constituted by gas-solid dispersions with high or low concentration of particles respectively, which move along the surface under action of gravity. They are characterized by high temperatures and can be partly fluidized. In general, they are controlled by topography; channelled along the valleys, they fill the depressions (Figure 5). Pyroclastic flows can be generated either by the collapse of the eruptive column, or by a directional explosion for the slipping of a part of the volcano, or by a lateral explosion at the base of a lava dome. Pyroclastic flows are made of a mixture of gases with dispersed solid particles of various sizes. The modelling of the phenomenon is very complex, because it depends on a number of factors difficult to catch among them the mass eruption rate, the volcano topographic profile, the magma properties, such as the water content and the temperature at the crater exit. Aiming at examining the evolution of a pyroclastic flow, Todesco et al. (2002) adopted a model based on the solution of the Navier-Stokes generalized equations for a multiphase mixture, the latter being represented as a two-phase mixture composed of a homogeneous melt phase, made up of magma and crystals, and a gas phase, made up of water vapor. The mechanical and thermal non-equilibrium effects between gas and various particulates phases are considered. Pyroclastic flows can produce high damages to the built environment in areas near to the vent. Although they would have a limited action range, the effects can be critical because of the combination of mechanical impact and thermal stress on the vertical surfaces of buildings (Zuccaro, 2010a-d). The main damages come from the impact on openings, particularly vulnerable. In these cases, although not resulting a static failure of the building, a fire risk is associated with the flow passage inside the building following the crash of the openings. In the case of Vesuvius and Campi Flegrei, pyroclastic flows can cause lateral pressure impact within a range of 0.5 and 10 kPa, and thermal stresses ranging between 150 and 450◦ C. In Campi Flegrei, due to the probable location of the vent near to densely populated areas (including the west area of Naples), the impact of pyroclastic flows would be particularly serious, while in the case of Vesuvius is expected a decay of the initial power due to the distance of the built areas from the vent. In the structural analyses, it is possible to schematize the action of the pyroclastic flows as a uniformly distributed static pressure (Petrazzuoli & Zuccaro, 2004), with temperature ranges between 200 and 350◦ C (Giurioli et al., 2008). In particular, with reference to a sub-PlinianVesuvian eruption, the dynamic pressures produced by Vesuvian sub-Plinian event are determined (Esposti Ongaro et al., 2002). The pressure was calculated as a function of the angle of flows propagation α at growing distance from the vent, in undisturbed atmosphere above the aerodynamic ground plane at 5 and 15 m, as it is shown in Table 1. Esposti Ongaro et al. (2002) also determined the pressure corresponding to an angle α equal to 30–45◦ , at 5 and 10 m above the ground and at 4–5 km from the vent (Table 4). The obtained results are related to 2D models such as to the volcano transversal section. In the Exploris Project (Neri et al., 2007) a 4D model was developed, where Vesuvius is schematized with its real geometric dimensions, besides, in the flows modelling the variable time is included. In this case, for a sub-Plinian eruption, a pressure equal to 1–3 kPa at 7.5 km from the vent, with a temperature equal to 250◦ C was calculated. In addition, with reference to a sector of the town of Torre del Greco (6 km from Vesuvius), Zuccaro & Ianniello (2004) have analyzed the interaction of pyroclastic flows with buildings in an urban settlement. The generated turbulences produce an increment factor of pressure varying over the range [−3; +2]. In particular, with an angle α = 90◦ , the pressure equals 3–5 kPa. 80

Table 4. Dynamic pressure (kPa) in function of the angle of flow propagation α at growing distance from the vent. Distance from the vent α [◦ ]

2 km

4 km

6 km

30 45 90 180 360

– – 11.0 4.0 1.80

5.0 3.0 1.0 1.0 0.1

– – 1.0 0.5 0.0

The experience from the 1997 Montserrat eruption (Canary Islands) has indicated that a building can survive under moderate pyroclastic flows pressure (1–5 kPa) if it remains intact, while if one or more openings fail, allowing hot gas and ash to enter, the entire building is likely to be destroyed (Baxter et al., 2005). In this case the contents of the construction and any timber structure are likely to catch fire; at the same time the principal structural walls and roofing will suffer a combination of internal and external pressures, which will cause partial or total failure (Spence et al., 2004). In general, the first elements to reach the collapse are the glass windows and the shutters. However they can be easily protected by more resistant panels. Nevertheless, the lateral resistance of a building to pyroclastic flow strongly depends on the design criteria applied to resist ordinary load conditions: of course an earthquake-proof building presents larger strength and stiffness capabilities than a not earthquake-proof building. 1.5.5.3 Flying fragments The explosive eruptions also are able to produce flying fragments of pyroclasts defined bombs and missiles. The largest clasts are exploded directly from the crater according to pure ballistic trajectories. On the contrary, the smaller clasts can be sustained by convection in the eruptive column. Then they are thrown in the atmosphere from the main flow to fall or be transported along the mountainside in gravitational currents. The word missile can also relate to flying debris, not involved in the eruption, set in motion by pyroclastic flows. The law which regulates the movement in a vacuum of a volcanic fragment with ballistic trajectory is (Dobran, 2006 & 2007) R = (u02 · sin 2θ)/g, where R is the block ejection distance, u0 is the ejection velocity, θ is the initial ejection angle of the fragment from the horizontal and g is the acceleration of free fall. This formula is applicable for very large blocks for which the air drag has a negligible effect on the clasts trajectory. In this case, the most efficient ejection angle is 45◦ . On the contrary, the presence of a stratified atmosphere ensures that the optimum elevation angle is less than 45◦ . In addition, the θ optimum take-off angle from the eruptive column, and not from the vent, ranges between 47◦ and 50◦ for the larger blocks and in the range between 38◦ and 40◦ for the smaller ones. A separate examination is given for missiles (particulates, debris, stones, loose flower pots, dustbins, etc) generated by pyroclastic flows. In fact they are incorporated into the main current and added to the destructive impact. These missiles can be related to the aerodynamics of flying debris with respect to cyclone and wind storm (Wills et al., 1998; Spence et al., 2005). Bombs and missiles cause damage which depends on the kinetic energy and the vulnerability of the struck object. A flying fragment can impact the roofing or the walls of a building, but, in particular, it can hit the most vulnerable parts, like the openings. If a building collapses, it is assumed that all occupants are killed. If buildings do not collapse, the main factor which governs vulnerability is the resistance of openings, especially the glass panes or the shutters which can prevent the hot ash to enter. Contrary, possible consequent fires and/or breathing difficulties for people inside can arise (Spence et al., 2005). Several studies aim at the evaluation of the speed of bombs and missiles, produced by explosive volcanic eruption, but the analysis of the effects of these flying objects buildings is not very much developed. Spence et al. (2005) have examined the window failure produced by missiles generated by pyroclastic flows. The probability of impact of flying debris on windows depends on the flow velocity, the flow density, the density of potential missiles in the area surrounding the volcano, as well as the surface and the orientation of windows. Missile impact causes failure when a fragment has a 81

sufficient kinetic energy to break the window. It is assumed that the energy required to break a glass panel equals the energy absorbed by the panel in its elastic deformation up to the point of failure. For a Young’s Modulus of 65,000 MPa for glass, the energy required to break the window ranges from 8 to 20 J for typical and large window panes of 3–4 mm thickness. Therefore, it can be assumed that, at any given flow velocity, missiles with a kinetic energy less than 8 J will break few windows, missiles with a kinetic energy between 8 and 20 J will break some windows, and missiles with a kinetic energy above 20 J will break many windows. 1.5.6 SECONDARY EVENTS 1.5.6.1 Lahars The lahars are a relevant risk factor for buildings and structures in volcanic areas. The same phenomenon may have specific characteristics depending on some variables (Zuccaro, 2010a-d). After an eruptive event, especially if explosive, the thermic change in the proximity of the volcano often produces rain. Combined with the pyroclast of poor coherence, with the volcanic high slope of (20–30◦ ) and the distinctive seismicity of the eruptive phase, the rain can cause the mobilization of the volcanic deposits and the consequent formation of mudslide and lahar. The term lahar has an Indonesian origin and indicates any type of muddy flow containing volcanic material. Lahar and mudslide are extremely dangerous because of their high kinetic energy, they being generally characterized by speed of the order of some tens kilometers per hour up to above 100 km/h (Carlino, 2001). Lahars with a high water content can be assimilated to the Newtonian fluids, whereas the lahar with a high concentration (high solid/water ratio) are assimilated to Bingham fluids. The Newtonian fluids freely move under the gravity until the critical stability condition arises, while the Bingham fluids, because of the greater viscosity, offer some resistance to the motion and so they need an additional strain to mobilize. At the same conditions in terms of gradient, granulometry, etc, the motion capacity of a lahar depends on the water content and the ability of the flow to lose or to absorb water during the way. The effects of lahars on the constructions are comparable to those ones produced by the debris flows. Damage to buildings caused by lahars can be connected to different factors. Hydrostatic and dynamic strength determine the amount of lateral forces that can bring to failure and collapse of technical elements such as openings and cladding. The density and velocity of the flow determines the magnitude of dynamic forces, while hydrostatic forces depend on the height and composition of the flow. Minor mudslides can cause abrasions of the external finishing of buildings and damage to surfaces and furnishings in case of penetration of the flow in the interior. Local effects may be caused by the transport of medium and large debris, rocks, but also uprooted trees, motorbikes and cars that can act as missiles on buildings exposed. Depending on the magnitude of the phenomenon and orographic conditions of the site, buildings of medium-low height can be buried by lahar. Further damage can be caused to structural parts of both masonry and reinforced concrete buildings, causing even serious cracks and damages, with structural failure involving foundations, due to erosion and soil liquefaction. Structural and non-structural metal elements can also be seriously damaged by the acidity of the flow. The response of structures and buildings technical elements to the action of lateral forces produced by lahars depends mainly on construction type and materials employed, as well as specific characteristics such as size in plan and elevation, number, size and position of openings, spatial distribution and presence of protective elements around the building able to divert the flow, etc. Faella and Nigro (2002) have analysed the structural and non-structural damage in the buildings impacted by the debris flows, during the hydrogeological disaster of May 1998 in Campania (Italy). The damage is significantly different in relation with: the position of the construction, the impact direction, the level of kinetic energy of the flow and the structural typology. This study, for masonry and reinforced concrete buildings, has identified the main collapse mechanisms and the debris hydrodynamic horizontal pressures, which assume the values of 150, 73.5 and 37.5 kNm−2 , for speeds of 10, 7 and 5 ms−1 , respectively. With particular reference to Vesuvius case, these pressures can be adopted as lahars actions on the constructions. This because the debris velocities furnished by Faella and Nigro result comparable to those ones calculated by Vallario (1994), with reference to of the possible lahar produced by a 82

Vesuvian eruption on the Cavallo riverbed (Torre del Greco city, south slope of the Vesuvius). In fact, they range between 3.94 and 10.14 ms−1 . However the lahars, as respect to the debris flows, present the additional variable of the temperature, which causes substantial degradation of mechanical properties of construction materials. Actually, the temperature of lahars is widely variable. It depends on the typology and the quantity of the erupted materials and on the time between the deposit and the mobilization. Obviously, at the same conditions, as far as the time passes, temperatures will be reduced. 1.5.6.2 Tsunami Tsunami is a Japanese word which means “wave (name) in the port (tsu)” and, since antiquity, it describes the phenomena of the rogue waves which produce devastating effects on the coast. It can present with an initial and temporary withdrawal of the waters, or with a flood which can show like a tide which rapidly comes in, like a waves trains or like a water wall. Among the cause of a tsunami, the volcanic eruptions are present. In particular, the anomalous waves can be produced by massive pyroclastic flows which reach the sea. That is the case of the explosive eruption of the Krakatua volcano (1883), in the Sunda Straits, between Sumatra and Java, that produced a large tsunami that killed more than 30,000 people living in the coastal villages of the Straits. Tsunami hazard in the Gulf of Naples is mostly related to the activity of the main volcanic structures found in the area (Vesuvius, Phlegrean Fields, Ischia) and also far to the south (Stromboli). With reference to Vesuvius, Historical information as well as numerical scenarios indicates that the tsunami hazard for the Gulf of Naples is not frequent, but not at all negligible. Combining this with the high vulnerability related to the huge population density and the very complex urbanization, it turns out that the tsunami risk assessment for the area is a problem whose treatment cannot be delayed any further. According to historical documents, anomalous sea oscillations and waves in the Gulf of Naples were observed not only concomitantly of the largest eruptions (79AD and 1631), but also to some the smaller events, such as the cases of 14 May 1698, 17 May 1813 and 4 April 1906 (Italian Tsunami Catalogue, Tinti et al., 2007; Tinti et al., 2010). According to Palermo et al. (2007), the actions produced by a tsunami on a construction can be grouped into two loading combinations: Initial Impact and Post-Impact Flow. The Initial Impact includes surges and debris impact force components. The surge force Fs is produced by the impact of the flood waves on the structures, while the debris force Fi is relating to impact structures due to significant debris (such as vehicles, components of buildings and drift wood) which the waves can transport. After the initial impact, a proposed second loading combination results, namely, the Post-Impact Flow. During this phase hydrodynamic (drag) forces FD are exerted on structures due to continuous flow of water around the structure. In addition, the inundation gives raise to hydrostatic forces FHS . The hydrostatic forces can occur on both the exterior and interior of the structure. The latter depends on the degree of damage sustained during the initial impact. Further, the structure is subject to debris from floating objects being transported by the moving body of water. Therefore, the second phase of loading includes Hydrodynamic and Hydrostatic forces, Debris Impact forces, and Buoyancy forces that result from the structure being submerged after the initial impact. 1.5.6.3 Volcanic earthquakes All volcanic eruptions are accompanied by local seismic activity, as it is testified by the swarms registered in occurrence of large eruptions (Benoit and Mc Nutt, 1996). The seismic events that characterize an eruptive phenomenon can be generally considered of low to medium intensity. Nevertheless, the cumulative damage caused by the sequence of earthquakes in various stages of the eruption produces a progressive increase in the level of expected damage (Zuccaro, 2010a-d). According to the sequence of phenomena characterizing the eruptive event, more conditions can occur and raise the damage caused by the earthquake. In particular, the ash fall creates a progressive overload on the roofs, and even when it doesn’t result in a partial collapse of the floor; it brings to an increase of reactive mass of the building, thus modifying the response to seismic action. The building types with high vulnerability, with particular reference to masonry structures, would then suffer more damage than for a single event comparable to the maximum intensity expected in case of a Sub-Plinian eruption. 83

While tectonic earthquakes are generally related to a shear-faulting mechanism, volcanic earthquakes may involve tensile, isotropic, and/or shear rock fractures, driven by the percolation of high-temperature fluids/gases or directly by the magma-ascent mechanism (Festa et al., 2004). In particular, seisms related to volcanic activity are of two general categories: volcano-tectonic earthquakes and long period earthquakes (Chouet, 1996). The first category, volcano-tectonic earthquakes, is produced, on one side, by stress changes in solid rock due to the injection or withdrawal of magma through the fractures and, on one other side, by tectonic displacements. They can cause land to subside and can produce large ground cracks. Volcano-tectonic earthquakes do not indicate that the volcano will be erupting but can occur at anytime. The second category, long period earthquakes, is only produced by the injection of magma through surrounding rock. Therefore they are a result of pressure changes during the unsteady transport of the magma. When pressure of the magma injection is high a lot of earthquakes are produced. This type of activity indicates that a volcano is about to erupt. Scientists use seismographs to record the signal from these earthquakes. This signal is known as volcanic tremor. The intensity of a volcanic earthquake is a function of the entity of the eruptive event. For example, in the case of Vesuvius, during several days before the 79AD Plinian event, big earth tremors. Afterwards, the seismic crisis preceding the 1631 Sub-Plinian Vesuvius eruption has been characterized by earthquake intensity equal to 4.0 degree on the Richter scale, temporally limited to some hours before the eruption (Cubellis and Marturano, 2006). After this eruption up to recent times the earthquakes were generally of low-moderate energy and related to eruptive activity. The most dangerous occurred on 15 June 1794 during the lateral eruption which destroyed the town of Torre del Greco. The shocks caused damage to buildings in the Vesuvian area and shattered window panes in Naples. With reference to the Vesuvius case, a comparison between tectonic and volcanic earthquakes has been conducted, through two real seismograms recorded in the Vesuvian area: the tectonic one of Irpinia (Avellino, Campania, Italy) on the 23 November 1980 and the volcanic Vesuvian one occurred in October 1999 (Mazzolani et al., 2009b and 2010). The first one with a magnitude of 6.9 and an epicentre about 30 km distant from the sea surface. The latter one with 3.6 magnitude and an epicentre 3.8 km distant from the sea surface. The substantial difference between these earthquakes is related to the peak maximum acceleration frequencies: the peaks of the tectonic and volcanic response spectra, equal to 0.37 g and 0.05 g respectively, occurs next to periods T equal to 0.40 s and 0.14 s respectively. According to the new technical Italian code (MD, 2008), the building fundamental period T1 can be calculated through the formula T1 = C1 · H 3/4 , where: C1 is equal to 0.05 and 0.075 for masonry and RC buildings, respectively, and H is the construction total height. So for masonry buildings the period T1 almost ranges from 0.11 to 0.28 s for heights between 3 and 10 m, while for RC buildings it almost ranges from 0.29 to 0.6 s for height between 6 and 16 m. This would implicate that masonry buildings suffered much more the volcanic earthquake, since its frequency is close to that of the building, while RC buildings are sensible to tectonic earthquakes.

1.5.7 MITIGATION STRATEGIES 1.5.7.1 Ash fall Mitigation strategies, beside the need to develop an operational plan for the removal of ash on roofs and transport networks, mainly concern the repairing and reinforcement of roofing systems in order to increase the load carrying capacity (see Table 5). Pitched roofs with wooden or steel structure, reducing the deposits of ashes, would be at risk only in proximal areas where the surface of the cover present disconnections or missing parts. In this case, given the adequate inherent fire resistance of commonly used coating materials (typically clay tiles or panels of steel sheet) is enough to replace the missing elements in order to prevent the passage of hot ashes under the roof covering. In case of flat roofs it is possible to identify two main types of intervention: the reinforcement of the roof slab in order to increase the resistance according to the expected overload, or the realization of a sloped roof over the existing one. 84

Table 5. Vulnerability of common roofing typologies. Vulnerabilty classes

Roofing type

Load (kPa)

A_rf

Weak pitched wooden roof

2,0

50%

B_ rf

Standard wooden flat roof Flat floor with steel beams and brick vaults Sap floors

3,0

50%

C1_rf

Flat floor with steel beams and hollow bricks R.C flat slab (more than 20 year old)

5,0

60%

C2_rf

R.C flat slab (less than 20 year old) Last generation R.C. flat slab

7,0

51%

D_ rf

Last generation R.R. pitched slab Last generation steel pitched roof

12,0

50%

Figure 6. Technical solution for the mitigation of ash fall impact on roofs through the employ of CFS structures (Alborelli, 2009).

Collapse probability

Figure 7. Technical solution for the protection of openings.

In the first case, it is necessary to define the characteristic flexural strength of different types of existing roofs in areas at risk (concrete and bricks, steel or wooden beams and hollow bricks or brick vaults, “Sap” floors, etc.), thus determining the capacity to withstand to overloads produced by ash. It is then possible to apply conventional technologies, such as integration of reinforced concrete slabs placed on the existing floors and connected to existing beams, or innovative solutions, including for instance the use of FRP (Fiber Reinforced Polymers) and FRCM (Fiber Reinforced Cementitious Matrix) systems for reinforcement of beams and joists. The main advantages of such interventions include the possibility of not modifying the existing roofing system. In the second case a very effective solution is to build truss or lattice structure in CFS (Cold Formed Steel) on top of the existing roof, in order to create a sloped surface. The mechanical properties and lightness of CFS structures allow the realization of a strong roofing system without a high overload on the underlying structure (Figure 6). The coating can be made of steel sheet, with the possibility of providing additional layers in order to offer additional benefits to the intervention of structural retrofit, such as the insertion of insulation or micro-ventilation system for energy conservation, or the integration of photovoltaic thin film for the production of electricity. Such actions may be also connected with housing refurbishment programs, allowing for instance the increase of building volume for intervention of volcanic and seismic mitigation. The realization of lightweight structures for protection from ash fall may be an appropriate solution not only for buildings but also for the several areas of historic and artistic interest (such as Pompeii, Herculaneum, Oplonti, Stabiae, etc.), which might be seriously compromised after an eruption of Vesuvius. In these areas, 85

however, the mitigation may be invasive in terms of visual impact, and it is possible to develop provisional removable shingles. An alternative to steel roofing is the realization of UHPC (Ultra High Performance Concrete) shells, characterized by very high mechanical properties, durability, resistance to high temperatures and fire, with very low thickness required (up to 2 cm for spans of 5 m), offering effective and innovative technical solutions in terms of aesthetics and design. 1.5.7.2 Pyroclastic flow Mitigation strategies mainly concern the reinforcement of infill panels in R.C. buildings and measures for the protection of openings (Figure 7). When reinforcing infill panels, the goal is to increase the impact resistance while withstanding the high temperatures produced by the flow. Currently used techniques for the seismic reinforcement of infill panels are generally effective to prevent them from breaking due to pyroclastic flow, however, as noticed above, the employ of currently used technologies that are particularly sensitive to temperature should be avoided. In the absence of specific constraints to envelope system modification, the goal of increasing infill panel impact resistance may be achieved by overlaying existing facades with coatings made of advanced materials offering high thermal and mechanical performances in very low thickness. It is the case of UHPC (Ultra High Performance Concrete) components, which can be cast in very large panels and show high durability and resistance to aggressive environment. These operations allow also to obtain additional performances, such as the increase of shear strength in the plane, where the panel is placed within the structural grid, or the increase of thermal resistance, where combined with a layer of insulation or with a ventilated facade system. The use of low thickness UHPC panels may also be suitable for the construction of temporary and removable systems to protect archaeological areas and sites of historical and artistic interest subject to the risk of pyroclastic flows. Protection of openings is an essential mitigation measure in relation to pyroclastic flows, as it allows minimizing the risk of fire related to penetration of the flow inside the buildings. At the same time the technical solutions provided should be able to withstand the mechanical stresses related to the pressure of the flow itself, but also to the potential presence of debris that can impact as “bullets” on openings surface. Borrowing technologies used in tropical areas for hurricanes protection it is possible to define different solutions, made with removable components or integrated into the shutting systems. In the first case, it is possible to overlay steel or Kevlar sheet to existing openings, anchored along the external perimeter. Protection systems integrated into the shutting systems, unlike the removable panels, are not always able to assure an effective response to the impact of the flow, but are suitable for medium ranges of temperature and pressure or for short exposition time. It is also possible to apply special protective films on glass surfaces that can provide protection from fire and explosion. Fire safety shutters, steel or aluminum associate the heat resistance with adequate mechanical strength. In some cases, a combination of protective films and special shutters should be provided, in order to reach the required levels of temperature resistance and mechanical strength. 1.5.7.3 Lahar Generally speaking, structures, infill panels and ground floor openings are the technical elements most at risk in case of lahars. The reinforcement of these elements yet does not guarantee the survival of the building in case of direct impact with mudslide and debris, especially in the case of compact urban areas, where a “tunnel effect” can increase speed and height of the flow after the passage inside particularly narrow roads. For this reason the most effective mitigation strategies are related to environmental engineering interventions, to be made in risk prone areas and designed to contain or divert lahars. Measures such as retention basins, alternative artificial canals, high-strength reinforced concrete containing structures, may be appropriate solutions to mitigate risk from lahars, reducing the entity of the phenomenon in residential areas and increasing the probability of survival of the buildings. 1.5.7.4 Earthquake Generally speaking, considering the high seismic vulnerability levels and the construction density in Vesuvius area, cost-effective mitigation measures should be provided. It is possible to choose 86

cheap and reliable technical solutions (such as iron chains in masonry buildings, the insertion of infill panels or resistant elements in soft floors of reinforced concrete buildings), but also to adopt, in case of seismic reinforcement, specific solutions able to respond effectively also to other volcanic phenomena, such as pyroclastic flows or ash fall. In this context, one solution is the construction of pitched roofs by overlapping light structures in CFS (Cold Formed Steel). This allows to chain vertical structures by increasing the resistance to seismic actions (box behaviour) and simultaneously prevent the deposit of ashes and the structural risks related to overloading of the roof, also in consideration of a possible earthquake following the ash fall phase. At the same time, should be avoided the employ of widely used reinforcement systems not satisfying the conditions of volcanic risk, such as FRP (Fiber Reinforced Polymers) in proximal areas, whose effectiveness is seriously reduced by the possible impact of pyroclastic flows. In fact, the high temperatures produced could affect the polymer matrix, whose physical and mechanical properties degrade in range above 60– 80◦ C, with the consequent failure of the system caused by the loss of adhesion of the reinforcement to the walls. In this cases alternative technologies should be adopted, compatible with the environmental conditions related to a volcanic event, such as FRCM (Fiber Reinforced Cementitious Matrix) systems, able to withstand high temperatures while preserving the mechanical properties. Global mitigation strategies related to seismic risk in case of a volcanic event may include planning for widespread interventions, defining the areas that require priority actions, such as the building curtains facing the main transport routes and escape routes identified by the Civil Protection Emergency Plan, in order to ensure safe evacuation routes during unrest phase, characterized by increased seismic intensity. 1.5.8 IMPACTS OF ASH FALL 1.5.8.1 Health The health effects of volcanic ash depend on the grain size, mineralogical composition and chemical coatings on the surface of the ash particles. The reporting of asthma-type symptoms following eruptions is variable and may depend on the ash characteristics and duration of exposure, as well as the social context and willingness to report such personal information. It is thought that pre-existing asthma or bronchitis suffers are at greater risk of suffering from respiratory impairment than those without a history of respiratory problems, however the health impacts of volcanic ash are inconclusive. 1.5.8.2 Structures Wet tephra is known to have a greater load than dry tephra, and various observations from historical volcanic eruptions in the 20th Century have shown variable critical thicknesses of tephra under which roofs have collapsed; observations vary from 75–300 mm thickness. Building styles, materials and general condition are important factors in the live load bearing capacity. Increasing the cross-sectional area of trusses and reducing the truss span, increases the threshold value that can be withstood by a building. The observations of roof pitch have ambiguous results. Accelerated corrosion of metal roofing is known to occur following ash fall events. 1.5.8.3 Agriculture & Environment Ash reduces water infiltration in the ground and increases surface albedo. When ash falls on leaves it reduces photosynthesis and can cause collapse and crushing of plants and crop. Ingesting ash may be harmful to livestock; causing abrasion of the teeth and in cases of high fluorine content, fluorine poisoning. Acid rain, a result of ash and rain, is capable of burning crops and leaves. 1.5.8.4 Water & Contamination Numerous studies have been undertaken on the composition of volcanic ash leachates and many soluble components have been detected. Aluminium, Iron and Manganese, Fluorine and Sulphate are amongst them. Ash also causes a lowering of the pH level. These contaminants are not thought to pose health risks, but rather cause impacts on potable water supplies, and associated issues of corrosion, scale deposition and staining may affect distribution networks. 87

Ash fall is known to increase the turbidity of water, which at heightened levels can prevent disin disinfection treatments from working effectively. Studies have also focused on recording the turbidity of water supplies following eruptions; which increases with the amount of ash entrained into water systems. Ash also clogs water networks; blocking irrigation systems and clogging the intakes at water processing plants. This build-up of ash in water infrastructure systems can cause extensive corrosion, abrasion and damage or failure. Ash clean-up commonly involves the hosing down of surfaces, and can therefore cause strain on water supplies. 1.5.8.5 Electrical distribution networks & computers Ash in combination with rain is known to cause power outages on electrical distribution systems. This is a result of electrical flashover which occurs due to the conductive properties of ash. Ash falling on overhead power lines can also cause breakages due to the weight of the ash. Computers have been tested for failure under ash fall conditions. Abrasion was evident and failures occurred, more prevalently if the conditions were also humid. 1.5.8.6 Aircraft Ash is abrasive and causes abrasion damage to several parts of the aircraft, it also blocks intakes and re-melts and accumulates in engines that are running at temperatures of hundreds of degrees Celsius. This can cause a loss engine power and can require entire engine replacement following an encounter with an ash cloud. Some mitigation and prevention measures have been developed including: avoidance of flying near ash clouds and development of ash warning systems to aviation industries, setting engines to low power during an encounter with ash, and covering grounded planes with protective sheeting across windows and openings. Costs of damage, rerouting, delays, cancellations and clean-up are extensive. 1.5.8.7 Land transport infrastructure Ash is known to reduce traction on roads and reduce visibility, making driving on roads more dangerous and disrupting traffic networks. This also restricts access for emergency services which is of critical concern during a crisis. However few studies have focused on the impact of volcanic ash on land transport infrastructure. 1.5.8.8 Emergency management Some effective techniques for the management of ash have been developed in preparedness and planning for ash fall events, public education and training, cleaning methods and apparatus and also preventative measures to ensure damage limitation. 1.5.8.9 Cost impact assessment The costs of ash fall disruption have been estimated where possible in a few studies. The cost implications depend on the industry affected, distribution of the ash and the duration of the event. However ash fall is known to cause extensive losses in business interruption, exemplified in the eruption of Eyjafjallajökull in Iceland in April 2010, with reported airline losses of €1.5–2.5 billion (The Daily Telegraph, 2010). Following the Ruapehu eruption in 1994/95 the Rangipo hydroelectric power station was damaged by ash carried in the river and cost an estimated $12 million NZD in loss of power generation and $6 million NZD in replacing damaged blades (Johnston et al., 2000). Cook et al. (1981) estimated crop losses from Mount St Helens ash fall at $100 million in 1980. Ash is also known to have caused damage to aircraft engines; mechanical parts; agricultural industries from crop or livestock losses; losses in the tourism industry; costs of emergency response and clean-up, and many more. Many losses are unquantifiable. 1.5.9 IMPACT OF PYROCLASTIC FLOWS The collapse of the sustained column and the consequent pyroclastic flow (PF) are very frequent phenomena in explosive eruptions. In this case the magmatic material erupted is composed by a 88

mixture of molten and solid pyroclasts in a continuous gas phase. This mixture is the product of gas exsolution and magma fragmentation processes that occur during the magma ascent from the deep magma reservoir to the ground surfaces. When this mixture of gas and pyroclasts penetrates into the cooler atmosphere it mixes up with the surrounding air forming a volcanic jet. Crater geometry, outlet pressure and velocity, temperature and gas content control the effectiveness of the mixing and, therefore, the global evolution of the explosive event. If the mixture at the top of the jet is reduced below the atmospheric density, then the eruption forms a convective buoyant plume called “Plinian” column. Otherwise the eruptive mixture collapses forming a ground PF that can propagate to great distance from the vent. The action on the vertical surfaces of buildings affected by the flow is a combination of impact and thermal stress, proportional to its mass and velocity. Pressure and temperature values vary depending on the characteristics of the eruption column and on the morphology of the invaded areas. In case of a Sub-Plinan I eruption the pressure can reach 10 kPa and the temperature can reach up to 400◦ C. However, considering the way of propagation of the flow within the territory, velocity and temperature values are not uniform, but weaker in lateral areas of the cloud, generally decreasing depending on distance from the vent. This means that not all buildings struck by the flow are destroyed, but it is possible to identify levels of damage as a function of impact characteristics and building vulnerability. 1.5.9.1 Building behaviour shouted by pyroclastyc flow 1.5.9.1.1 Background Damage resulting from the impact of pyroclastic flows on buildings depends on the combination of several factors: the duration of the phenomenon, the temperature of the flow and pressure produced by the impact (Zuccaro, 2010a-d). In general, the impact of pyroclastic flows can be classified into three main categories: a) The values of pressure and temperature are likely to damage the structure, until partial or total collapse b) The values of pressure and temperature are not likely to damage the structure, but there is a breakthrough of non-structural parts (window frames or infill panels) that allows the penetration of the flow into the building c) For lower values of pressure and temperature none of the building technical elements is expected to collapse, but the difference between the external and internal pressure causes the infiltration of the flow inside the building. While it is clear that in the first case the damage is very serious, it should be noted that in the other two cases, the flow infiltration can lead to the destruction of the building, mainly in case of breakage, as the strong internal pressure caused by flow infiltration can “inflate” the building causing the break of the roof or windows to the outside. It has to be considered, finally, the possibility that the high temperature of the flow entering the building could trigger fires, which could destroy the building even in the absence of mechanical damage. For these reasons we must distinguish three different typologies of vulnerability against three different types of expected damage. a) the vulnerability of the major elements (masonry walls, frame) b) the vulnerability of non-bearing (coverage, cladding) c) understanding vulnerability as “permeability” to infiltration, and expressed as ACH (Air Change for Hour) 1.5.9.1.2 Vulnerability of bearing elements The most significant parameter in this case is the dynamic pressure, whereas the temperature is less decisive. The evaluation of the building structure vulnerability to pyroclastic flow actions requires the estimation of the limit horizontal pressure at the collapse state of the standard buildings. A vulnerability analysis has been carried out by means of fundamental theorems of limit state analysis applied to R.C. frames and to the masonry walls, ((Zuccaro, 2010a–d, Spence et al 2004). It should be considered that the lateral pressure caused by the flow is quite different from seismic action. Pyroclastic flow action is not cyclical; therefore the ductility as energy dissipation capacity is 89

Table 6. Pyroclastic Flows – Structural Classification. Type

Description

Ap

Weak Masonry Buildings of 3–4 storeys with deformable floor. Weak or strong Masonry Buildings with more then 4 storeys. Medium Masonry Buildings of 1–2 storeys with deformable floor. Strong Masonry Buildings of 3 or more storeys with rigid floor. Strong Masonry Buildings of 1–2 storeys with rigid floor. Non aseismic r. c. buildings of more than 6 storeys (High). Non aseismic r. c. buildings of 4–6 storeys (Medium). Non aseismic r.c. buildings of 1–3 storeys (Low)

Bp Cp Dp Ep Fp

less important. Also, unlike the case of earthquake, mass is not directly proportional to lateral action, but plays a stabilizing function. The slenderness of the building is a factor strongly conditioning the level of vulnerability. It has been investigated the behavior of several sample buildings loaded by lateral increasing pressure, exchanging the typological and geometrical characteristics of the buildings, and computing the collapse values. The results of numerical analysis show a significantly different behavior between masonry and reinforced concrete structures, thus suggesting the definition of two separate vulnerability scales). Were therefore identified three classes of vulnerability (A, B, C) for masonry structures and three for R.C. structures (D, E, F), defining the buildings assignment criteria and the collapse probability as a function of lateral pressure by flow (Table 6). 1.5.10 VULNERABILITY, RISK, AND DAMAGE ASSESSMENT Identifying and assessing hazards and risks consist of three steps and questions: a) where and how does the hazardous volcanic process occur? This requires the study of thematic maps (geology, topography, population, city plans), archives on past catastrophic events, aerial photographs and satellite images, geodetic surveys and DEMs, and mapping the past extent and path of the volcanic flows and tephra-fall deposits on and around the active volcano. b) How large and how often does any given volcanic hazardous phenomenon occur? What is its magnitude and frequency? Mapping the extent of volcanic deposits, estimating the volume of the deposits, and assessing the path of the volcanic flow are essential tasks for computing magnitude and frequency. This implies detailed hazard-zone mapping for each hazardous process (lava flow, tephra fall, pyroclastic flow, lahars, etc.) and for each of the eruption scenarios, which is based on the past and present behaviour of the volcano. Mapping is best carried out by using statistical approach and modelling. Geomorphic surveys with the aid of satellite imagery form a logical starting point for natural hazard zoning. Geomorphic hazard zonation recognizes old deposits, maps flow paths and delineates hazard zones, which are primary inputs in elaborating eruption scenarios. The second step requires modelling based on semi-empirical codes and on numerical codes, which enable us to delineate the areas likely to be affected in the case of an eruption or a non volcanic crisis (e.g. debris avalanche, rain-triggered lahars, flash floods in cities, etc;). An alternative estimate of hazard zones can be obtained with the aid of mathematical models that simulate the evolution of volcanic phenomena and compute the effects at ground level, allowing the estimation of the area affected by an event according to a certain scenario. Geomorphic and hydrologic parameters are critical input requirements for the use of DEMs and GIS in longterm planning. The use of DEMs and of simulation models such as LAHARZ and FLOW3D have enabled Iverson et al. (1998), Pareschi et al. (2000) and Sheridan et al. (2001) to gauge volcanic flow hazards in densely populated areas around Mt Rainier, Vesuvius and Popocatepetl volcanoes, respectively. c) The third step of risk assessment requires the development of a series of scenarios in which eruption magnitudes, hazard types, composite risk zonation and the vulnerability of people 90

and infrastructure are adequately considered. Eruption scenarios are useful for preparation of emergency plans and long-term land-use planning. How can we protect people, communities, and elements at stakes? The volcanic risk should be (see Table 7): 1) analysed in terms of frequency and magnitude in order to determine expected damages on housing and infrastructure or lifelines; 2) evaluated in terms of a cost and benefit analysis (the value of any given element at risk with and without protection); 3) counteracted by land use regulations and careful city planning; 4) dealt with civil works for protection (e.g. dams), mitigation procedures and contingency planning (surveillance network, shelters, roads, radio links, etc.) in case of emergency. 1.5.11 STRUCTURAL ASPECTS 1.5.11.1 Passive and active actions Passive protection consists in educating people how to behave in the case of eruption or earthquake, in preparing people to evacuate in advance of a threatening eruption, and in increasing the knowledge of volcanic activity (education programme at school and through workshops) and the awareness of danger. Active protection consists in designing civil works against the effects of volcanic flows: diverting lava flows (e.g. Etna in 1983), sabo dams filtering lahars, shelters or bunkers against pyroclastic flows, long-lasting, reinforced (steel roof) shelters for protecting people away from the harmful effects of tephra fallout, etc. 1.5.11.2 Strategies for reducing the effects of volcanic phenomena or damages Blong (2000) provides a few strategies and strengthening designs in response to the principal volcanic hazard types (see Table 8). 1.5.12 RESEARCH ACTIVITY AND/OR GUIDELINES 1.5.12.1 With respect to the state-of-the-art in volcanology A better understanding of the eruptive behaviour has been gained through well equipped volcanoes, which have been used as laboratories where surveillance techniques are combined with detailed petrological studies of eruptive products and process modelling (e.g., case study of Montserrat). New petrological and geochemical tools have enabled researchers to better decipher the ascent of magmas within the plumbing system of the volcano and to combine this data with experimental petrology. Statistical (deterministic and stochastic) approach of hazard and risk assessment: quantifying long and short-term volcanic hazard and building up a common strategy (e.g. the probabilistic volcanic hazard assessment PVHA, Marzocchi et al., 2007). Analogue and numerical modelling: Iverson et al. (1998) automated hazard zone delineation by embedding the predictive equations in a GIS computer program that uses a DEM of topography. The simulation model LAHARZ provides a rapid, automated means of applying predictive equations to regions around edifices and comparing the results with the hazard-zone boundaries established in the field by mapping flow deposits. Pareschi et al. (2000) used a computer simulation approach to deal with ongoing volcanic hazards controlled by topography, such as lava flows. A maximum slope-statistical approach allows the authors to assign the lava vent and to estimate the zonation of hazards using a map superimposed on a geo-referenced image of Mt Etna and other GIS layers. Sheridan et al. (2001) have used DEM- and GIS-based computer models for simulating lahars and pyroclastic flows to gauge volcanic hazards at Popocatepetl in Mexico, in addition to a detailed survey of the past eruptive history and a close monitoring of the present activity. The assessment of physical vulnerability has been undertaken at the scale of a city and city block by using either models of flow impacts on housing and lifelines (e.g. Vesuvius) or in situ geotechnical tests of construction material within and near buildings (Delaite et al., 2005). 91

Table 7. Elements to be accounted for in vulnerability, risk, and damage assessment (housing, infrastructure, lifelines, people and civil authorities) in case of eruptive crisis. Vulnerable elements

Description

Meaning

Housing/Land use Residential; educational (primary, secondary, university); commercial (supermarkets, shops); institutional (city hall, district, region); religious sites, cultural and sporting facilities, etc.

Type Construction material; construction quality, number of floors, roof type, wall (principal), doors, windows; number of dwellers; cost of construction.

Role Value with respect to local and regional development; role in district or in city block; role of authorities, as perceived, as exerted; communication network and decision taking process.

Infrastructure Roads (sealed and unsealed), bridges, railway, airport, control points.

Types and tests Material type to be identified (size), mechanical tests: impact strain (uniaxial, punctual, dynamic pressure), yield strength.

Role Value with respect to local material and to mitigation procedures in case of expected or measured damage.

Networks Fluids (gas, electricity, phone, oil), Internet, network of decision making process and chain of command (council, authorities, city hall and region council)

Dysfunctions In case of eruptive or non-eruptive crisis: failure of networks, missing or ill-given orders for evacuation

Factors Physical (e.g. effects of lahars or pyroclastic flows) Technical (dilapidated, defects) Political: failing authorities or failing chain of command

“Natural” areas Gardens and parks Sporting areas (golf, tennis court, fields and running tracks)

Public use Distinct effects according to season, weekdays, day and/or night time

Factors and consequences Temporary or almost permanent occupation, physical abilities of dwellers in case of alert: consequences on injuries and deaths.

People Men, women, children, elderly, social and professional categories.

Characteristics Pattern of spatial distribution, social and economic pattern, age, level of education and culture

Assessment method Survey and interviews for assessing knowledge and perception of risk and level of preparedness and consciousness

Civil authorities National institutions, decentralized state services (actions for mitigation or emergency procedures), territorial and city councils, civil defence bodies

Characteristics Existing tools for management and education: procedures, policies, relief planning, warning dissemination to exposed people, information for mitigation procedures

Assessment methods Social survey and interviews among the decision makers and civil religious authorities, and local leaders.

92

Table 8. Principal volcanic hazard types. Volcanic hazard

Risk reduction strategy

Tephra fall

Use steeper roof pitches (>45◦ ), short spans in roof structure, simple roof designs, and roof-sheeting profiles with fluorocarbon polymer coatings noted for corrosion resistance and low frictional resistance; Clean roofs to prevent excessive tephra loads on buildings; Utilize underground electrical supply.

Lava flows

Spray water on advancing lava fronts; Divert lava flows using bombing with guidance systems; Built earth barriers and artificial channels.

Lahars

Drainage of crater lakes; Revegetate unconsolidated tephra; Reconsider bridge clearances; Install crossing gates on roads across lahar channels; Use lahar flow warning systems, retention basins, engineered channels, and land-use planning

Pyroclastic flows

Create refuges in air-tight bellow-ground cellars; Divert distal flows using earth barriers; Increase awareness on lethal effects of decoupled pyroclastic surges from flows.

Meanwhile the assessment of social and economic factors that make people vulnerable has been undertaken by surveys and interviews in communities living around active volcanoes such as Merapi and Pinatubo or Mayon. 1.5.12.2 Suggestions for the implementation of design codes and guidelines – In situ geotechnical tests for assessing the resilience of housing and bridges to volcanic flows; – contingency planning and mitigation procedures to be implemented in large cities in the developing world (Indonesia, Andes, Philippines. . .); – education programs for increasing awareness and improving consciousness among civil authorities of large cities and regional councils in the developing world (e.g. SE Asia, Andean countries). 1.5.13 EXAMPLE OF APPLICATION 1.5.13.1 A case study: The eruptive crisis of Ubinas volcano, Peru, 2006–2007 The most recent explosive activity of Ubinas volcano (Rivera et al., 2008) started on 27 March 2006. In response to the volcanic crisis, members of three national Institutions (INGEMMET, IGP, UNSA) as well as the regional Civil Defence offices in Moquegua (RCCDM) and Arequipa formed a joint scientific committee. With foreign help, a preliminary hazard-zone map and a contingency map were produced in early April 2006. The hazard-zone map is based on two eruption scenarios: 1) a small eruption similar to the 1990–1998 vulcanian episode of Sabancaya, 2) a moderate event such as theAD 1677 scoria-and-ash fall and flow-producing eruption at Ubinas. Monitoring, consisting of a network of seismometers, EDM, and geochemical survey of thermal springs, has been undertaken by a pool of Institutions on a temporary basis until May 2006 and on a permanent basis ever since. The scientific committee successfully offered a three-stepped response to the increase in eruptive activity: 1) The appearance of an incandescent lava plug in the vent on 20 April prompted the scientific committee to ask RCCDM to evacuate 150 people from the nearest hamlet of Querapi (situated at the foot of the unstable south flank) to the first shelter (village of Anascapa) 8 km away. 93

2) A substantial increase in eruptive activity between 27 April and 2 June led the scientific committee to increase the alert level from yellow to orange and implemented the evacuation plan based on the contingency map. RCCDM further issued the order to evacuate five villages within 12 km of Ubinas. Between 9–11 June 2007, 1000 people were relocated to the second shelter (Chacchagen) 20 km away from the volcano. 3) After ∼9 months in Chacchagen, the refugees returned to their villages in March 2007, as the population could not cope with less economic resources and a tense situation. The population was also disappointed by the fact that the planned relocation on the remote coast near Moquegua has not been implemented in 2007 (the political decision has not been taken by the government yet). Despite economic and social drawbacks, the challenging crisis of the most active volcano in Peru was the first opportunity for Peruvian institutions to successfully cooperate in, and gain lessons from managing volcanic crises. 1.5.13.2 Second case study: Physical vulnerability in the city of Arequipa, Peru Arequipa (Martelli et al., 2008) is the second largest city in Peru with a population exceeding 860,000. Rapid population growth since the 1940s has resulted in urban growth onto the southwest flank of the volcano, Rio Chili River terraces and adjacent to tributaries within 9 km of El Misti summit. With an expanding city into more hazardous prone areas it is necessary to assess the vulnerability of buildings and infrastructure in response to the threat posed from volcanic mass flows. Previous studies using Titan2D and LaharZ have attempted the delineation of debris flow inundation zones from El Misti (Delaite et al., 2005). Characteristics such as pending and short run out distances were unrealistic in earlier Titan2D simulations. Four main terraces of the Rio Chili River from Chilina to the Puerte Bolognesi Bridge (approximately 5 km2 ) were surveyed to obtain detailed topographical data. A DEM was then computed using a DGPS data, aerial photographs and stereophotogrammetry. Lahar volumes ranging from 0.01.106 m3 to 11.106 m3 with solid fractions of 0.3–0.5 were computed. Modelled results are enhanced with a new DEM; however further analysis will need to be undertaken as to whether the simulations are more realistic. Quantitative descriptions of buildings at building level identified nineteen land-use patterns and ten construction types. Most new construction comprised unreinforced masonry panels (perforated red brick and mortar) with cast-in-situ reinforced concrete frames (horizontal and vertical), and flat or pitched reinforced concrete slab roofs. Large glass windows are throughout with aluminium or wood framing and often secured with steel bars. Doors are solid and wooden with steel security screen/bars. Conversely, Type I construction comprised old stone/ignimbrite base with unreinforced masonry panels (ignimbrite, brick or adobe, with poor quality mortar). The walls are not confined by either reinforced horizontal or vertical cast-in-situ concrete, and in most cases appear unstable. Wooden rafters support corrugated iron roofs, secured with heavy objects such as rocks. Less than 50% of the population surveyed resides in dwellings less than Type C, however, the majority of those are situated in areas that are more hazardous (e.g. Rio Chili lower terraces) areas. 1.5.14 FURTHER DEVELOPMENTS Identification of further needs for the research and suggestions for possible developments: a geospatial platform for research and training that helps in the decision- and policy-making process (© SUNY at Buffalo, New York: M.F. Sheridan and C. Renschler). The impact and consequences of extreme geophysical events, like mudflows, on landscape properties and processes can be continuously assessed by a well-coordinated interdisciplinary research and outreach using the Geospatial Project Management Tool (GeoProMT© ) applied to risk assessment and resilience. Communication between various involved disciplines and stakeholders is a key to the successful implementation of an integrated risk management plan. As the amount of spatiotemporal data representing environmental properties at various scales increases, there is a lag in effective communication among participating disciplines that use this detailed information to predict landscape processes. These issues become apparent at the level of decision support tools for extreme events/disaster management in natural and managed environments. GeoProMT© is a collaborative platform for research and training to document and communicate the fundamental steps in transforming information for extreme events at various scales for analysis and management. 94

GeoProMT© is an internet-based interface for the management of shared geo-spatial and multitemporal information such as measurements, remotely sensed images, and other GIS data. This tool enhances collaborative research activities and the ability to assimilate data from diverse sources by integrating information management. This facilitates a better understanding of natural processes and enhances the integrated assessment of resilience against both the slow and fast onset of hazards and risks. Fundamental to understanding and communicating complex natural processes are: a) representation of spatio-temporal variability, extremes, and uncertainty of environmental properties and processes in the digital domain, b) transformation of their spatiotemporal representation across scales (e.g., interpolation, aggregation, disaggregation.) during data processing and modelling in the digital domain, and designing and developing tools for: c) geo-spatial data management d) geo-spatial process modelling and effective implementation, and e) supporting decision- and policy-making in natural resources and hazard management at various spatial and temporal scales of interest. GeoProMT© is useful for researchers, practitioners and decision-makers because it provides an integrated environmental system assessment and data management approach that considers the spatial and temporal scales and variability in natural processes. Particularly in the occurrence or onset of extreme events it can utilize the latest data sources that are available at variable scales, combine them with existing information, and update assessment products such as risk and vulnerability assessment maps. Because integrated geo-spatial assessment requires careful consideration of all the steps in utilizing data, modelling and decision-making formats, each step in the sequence must be assessed in terms of how information is being scaled. At the process scale various geophysical models (e.g. TITAN2D, LAHARZ, or many other examples) are appropriate for incorporation in the tool. Examples that illustrate the application of GeoProMT include: 1) Working with geoscientists, public officials, and civil protection authorities to understand and improve the new volcanic hazard map of El Misti Volcano as it presents a threat to Arequipa, Peru; 2) Improving and evaluating a new hazard map and mitigation plan for potential mudflows associated with potential future events around La Soufrière of Guadeloupe; 3) Developing a plan for monitoring mudflows around Semeru, Java, to calibrate computational models, like TITAN2D and others, for more accurate simulation outcomes. In all three cases, GeoProMT will be used for education, training, and scientific evaluation of data. It will provide an improved new technique for remote transmission of accurate geospatial information between scientists, officials, and responsible authorities in a real-time learning environment. REFERENCES Alborelli, E. 2009. Messa in opera di misure di mitigazione per edifici soggetti a precipitazione di materiale piroclastico. Tecnologie sostenibili per la riqualificazione delle coperture. Degree tesis, Università degli Studi di Napoli Federico II, Facoltà di Architettura. Baxter, P.J., Cole, P.D., Spence, R., Zuccaro G., Boyd R. and Neri, A. 2005. The impacts of pyroclastic density currents on buildings during the eruption of the Soufrière Hills volcano, Montserrat. Bulletin of Volcanology. 67: 292–313. Benoit, J.P. and McNutt, S.R. 1996. Global volcanic earthquake swarm database and preliminary analysis of volcanic earthquake swarm duration. Annali de Geofisca: 39, 221–229. Blong, R., 1981. Some effects of tephra falls on Buildings. In: Sparks, S. & Self, S. (Eds), “Tephra Studies”. Blong, R., 1984. Volcanic hazards – a sourcebook on the effects of eruptions. Sydney: Academic Press. Blong, R., 2000. Volcanic hazards and risk management. In Sigurdsson, H., Houghton, B., McNutt, S.R., Rymer, H. and Stix, J. (eds), Encyclopedia of volcanoes. San Diego, CA, Academic Press: 1215–27. Blong, R., 2003. Building damage in Rabaul, Papua New Guinea, 1994. Bulletin of Volcanology, Vol. 65. Carlino S. 2001. The floods and the mudslides after the Vesuvius eruption. History and risk. Interventi di ingegneria naturalistica nel Parco Nazionale del Vesuvio. Ente Parco nazionale del Vesuvio (Napoli). pp.43–69 8 (in Italian).

95

Casadevall, T.J., Delos Reyes, P.J. & Schneider, D.J., 1996. The 1991 Pinatubo eruptions and their effects on aircraft operations, Philippine Institute of Volcanology and Seismology & University of Washington Press. Chester, D.K., Degg, M., Duncan, A.M. and Guest, J.E., 2001. The increasing exposure of cities to the effects of volcanic eruptions: a global survey. Global Environmental Change, 2: 89–103. Chouet, B.A. 1996. Long-period volcano seismicity: its source and use in eruption forecasting. Nature 380, 309–316, doi:10.1038/380309a0. Colombrita R. 1984. Methodology for the construction of earth barriers to divert lava flows: the Mt. Etna 1983 eruption. Bull. Volcanol., 47-4 (2), 1009–1038. Cook, R.J., Barron, J.C., Papendick, R.R., Williams, G.J., 1981. Impact on agriculture of the Mount St Helens eruptions. Science, Vol. 211, p. 16–22. Cowie HA, Baxter PJ, Hincks T, Searl A, Sparks RSJ, Tran CL, Aspinall W, Woo G., 2003. Risk assessment for silicosis and exposure to volcanic ash on Montserrat. Report to the UK Department for International Development, UK Department for International Development, London, p 49, Intergovernmental Oceanographic Commission (of UNESCO): 1998, Post-Tsunami Survey Field Guide (1st ed.), Manuals and Guides #37, Paris, France. Cubellis E., Marturano A. 2006. Analysis of historical and present earthquakes at Vesuvius for seismic hazard evaluation. XY0701 EGU 2006. Session NH5.03: Volcanic Hazard and Risk, Vienna, Austria, July 02, 2006. De Gregorio D., Faggiano B., Formisano A., Mazzolani F.M., 2010. Air fall deposits due to explosive eruptions: action model and robustness assessment of the Vesuvian roofs. Proceeding of the International Conference COST Action C26 “Urban habitat constructions under catastrophic events”, Naples, Italy, 16–18 September 2010. (in press). Delaite, G., Stinton, A.J., Sheridan, M.F., Thouret, J.C., Burkett, B., 2004. A comparison of TITAN2D and LAHARZ simulated debris flow hazards at El Misti Volcano, southern Peru. European Geophysical Union, Nice, France, April 26, 2004. Dobran F. 2006. VESUVIUS. Education, security and prosperity. Development in volcanology. Flavio Dobran Eds. Elsevier. Dobran F. 2007. Urban Habitat Constructions Around Vesuvius. Environmental Risk and Engineering Challenges. Proc. of COST Action C26 Seminar on Urban Habitat Constructions Under Catastrophic Events, Prague, 30–31 March 2007. Dumaisnil C., Thouret J.C.1 & Muzeau J.P., 2008. Volcanic hazard, Proc. International Symposium COST C26: Urban Habitat Constructions under Catastrophic Events, Malta, p. 371–382. Esposti Ongaro T., Neri A., Todesco M., Macedonio G. 2002. Pyroclastic flow hazard assessment at Vesuvius (Italy) by using numerical modeling. II. Analysis of flow variables. Bull. Volcanol. (2002) 64:178–191. DOI 10.1007/s00445-001-0190-1. Faella C., Nigro E. 2002. Debris flow effects on constructions. Damage analysis, collapse mechanisms, impact velocities, code provisions. Internal Report COST-C12/WG2. Volos, Greece, 14–15 June, 2002. Festa, G., Zollo, A., Manfredi, G., Polese, M. and Cosenza, E. 2004. Simulation of the earthquake ground motion and effects on engineering structures during the pre-eruptive phase of an active volcano. Bull. Seism. Soc. Am. 94: 6, 2213–2221. Frameworks Architects, Blong, R., Ove Arup & Partners Pacific 1996. Ralum civic centre and Kokopo commercial business centre buildings volcanic impact report, Vol. 2. World Bank Project, Gazelle Restoration Authority, Rabaul, 1996. Gurioli L., Zanella E., Cioni R., Lanza R. 2008. Determinazione paleomagnetica delle temperature di messa in posto di flussi piroclastici dell’eruzione del 79 d.c. del Vesuvio. GNGTS-Atti del 18◦ Convegno Nazionale. December 12th, 2008. Gordon, K.D., Cole, J.W., Rosenberg, M.D., & Johnston, D.M., 2005. Effects of Volcanic Ash on Computers and Electronic Equipment, Natural Hazards, Vol. 34, 2005. Horwell, C.J. & Baxter, P.J., 2006. The respiratory health hazards of volcanic ash: a review for volcanic risk mitigation”, Bulletin of Volcanology, Vol. 69. Inbar, M., Ostera, H.A., Parica, C.A., Remesal, M.B., & Saliani, F.M. 1995. Environmental assessment of 1991 Hudson volcano eruption ashfall effects on southern Patagonia region, Argentina. Environmental Geology, Vol. 25. Iverson, R.M., Schilling, S.P. and Vallance, J.W., 1998. Objective delineation of lahar-inundation hazard zones. Geological Society of America Bulletin, 110: 972–84. Johnston, D.M. 1997a. Physical and Social Impacts of past and future volcanic eruptions in New Zealand. Unpublished PhD thesis. 1997. Johnston, D.M. 1997b. The impacts of recent falls of volcanic ash on public utilities in two communities in the United States of America. Institute of Geological and Nuclear Sciences Report, pp. 21. Johnston, D.M., Houghton, B.F., Neall, V.E., Ronan, K.R., & Paton, D. 2000. Impacts of the 1945 and 1995-6 Ruapehu eruptions, New Zealand: An example of increasing societal vulnerability. GSA Bulletin, Vol. 112.

96

Johnston, D.M., Stewart, C., Leonard, G.S., Hoverd, J., Thordarsson, T., Cronin, S., 2004. Impacts of volcanic ash on water supplies in Auckland: part I. Institute of Geological and Nuclear Sciences Science Report, 2004/25. Marsella M., Coltelli M., Napoleoni Q., Scifoni S. 2008. Lava flow simulation for the design of the barriers. Study case: Etna 2001. INGV sezione di Catania. www.ct.ingv.it (in Italian). Martelli K., Thouret J.-C., Vargas R., Van Westen C., Fabre D., Sheridan M.F., 2008. Building and infrastructure vulnerability from volcanic mass flows in the city of Arequipa (Peru). IAVCEI General Assembly, Reykjavik, Iceland 17–22 August 2008 (Abstract). Marzocchi W., Neri A., Newhall C.G., Papale P., 2007. Probabilistic volcanic hazard and risk assessment. EOS, Trans., Amer. Geoph. Union, 88, 32: 318. Mazzolani, F.M., Faggiano, B. and De Gregorio, D. 2008. Actions in the catastrophic scenarios of a volcanic eruption. Proceeding of COST Action C26 Symposium on “Urban habitat construction under catastrophic events”, Malta, 23–25 October 2008. Datasheet n◦ 5.1: 449–467. ISBN 978-99909-44-40-2. Mazzolani F.M., Faggiano B., De Gregorio D., 2009a. The catastrophic scenario in explosive volcanic eruptions in urban areas. Proceeding of Protection of Historical Buildings, PROHITECH 09, Rome, Italy, 21–24 June. Vol. 2: 1529–1534. ISBN: 978-0-415-55805-1. Mazzolani F.M., Faggiano B., Formisano A., De Gregorio D., 2009b. Vulnerability evaluation of RC structures in the Vesuvian area. Proceeding of Protection of Historical Buildings, PROHITECH 09, Rome, Italy, 21–24 June. Vol. 2: 1523–1528. ISBN: 978-0-415-55805-1. Mazzolani F.M., Indirli M., Zuccaro G., Faggiano B., Formisano A. and De Gregorio D. 2009c. Catastrophic effects of a Vesuvian eruption on the built environment. Proc. PROTECT 2009, 2nd International Workshop on Performance, Protection & Strengthening of Structures under Extreme Loading, Shonan Village Center, Hayama, Japan, 19–21 August 2009. Mazzolani F.M., Faggiano B., Formisano A., De Gregorio D., Nunziata C., Mandara A., 2010. Volcanic and tectonic earthquakes effects in the Vesuvian urban habitat. Proceeding of the International Conference 14th ECEE, European Conference on Earthquake Engineering, Ohrid, Republic of Macedonia, August 30–September 03. Paper n. 1179. (in press). MD., 2008. Ministerial Decree, Technical codes for constructions. Official Gazette of the Italian Rep., January 14th. (in Italian). Neri A., Esposti Ongaro T., Macedonio G., De’ Vitturi M., Cavazzoni C., Erbacci G., Baxter P. 2007. 4D simulation of explosive eruption dynamics at Vesuvius. Geophysical Research Letters, Vol. 34, L04309, doi: 10.1029/ 2006GL028597. Newnham, R.M., Dirks, K.N., & Samaranayake, D., 2010. An investigation into long-distance health impacts of the 1996 eruption of Mt Ruapehu, New Zealand. Atmospheric Environment, Vol. 44. Palermo, D., Nistor, I., Nouri, Y., and Cornett, A. 2007. Tsunami-Induced Impact and Hydrodynamic Loading of Near-Shoreline Structures. Proc. PROTECT 2009, 1nd International Workshop on Performance, Protection & Strengthening of Structures under Extreme Loading, August 20–22, 2007 Whistler, Canada. Pareschi, M.T., Cavarra, L., Favalli, M., Giannini, F. and Meriggi, A., 2000. GIS and volcanic risk management. Natural Hazards, 21: 361–79. Petrazzuoli S.M. & Zuccaro G. 2004. Structural resistance of reinforced concrete buildings under pyroclastic flows: a study of the Vesuvian area. Journal of Volcanology and Geothermal Research, 133 (2004) 353–367. Rivera M., Mariño J., Thouret J.-C., Fuentes J., Cacya L., ArguedasA., Lautze N., AguilarV., 2008. Management of the volcanic crisis during the most recent Ubinas eruptive activity. IAVCEI General Assembly, Reykjavik, Iceland 17–22 August 2008 (Abstract). Sheridan, M., Hubbard, B., Bursik, M.I., Abrams, M., Siebe, C., Macias, J.L. and Delgado, H., 2001. Gauging short-term volcanic hazards at Popocatepetl. EOS, Trans. Amer. Geoph. Union, 185: 187–88. Schriever, W.R. & Hansen, A.T., 1964. Snow loads and strength of small roofs in Canada. Forest products Journal. Vol. 14, Issue 3, p129–136. Spence, R.J.S., Antonios, P., Baxter, P.J., Coburn, A.W., White, M., Dayrit, M., & Field Epidemiology Training Team, 1996. Building Damage Caused by the Mount Pinatubo Eruption of June 15, 1991. Philippine Institute of Volcanology and Seismology & University of Washington Press, 1996. Spence R.J.S., Baxter P.J., Zuccaro G. 2004. Building vulnerability and human casualty estimation for a pyroclastic flow: a model and its application to Vesuvius. Journal of Volcan. and Geothermal Research 133 (2004) 321–343. Spence, R.J.S., Kelman, I., Baxter, P.J., Zuccaro, G., & Petrazzuoli, S. 2005. Residential building and occupant vulnerability to tephra fall. Natural Hazards and Earth System Sciences, Vol. 5. Stewart, C., Johnston, D.M., Leonard, G.S., Horwell, C.J., Thordarson, T., & Cronin, S.J., 2006. Contamination of water supplies by volcanic ashfall: A literature review and simple impact modelling, Journal of Volcanology and Geothermal Research, Vol. 158.

97

Stewart, C., Pizzolon, L., Wilson, T., Leonard, G., Johnston, D. Cronin, S., 2009. Can volcanic ash poison water supplies? Integrated Environmental Assessment and Management, 5(3): 713–716 Sword-Daniels V., Rossetto T., Twigg J., Johnston D., Wilson T., Cole J., Loughlin S., & Sargeant S., 2010. Review of the impacts of volcanic ash fall on urban environments. Proceedings of COST Action C26 Final International Conference on Urban habitat construction under catastrophic events, Naples, 16–18 September 2010. The Daily Telegraph, 2010. The Daily Telegraph, 27th April 2010. Volcanic ash crisis cost airlines £2.2 billion. Thouret J-C., 2010. Volcanic hazards and risks: a geomorphological perspective, Chapter 3, in Geomorphological Hazards and Disaster Prevention, Irasema Alcántara-Ayala & Andrew S. Goudie Eds., Cambridge University Press, ISBN 978-0-521-76925-9 Hardback. Tilling, R.I. 1991. Reducing volcanic risk: Are we winning some battles but losing the war? Earthquakes and Volcanoes, 22 (3), p. 133–137. Tilling, R.I., 2005. Volcano hazards. In: Marti, J. and Ernst, G.G.J., (eds), Volcanoes and the environment, Cambridge Press, 55–89. Tinti S., Maramai A., Graziani L. 2007. The Italian Tsunami Catalogue (ITC), Version 2. http://www.ingv.it/ servizi-e-risorse/BD/catalogo-tsunami/catalogo-degli-tsunami-italiani. Tinti S., Zaniboni F., Armigliato A. and Pagnoni G., 2010. Tsunami hazard and risk evaluation in the Gulf of Naples: State of the art and perspectives. Proceedings of COST Action C26 Final International Conference on Urban habitat construction under catastrophic events, Naples, 16–18 September 2010. Todesco M., Neri A., Esposti Ongaro T., Papale P., Macedonio G., Santacroce R., Longo A. 2002. Pyroclastic flow hazard assessment at Vesuvius (Italy) by using numerical modeling. I. Large-scale dynamics. Bull. Volcanol. (2002) 64:155–177. DOI 10.1007/s00445-001-0189-7. Vallario A., 1994. Potenziale rischio idrogeologico al Somma-Vesuvio. Eruzione vesuviana del 1944. Ricordo di un evento eruttivo 50 anni dopo. Comune di San Sebastiano al Vesuvio. Wills J., Wyatt T., Lee B. 1998. Warnings of high winds in densely populated areas. Book 4 of the IDNDR Flagship Programme – Forecasts and Warnings, Thomas Telford, London. Wilson, T., Cole, J., Stewart, C., Dewar, D., & Cronin, S. 2008. Assessment of long-term impacts on agriculture and infrastructure and recovery from the 1991 eruption of Hudson Volcano, Chile. Christchurch, 2008. Wilson, T.M. et al. 2009a. Vulnerability of Pastoral Farming Systems to Volcanic Ashfall Hazards. PhD thesis, 2009. Wilson, T., Daly, M., & Johnston, D. 2009b. Review of Impacts of Volcanic Ash on Electricity Distribution Systems, Broadcasting and Communication Networks. Auckland Engineering Lifelines Group. Auckland Regional Council Techinical Publication, No. 051, 79p. Zuccaro G., Ianniello D. 2004. Interaction of pyroclastic flows with building structures in an urban settlement: a fluid-dynamic simulation impact model. Journal of volcanology and geothermal research 133, 345–352. Zuccaro G. & Cacace F., 2010a. Seismic impact scenarios in the volcanic areas in Campania. Proceedings of COST Action C26 Final International Conference onUrban habitat construction under catastrophic events, Naples, 16–18 September 2010. Zuccaro G., & Leone M.F. 2010b. Building technologies for the mitigation of volcanic risk. Proceedings of COST Action C26 Final International Conference on Urban habitat construction under catastrophic events, Naples, 16–18 September 2010. Zuccaro G., Cacace F. & Nardone S., 2010c. Human and structural damage consequent to a Sub-Plinian like eruption at Mount Vesuvius. Proceedings of COST Action C26 Final International Conference on Urban habitat construction under catastrophic events, Naples, 16–18 September 2010. Zuccaro G., Cacace F. & Rauci M., 2010d. Vulnerability functions for building structures under pyroclastic flow actions. Proceedings of COST Action C26 Final International Conference on Urban habitat construction under catastrophic events, Naples, 16–18 September 2010.

98

Urban Habitat Constructions under Catastrophic Events (Final Report) – Mazzolani (Ed). © 2010 Taylor & Francis Group, London, ISBN 978-0-415-60686-8

1.6 Tsunami hazard and risk evaluation in the Gulf of Naples: State of the art and perspectives S. Tinti, F. Zaniboni, A. Armigliato & G. Pagnoni University of Bologna, Dept. of Physics, Sector of Geophysics, Bologna, Italy

1.6.1 INTRODUCTION The Gulf of Naples (southern Italy, see Figure 1) is characterised by a non-negligible tsunami hazard. The Italian Tsunami Catalogue (Tinti et al., 2004, hereafter referred to as ITC) indicates that the largest part of the documented events are associated to the volcanic activity of Vesuvius: the two most famous and catastrophic eruptions occurred in 24 August 79 AD and in 17 December 1631 and they are both reported to have been accompanied by significant sea In addition to the two previous sources, that can be considered as local, a far-field potential tsunamigenic source relevant to the Gulf of Naples is the continuous volcanic activity of Stromboli, the northernmost island of the Aeolian archipelago. The partial or total collapse of the north-western flank of the volcanic edifice, known as Sciara del Fuoco, represents a threat in terms of tsunami impact even for the Campania coasts and the Gulf of Naples in particular. In terms of risk, the exposure of the Gulf of Naples to hazardous phenomena is nowadays severely increased by the extremely dense and often unorganised urbanisation. The risk related to the volcanic hazard indeed appears to be the most critical, but tsunamis also represent an important aspect in the risk mitigation and civil protection perspective. We present the state of the art of the research on the tsunami hazard assessment for the Gulf of Naples, which is based prominently on the numerical modelling of the tsunami generation,

Figure 1.

Geographical sketch of the Gulf of Naples.

99

propagation and impact of the tsunami waves along the coastline of the Gulf. Two main generating mechanisms are taken in consideration. The first, related to the volcanic activity of Vesuvius, is treated by considering pressure pulses moving from the Vesuvius itself toward the open sea. As regards Ischia (near-field) and Stromboli (far-field), the generation of tsunamis by potential mass failures along the islands flanks is simulated by means of a Lagrangian block model, called UBO-BLOCK, developed by the Tsunami Research Team operating at the University of Bologna and existing in 1-D (Tinti et al., 1997) and 2-D (e.g. Tinti et al., 2006) versions. The tsunami propagation is then simulated with the finite-element numerical code UBO-TSUFE (Tinti et al., 1994), developed by the same research group and implementing the non-linear shallow-water equations. We will finally introduce some clues on the tsunami risk assessment for the area, which has never been faced before and that may take advantage from innovative instruments recently developed in the framework of EU-funded projects such as TRANSFER and SCHEMA.

1.6.2 TSUNAMI HAZARD ASSOCIATED WITH VESUVIUS PYROCLASTIC FLOWS 1.6.2.1 Historical information Some of the most fatal and costly natural catastrophes that occurred in the past few decades were the result of extreme wind incidents. Hurricanes, typhoons and tornadoes were reported and tracked around the globe. The social and economic impact associated to these natural disasters initiated a drastic response from a number of political and academic institutions. Both structural and wind engineers invested a significant amount of effort to better understand the impact of such events on structures. Several studies were initiated after major hurricanes and had as main objective to assess the induced damage to both low-rise and high rise structures. The assessment revealed in most cases local damages for tall buildings (façade and non-structural elements failure – see Figure 2), whereas low-rise buildings have in many cases collapsed or torn apart even during moderate intensity cyclone events (Figure 3). According to the ITC the Gulf of Naples has been affected by a number of tsunamis mostly associated with the volcanic activity of Vesuvius. The most famous eruption is probably the one occurred on 24 August 79 AD, when a VEI = 5 eruption destroyed Pompei, Hercolaneum and other cities. A tsunami was observed on the second day of the eruption and its main manifestation consisted in a sea retreat. But the most conspicuous tsunami related to the Vesuvius activity was probably due to the last sub-plinian eruption occurred on 17 December 1631 (VEI = 4). The water oscillations were quite relevant and in the harbor of Naples waves of 2–5 m in amplitude were

Figure 2. Amplitude of the forcing impulse at different times and dynamic pressure as a function of time plotted in the points drawn as stars in the upper-left panel.

100

observed (Rosi et al., 1993). The sea was seen to withdraw and then to inundate the coast for three times: the largest retreat (about 1 mile) was reported in Sorrento. The real cause of the tsunami is uncertain: maybe it was the response to a concurrent earthquake that is reported to have occurred some time before the tsunami (but the time delay, which is crucial, cannot be quantified precisely) or it was the effect of the pyroclastic flows that are known to have reached the sea in at least five different points along the coast. In the following we present the modeling introduced in Tinti et al. (2003a) based on the hypothesis that the tsunami was due to the light component of the pyroclastic flow. 1.6.2.2 Numerical simulations: The tsunami generation mechanism Estimates on forcing impulses are based on simulations by Esposti Ongaro et al. (2002) who, however, compute pyroclastic flows on a domain that does not reach the sea. Seaward extrapolation is made here by using a very simple propagation law. Forcing is supposedly a pressure impulse that propagates seaward with constant radial velocity over a sector centered on the Vesuvius crater. The pulse duration is assumed to be constant, and its amplitude decreases with the distance from the crater, according to a decay power law (exponent is −1 in the example shown in Figure 2). The sector considered here covers the whole Gulf of Naples. The adopted pulse radial speed is of 15 m/s. The bottom-right panel of Figure 2 shows time histories of the forcing pulses corresponding to the locations marked with numbered stars in the upper-left map of the same figure. It is seen that, according to our hypothesis, the pulse is rather peaked with a sharp front, and does not change its shape and duration (about 3 minutes) while propagating seaward. Pressure pulse determines a sea level deformation that is responsible for the tsunami. The forcing sea deformation is sketched at 2, 5 and 10 minutes in Figure 2, with the time being measured since the arrival of the pyroclastic flow at the coastline closest to the crater. 1.6.2.3 Numerical simulations: Tsunami propagation Tsunami simulations are computed by means of a numerical scheme (UBO-TSUFE code) based on finite triangular-element mesh allowing one to reproduce quite well the complex domain of the Gulf of Naples that is bounded by complicated peninsulas and pointed by numerous islands. We solve the Navier-Stokes equations in the shallow-water approximation and in conditions of dynamic forcing that have resemblance with the equations governing tsunamigenesis induced by submarine moving bodies. We limit our attention here to the maximum energy density (in KJ/m2 ) computed at the grid nodes in the first 30 minutes of propagation (Figure 3, left panel): the zones with the highest tsunami energy concentration are placed at the northern and southern corner of the gulf, corresponding to the areas of Naples and of Castellammare. This is confirmed by the plot of the maximum positive and maximum negative water elevations computed along the coastline (Figure 3, right panel).

Figure 3. Maximum energy density after 30 minutes (left) and maximum water elevation along the coast (right). In the left plot, the stations are numbered as follows: 1-Miseno, 2-Pozzuoli, 3-Bagnoli, 4-Posillipo, 5-Napoli, 6-Torre del Greco, 7-Torre Annunziata, 8-Castellammare, 9-Sorrento, 10-Positano.

101

It must be noticed that the largest calculated tsunami amplitudes are not larger than 40 cm for the particular case shown here, which is smaller by about a factor of 10 than the sea level oscillations observed during the 1631 eruption. Tinti et al. (2003b) performed a sensitivity analysis aimed at quantifying the effect of varying the pulse speed, amplitude and duration as well as the particular sector along which the pressure impulse propagates. Although the variation of some parameters, and in particular of the pulse amplitude and duration, produce relevant effects on the computed tsunamis, the largest computed tsunami amplitudes are still at least 5 times smaller than the historical reports for the 1631 event. A more in-depth understanding of the generating mechanism of the tsunami waves in the gulf of Naples by the Vesuvius pyroclastic flows represents a research line of primary importance for the next future.

1.6.3 ISCHIA: LANDSLIDES FROM MT. EPOMEO 1.6.3.1 The Ischia Debris Avalanche (IDA) scenario The island of Ischia, located to the north-western end of the Gulf of Naples (Figure 4), is characterized by an important volcanic activity, well documented in the literature (Gillot et al., 1982; Orsi et al., 1991). Its main manifestation is the Mount Epomeo uplift, started around 33 ky BP, with an average rate of 20–25 mm/y, causing the progressive steepening of the relief flanks, whose top presently raises up to 787 m, hence increasing the gravitational instability of masses along the slopes and occasionally causing, in addition to seismic shaking, the detachment of sliding bodies. This mechanism is confirmed by the high number of terrestrial slides recognized on the island of Ischia, radiating from Mount Epomeo, described in de Vita et al. (2006). Occasionally the sliding masses can reach the sea, causing tsunamis with relevant effects, due also to the steep slopes enhancing slide velocity. A number of bathymetric and geophysical marine surveys have evidenced the presence of a large number of submarine deposits around the coasts of Ischia. De Alteriis & Violante (2009) describe evidences found off the northern and western coasts, where debris deposits of some million m3 in volume were recognized, even though it was not possible to link them to subaerial scars. But the most spectacular deposit was found off the southern coast, covering an area of more than 200 km2 , starting from the toe of the continental slope, at about 600 m, and extending down to the sub-horizontal area south of Ischia, at about 1000 m depth. This deposit, well studied and characterized in Chiocci & de Alteriis (2006), has

Figure 4. Topo-bathymetric map of the Ischia island. Mount Epomeo exhibits an evident scar along its southern flank that can be seen as a caldera and flank collapse. The dotted line encloses the IDA debris flow area.

102

been correlated to the big scar cutting the southern flank of Mount Epomeo. The hypothesis of a single big event, called Ischia Debris Avalanche (IDA), occurring in Greek-Roman time (de Alteriis et al., 2010) is here investigated. It involved a consistent part of the Mount Epomeo edifice, both in its subaerial and submerged parts. The IDA can probably be considered as the worst-case scenario for the generation of tsunamis affecting Ischia and the Gulf of Naples. It is briefly summarized in the following two subsections. 1.6.3.2 IDA scenario simulation The simulation of the IDA has been carried out through codes developed by the University of Bologna research team. First of all, on the basis of Chiocci and de Alteriis (2006) hypothesis, the initial sliding mass and the sliding surface have been reconstructed (Figure 5): the Mount Epomeo edifice has been obtained hypothesizing a conical shape, filling the scar along the southern flank from the top down to 600 m sea depth. In this way we obtained a mass of over 3 km3 . The simulation of the landslide has been performed through the code UBO-BLOCK1 (see Tinti et al., 1997 for details), that divides the mass into a chain of blocks sliding along a predefined path, on which all the forces acting on the mass are projected, and on which the motion equations are numerically solved. The simulation shows that the sliding mass stops at around 1000 m depth after around 700 s, for a 30 km runout, reaching considerable peak velocities: 60 m/s with maximum of 70 m/s for some blocks, after 120 seconds (Figure 5). The maximum Froude number, measuring the efficiency in tsunami generation through the ratio between the slide horizontal velocity and the wave phase velocity, is reached before the velocity peak, after around 60 seconds: in that moment it is close to the critical value, 1, meaning that the mass and the generated wave move with similar velocities and the energy transfer from the slide to the tsunami is maximum. When moving in deeper waters, the mass loses efficiency in generating tsunami and the Froude number decreases. 1.6.3.3 Tsunami generation and propagation The knowledge of the slide dynamics at each time step allows us to evaluate the tsunamigenic impulse, computed through a transfer function filtering it through the sea depth. The tsunami propagation is simulated by means of the UBO-TSUFE code. Figure 6 shows the maximum water elevation over each node of the computational domain: higher values are in black-dark grey. We can notice that around the southern and western coasts of Ischia more than 20 m waves are attained. A strong tsunami energy beam is directed towards SE, hitting the western coast of Capri and then the Sorrento peninsula again with more than 20 m waves, while another beam heads for NE and hits Capo Posillipo, partially protecting the city of Naples from the full tsunami impact. Notice how the Salerno Gulf is protected by the Sorrento peninsula.

Figure 5. Profile of the simulated slide, showing the initial reconstructed sliding body (dark grey), the final simulated deposit (black) and the sliding surface (light grey). In the inset graph the average velocity and the Froude number (dotted line) vs. time are reported.

103

Figure 6.

Maximum water elevation (white-grey-black scale) over the computational domain.

Tinti et al. (2010) found out that the tsunami hits Procida and Capri in 6–8 minutes, entering the Gulf of Naples and being strongly slowed down by the bathymetry; in 15–20 minutes it involves all its coasts, hitting the city of Naples 15 minutes after the slide with 7–8 m high waves. The tsunami propagates with a strong positive first signal, meaning sea ingression onshore, followed by a series of crests and troughs with a period ranging around 5–6 minutes; not always the first wave is the highest. The sea continues to be perturbed by the tsunami for a lot of hours, due also to the particular morphology of the Gulf of Naples, enhancing multiple reflections of the wave, trapping the tsunami energy inside it and not favouring its dispersion into the open sea. It is worth noting that the pattern of propagation depicted in Figure 6 can be expected also for slides of two or three orders of magnitude lower with respect to IDA, that are not so unusual along the Mount Epomeo flanks, as evidenced by the bathymetric surveys.

1.6.4 TSUNAMI HAZARD FROM FAR-FIELD SOURCES: THE STROMBOLI HOLOCENE COLLAPSE SCENARIO 1.6.4.1 The Sciara del Fuoco Holocenic landslide The Stromboli volcano is located in the Aeolian archipelago (SE Tyrrhenian sea). It is characterized by a persistent activity, consisting in periodic moderate explosions alternating to strong paroxysmal phases, with lava effusion, occurring with some year intervals (Rosi et al., 2000). The main part of the ejected material accumulates along the Sciara del Fuoco, a big scar extending along the northwest flank of the volcano, where periodically masses of deposited material reach gravitational instability and slide into the sea, sometimes generating tsunamis. This is the case for example of the 2002 Stromboli tsunamis, caused by two landslides (the first submarine, the second subaerial), with volumes in the order of a few millions m3 : the generated tsunamis had relevant effects on the coasts of Stromboli, hitting some areas with more than 10 m high waves (e.g. Tinti et al., 2005; Tinti et al., 2006). The impact on the other islands of the Aeolian archipelago and in general of the southern Tyrrhenian coasts was less destructive, although tsunami effects were experienced as far as the Cilento coasts (Maramai et al., 2005). The history of the volcano is also characterized by a series of catastrophic events, the last of which was the Sciara del Fuoco slide, that occurred in the Holocene. That event, with a volume ranging about 0.5–1 km3 (Kokelaar & Romagnoli, 1995; Tibaldi, 2001) surely had more regional effects compared to the 2002 events. Numerical simulations of the Sciara del Fuoco Holocenic were performed by Tinti et al. (2000) and Tinti et al. 104

Figure 7. Maximum water elevation (white-grey-black scale) over the computational domain. Some preferential direction for the tsunami energy distribution are visible.

(2003c): the sliding body, half subaerial and half submerged, attained during its descent really high velocity (more than 60 m/s peak), provoking along the coasts of Stromboli 20–40 m wave height, 2–4 m wave in the other Aeolian archipelago islands and 1–4 m tsunami heights along the coasts of Calabria. 1.6.4.2 Main tsunami propagation features in the southern Tyrrhenian sea The use of a numerical finite-difference scheme allowed us to simulate the evolution of the tsunami generated by the flank collapse at Stromboli in a much wider portion of the Tyrrhenian basin. Figures 7a–b show two different snapshots taken at 25 and 35 minutes after the landslide onset, chosen for their relevance as regards the tsunami impact along the Campania coasts. Although not immediately observable in the shaded relief maps, the leading front of the tsunami propagating toward deep waters is characterized by a negative sign. Further, as expected, tsunami amplitude (and hence its energy) generally decays with distance, so that smaller and smaller tsunami effects are suffered by coastlines placed at increasing distances from Stromboli. Within 25 minutes after the landslide occurrence, the tsunami attacks and travels beyond Ustica island to the west and reaches the southern coasts of Campania (Cilento) to the north, while after 35 minutes the Gulf of Naples as well as the islands of Capri and Ischia experience sea level disturbances which are computed in tens of cm up to 1–1.5 m depending on the particular coastal portion. The field of maximum tsunami amplitude (Figure 8) clearly shows that tsunami energy is radiated in a highly heterogeneous pattern, mainly determined by the very complex bathymetry of the Tyrrhenian basin. In particular, it can be appreciated that tsunami energy tends to be trapped along specific directions. In particular, relevant energy focussing appears to take place along a northward-trending direction linking Stromboli to southern Campania. 1.6.4.3 Conclusions We have summarized the present knowledge on the tsunami hazard concerning the Gulf of Naples. We emphasize that tsunami hazard in this area is related for the largest part to the activity of volcanoes placed both in the near-field (Vesuvius and Ischia) and in the far-field (Stromboli). While the capability of estimating tsunami hazard in different regions has continuously being improved in the last couple of decades through the progress in the historical and geological knowledge on one side and in the numerical modeling tools on the other (see for instance the outcomes of the EUfunded TRANSFER project), the tsunami risk estimation and mitigation problem has been tackled in a systematic way only in very recent years. In general terms, the risk estimation is the result of the combination of detailed inundation maps coming from the hazard analysis, and of the vulnerability 105

Figure 8. Maximum water elevation field over the computationaldomain. Bathymetry is plotted in shaded relief in the background.

assessment. One of the approaches existing for the second aspect has been developed by the EUfunded SCHEMA project: it distinguishes between primary (type and material) and secondary (ground, age, foundation, orientation, etc.) criteria for buildings, and it adopts a building damage matrix, basically depending on building type and water inundation depth. This approach has been tested on different test sites in SCHEMA, there including the town of Catania, and could represent a useful tool to assess the tsunami risk for the Gulf of Naples area. REFERENCES Chiocci, F.L. & de Alteriis, G. 2006. The Ischia debris avalanche: first clear submarine evidence in the Mediterranean of a volcanic island prehistorical collapse. Terra Nova 18: 202–209. De Alteriis, G. & Violante, C. 2009. Catastrophic landslides off Ischia volcanic island (Italy) during pre-history. In C. Violante (ed.), Geohazard in rocky coastal areas, Special issue of the Journal of the Geological Society of London, 322: 73–104. De Alteriis, G., Insinga, D., Morabito, S., Morra, V., Chiocci, F.L., Terrasi, F., Lubritto, C., Di Benedetto, C. & Pazzanese, M. 2010. Age of submarine debris avalanches and tephrostratigraphy offshore Ischia island, Tyrrhenian sea, Italy. Marine Geology (submitted). De Vita, S., Sansivero, F., Orsi, G. & Marotta, E. 2006. Cyclical slope instability and volcanism related to volcanotectonism in resurgent calderas: the Ischia island (Italy) case study. Engineering Geology, 86: 148–165. Esposti-Ongaro, T., Neri, A., Todesco M. & Macedonio, G. 2002. Pyroclastic flow hazard assessment at Vesuvius (Italy) by using numerical modeling. II. Analysis of flow variables. Bulletin of Volcanology 64: 178–191. Gillot, P-Y., Chiesa, S., Pasquaré, G. & Vezzoli, L. 1982. <33,000-yr K-Ar dating of the volcano-tectonic horst of the Isle of Ischia, Gulf of Naples. Nature 299: 242–245. Kokelaar, P. & Romagnoli, C. 1995. Sector collapse, sedimentation and clast population evolution at an active island-arc volcano: Stromboli, Italy. Bulletin of Volcanology 57: 240–262. Maramai, A., Graziani, L., Alessio, G., Burrato, P., Colini, L., Cucci, L., Nappi, R., Nardi, A. & Vilardo, G. 2005. Near- and far-field survey report of the 30 December 2002 Stromboli (southern Italy) tsunami. Marine Geology 215: 93–106. Orsi, G., Gallo, G. & Zanchi, A. 1991. Simple shearing blockresurgence in caldera depression. A model from Pantelleria and Ischia. Journal of Volcanology and Geothermal Research 47: 1–11.

106

Rosi, M., Bertagnini, A. & Landi, P. 2000. Onset of the persistent activity at Stromboli volcano. Bulletin of Volcanology 62: 294–300. Rosi, M., Principe, C. & Vecci, R. 1993. The 1631 Vesuvius eruption – Reconstruction based on historical and stratigraphical data. Journal of Volcanology and Geothermal research 58: 151–182. Tibaldi, A. 2001. Multiple sector collapses at Stromboli volcano, Italy: How they work. Bulletin of Volcanology 63: 112–125. Tinti, S., Gavagni, I. & Piatanesi, A. 1994. A finite-element numerical approach for modeling tsunamis. Annals of Geophysics 37: 1009–1026. Tinti, S., Bortolucci, E. & Vannini, C. 1997. A block-based theoretical model suited to gravitational sliding. Natural Hazards 16: 1–28. Tinti, S., Bortolucci, E. & Romagnoli, C. 2000. Computer simulations of tsunamis due to flank collapse at Stromboli, Italy. Journal of Volcanology and Geothermal research 96: 103–128. Tinti, S., Pagnoni, G. & Piatanesi, A. 2003a. Simulation of tsunamis induced by volcanic activity in the gulf of Naples. Natural Hazards and Earth System Sciences 3: 311–320. Tinti, S., Pagnoni, G. & Zaniboni, F. 2003b. Study of tsunamis induced by pyroclastic flows from Vesuvius, Italy. Geophysical Research Abstracts 5: 11618. Tinti, S., Pagnoni, G., Zaniboni, F. & Bortolucci, E. 2003c. Tsunami generation in Stromboli island and impact on the southeast Tyrrhenian coasts. Natural Hazards and Earth System Sciences 3: 1–11. Tinti, S., Maramai, A. & Graziani, L. 2004. The new catalogue of Italian tsunamis. Natural Hazards 33: 439–465. Tinti, S., Manucci, A., Pagnoni, G., Armigliato, A. & Zaniboni, F. 2005. The 30th December 2002 landslideinduced tsunami in Stromboli: Sequence of the events reconstructed from the eyewitness accounts. Natural Hazards and Earth System Sciences 5: 763–775. Tinti, S., Pagnoni, G. & Zaniboni, F. 2006. The landslides and tsunamis of 30th December 2002 in Stromboli analysed through numerical simulations. Bulletin of Volcanology 68: 462–479. Tinti, S., Chiocci, F.L., Zaniboni, F., Pagnoni, G. & de Alteriis, G. 2010. Numerical simulations of the tsunami generated by a past catastrophic landslide on the volcanic island of Ischia. Marine Geophysical Researches (submitted).

107

Chapter 2: Analysis of behaviour of constructions under catastrophic events

Urban Habitat Constructions under Catastrophic Events (Final Report) – Mazzolani (Ed). © 2010 Taylor & Francis Group, London, ISBN 978-0-415-60686-8

2.1 Analyses of structures under fire D. Bacinskas, E. Geda, V. Gribniak & G. Kaklauskas Vilnius Gediminas Technical University, Lithuania

G. Cefarelli, B. Faggiano, A. Ferraro, F.M. Mazzolani & E. Nigro University of Napoli “Federico II”, Italy

C. Couto, N. Lopes & P. Vila Real University of Aveiro, Portugal

M. Hajpál & Á. Török Budapest University of Technology and Economics, Hungary

M. Kaliske University of Leipzig, Germany

L. Kwasniewski Warsaw University of Technology, Poland

D. Pintea & R. Zaharia University of Timisoara, Romania

2.1.1 INTRODUCTION The main objective of a fire-structure analysis is to predict the effects of fires in buildings, e.g. the fire resistance and the structure’s performance under heating and cooling caused by fire. The results of such analysis can be applied in the design of fire protection systems, in the evaluation of fire safety and as an addendum of experiments. Advanced calculation techniques can be helpful in the areas where experiments encounter difficulties such as testing large specimens, implementation of loading and boundary condition, measurements and interpretation of specimen’s behavior. A computational model used for fire-structural (member or global) analysis should properly represent the considered problem in terms of: type of analysis and solution methods, geometry, temperature dependent material properties, mechanical boundary conditions and loading, thermal conditions. The fire resistance analysis of reinforced concrete (RC) structures faces additional challenges and constitutes an important part in their design. From the constructional point of view, buildings and structures at fire have to carry mechanical loadings and thus provide safe people evacuation (rescue) and safe firemen work. High temperatures have a very significant adverse effect on thermomechanical properties of RC members. High temperature substantially reduces strength of concrete and steel, and causes significant increase in cracking, strains and deflections. Load bearing capacity of structure decreases and may fail at critical points. 2.1.2 TYPE OF ANALYSIS AND SOLUTION METHODS Depending on the simulated test scenario, three types of analysis can be considered: structural, thermal or coupled structural-thermal. Structural stress analysis should be able to take into account strains due to elastic and plastic deformation and due to thermal elongation if coupled structural-thermal analysis is performed. Creep strains can usually be omitted for transient analysis. Incremental, transient structural analysis should be based on explicit or implicit methods for 111

time integration. Application of explicit methods in coupled structural-thermal fire analysis is not feasible due to consideration of relatively long time intervals. For the thermal calculations usually unconditionally stable implicit time integration is applied (Hallquist 2006). One can choose between general purpose commercial programs and research oriented specialized unique programs developed by academia. In both cases, the majority of today’s computer programs, dedicated to structural analysis, are based on the Finite Element (FE) Method. Well validated nonlinear codes are preferred over specialized, experimental computer programs. A chosen code should cover all analysis aspects, important for the considered case.

2.1.3 THERMAL PROPERTIES OF MATERIALS In fire conditions the temperature dependent properties shall be taken into account. The thermal and mechanical properties of steel, concrete, aluminium should be determined from the following clauses. 2.1.3.1 Steel (EN 1993-1-2) The relative thermal elongation of steel l/l is given in formulae (3.1 a-c) from EN-1993-1-2). In these formulae the thermal elongation of steel is computed as function of the steel temperature θa . EN 1993-1-2 gives formulae (3.2 a-d) for computing the specific heat of steel ca as function of the steel temperature θa . The thermal conductivity of steel λa is given by the formulae (3.3 a-b) as function of the steel temperature θa . The graphical representation of these formulae is also given for each of the thermal properties. The thermal conductivity of steel as function of the temperature is presented in Figure 1. 2.1.3.2 Aluminium alloys (EN 1999-1-2) The formulae for computing the relative thermal elongation (strain) of aluminium alloys l/l are given in paragraph “3.3.1.1 Thermal elongation” from EN 1999-1-2 as function of the aluminium temperature θal . The formulae for computing the specific heat of aluminium cal as function of the aluminium temperature are given in paragraph “3.3.1.2 Specific heat”. The variation of the specific heat of the aluminium alloys with the temperature is presented in Figure 2. Similarly the computation of the thermal conductivity of aluminium alloys as function of the aluminium temperature is given in paragraph “3.3.1.3 Thermal conductivity”. 2.1.3.3 Concrete with siliceous and calcareous aggregates (EN 1992-1-2) The thermal strain of concrete εc (θ) is given in formulae as function of concrete temperature for siliceous and calcareous aggregates in paragraph “3.3.1 Thermal elongation” (EN 1992-1-2). The formulae for computing the specific heat cp (θ) of dry concrete (u = 0%) with siliceous and

Figure 1. Thermal conductivity of steel at elevated temperature.

112

calcareous aggregates is given in paragraph “3.3.2 Specific heat” (EN 1992-1-2) as function of the concrete temperature. Where the moisture content is not considered explicitly in the calculation method, the function given for the specific heat of concrete with siliceous or calcareous aggregates may be modelled by a constant value, cp.peak , situated between 100◦ C and 115◦ C with linear decrease between 115◦ C and 200◦ C. The thermal conductivity λc of concrete may be determined between lower and upper limit values, given in paragraph “3.3.3 Thermal conductivity” as function of the concrete temperature. The thermal conductivity of concrete is presented in Figure 3. 2.1.3.4 Natural stones The heating causes a colour change of stones (Fig. 4a). Not only colour but also other external signs of heat are observed. Limestone samples are cracked at lower temperatures while at higher temperature the samples collapsed or exploded (Fig. 4b). According to the thermal decomposition of carbonates this processes is dedicated to the formation of new mineral phases (portlandite).

Figure 2.

Specific heat of aluminium alloys as a function of the temperature.

Figure 3. Thermal conductivity of concrete.

Figure 4. a) Visible colour changes of different stone types before heating and after heating from 150◦ C to 750◦ C. T-Tardos compact limestone, F-Sütto travertine, D-Sóskút coarse limestone, Rt-Egertihamér rhyolite tuff, V-Balatonrendesi sandstone, b) Crack formation nad disintegration of cylindrical sample of Sóskút coarse limestone sample. after heating on 900◦ C (after Hajpál 2008).

113

Figure 5.

a) Scaling at window edges in Lobenfeld b) Rounding of edges in Dresden.

Figure 6.

Stress-strain relationship for carbon steel at elevated temperatures.

The most important kind of decay of stones due to fire are scaling off (Fig. 5a), spalling, cracking, rounding off the edges (Fig. 5b). Fire can completely destroy ornaments and can damage carved forms. Fire damaged stones are often replaced by new ones (Hajpál 2000).

2.1.4 MECHANICAL PROPERTIES OF STRUCTURAL ELEMENTS 2.1.4.1 Carbon steel (EN 1992-1-2, EN 1993-1-2) For heating rates between 2 and 50 K/min, the strength and deformation properties of structural steel at elevated temperatures should be obtained from the stress-strain relationship given in Figure 6. Table 3.1 (EN1993-1-2) gives the reduction factors for the stress-strain relationship for steel at elevated temperatures. These reduction factors are defined as follows: effective yield strength, relative to yield strength at 20◦ C:

proportional limit, relative to yield strength at 20◦ C:

slope of linear elastic range, relative to slope at 20◦ C:

114

Figure 7. Mathematical model for stress-strain relationships of concrete under compression at elevated temperatures.

2.1.4.2 Aluminium alloys (EN 1999-1-2) For thermal exposure up to 2 hours, the 0,2% proof strength at elevated temperature of the aluminum alloys listed in Table 1 (EN 1999-1-2), follows from:

where fo,θ is 0,2 proof strength at elevated temperature fo is 0,2 proof strength at room temperature according to EN 1999-1-1. For intermediate values of aluminium temperature linear interpolation may be used. The modulus of elasticity of all aluminium alloys after two hours thermal exposure to elevated temperature Eal,θ should be obtained from Table 2 (EN 1999-1-2). 2.1.4.3 Concrete (EN 1992-1-2) For heating rates between 2 and 50 K/min, the strength and deformation properties of compressive concrete at elevated temperatures should be obtained from the stress-strain relationship given in Figure 7. The stress-strain relationships given in Figure 7 are defined by two parameters: the compressive strength; the strain εc1,θ corresponding to fc,θ Table 3.1 (EN 1992-1-2) gives for elevated concrete temperatures θc , the reduction factor kc,θ to be applied to fc in order to determine fc,θ and the strain εc1,θ . For intermediate values of the temperature, linear interpolation may be used. The parameters specified in Table 3.1 may be used for normal weight concrete with siliceous or calcareous (containing at least 80% calcareous aggregate by weight) aggregates. Values for εcu1,θ defining the range of the descending branch may be taken also from Table 3.1 (EN 1992-1-2). For thermal actions in accordance with EN 1991-1-2 Section 3 (natural fire simulation), particularly when considering the descending temperature branch, the mathematical model for stress-strain relationships of concrete specified in Figure 7 should be modified. The tensile strength of concrete may be assumed to be zero, which is on safe side. If it is necessary to take account of the tensile strength, when using the simplified or general calculation method may be used. The reduction of the tensile strength of concrete is allowed for by the coefficient kct,θ for which fct,θ = kct,θ fct . In absence of more accurate information kct,θ values specified in Figure 3.2 (EN 1991-1-2) should be used. The main factors affecting the compressive and tensile strength of concrete are mix proportions, water/cement ratio, aggregate/cement ratio, type of aggregate (Schneider & Horvath 2003). Comparison of experimental and predicted strengths of compressive and tensile strength of concrete subjected to fire is presented in Figure 8 and 9, respectively. The strain components at any stress level can be modelled using the superposition theory whereby the total strain is considered to be the sum of various strain components (Schneider & Horvath 2003):

115

Figure 8. Comparison of experimental (1–3) and predicted (4–9) strengths of compressive concrete subjected to fire: (1) T. T. Lie, (2) H. L. Malhotra, (3) M. S. Abrams, (4) T. T. Lie and T. D. Lin, (5) K. D. Hertz, (6) T. T. Lie et al., (7) STR 2.05.11:2005 (Lithuanian Code), (8) L. Li ir J. A. Purkiss, (11) EC2.

Figure 9. Comparison of experimental (1–9) and predicted (10–13) strengths of tensile concrete subjected to fire: (1) T. T. Lie, (2) Y. Anderberg and S. Thelandersson, (3), (4) Blundell, (5), (6) Z. P. Bažant and J. C. Chern, (7), (8) Sager, (9) U. Schneider, (10) Z. P. Bažant and J. C. Chern, (11) J. Xiao and G. Konig, (12) M. J. Terro, (13) EC2.

where εtot is the total strain, εσ is the stress–related strain, εth is the thermal strain, εtr,cr is the transient creep strain often called load induced thermal strain, θ is the temperature, t is the time, σ is a stress, σ is the stress history. The stress–related strain is a function of the applied stress and the temperature. It may be split into elastic and plastic part. The thermal strain is the free thermal expansion resulting from fire temperatures. It is mainly influenced by the type and amount of aggregate. Calculation of thermal strain for concrete is given in the section 4.3 of the technical sheet. The main factor affecting the thermal strain is the type of aggregate. The coarse aggregate fraction plays a dominant role (Schneider & Horvath 2003). Transient creep strain or load induced thermal strain (lits) develops during first heating under load. It is unique to concrete amongst structural materials. Lits is much larger than the elastic strain, and contributes to a significant relaxation and redistribution of thermal stresses in heated concrete structures. Any structural analysis of heated concrete that ignores lits will, therefore, be wholly inappropriate and will yield erroneous results, particularly for columns exposed to fire. This phenomenon is still not fully appreciated by structural engineers and should be incorporated more fully into standards and design codes, Khoury (2000). The main factors affecting the transient strain are type of aggregate, aggregate/cement ratio, curing conditions, loading level (Schneider & Horvath 2003). Mathematical models for transient thermal strain calculations are reviewed by Youssef & Moftah (2007). It is evident that one could add a shrinkage strain component to Equation 2. However, since all experimental high temperature data are reported from unsealed test conditions the shrinkage component can be viewed as being included in the thermal strain. Furthermore, shrinkage is assumed to be independent of loading (Nielsen et al. 2004). The other phenomenon of concrete at high temperatures is explosive spalling, which results in loss of material, reduction in section size and exposure of the reinforcing steel to excessive temperatures. Spalling is the violent or non-violent breaking off of layers or pieces of concrete from the surface of a structural element when it is exposed to high and rapidly rising temperatures (Khoury 2000). 116

Figure 10.

Uniaxial compressive strength of different stone types as function of the heating temperature.

Figure 11. Compressive stress as function of the negative axial deformation of a coarse limestone at after heating on different temperature.

When the moisture content of the concrete is less than 3% by weight explosive spalling is unlikely to occur. Above 3% more accurate assessments, moisture content, type of aggregate, permeability of concrete and heating rate should be considered. Spalling can be grouped into four categories: (a) aggregate spalling; (b) explosive spalling; (c) surface spalling; (d) corner/sloughing-off spalling. The first three occur during the first 20–30 min into a fire and are influenced by the heating rate, while the fourth occurs after 30–60 min of fire and is influenced by the maximum temperature. The main parameters affecting the spalling effect are content of moisture in concrete, the heating condition, compressive stresses, thickness of concrete, position of reinforcement, mix proportion, fibre volume (Schneider and Horvath 2003). The prediction of spalling is now becoming possible with the development of thermo-hydro-mechanical nonlinear finite element models capable of predicting pore pressures (Khoury 2000). 2.1.4.4 Stones Natural stones are considered as less sensitive materials to fire. According to testing of different natural stone types at various temperatures, it has been proved that fire can cause rapid and irreversible physical changes (Hajpál & Török 2004, Hajpál 2008). These alterations negatively influence the strength and static behaviour of the whole monument (Hajpál 2008). Test results have shown that the compressive strength of various lithologies depends on the heating temperature. It can be observed, that the heating does not cause a decrease in the strength for all rock types. The Balatonrendes and Ezüsthegy sandstone and also the Egertihamér rhyolite tuff have higher strength after the heating at 900◦ C than at room temperature The limestone types lost their strength only at elevated temperatures (Fig. 10). Strength parameters and axial deformation of limestones do not change uniformly. The tests have demonstrated the differences of compressive stress and axial deformation with increasing temperature (Fig. 11). 117

Figure 12.

Indirect tensile strength as function the heating temperature by different stone types.

Figure 13.

Determination of fire resistance for assumed loading, Kosiorek (2002).

Indirect tensile strength of limestones shows slight increase up to 150◦ C, which is followed by a decrease, while the tensile strength of sandstones and rhyolite tuff do not reflect a clear trend with increasing temperature (Fig. 12).

2.1.5 LOADING, BOUNDARY AND INITIAL CONDITIONS In engineering practice there are typically two loading and heating scenarios taken into account. One scenario considers increasing static loading in constant elevated temperature. This scenario is used to determine critical loading for selected temperatures. In the second scenario the structure is analyzed under constant mechanical loading but at increasing temperature. The objective for the case of structural steel members is to determine the critical temperature and time. Using repeatable calculations following both scenario diagrams, shown in Figure 13, load-temperaturetime relationships can be formulated and for assumed loading L the fire resistance can be determined in terms of critical temperature Tcr or critical time tcr . In the third scenario, e.g. following experiment or actual fire, both temperature and loading are time depended. Loading can be temperature dependant due to thermal elongation. Defined mechanical boundary conditions, loading and interactions should be relevant with the actual features of the analyzed member or structure. Depending on the type of analysis, the mechanical loading can be represented by pressures and forces, or prescribed displacements. If necessary, the time or temperature dependent boundary conditions can be represented by contact definitions.

2.1.6 THERMAL CONDITIONS Depending on the considered loading scenario, thermal conditions can be modelled applying variety of time and temperature dependent boundary conditions including prescribed temperature fields, insulation, flux, convection, and radiation. Direct thermal loading. Prescribed temperatures as functions of time are applied to the model nodes (including internal). Relevant for structural analysis without heat conduction. 118

Constant or time dependent prescribed temperatures applied to selected nodes, on external model surfaces. In this simplified approach heat transfer between surroundings and the model external surfaces is not analysed. Heat transfer inside the model is included. Applicable for insulation layers. Full insulation. The heat transfer on the model surface is fully inhibited. Prescribed flux, applied to external model surfaces. The flux can be time or temperature dependent or constant through analysis. Can be used for thermal or coupled structural-thermal analysis. This option requires reliable data specifying the flux magnitude. Heat transfer between a member and surroundings, defined in terms of convection and radiation. Convection and radiation can be defined for selected model external surfaces. Applicable for thermal and coupled structural-thermal analysis. Transient convection can be expressed by the following formula (Shapiro 2005), (EN1991-1-2, 2002)

•

where: hnet,c – is net convective heat flux in [W/m2 ] αc – convective heat transfer coefficient [W/m2 K], can be constant or temperature dependent. Depends on the model’s material, surface finish, fire protection and type of the surrounding gas (EN1991-1-2, 2002). TM – calculated current temperature on the model surface [K], TS – prescribed temperature of the surroundings can be constant or time dependent (e.g. nominal temperature curve) [K]. Radiative transfer between gas and member can be expressed as (Shapiro 2005), (EN1991-1-2, 2002):

•

where hnet,r – is net radiative heat flux in [W/m2 ] εm – surface absorptivity (emissivity) coefficient can be constant or time/temperature dependent (Hallquist 2006). Depends on the model’s material, surface finish, and fire protection, σSB – is the Stefan Boltzmann constant [5, 67 × 10−8 W/m2 K4 ], TM and TS – the same as above [K]. Compound convective heat transfer can also be used with resultant heat transfer coefficient including radiation effects (EN1991-1-2, 2002). 2.1.7 BEHAVIOUR AND MODELLING ASPECTS 2.1.7.1 Stability check for unbraced steel rigid frames in case of fire At normal temperature design, where it is necessary to consider the influence of the deformed geometry of the structure (2nd order effects) to verify the stability of columns belonging to a structural framed system, when global frame imperfections are considered but member imperfections are not taken into account, two procedures can be adopted (EN 1993-1-1, 2005, ECCS 2006): i) to perform a 2nd order analysis including the effects of lateral displacements and check of the member instability with non-sway buckling lengths; and ii) to perform a 1st order analysis and check of the member instability with sway buckling lengths. For the first procedure, it should be noted that nonsway effective lengths can be used because no sway will occur in addition to that which causes the second-order effects calculated by a P- second-order analysis. For sake of simplicity, when using the first methodology, the buckling length of a member may be taken as its system length, which is safe and suggested by the Eurocode 3 (CEN 1993-1-1, 2005) for normal temperature design. In fire situation, Eurocode 3 (CEN 1993-1-2, 2005) states that, using simple calculation methods, a global analysis of the frame should be done as for normal temperature and the “buckling length lfi of a column for the fire design should generally be determined as for normal temperature design”. However, in the case of a braced frame in which each storey comprises a separate fire compartment with sufficient fire resistance, the buckling length, lfi , of a continuous column may be taken as 0,5L in an intermediate storey and 0,7L in the top storey, where L is the system length in the relevant 119

storey. For unbraced structures no specific guidance is given by the Eurocode. For these cases this work shows that considering a buckling length of the columns in a sway mode, independently of the 2nd order effects being negligible or not (so-called P- effects) at normal temperature, leads to good results, for the case of regular multi-storey buildings. Only few studies were made on that subject. Publication no. 159 from Steel Construction Institute (SCI 1996) proposes for the case of columns in sway frames in fire conditions that the effective slenderness ratio may conservatively be taken as λθ = 1.25λ (considering the buckling length equal to the system length). A publication from ECCS (1983) suggests that if a global analysis of the frame is not performed to take account of instability effects at elevated temperature, default critical temperature of 300◦ C should be considered, which is too conservative. A buckling length equal to the system length is also suggested by Wang (1997). A global analysis including the instability effects at elevated temperature is rather complex to be used with simple calculation models, therefore simple and safe procedures should be available for design purpose. The methodology for fire design with simple calculation models consists on evaluating the internal forces in the structure as for normal temperature considering the accidental load combination for fire situation and then checking the fire resistance of each member separately. This was the procedure adopted in the parametric study carried out in this work where the simple calculation model was performed throughout the software Elefir EN (Vila Real & Franssen 2010) and the advanced calculation model SAFIR (Franssen 2005) was used for comparison. Part 1.2 of Eurocode 3 states that the buckling length lfi of a column for the fire design situation should be determined as for normal temperature design. It is not clear if it should be used the same procedure but considering the mechanical properties of steel, namely the Young’s modulus, at elevated temperature. If elevated temperature should be used, the process is not an easy task for design purposes. Due to this difficulty, the Wood method (ECCS 2006) at normal temperature has been used in this work to evaluate the buckling length ratio (lcr /L) of the columns. The buckling lengths at elevated temperature were considered with the same value as at normal temperature, i.e., lfi = lcr . According to the Wood method the buckling length of a column in a non-sway or sway mode may be obtained from Figure 14 and Figure 15 respectively, function of the distribution factors η1 and η2 , which are given by:

Figure 14.

Buckling coefficient lcr /L for non-sway frames.

Figure 15.

Buckling coefficient lcr /L sway frames.

120

where Kc , K1 and K2 are the flexural stiffness coefficients (EI/L) for the adjacent length of columns, and Kij are the effective beam flexural stiffness coefficient. For beams with double curvature Kij = 1.5 EI/L and for the case of single curvature Kij = 0.5 EI/L. In this work for checking the fire resistance of unbraced steel frames by simple calculation models, the internal forces were obtained at normal temperature performing a first order analysis and the member instability was checked using sway buckling lengths. Buckling lengths equal to the system length were also used for comparison. Beams were assumed to be heated on three sides and all the columns on four sides by the standard fire curve ISO 834. Starting from an unbraced steel frame of a three bay – three storey office building shown in Figure 16, several combinations of different numbers of bays and storeys were considered in the parametric study. This frame has been analysed at normal temperature in the publication no. 119 from ECCS (2006). The members are made of hot-rolled profiles of steel grade S235 being the external columns in HE 220 B, the internal columns in HE 260 B, the intermediate beams in IPE 450 and the top beams in IPE 360. The structure was assumed to be braced in the out of plane direction and unbraced in the plane of the frame. Out of plane column buckling is prevented and lateral restraint is assumed to be provided to the beams by the concrete floor and roof slabs. The columns are continuous throughout of the full height of the building. The load combinations for accidental fire situation used are listed in Table 1, where G refers to the permanent loading, W to the wind loading and I1 , I2 and I3 to the imposed loads alternance. Frame imperfections due to unavoidable initial out-of-plumb were taken into account prescribing a notional horizontal force that was applied at each story level (EN 1993-1-1, 2005). A parametric study has been performed considering several combinations of bays and storeys from 1×1 to 3×3 in a total of 9 unbraced frames as shown in Figure 17. The frames were considered to be pinned or fixed at the supports and the seven load combinations presented in Table 1 were considered. The fire scenarios used with the advanced calculation model were the standard fire acting in each storey separately from the ground floor to the upper floor. The results of the parametric study plotted in Figure 18 show that the proposal made to consider the sway buckling lengths in the case of unbraced frames is mostly on the safe side when compared with the advanced calculation method. This figure also shows that if the system length is considered for the buckling length the results are too unsafe.

Figure 16.

Frame geometry. Table 1. Load combination cases. Load combination

Accidental combination

Case 1 Case 2 Case 3 Case 4 Case 5 Case 6 Case 7

Gk + 0.2Wk Gk + 0.5I1 Gk + 0.5I2 Gk + 0.5I3 Gk + 0.2Wk + 0.3I1 Gk + 0.2Wk + 0.3I2 Gk + 0.2Wk + 0.3I3

121

Figure 17.

Frame geometry used in the parametric study.

Figure 18. Comparison between simple calculation model (Elefir-EN) and advanced calculation model (SAFIR). a) lfi = L; b) lfi = kL (k obtained with Wood method for-sway frames).

Figure 19. ture 20◦ C.

Contours of Mises stress under increasing mechanical loading at constant ambient tempera-

2.1.7.2 Computer simulation of a steel connection at elevated temperature Experiments are the most reliable source of information on responses of real structures, and the only method of final validation of the finite element (FE) analysis. However, the high cost of full scale laboratory tests and difficulties with collecting extensive data lead to growing interest in analytical and computational methods. With the increasing computing capabilities nowadays, it is possible using the Finite Element Method to simulate complex real cases and consider wide range of parameters. A reliable, analytical investigation can reduce costs dramatically and allow for faster introduction of new design improvements and maintenance decisions. As an example Figures 19–22 show results of a feasibility study on a coupled structural thermal analysis of a beam to column connection subjected to fire, Tybura & Kwasniewski (2008). Numerical results in the form of moment-rotation characteristics are compared with the published data for a selected flush end-plate connection, Al-Jabria et al. (2006). The Finite Element (FE) analysis is conducted using commercial program LA-DYNA® , Hallquist (2006). The considered connection consist of two 254 × 102UB22 beam segments connected to a 152 × 152UC23 column 122

Figure 20.

Moment versus rotation curve at constant temperature 20◦ C.

Figure 21. Temperature versus rotation.

Figure 22.

Contours of Mises stress for bending moment 17 kNm and increasing temperature.

using 8 mm thick flush end plate and six M16 bolts. The test setup and all di-mensions are provided in Al-Jabria et al. (2006). Several three dimensional simplified FE models were developed using 4 – node shell elements for the purpose of a global analysis intended for a large scale structure. The model, shown in Figure 19, represents one forth of the configuration. This model is appropriate for symmetrical loading and temperature conditions. The bolts are represented by 1D beam elements connected to rigid shell elements in end plate and column flange, used to better distribute forces around bolts heads and nuts. A temperature dependent elastic – plastic material model with strain hardening, was applied. It allows for relating material parameters such as: elastic modulus, Poisson’s ratio, coefficients of thermal expansion, yield stress, and plastic hardening modulus to temperature, represented in a discrete way at selected points. All components of the tested connection, except the bolts, were made of steel S275. The stress-strain relationships of the steel S275 at elevated temperatures, calculated based on the Eurocode 3 (CEN (2005a)). For coupled structural thermal and thermal only analyses, thermal properties such as heat capacity and thermal conductivity are specified in the additional material model, called “thermal_isotropic”, Hallquist (2006). All thermal parameters can also be 123

defined as temperature dependent. For the bolts assumed yield and ultimate stresses were 480 and 600 MPa, respectively, Al-Jabria et al. (2006). Depending on the considered case, the loading can be performed as a predefined displacement or concentrated force applied to a selected node rigidly connected with the beam segment. The top and the bottom the column and the middle section of the column web are constrained. For the model of one forth of the test configuration additional constraints are applied on the vertical symmetry plane. Included in the FE model segments of the beam and the column represent regions with the maximum deformation, where the influence of the assumed boundary conditions can be neglected. In the FE model the beam is connected to the column through contact between column flange, end plate and bolts. Due to internal complexity of con-tact algorithms incorporated in the FE programs, an analysis including contact is usually challengeable, can affect results and even lead to problems with convergence. The calculated results were compared with the experimental data presented in the paper, Al-Jabria et al. (2006). All structural analysis presented here are based on static calculation using implicit solver, where instead of physical time a loading parameter is applied, Hallquist (2006). The temperature is applied uniformly to all the nodes, simulating furnace test conditions. Depending on considered case during the simulation temperature is constant or increases with the loading parameter. Curves in Figure 20 show comparison of calculated and experimental relationships between moment and rotation, for increasing loading at ambient temperature 20◦ C. Figure 19 presents contours of Mises stress for the same loading case. Curves in Figure 21 show calculated relationships between temperature and rotation, for four loadings producing moments at the connection M = 4, 8, 13, and 17 kNm. The loading is applied gradually at the begging of the simulation and then kept constant while the temperature is increased from 100 to 800◦ C. Points in Figure 21 represent experimental values. Comparison with the experiment shows higher resistance of the FE model, mainly due to overestimated material parameters for larger strains. Figure 22 shows contours of effective Mises for bending moment 17 kNm and increasing temperature. In all figures the Mises effective stress is mapped with the same color scale for the range from 0 to 275 MPa. The problem of moment resistance degradation in elevated temperatures was simplified herein in terms of material representation and heat transfer. Chosen material model allows only for coarse piecewise linear approximation of stress – strain relationships through specification of hardening modulus. This approach leads to overestimated stresses for higher strain values and results with higher loading values comparing to the experimental data. Due to high value of thermal conductivity of steel temperature distribution during fire can be assumed as uniform and heat transfer does not have to be considered. For concrete and composite (concrete and steel) structures such approach can be insufficient. 2.1.7.3 Fully coupled temperature-displacement analyses of steel portal frames under fire The behaviour of steel structures under fire needs particular attention since the structural steel undergoes considerable deterioration in presence of high temperatures, such as the reduction of both resistance and stiffness of steel. This can cause the collapse of structures that are safely designed for ordinary load combinations, in which the fire scenario is disregarded. Consequently, the behaviour in fire of steel structures requires deep investigations from both experimental and numerical points of view. Recently a numerical study aimed at investigating the behaviour of steel structures under fire based on the use of fully coupled temperature-displacement finite element analyses, carried out by means of the advanced computer program ABAQUS has been presented (Faggiano et al. 2007a). The used method allows to consider at the same time the mechanical and thermal aspects of the problem. The mechanical and thermal problems are faced up in a unique model, in which the actual phases of the modelled phenomenon, say the sequential application to the structure of the design loads and, then, of the fire scenario, are reproduced in a step-by-step analysis. Such approach differs from the usually adopted one, which consists, for the sake of simplicity, in performing the heat transfer analysis and the mechanical one separately (uncoupled analyses): the first one allows to evaluate the temperature-time law within the structural elements exposed to fire, completely neglecting the stress-displacement aspect; the second one consists in the usual structural analysis, 124

Figure 23.

Finite element meshes of the models (Faggiano et al., 2007a).

in which the structure is subjected to the external loads; at the end of the structural analysis, the temperature-time variation, obtained from the preliminary heat transfer analysis, is imposed to the structural members, so allowing the calculation of the fire resistance of the structure. On the contrary, in the case of fully coupled temperature-displacement analyses, the used finite elements are endowed with both displacement and temperature degrees of freedom, so that the mechanical and thermal equations are written simultaneously and the mutual interactions between the two aspects of the problem can be easily caught. The study dealt with simple steel portal frames, focusing on the main geometrical and mechanical parameters that influence the fire resistance of the considered structures, such as the span over height ratio, the massivity ratio of the structural members, the steel grade and the exploitation degree of the material (Faggiano et al., 2007a). However such a methodology was already applied for the investigation of the behaviour of steel structures exposed to fire after being damaged by an earthquake (Faggiano et al., 2007b). Other studies on the subject, based on different approaches, are presented in (Della Corte and Landolfo, 2001; Della Corte et al., 2003a, b; Della Corte et al., 2005; Faggiano et al., 2005). 2.1.7.4 Simplified tool for analysis of RC members under fire Lack of experimental and theoretical investigations on behavior of RC members under high temperatures hampers development of the constitutive laws. Compressive strength does serve as a sufficient parameter to characterize thermal, physical and mechanical properties of concrete at elevated temperatures. Geda (2010) performed a comprehensive study aiming at numerical analysis of thermo-mechanically loaded RC members. The analysis has shown that: 1) A universal thermal and physical-mechanical model for concrete has not been proposed until now. Due to limited experimental and theoretical investigation of behavior of RC members, the material models are not accurate. 2) Compressive strength of concrete is not a sufficient parameter characterizing thermal, physical and mechanical properties of concrete at elevated temperatures. 3) Most physical and thermo-mechanical parameters are most accurately characterized by the Eurocode 2 (CEN 2004). 4) Some characteristics of concrete were best predicted by other models: modulus of elasticity by Xiao & Konig (2004), strain corresponding to maximal stress by Khennane & Baker (1993), limit strain by Terro (1998), transient creep strain by Anderberg & Thelandersson (1976). The behavior of RC structures at elevated temperatures is very complex. With rising temperature, thermal, physical and mechanical properties of concrete and reinforcement significantly change. Analytical and computation methods (Huang et al., 1999, Bratina et al., 2007, Capua Di & Mari 2007, and Kodur & Dwaikat 2008) have been extensively developed in the field of RC building exposed to high temperature or accidental fire. However, in the analysis an engineer usually employs various formulae for the fire resistance of structures offered by building codes (CEN 2004), without really understanding the thermo-mechanical behavior of a structure during fire (Bratina et al., 2007). On the other hand, advanced non-linear mechanical models based on the 2D or 3D finite element (FE) method (Huang et al., 1999, Cervenka et al., 2005, and Capua Di & Mari 2007) which were rapidly progressing within last three decades are based on universal principles and can include all possible effects. However, such methods are highly demanding in terms of the computational recourses. Besides, the constitutive laws taking into account the high temperature effects are not accurate enough. Recent fires with fatalities stimulate new investigation wave for providing fire resistance in reinforced concrete structures subjected to high temperature. Analytical and numerical methods are widely used for analysis of fire resistance in reinforced concrete structures. However, 125

Figure 24.

Constitutive law for cracked tensile concrete.

advanced FE methods based on non-linear material models are highly expensive in terms of the computational time. Therefore, their application is limited to simple cases. A numerical procedure, based on Layer section model and smeared crack approach, aiming at deformation analysis of bending RC members, has been developed and improved at Vilnius Gediminas Technical University. The procedure assures higher accuracy of deflection predictions in comparison to the design code methods. An efficient combination of accuracy and simplicity has been achieved in the Layer section model. This allowed incorporating it into a simple engineering technique based on classical principles of strength of materials extended to layered approach and use of full material diagrams. In this study, an attempt has been made to extend application of the Layer section model to stress and strain analysis of RC bending members subjected to high temperature, taking into account non-linear physical and thermo-mechanical materials properties. Proposed numerical procedure is developed to assess the stress-strain state, load bearing capacity and failure time of RC members. This approach is very effective in terms of computer resources, i.e. the calculation time decreases hundreds of time in comparison to standard non-linear FE programs (MSC.MARC, DIANA, and ATENA). The Layer section model for RC members subjected to elevated temperatures is based on the following assumption and approaches: 1) Smeared crack approach, i.e. average stresses and strains are used. 2) Linear distribution of strain within the depth of the section, i.e. the Bernoulli hypothesis is adopted. 3) Perfect bond between concrete and reinforcement is assumed; reinforcement slippage occurring at advanced stress-strain states is included into stress-strain diagram of tensile concrete. 4) Temperature is increasing, i.e. the cooling-down stage is not considered. 5) Thermal strain as well as transient creep strain are assessed as equivalent axial forces and bending moments. Behavior of compressive concrete and reinforcement is modeled according to Eurocode 2 (CEN 2004). Behavior of tensile concrete is modeled by the bilinear stress-strain relationship shown in Figure 24 with tensile strength and modulus of elasticity taken from Eurocode 2 (CEN 2004). In this figure, θ is the temperature arising in the layer. The descending branch of the diagram is characterized by the ultimate strain (Kaklauskas 2004):

Here p is the tensile reinforcement ratio (%); fct and Ec are the tensile strength and the modulus of elasticity of concrete, respectively. The latter two parameters are derived according to Eurocode 2 (CEN 2004). A cross-section as shown in Figures 25a and 25b is divided into a number of layers corresponding to either concrete or reinforcement. Temperature gradient (Fig. 25c) within the section was determined in the using the approaches and assumptions of heat transfer theory. Nonlinear distribution of high temperatures was replaced by equivalent axial force and bending moment. Variable mechanical and physical properties of every layer can be evaluated in the analysis duo to loading and temperature effect. Cross-section of RC members is replaced by transformed concrete sections. This is performed by multiplying area of i-th layer by ratio of modulus of elasticity Ei (θ, εσ )/Ec (20◦ C), where Ei (θ, εσ ) is the temperature-dependent secant modulus of elasticity of 126

Figure 25. Stress and strain caused by non-linear temperature gradient (Bacinskas et al., 2008): RC cross-section (a); Layer section model (b); Temperature gradient (c); Distribution of strain (d) and stress across the section (e).

i-th layer, Ec (20◦ C) is the modulus of elasticity of concrete at normal temperature. Geometrical characteristics of transformed cross-section (area Ac,eff , first moment Sc,eff and second moment of inertia Ic,eff ) determined about the top edge of the section:

Here n is the total number of layer; bi and ti are the width and the thickness of the i-th layer, respectively; yi is the distance of i-th layer from the top of the section. Concrete total strain can be expressed as follows:

Here εσ (T , σ) is the stress-induced strain; εth (θ, θ0 ) is the thermal strain; and εttc (θ, σ) is the strain due to transient thermal creep. Thermal-induced strain in the i-th layer can be expressed as follows:

Here αi is the coefficient of thermal expansion of i-th layer. It is recommended to calculate this coefficient according to Eurocode 2 (CEN 2004) recommendation. Transient thermal creep strain for tensile concrete layers is neglected. The strain for compressive concrete can be calculated according to Anderberg & Thelandersson (1976):

Here β is the material coefficient, which varies from 1.8 up to 2.35; is the compressive stress in the i-th layer; fc (20◦ C) is the compressive strength of concrete under normal condition. Transient thermal creep strain is taken negative. It should be noted that temperature in the cross-section of the element under fire is distributed non-uniformly. Furthermore, physical and mechanical properties of layers are varying. Free extension of every layer is limited by adjacent layers. Therefore, bending stresses (Fig. 25e) arise in the element cross-section due to layer mutual interaction. Influence of strains εth (θ, θ0 ) and εttc (θ, σ) on the stress-strain state of RC member can be evaluated by introducing equivalent axial forces and bending moments. Equivalent axial force can be defined as a sum of appropriate stresses:

127

Equivalent bending moments can be determined analogically:

Given equivalent actions are applied additionally to the RC element subjected to external loads Next and Mext :

Then strain at top fiber and curvature for the section under consideration can be determined:

Total strain at any point of the section is defined in terms of above parameters:

From expression (3), stress-induced strain in i-th layer can be determined as follows:

Then stress in i-th layer is derived using the following equation:

In Equation 16, transient creep strains for layers, which correspond to reinforcement and tensile concrete, are assumed to be equal to zero. Proposed numerical procedure for deformational analysis of RC members subjected to thermomechanical loading has been performed iteratively by the following steps: 1. Geometrical characteristics are calculated for the transformed cross-section by Equation 7. In the first iteration, linear materials properties are assumed both for concrete and steel layers taking into account temperature effects. 2. Equivalent axial forces and bending moments are calculated by Equations 11 and 12. Total actions applied to the element are obtained using Equation 13. In the first iteration, transient thermal creep-induced actions are assumed to be equal to zero. 3. Strain at top fibre and curvature are calculated using Equation 14. 4. Total strain is derived by Equation 15 for each of layers. 5. For the assumed constitutive law of reinforcement and concrete, stress is calculated using equation (12) for each of layers. Secant deformational modulus in the layer is determined as a ratio between given stress and stress-induced strain, obtained by Equation 16. 6. Obtained deformational modulus is compared with previously assumed or calculated one for each of the layers. If the agreement is not within the assumed errors limits, a new iteration is started from step 2. Over vice, obtained values of strains, stresses and curvatures are assessed. For deflection calculation which is performed by Mohr’s integral technique, analogous computations are carried out for other sections of the member. Proposed model can assess the stress-strain state, load carrying capacity and failure time of RC members. This model is simple and versatile. Its simplicity is due to use of classical formulas of mechanics of materials. Application of a uniform model in the short- and long-term analysis (including shrinkage and creep) for both ordinal and pre-stressed RC members under normal and high temperatures characterizes the versatility of the model. The proposed algorithm is very effective in terms of computer resources, i.e. the calculation time decreases hundreds of times (from hours to few seconds) in comparison to standard non-linear finite element programs. 128

Figure 26.

Normalized deflections of Slabs 1 (a) and 2 (b) subjected to ISO 834 fire conditions.

The Layer section model has been applied to perform stress and strain analysis of flexural RC members subjected to high temperature, taking into account non-linear physical and thermomechanical material properties. A number of numerical studies performed by the authors (see for instance: Gribniak et al., 2006, Kaklauskas et al., 2007, Bacinskas et al., 2008, and Geda 2010) show satisfactory accuracy and computational effectiveness of the proposed procedure. To illustrate application of the procedure, a comparison between the computed (using the described procedure) and the measured RC slab deflections is presented in the next section. This section presents a comparison between the predicted and measured RC slab deflections reported by Cook (2001). It includes results of modeling of two floor slabs (namely Slab 1 and Slab 2) exposed to heating conditions specified in ISO 834. The specimens were 4700 mm long, 150 mm high and 925 mm wide. The reinforcement cover is 25 mm. The slabs were cast of concrete mixes with siliceous aggregated and designed to have characteristic cube strength of 30 MPa. The reinforcing steel bars were of high yield ribbed bar having yield strength of 460 MPa. Slabs were reinforced with 10 bars of 8 mm diameter. As shown in Figure 26a, Slab 1 was subjected to high temperature without mechanical loading and Slab 2, shown in Figure 26b, was also subjected external loading (distributed load q = 1.5 kN/m2 ). The temperature distributions within cross-section of both slabs after 30, 60, 90 and 120 min fire exposure were presented in (Cook 2001). These temperature profiles were used for the respective time-deflection analysis of slabs. As temperature dependent material properties were not given in the reference, they were assumed according to Eurocode 2 (CEN 2004). The modeled time-deflection diagrams are presented in Figure 26 along with the experimental curves. The time-deflections diagrams are presented in terms of normalized deflections, where f is the deflection of slab after 120 min of fire exposure. It can be seen from Figure 26 that the shape of the experimental load-deflection diagrams was well captured in the present analysis. Agreement of the calculated and measured deflections is within reasonable limits. In both analyses, the deflections were underestimated, but the maximal error has not exceeded 35%. In this study, an attempt has been made to extend application of the Layer section model to stress and strain analysis of flexural RC members subjected to high temperature. A powerful calculation technique has been developed. Variation of material properties within the section due to different loading and temperature gradient was assessed in the analysis. Restrained thermal deformations as well as transient thermal creep were modeled by means of fictitious equivalent forces. The proposed algorithm is very effective in terms of computer resources, i.e. the calculation time decreases hundreds of times (from hours to few seconds) in comparison to standard non-linear finite element programs. Comparison of the experimental and modeling results has shown that the proposed model has satisfactorily captured the load-deflection behavior of the precast concrete slabs. 2.1.7.5 Aluminium alloys structures The prediction of the mechanical response of aluminium alloy structures exposed to fire is complicated for two principal reasons: 1) the difficulty of developing accurate structural analyses in post-elastic field, taking correctly into account the mechanical features of the basic material, such as the strain-hardening and the limited deformation capacity; 2) the inadequate knowledge of the material behaviour under high temperatures. As a consequence, first of all the specific mechanical 129

Figure 27. EC9 aluminium alloy’s mechanical properties as function of temperatures (f[N/mm2 ], et[%]) (Faggiano et al., 2005).

Figure 28.

Study cases (Faggiano et al., 2005).

properties and the whole stress-strain curve of the material as a function of temperature have to be accurately defined. Moreover, the methods of structural analysis in fire conditions should hold in due account the influence of the shape of the material constitutive law and thus of the kinematic strain hardening on the global behaviour of the structure. Therefore, for allowing practical analysis of complex structures in fire conditions through advanced methods, such accurate material models should be implemented in finite element programs. In this context, a wide examination of the results of experimental tests (ASM Specialty Book, 1993) carried out on different aluminium alloys exposed to high temperatures has been presented (Faggiano et al., 2004a), aiming at characterizing the behaviour under fire in relation to the series and treatments (work hardening state (H), hardening state due to heat treatment (T), annealed state (O)) of the aluminium alloys. The variation laws of the following characteristic parameters has been drawn: the elastic modulus (E), the elastic limit stress conventionally defined as 0.2% proof strength (f0,2), the ultimate strength (ft) and the ultimate deformation (et). Then, a mechanical model, which appropriately represents the peculiarity of such materials subjected to high temperatures, has been proposed, based on the well known Ramberg – Osgood law. The obtained simplified constitutive law has been introduced in a finite element program for the calculus under fire of structures (Franssen, 1998), with specific reference to the aluminium alloys selected for structural uses by the Eurocode 9 (CEN, prEN 1999-1-2, 2003). Finally, the results of the structural analyses in fire conditions obtained for a simple portal frame and carried out for all the EC9 aluminium alloys have been presented, clarifying the impact of the material modelling on the global response of the structure exposed to fire, evaluated in terms of time up to collapse for a conventional fire scenario (Faggiano et al., 2003; 2004a,b, 2005). As a first result of the study, it has been pointed out that simplified mechanical models, such as the elastic-perfectly plastic one, generally are not able to correctly characterize the material 130

Figure 29.

Global analysis for the various kinds of designed frames subjected to fire.

behaviour at the high temperatures, since they disregard the beneficial effect due to continuous material hardening, which is somewhat effective in balancing strength decay due to high temperatures. Therefore, in order to take specifically into account the effect of the strain hardening, the more comprehensive mechanical model for the aluminium alloys proposed is able to represent in an appropriate manner all the peculiarities of such materials exposed to high temperatures. The structural analysis in fire conditions of a study case related to a simple portal frame has pointed out the remarkable effect of material modelling of aluminium alloys, since the adoption of elasticperfectly plastic model results very conservative and not convenient for a material which exhibits a so rapid strength decay with high temperature. Finally, not treated alloys (O) give rise to the best behaviour under fire due to the beneficial effect of material strain hardening and to the fact that the strength degradation at high temperature is softer than for treated alloys. 2.1.7.6 Composite steel-concrete frames The advanced calculation models allow to evaluate the structural fire behaviour of single members, substructures and entire structures (EN 1994-1-2). The topic of this contribution is the application of advanced calculation models for structural fire analysis of composite steel and concrete frames in order to compare the results of member, substructure and global analyses in terms of fire safety assessment (Nigro et al., 2008, 2009). The influence of some aspects of structural response developing during the fire exposure, generally neglected in the member analysis, on the assessment of the structural fire safety is pointed out, such as: indirect fire actions, large displacements, geometric and mechanical non-linearities. Two composite steel-concrete frames, with four storeys and three spans, are designed for two different seismic zones according to the recent Italian Technical Code (2008). The beams are composite comprising steel beam with no concrete encasement and the columns are partially encased. Each frame has different over-strength of the columns with respect to the beam, due to the capacity design rules and damage limit state requirements of the seismic design. Moreover, in order to improve the fire resistance, composite steel beams with partial concrete encasement are also adopted for both frames. More details of the frame cross-sections are reported in Figure 29. Each frame is subjected to different fire scenarios with the ISO 834 standard time-temperature curve (EN 1991-1-2); for each fire scenario the structural fire behaviour of entire structures, single members and various possible substructures is analyzed. The structural analyses are carried out by means of the non-linear software SAFIR2007 developed at the University of Liege (Franssen 2008). The substructures are different for their limits and for the boundary conditions in order to highlight their influence on the assessment of the structural fire safety. Indeed, the choice of 131

Figure 30.

Substructure analysis results for fire scenario 2.

substructure, its limits and boundary conditions is not simple and it is depending on both the fire scenario and the structural geometry (Franssen 2005). Some criteria, making easy the choice of the substructures which need to be analyzed for assessing the structural fire safety, are applied in the following for the designed steel-concrete frames (Nigro et al., 2009). In Figure 29 the global analysis results (collapse time, failure section) for two analyzed fire scenario are summarized: a) fire on the overall first floor; b) fire limited to the central span of the first floor. The comparison between the frames shows that the collapse time of the frame designed for seismic zone 4 is quite similar to those of the other frame, designed for seismic zone 2. This is a consequence of the internal forces’entity produced by constrained thermal expansions: those forces (named generally indirect effects) have a higher value for the frame designed for seismic zone 2 (Fig. 29). Moreover, in the case of composite steel beam with partial concrete encasement the significant improvement of fire resistance time is remarked. In order to reduce the computational time, the substructure analysis can be used. However, the selection of a specific substructure affects the analysis’ results. In Figure 30 is reported a comparison between the global and substructure analysis results (collapse time) for fire scenario 2. Meaningful is the case of b2 and c2 substructures subjected to fire scenario 2. The translational restraints in horizontal direction for nodes I and N allow a better development of the catenary action (Usmani et al., 2001) along the heated beam. It produces an overestimation of the structure fire resistance time with respect to the global analysis results. The simplest substructure is the single member. Single member analysis allows to consider the structures like an assembly of single elements (beams and columns): therefore, those analysis type does not allow to take account of the effects of the structural redundancy. In Figure 31 it is reported a comparison between the global and simple member analysis results (collapse time and failure section) for two fire scenario. The results of single member analysis are conservative when the element that collapse in the global analysis is a beam, because the catenary effect is neglected in the single member model. Instead, the single member analysis is 132

Figure 31.

Single member analysis vs global analysis results.

not conservative when the element that collapses in the global analysis is a column, due to indirect actions produced by constrained thermal expansions. 2.1.8 FIRE RESISTANCE OF STRUCTURES AFTER EARTHQUAKE 2.1.8.1 State of the problem The behaviour in fire of structures which have been damaged by earthquakes represents an important investigation field since in many cases fires break out after a seismic event, giving rise to a real catastrophe. In fact negative effects of fires on structures and human lives may be comparable to those of the earthquake itself. Moreover, even in case no fire develops immediately after an earthquake, the possibility of delayed fires affecting the structure must be adequately taken into account, since the earthquake induced damage makes the structure more vulnerable to fire effects than the undamaged one. This is because the consequence of fire on a structural system is mainly a gradual decay of the mechanical properties as far as temperature grows. It is apparent that the more the structural behaviour is degraded after an earthquake the more time up to collapse due to fire is short. In view of the development of a comprehensive methodology of performance-based design of buildings, the fire resistance performance should be taken into account considering also the earthquake-induced damage for those buildings located in seismic areas. This consideration leads to the conclusion that the fire-safety codes should distinguish between structures located in seismic and non-seismic areas, by requiring more stringent fire resistance provisions for those buildings potentially subjected to seismic actions. 2.1.8.2 Preliminary studies In recent years, a number of studies on the behaviour of steel structures damaged by earthquakes and exposed to fires has been carried out (Della Corte et al., 2001, 2003a, b, 2005; Faggiano et al., 2005, Zaharia & Pintea, 2009). In particular some numerical analyses were devoted to investigate the effects of structural earthquake-induced damage on the fire resistance of MR steel frames. Modelling the behaviour of buildings subject to fires following earthquakes is a challenging but very difficult task for a structural engineer. In fact, not only knowledge about the mechanical response of the structure to the external action, but also dominance of several interdisciplinary issues, like modelling of seismic and fire actions is required. Grossly, the following general modelling aspects could be identified: a) modelling of the seismic action; b) modelling of the structural response during the earthquake; c) modelling of the fire action; d) modelling of the thermo-mechanical behaviour of the structure subject to fire. A key aspect of the study has been the interpretation of the earthquake-induced damage, which has been done by means of a simple modelling scheme. In particular, structural damage has been schematised as the combination of two damage types: a ‘geometrical damage’, which consists of the residual deformation of the structure, and a ‘mechanical damage’, which consists of the reduction of the main mechanical properties of the structural components (stiffness and strength degradation). Figure 32 efficiently synthesises this scheme, evidencing that the structure after the earthquake could be subjected to significant residual P-Delta effects, which, together with the reduced lateral strength of the frame, could induce an important reduction of the frame fire resistance. 133

Figure 32.

Residual P-Delta effects and local plastic deformation due to earthquake (Della Corte et al., 2005).

Figure 33. Fire resistance rating reductions of study MR steel frames subjected to earthquakes (Della Corte et al., 2005).

This schematisation allows for a rational evaluation of the mechanical state of the structure after the earthquake and of its mechanical behaviour under external actions succeeding the earthquake. In addition, it is a modelling very useful approach for parametrical analyses. Figures 33a through 33d illustrate the normalized fire resistance rating reduction obtained for the four examined cases of MR steel frames (Perimeter and Spatial frames; designed at Ultimate Limit States, ULS, and Serviceability Limit State, SLS), as a function of the normalized spectral acceleration Sa,e , for a number of acceleration records. The fire resistance rating reduction usually becomes non-negligible for very rare earthquakes, i.e. earthquakes having a mean return period larger than 475 years. 2.1.8.3 Refined approach More comprehensive numerical simulations overcoming some conceptual and numerical limitations of the simplified models have been developed, preliminary to simple portal frames, in order to have a more accurate representation of seismic damage within the structure (Faggiano et al., 2007a). At this aim, the finite element multi purpose computer program ABAQUS v.6.5 (2004) has been used, which allows to perform coupled thermal-displacement analyses, so giving the possibility to reproduce, in a step-by step process, the actual phases of the modelled phenomenon, from the application of the vertical loads and the earthquake induced damage up to the exposure of the structure to fire. Therefore the analysis procedure is articulated in three different phases: 1. Seismic analysis of structures; 2. Identification of the performance levels, according to the mentioned SEAOC indications; 3. Analysis under fire of the structures already damaged by earthquake, starting from each previously defined performance levels. 134

The seismic damaged states of the structures, characterizing the performance levels, are considered as initial configurations for the fire analysis, aiming at the evaluation of the effect of the seismic induced damage on the fire resistance and the collapse mode of the study structures.

2.1.9 FURTHER DEVELOPMENTS Most of the current research work on structures subjected to elevated temperatures is dedicated to steel structures. The experimental and numerical studies show importance and complexity of beam to column connections in structural analysis Galambos (2000). The flexibility and strength of a connection pay important role in overall behavior of many steel structures. At the same time the great variety of joint types require complex and unique analyses. The connections are also critical for the resistance of steel structures subjected to elevated temperatures Franssen & Zaharia (2006). Precise numerical analysis is complex as it should take into account many parameters such as contact between bolts, column flange, and end plate, stress concentration around bolts, prestressing forces. Material degradation and elongation caused by elevated temperature additionally complicates the study. For concrete structures there are important and complex at the same time, thermo-hydromechanical phenomena resulting in nonlinear interaction among additional effects like transient creep strain, load induced thermal strain, shrinkage, pore pressures and spalling. The prediction of behaviour of concrete structures and structural elements imposes the main challenge for future research. Another issue is the question about predictive capability for analysis of structures under fire, especially of the nonlinear global FE analysis intended to replicate real fires. The FEA model verification and validation (V&V) is more often recognized as a procedure warranting modeling accuracy (Oberkampf et al., 2004). The calculation verification is intended to estimate the numerical errors due to discretization approximations. Validation, through mainly comparison with experiments, evaluates the accuracy with which the mathematical model depicts the actual physical event. There is a need for experimental benchmark problems which could be used for the FE model validation. The test conditions in terms of loading, thermal and mechanical boundary conditions, and measurements should be clearly specified and easy to follow. REFERENCES Al-Jabria, K.S. Seibib, A. & Karrechc, A. 2006. Modelling of unstiffened flush end-plate bolted connections in fire. Journal of Constructional Steel Research. 62: 151–159. Anderberg, Y. & Thelandersson, S. 1976. Stress and Deformation Characteristic of Concrete at High Temperatures, Part 2: Experimental Investigation and Material Behaviour Model. Bulletin 54. Lund: Lund University of Technology. 84 p. ASM Specialty Handbook, 1993. Aluminium and aluminium alloys. Edited by J.R. Davis &Associates. Bacinskas, D. Kaklauskas, G. & Gribniak, V. 2008. Layered section analysis of RC slabs subjected to fire, in Proc. of the International fib Conference Fire Design of Concrete Structures. Coimbra: University of Coimbra, 311–318. Bratina, S. Saje, M. & Planinc, I. 2007. The effects of different strain contributions on the response of RC beams in fire, Engineering Structures 29(3): 418–430. Capua Di, D. & Mari, A.R. 2007. Nonlinear analysis of reinforced concrete cross-sections exposed to fire, Fire Safety Journal 42(2): 139–149. CEN, European Committee for Standardization, 2002. EN 1990 Eurocode 0: Basis of Structural Design. Brussels, Belgium. CEN, European Committee for Standardization, 2002. 1991-1-2 Eurocode 1: Actions on structures – Part 1-2; General Actions – Actions on structures exposed to fire. Brussels, Belgium. CEN, European Committee for Standardization, 2004a. Eurocode 2: Design of Concrete Structures – Part 1: General Rules and Rules for Buildings, EN 1992-1-1:2004. Brussels, Belgium. CEN, European Committee for Standardization, 2004b. EN 1992-1-2, Eurocode 3: Design of concrete structures – Part 1-2: Structural fire design. Brussels, Belgium. CEN, European Committee for Standardization, 2005a. EN 1993-1-2, Eurocode 3: Design of steel structures – Part 1-2: Structural fire design. Brussels, Belgium.

135

CEN, European Committee for Standardisation, 2005b. EN 1993-1-1, Eurocode 3, Design of Steel Structures – part 1-1. General rules and rules for buildings. Brussels, Belgium. CEN, European Committee for Standardization, 2005c. EN 1994-1-2, Eurocode 3: Design of composite steel and concrete structures – Part 1-2: Structural fire design. Brussels, Belgium. CEN, European Committee for Standardization, 2005d. prEN 1993-1-4, Eurocode 3: Design of steel structures – Part 1-4: Supplementary Rules for Stainless Steels. Brussels, Belgium. CEN (European Communities for Standardisation), prEN 1999-1-2, 2003. Eurocode 9: Design of aluminium structures – Part 1-2: General rules-Structural fire design. Brussels, Belgium. Cervenka, J. Surovec, J. Fellinger, J. Feron, C. Wageneder, J. Kaklauskas G. & Corsi, F. 2005. FE Simulations of a Suspending Ceiling. Comparison of Fire Test Analyses, UPTUN Report 441, Brussels: UPTUN. 30 p. Cooke, G.M.E. 2001. Behaviour of precast concrete floor slabs exposed to standardised fires, Fire Safety Journal 36(5): 459–475. Della Corte G., Landolfo R., 2001. Post-earthquake fire resistance of steel structures, Safety and reliability, towards a safer World, Proceedings of the European Conference on Safety and Reliability – ESREL 2001, Torino, Italy. Della Corte G., Landolfo R., Mammana O., 2003a. Fire resistance of MR frames damaged by earthquakes, Proceedings of the Fourth International Conference on the Behaviour of Steel Structures in Seismic Areas (STESSA 2003), Naples, Italy. Della Corte G., Landolfo R., Mazzolani F.M., 2003b. Post-earthquake fire resistance of moment-resisting frames, Fire Safety Journal, Vol. 38, pp. 593–612. Della Corte G., Faggiano B., Mazzolani F.M., 2005. On the structural effects of the fire following earthquake, Proceedings of the Final Conference COST C12 “Improving buildings’ structural quality by new technologies”, Innsbruck, Austria. ECCS. European Convention for Constructional Steelwork. 1983. European Recommendations for the Fire Safety of Steel Structures. Elsevier. ECCS. European Convention for Constructional Steelwork. 2006. Rules for Member Stability in EN 1993-1-1: Background documentation and Design Guidelines. ECCS publication no. 119, Brussels. Faggiano B., De Matteis G., Landolfo R., Mazzolani F.M., 2003. On the behaviour of alluminium alloy structures exposed to fire. Proceedings of the XIX National Congress C.T.A. “III week of steel Constructions”, 28–30 September, Genova, Italia, 393–406 pp. Faggiano B., De Matteis G., Landolfo R., Mazzolani F.M., 2004a. The influence of material modelling on the fire resistance of aluminium alloy structures. Proceedings of the 9th International Conference on Aluminum Structural Design (INALCO 2004), Cleveland, Ohio (USA), 2–4 June. Faggiano B., De Matteis G., Landolfo R., Mazzolani F.M., 2004b. Effects of high temperatures on the resistance of aluminium alloy structures. Proceedings of the 7th International Conference on Modern Building Materials, Structures and Techniques, Vilnius, Lithuania, 19–21 May. Faggiano B., De Matteis G., Landolfo R., Mazzolani F.M. 2005a. “On the fire resistance of aluminium alloy structures”. Proceedings of the Final Conference of the COST C12 “Improvement of buildings’ structural quality by new technologies”, Innsbruck, Austria, 20-22 January, A.A. Balkema Publishers, Taylor & Francis Group plc, London, UK, ISBN: 04-1536-609-7, pp. 267–275. Faggiano B., Della Corte G., Mazzolani F.M., Landolfo R., 2005b. Post-earthquake fire resistance of moment resisting steel frames, Proceedings of the Eurosteel Conference on Steel and Composite Structures, Maastricht, The Netherlands. Faggiano, B. Esposto, M. Mazzolani, F.M. 2007a. “Fully coupled temperature – displacement analyses of steel portal frames under fire”. Atti del XXI Congresso CTA, Costruire con l’acciaio, Catania, Italia, a cura di A. Ghersi, Flaccovio Editore, ISBN 978-88-7758-787-9, pp. 561–568. Faggiano B., Esposto M., Mazzolani F.M., Landolfo R., 2007b. Fire analysis on steel portal frames damaged after earthquake according to performance based design, COST Action C26 “Urban habitat constructions under catastrophic events”, Proceedings of Workshop in Prague, pp. 35–40. Franssen, J-M. 2005. SAFIR, A Thermal/Structural Program Modelling Structures under Fire, Engineering Journal, A.I.S.C., Vol 42, No. 3, 143–158. Franssen, J-M. & Zaharia, R. 2006. Design of Steel Structures Subjected to Fire: Background and Design Guide to Eurocode 3. Fire Safety Journal, Volume 41, Issue 8: 628–629. Galambos, T.V. 2000. Recent research and design develop-ments in steel and composite steel–concrete structures in USA. Journal of Constructional Steel Research, Volume 55, Issues 1–3: 289–303. Geda, E. 2010. Layer Model for Reinforced Concrete Members under Fire. PhD dissertation, Vilnius Gediminas Technical University, Vilnius, Lithuania. 118 p. (in Lithuanian). Gribniak, V. Bacinskas, D. & Kaklauskas, G. 2006. Numerical simulation strategy of bearing RC tunnel members in fire, The Baltic Journal of Road and Bridge Engineering 1(1): 5–9.

136

Hajpál, M. 2000. Investigation of burnt sandstone, Proceedings 1st international conference on Fire Protection of Cultural Heritage 2000 Thessaloniki, pp 349–361. Hajpál, M. & Török, Á. 2004. Physical and mineralogical changes in sandstones due to fire and heat. Environmental Geology, 46, 3, 306–312. Hajpál, M. 2008. Heat effect by natural stones used by historical monuments, Proceedings 11th International Congress on Deterioration and Conservation of Stone (STONE 2008). Huang, Z. Burgess, I.W. & Plank, J.R. 1999. Nonlinear analysis of reinforced concrete slabs subjected to fire, ACI Structural Journal 96(1): 127–135. Hallquist J.O. 2006. LS-DYNA Keyword Manual. Livermore: Livermore Software Technology Corporation. Italian Technical Code for the Constructions 2008 (in Italian). “Norme Tecniche per le Costruzioni”, Supplemento Ordinario della Gazzetta Ufficiale della Repubblica Italiana del 4 febbraio 2008, n. 29. Kaklauskas, G. 2004. Flexural layered deformational model of reinforced concrete members, Magazine of Concrete Research 56(10): 575–584. Kaklauskas, G. Bacinskas, D. Gribniak, V. & Geda, E. 2007. Mechanical simulation of reinforced concrete slabs subjected to fire, Technological and Economic Development of Economy 13(4): 295–302 (in Lithuanian). Khennane, A.; Baker, G. 1993. Unaxial model for concrete under variable temperature and stress, ASCE Journal Engineering Mechanics 119(8): 1507–1525. Khoury, G.A. 2000. Effect of Fire on Concrete and Concrete Structures. Progress in Structural Engineering and Materials 2(4): 429–447. Kodur, V.K.R. & Dwaikat, M.B. 2008. A numerical model for predicting the fire resistance of reinforced concrete beams, Cement and Concrete Composites 30(5): 431–443. Kosiorek, M. 2002. Fire safety in regulations and designing; Fire Protection, 1/2002 (1). König, J. 2005. Structural fire design according to Eurocode 5-design rules and their background. Fire and materials 29: 147–163 Wiley InterScience. Nielsen, C.V. Pearce C. J. & Bicanic N. 2004. Improved Phenomenological Modelling of Transient Thermal Strains for Concrete at High Temperatures. Computers and Concrete 1(2): 189–209. Nigro E., Ferraro A., Cefarelli G. 2008. Structural fire analysis of composite steel and concrete frames (in Italian), Costruzioni Metalliche, n. 6, December 2008, ACAI, pp. 51–64. Nigro E., Pustorino S., Cefarelli G., Princi P. 2009. Progettazione di strutture in acciaio e composte acciaiocalcestruzzo in caso di incendio, Ed. Hoepli, Milano. Oberkampf, W.L. Trucano, T.G. & Hirsch C. 2004. Verification, validation, and predictive capability in computational engineering and physics, Appl. Mech. Rev. 57 (5), 345–384. Shapiro, A. 2005. Heat Transfer in LS-DYNA, 5th European LS-DYNA Users Conference New Applications and Developments, Birmingham, UK. Schneider, U. & Horvath, J. 2003. Behaviour of Ordinary Concrete at High Temperatures. Research Report, Vol. 9. Institute for Building Materials, Buildings Physics and Fire Protection, Vienna University of Technology. SCI, Steel Construction Institute. 1996. Structural Fire Design to EC3 & EC4, and comparison wit5h BS 5950. Technical Report, SCI publication no. 159. Terro, M.J. 1998. Numerical modeling of the behavior of concrete structures in fire, ACI Structural Journal 95(2): 183–193. Török, Á. & Hajpál, M. 2005. Effect of Temperature Changes on the Mineralogy and Physical properties of Sandstones. A Laboratory Study. International Journal for Restoration of Buildings and Monuments, 11, 4, Freiburg, 211–217. Tybura, A. & Kwasniewski, L. (2008) Computer Simulation of the Steel Connection at Elevated Temperatures, Proceedings of XIII Slovak – Polish – Russian seminar Theoretical Foundation of Civil Engineering, MGSU, Moscow, Russia, Warszawa-Wroclaw, Poland, 02-06.06.2008, pp. 231–236. Usmani A.S., Rotter J.M., Lamont S., Sanad A.M., Gollie M. 2001. Fundamental Principles of Structural Behaviour Under Thermal Effects. Fire Safety, 2001(36): 721–724. Vila Real, P.M.M. & Franssen, J.-M. 2010. Elefir-EN, http://elefiren.web.ua.pt. Wang, Y.C. 1997. The effects of frame continuity on the behaviour of steel c columns under FIRE conditions and FIRE resistant design proposals, J. Construct. Steel Res., Vol. 41, No. 1, pp. 93–111. Youssef, M.A. & Moftah, M. 2007. General Stress–Strain Relationship for Concrete at Elevated Temperatures. Engineering Structures 29(3): 2618–2634. Xiao, J. & Konig, G. 2004. Study on concrete at high temperature in China an overview, Fire Safety Journal 39(1): 89–103. Zaharia, R. & Pintea, D., Fire after earthquake analysis of steel moment resisting frames, International Journal of Steel Structures, December 2009, Vol 9, No 4, 275–284.

137

Urban Habitat Constructions under Catastrophic Events (Final Report) – Mazzolani (Ed). © 2010 Taylor & Francis Group, London, ISBN 978-0-415-60686-8

2.2 Evaluation of structural response under exceptional seismic actions M. Fischinger University of Ljubljana, Ljubljana, Slovenia

G. Della Corte University of Naples “Federico II”, Naples, Italy

2.2.1 STATE OF THE ART 2.2.1.1 Introduction The main objective of the COST Action C26 is to increase the knowledge on the behaviour of constructions located in urban habitats and subjected to exceptional and catastrophic events. One of the key tasks to achieve this goal is to define suitable tools for predicting the ultimate response of such constructions under extreme conditions, occurring when both loading and structural resistance are combined in such a way to reduce the safety level below acceptable values (Mazzolani, 2008). Building and structural elements subjected to exceptional seismic loads are expected to enter into highly inelastic range exhibiting strong stiffness and in particular strength degradation as well as failure modes, which up-to-date codes try to prevent. While it has been repeatedly demonstrated that traditional inelastic models provide reliable and efficient tools to analyze inelastic seismic response under ”normal“ conditions, the question arises, if such models are able to provide adequate information in the analysis of the exceptional events when structures are entering into post-critical region and into normally prohibited failure modes. This question is becoming of the utmost importance at the time of the introduction of the new methodologies of the performance based design and seismic risk evaluation (ATC 63, 2009; Pascu, 2008, Deierlein, 2004, Fardis, 2004; see also WG2 report “Performance based evaluation and risk analysis”). As defined by PEER researchers (i.e. Deierlein, 2004) “performance based earthquake engineering seeks to improve seismic risk decision-making through assessment and design methods that have a strong scientific basis that express options in terms that enable stakeholders to make informed decisions”. The basic concept of this procedure, illustrated in Figure 1, shows that to apply the procedure one needs in-depth knowledge about the behaviour of structures and their components across a large span of expected performance behaviour well into the post-critical zone. Performing related inelastic analyses and in particular extensive risk analysis one needs reliable but efficient and robust analytical tools. Macro elements (predominantly addressed in this report), defined here as models consisting of an assemblage of inelastic springs controlled by robust hysteretic rules, might be efficient tool to perform this task. However, if such models are to be used to provide adequate information in the analysis of the exceptional events when structures are entering into post-critical region and normally prohibited failure modes, a lot of research is still needed. First of all the reliable models of any kind and in particular the macro models to analyse highly inelastic response need a solid experimental background. Worldwide, the research is strongly supported by consortia of large scale experimental facilities working in the frame of the extensive research efforts like SERIES/ECOLEADER in Europe, Network of Earthquake Engineering Simulation – NEES in the USA and E-defence in Japan. These efforts have recently contributed to “strong scientific basis” needed for the implementation of the presented methodology and related analytical tools. Within, or in parallel with these systematic experimental campaigns experiments have been performed by most COST C26 partners, as overviewed and cited in the following sections. 139

Figure 1. The PEER concept of the Performance-Based methodology.

Problems of seismic retrofit are of special importance within the Action. By definition, the structures requiring seismic rehabilitation lack adequate seismic capacity as well as proper design in the past. Several COST C26 participants have addressed this topic, in particular within the WG2 sub-task “Innovative protection technologies and case studies”. New experimentally supported analytical models and tools have been proposed. Of particular importance is the research related to the (masonry) historical monuments. Several COST C26 participants have also shown the importance and benefit of the adequate seismic design and modelling in enhancing the structure’s ability to withstand impact and explosion loads related to the natural and man-made (terrorist) actions. More refined (FEM based) models have been typically used in the related analyses. It is the ambition of this report to provide a short state-of-the-art report on the analysis of behaviour of constructions under catastrophic seismic events supported by the selected relevant contributions of the COST C26 members. It has been decided to organize the report into two basic Sections: “Experimental results” and “Analytical studies”. Each of these two main Sections deal with all the structural materials for which tests to collapse are available (basically, the three main structural materials: concrete, steel and masonry). Within each topic the overview of the selected models and procedures is given and the supporting studies are presented. The transversal relation with the work performed within the frame of the other WG2 sub-topics as well as with the work of the other working groups and related international actions is also presented. 2.2.1.2 Experimental tests Generalities While a lot of experimental collapse tests at the structure component level have been carried out over the past decades, a few tests have been carried out on complete structures subjected to strong earthquake shaking through global collapse. However, recent emphasis on performance-based seismic engineering has increased the desire to have experimental validation of analytical methodologies which are being developed to assess structural response through collapse. In response to this, tests have been carried out in the last ten years, either on small structures or on small-scale models of real structures or even on full-scale buildings. Such tests are referenced to and summarized in the following. Reinforced concrete structures The goal of the traditional design was to provide RC members with sufficient inelastic deformation capacity without considerable strength deterioration. Typical pioneers of this (capacity design) concept have belonged to the so-called New Zealand school, i.e. Park and Paulay (1975) and later Paulay and Priestley (1992). In support of such approach the tests did not need to enter far into post-critical range. Moreover, the experiments were predominantly limited to traditional structural systems and well-reinforced standard members like columns, beams, and walls. All these have been reflected in the major experimental data bases, i.e. (PEER, 2007) and (Panagiotakos and Fardis, 2001). Only recently the interest has shifted towards near-collapse behaviour. This information is urgently needed for several important reasons: (a) to support development of the PBD procedures, 140

Figure 2.

Different hysteretic curves for gravity load collapse curves (Elwood and Moehle, 2002).

Figure 3.

Full-scale collapse test (Matsumori et al., 2007).

Figure 4.

Collapse of a short column.

(b) to provide the basis for the seismic risk studies, (c) to gain adequate information of innovative structural systems, (d) to provide data for seismic rehabilitation of the systems designed prior to the introduction of the modern seismic design principles and to (e) to foresee the behaviour of structural systems at exceptional events. The majority of the experiments have been performed by the major experimental centres within the NEES and EU consortia, E-defense in Japan, as well as in Taiwan and New Zealand. Short overview of the most important results is given below. Early works in Japan and the USA are well documented in the 4 proceedings of the U.S.-Japan Workshops on Performance-Based Earthquake Engineering Methodology for Reinforced Concrete Building Structures (Moehle and Kabeyasawa, 2000–2004). Of particular interests are works on gravity load collapse of the older design building frames during earthquakes. A number of cyclic tests were done by Moehle et al. (2000). Typical behaviour is shown in Fig. 2 for (a) moderate flexural ductility, followed by loss of lateral resistance due to the shear failure, (b) low flexural ductility interrupted by loss of lateral resistance due to the shear failure, but sustained vertical load capacity to relatively large displacements and (c) low flexural ductility interrupted by sudden shear and gravity load failure. It is obvious that such behaviour is difficult to monitor by standard numerical models (see section 1.3.2). Research with similar objectives, but performed by shaketable test was reported by Elwood and Moehle (2002). Tests were done on specimens comprised with three columns interconnected by a beam. The central column was shear critical and the external 2 columns were ductile. So the process of the dynamic shear and axial load failures in RC columns was studied in cases when alternative load path was provided for load redistribution. The tests showed that system failure rather than component failure should be considered. The more recent collaborative Japan-U.S. research effort is well documented in the 2 proceedings of the NEES/E-defense Workshops on Collapse Simulation of Reinforced Concrete Building Structures (Moehle and Kabeyasawa, 2005–2007). Of particular interest is the work of the Japanese researchers on the shake table test on a full-scale six-story reinforced concrete building at E-defense (Matsumori et al. 2007; Fig. 3). Non-ductile mechanism of the full-scale structure was investigated all the way to the collapse (Fig. 4). Although the static calculation had foreseen weak beam – strong 141

column response, the structure failed in shear at the short columns and the structural wall in the first story at the 100% JMA Kobe earthquake input. Following the Chi-chi earthquake experimental activity strongly intensified in Taiwan. For example, shaking table tests using near-fault input motions were used to study post-peak behaviour involved in global collapse mechanism (Wu et al., 2006). The core research in Europe was done by the consortium of the large-scale experimental facilities. Two specific topics (a) lightly reinforced structural walls and (b) precast industrial buildings are mentioned here. Within the CAMUS studies (i.e. Combescure and Chaudat 2002) performed at the shaking table at Saclay a series of 5-story RC cantilever walls in the scale 1:3 were investigated. In CAMUS1 very lightly reinforced wall according to the French design practice was studied. In CAMUS 3 the wall was designed according to the EC8 provisions. Parallel benchmark studies were organized. As an outgrowth of the CAMUS project Slovenian researchers tested a 1:3 model of a 5-story wall consisted of two coupled T-shaped piers (see Section 2.2.1 for further reference). The piers were reinforced by light (minimum reinforcement) according to the Slovenian building practice. Some free edges of the wall piers were confined and some were not. Very simple and weak diagonal reinforcement consisting of 2 crossed bars was used in the coupling beam. Tests were performed at LNEC in Lisbon. Precast buildings have been frequently used in many European countries. Predominant type of the precast (industrial) building consists of an assemblage of columns tied together with floor diaphragms. However until recently neither such system as the whole, nor the connections were fully tested. In the case of a strong earthquake, loss of lives, the damage to buildings and equipment, and in particular the indirect damage due to the interruption of the production, could be catastrophic. Research within the PRECAST project (Toniolo 2007, Fischinger et al. 2008a) has shown that very large rotations at the top of the slender columns in precast industrial buildings are expected in the case of the strong earthquakes. Question arose if typical dowel connections could survive such large rotations. This problem was addressed within the SAFECAST project (Fischinger et al. 2010). As this research is directly linked to the COST C26, further explanations are given in Section 2.2.1. Steel structures Among the first tests specifically devoted to assess collapse conditions of steel structures those carried out by Vian and Bruneau (2001, 2003) are worthy to be mentioned. They made tests through collapse of small specimens on small shaking tables with a selected acceleration input. The specimens were made up by placing selected weights on a slab supported on four steel columns. 15 specimens were examined, with square tubular steel columns. Mechanical parameters of specimens were selected in order to be comparable to values typical of real structures (hence, the adopted acceleration record was not time-scaled). Each specimen was tested several times, by increasing the table accelerations, up to specimen collapse. The Authors focused the attention on the sensitivity of the collapse “capacity” to the value of the stability factor θ, which is the ratio between the total story weight and the product of the initial (elastic) story stiffness and story height. Limitations on values of θ are usually used by Structural Codes (e.g. Eurocode 8) to limit the reduction of the firstorder story shear strength due to P-Delta effects. Basically, the Authors’ main conclusion is that the stability coefficient θ has a significant effect on the response of the structure near to collapse. This conclusion is based on plots relating one of three parameters (spectral acceleration, ductility or story drift ratio) measured in the next to last test versus the stability coefficient. Based on these plots the Authors interpolate some functions, showing an inverse relationship between the collapse capacity (measured by means of one of the three mentioned parameters) and the stability coefficient. Vian and Bruneau also implemented some numerical models to reproduce the response observed experimentally, but they stated that the results were to be considered preliminary and refinements would have been necessary to better capture the observed response. Shaking table tests on small specimens very similar to those used by Vian and Bruneau were carried out by Kanvinde (2003). However, Kanvinde subjected multiple nominally identical specimens to the same table acceleration input. The intensity of the table input was defined by means of the elastic spectral acceleration corresponding to the first period of vibration of the structure and a viscous damping ratio of 1%. The level of shaking (i.e. the value of the spectral acceleration) was selected to be close to the one predicted to produce collapse of the specimen. These results are very 142

Figure 5. 2004).

2D frame model (a); Klevis and round bars to emulate plastic hinges (b), (Rodgers and Mahin,

interesting, since they show how a given ground acceleration input applied to nominally identical specimens can produce significantly different displacements when the structure is near to collapse. This is because the relationship between spectral acceleration and drift can be quite flat near to collapse, i.e. under a given spectral acceleration intensity many different drift responses can occur. In other words, the displacement response near to collapse is very sensitive to small perturbations, what is typically of unstable systems. Kanvinde also implemented a numerical model of the tested structures, illustrating how available models may be able to capture the spectral acceleration and associated drift response to collapse. In addition, Kanvinde showed that for non-degrading structural systems the static stability limit state, obtained by means of a pushover analysis, may give a reasonable estimate of the drift capacity under earthquakes. Among the first significant tests on full-scale structures, those carried out by Nakashima et al. (2006) can be mentioned. The Authors built up in the laboratory a full-scale 3-story momentresisting steel frame and tested it to near collapse. The main objective of those tests was to assess the range of applicability of non-deteriorating component models. The Authors concluded that such models are able to reproduce reality up to interstory drift angles of about 1/25. Beyond this value the consideration of degradation of stiffness and strength at plastic hinge locations becomes essential. A more comprehensive study on the effects of different hysteresis behaviours on global frame response up to large permanent drifts and collapse was carried out by Rodgers and Mahin (2004, 2008), by means of shaking table tests on small-scale models. Special attention was devoted to the effects of brittle or ductile fractures. The models were two-story moment resisting steel frames, with plastic deformations confined into special mechanical connections made with a klevis and round bars (Fig. 5). One of the advantages of using the special mechanical connections was to mimic different hysteresis behaviors with good predictability, thus permitting an understanding of the effect of generic hysteresis responses independent of the connection detail. One of the most basic and clear conclusion of the research is that strength degradation could exacerbate the effect of geometric non linearity and then have an important effect on collapse. Among the different types of degradation, that associated with a post-critical negative stiffness in the restoring force-deformation relationship had the most important consequences. Results of shaking table tests on a 3D two-story one-bay steel frame model are reported by Shimada et al. (2008). Square hollow sections were used and the model was designed to collapse in a column-sway mode under increasing levels of acceleration input. Test results clearly shows the importance of deterioration and P-Delta effects and the possibility to define collapse as the condition when the system total restoring force becomes zero. Such a result is an extension of the conclusion obtained by Kanvinde (2003), including the deterioration of strength at component level. Yamada et al. (2008) and Suita et al. (2008) presented results of full-scale shaking table tests of a 4-story steel building subjected to a given acceleration record scaled to four levels of intensity, up to global collapse. The shaking was 3D (i.e. all the three components of acceleration were used). The building was designed according to the Japanese code, with square hollow steel shapes for columns 143

Figure 6. 3D full-scale building model (a); Sidesway first-story collapse (b), (Yamada et al., 2008, Suita et al., 2008).

Figure 7.

Full-scale 2D frame model (a); Sidesway collapse (b), (Lignos et al., 2008).

and wide-flange I shapes for beams. Composite action with the RC slab was considered. The building was completed with all the non-structural components (Fig. 6a). The building exhibited a column-sway collapse mode at the first story (Fig. 6b). At collapse, interstory drift angles at the first story started to increase almost proportionally in the X andY directions, but eventually the building swayed mostly in the Y direction. The final values of interstory drift angles were 0.19 rad in the Y direction and 0.08 rad in the X direction, after which the building settled on the safe guard frame. The earthquake shaking intensity bringing the building to collapse was 2.5 times larger than the design level 2 earthquake intensity, which is the one used in Japan to assign strength to members. The evolution of damage from small intensity shaking through collapse showed that plastic deformations at moderate earthquake intensities may appear much more uniformly distributed than at collapse. The progressive change from a beam-sway to a column-sway mechanism is attributed by the Authors to degradation of column plastic hinges. Lignos et al. (2008) presented results of shaking table collapse tests of two 1:8 scale models of a two-bay 4-story moment-resisting steel frame with reduced beams sections (RBS). The frames were designed according to codified US design rules. Each model was subjected to 5 levels of shaking, with increasing spectral accelerations. The first four levels of shaking did correspond to the usually considered four structural performance levels (service, design, maximum considered and collapse). However, the frame did not collapse under the fourth level of shaking (what was predicted to occur) and an additional level of shaking was selected to bring the structure to global sidesway collapse (Fig. 7a). Extensive measurements of engineering demand parameters (story inertia forces, story shear forces, accelerations, velocities, displacements, plastic hinge rotations and moments) were made throughout the tests. The tests were complemented with component level testing of plastic hinges. The latter were artificially reproduced by means of two steel flange plates with controllable degrading response (Fig. 7b), in a similar fashion to the klevis-based connections used by Rodgers and Mahin (2004, 2008). The experimental results shows that the frame was able to sustain accelerations up to 2.2 times the design level. The Authors also carried out numerical 144

analysis and showed that a good prediction of frame response is possible with relatively simple numerical models, under the condition that degradation phenomena in plastic hinges are accurately reproduced. The testing procedure used is similar to the frequently adopted incremental dynamic analysis (IDA, Bazzurro and Cornell 1994a, 1994b, Shome and Cornell 1999, Luco and Cornell 2000, Vamvatsikos and Cornell 2005), but the difference is that the frame is damaged at the starting of each new step in the procedure. Understanding of damage accumulation effects was however one of the objectives of the study. The Authors carried out also numerical studies which are described and commented on in the next Section. Looking at the above experimental test results, the paramount role of strength degradation in determining global collapse appears very clearly. Global sidesway collapse may occur when global story P-Delta effects offset the first-order story shear strength. Less obvious is the effect of stiffness degradation, which is however also to be considered by the analytical models in order to obtain realistic predictions of deformation demand.

2.2.1.3 Analytical studies Generalities Though the mentioned experimental results are very useful to understand the basics of dynamic instability during earthquakes, the obvious limitation is that results are limited to a few case structures and ground acceleration input. Often, they are carried out on very simple specimens which do not reproduce the stiffness and strength degradation of real-world structures. Reinforced concrete structures All experimental studies reviewed in Section 1.2.2 were accompanied by numerical studies in support of the development of the numerical models suitable to describe the response well into the post-critical range, which is typical for catastrophic events. Most of these models are empirically calibrated macro models. Based on the works on gravity load collapse of the older design building frames during earthquakes Elwood and Moehle (2002) developed shear and axial load failure model for reinforced concrete frames. The model, which is based on shear friction was incorporated into OpenSees. For the purpose of reproducing the collapse process observed from the full-scale shake table tests on the 6-story RC wall frame structure (see Section 1.2.2) a non-linear response history was studied (Kim et al., 2007). Shear critical structural wall response was simulated by the model developed by Chen (2000). Shear panel consisted of one isoparametric element based on the smeared-rotating crack approach. Two side columns were modelled by one-component model. Many researchers have used multiple-vertical-spring-element model (MVLEM) to describe inelastic response of RC structural walls. The advantage of this model is the ability to describe the shift of the neutral axis and the related elongation of the centroidal line. As such the model can well describe the 3D interaction between structural elements well into inelastic region. A 3D version of the element (Fig. 8) was proposed by Kante (2005) and incorporated into OpenSees. Using this element (Fischinger et al., 2008b) the researchers of ULJ got the NEES prize for the best prediction of the seismic response of the full-scale 7-story building slice with rectangular RC structural wall (Figure 9) tested on the shaking table at UC San Diego in the frame of the NEES project (Panagiotou et al., 2006). Inelastic response of RC columns is frequently described by lumped plasticity models based on the empirically determined length of the plastic hinge and simplified hysteretic rules (i.e. Takeda). Typically, there are two major limitations of such models (a) they are not able to describe realistically the deterioration processes in RC elements in the post critical range, (b) they are calibrated for well known structural systems. As such they are of limited use for the analysis of the structural behaviour during catastrophic events. Several alternative models were studied in (Fischinger et al., 2008a). The peak-oriented hysteretic model that accounts for history-dependent strength and stiffness deterioration, developed by Ibarra et al. (2005), proved to be efficient. In this model the behaviour is first described by a monotonic backbone curve (Fig. 10). Pre-capping and post-capping cyclic strength deterioration, based on the energy dissipation criterion, is then considered (Fig. 11). 145

Figure 8. element.

Multiple-verical-line-model

Figure 9. Full-scale 7-story RC wall tested at UC San Diego (Panagiotou et al., 2006).

Figure 10. model.

Monotonic behaviour of the Ibarra Figure 11.

Cyclic behaviour of the Ibarra model.

The rate of deterioration is based on the parameter:

where Ei is the hysteretic energy dissipated in excursion i; Ej is the hysteretic energy dissipated in all previous excursions; Et = λMy θy is the reference hysteretic energy capacity; λ is the normalized energy dissipation capacity (larger values of λ indicate a larger dissipation capacity and slower cyclic strength deterioration); c is a cyclic deterioration calibration term (a c value of 1.0, used in this study, causes a constant rate of deterioration). Haselton (Haselton, 2006) has calibrated the Ibarra hysteretic model to a large number of experimental tests of RC columns included in the PEER database (2007). Based on these calibrations, he proposed empirical equations which could be used to predict the modelling parameters which could then be directly applied in the Ibarra hysteretic model. The application of the model is demonstrated in Section 2.3.1. Steel structures The last ten-years experimental tests through collapse of steel moment-resisting frames were paralleled by analytical studies, which were based on some refined hysteresis models in order to capture degradation of plastic hinges. First analytical studies were carried out on a deterministic basis, while eventually considering variability of structural response with the ground motion acceleration history on a statistical basis. The main objective of those first studies was to understand the physical phenomena leading to collapse. More recently, the development of performance-based earthquake 146

Figure 12.

Effect of degradation on drift (a) and plastic hinge rotation (b) demand (Della Corte et al., 2002).

engineering led to a more comprehensive and rigorous treatment of the uncertainties involved in the process of collapse safety assessment. Among the first studies, the work presented by Mehanny and Deierlein (2000) deals with a special case of a composite steel beams and concrete columns structural system. However, many of the concepts discussed and the procedures proposed have general importance. The Authors clearly considered the significance of damage accumulation due to the repetition of inelastic deformations on the collapse assessment. In fact, Authors propose to carry out a time history analysis (THA) under the selected ground acceleration record, in order to calculate a damage index of plastic zones. Then, the damage index is used to calculate reduced members stiffness and strength. The latter are finally used to build a new model which is subjected to increasing levels of gravity loads. The structure is stable if the gravity load multipliers at the end of the THA is larger than 1. The collapse limit state is reached for the smaller earthquake intensity under which the gravity load multiplier first attains a unit value. This methodology is perfectly coherent with the definition of the collapse limit state as the one in which the structure loses its capacity to sustain gravity loads. However, the proposed method of analysis has one important limitation due to the assessment of damage on the basis of a non degrading structural model. In fact, it has clearly been shown that explicit consideration of damage in THA is essential to correctly calculate the displacement demand when the structure enters the degrading range of response (Della Corte 2001, Della Corte et al., 2002, Krawinkler 2002) (Fig. 12). In other words, there is strict dependency of displacement demand on degradation of the restoring forces. Besides, the method proposed by Mehanny and Deierlein has the disadvantage to require two models to be built, one for the THA and one for the subsequent pushover under gravity loads. However, the method is able to capture those situations were collapse may occur with a non-flat (or hardening) response in terms of relationship between spectral acceleration and maximum transient drift. This possibility has been commented in Della Corte (2001), where a procedure very similar to the one proposed by Mehanny and Deierlein is suggested, but with the following differences: 1) the THA is carried out using degrading structural models; 2) the inelastic pushover analysis under gravity loads is replaced by a simpler elastic lateral force plus gravity loads analysis. In particular, the second step of the analysis is carried out to capture also the residual lateral strength of the structure and to define a seismic damage index varying between 0 (no damage) and 1 (global collapse). Both the first and the second step are carried out using the same structural model, hence the additional time required to carry out the pushover static analysis is marginal. It is noted that, once a model is built for both dynamic and static analysis, also the more rigorous inelastic pushover analysis under gravity loads could be used. The idea behind substituting the inelastic pushover with an elastic lateral load analysis is that inelastic demands could be estimated through simplified tools, e.g. demand spectra, based on and verified with many THA. Then, the assessment of collapse safety could be carried out with a relatively simple procedure. More recent research results (Ibarra and Krawinkler 2005) have clearly highlighted the importance of a probabilistic framework to asses collapse safety of structures. This has somewhat shifted the attention towards different aspects of the problem, such as estimating the mean annual frequency of exceedance of earthquake intensities inducing a given damage state and the incorporation of ground motion (i.e. frequency content) and system model uncertainties in the process of collapse assessment. Indeed, the development and application of probabilistic approaches to collapse assessment as proposed in the work by Ibarra and Krawinkler (2005) clearly shows how large can be the 147

Figure 13.

Pre-test (a) and post-test (b) analytical predictions (Lignos et al., 2008).

variance of the annual frequency of exceeding the collapse limit state when record-to-record variability and system mechanical properties uncertainties are included in the analysis. Besides to the emphasis on the need to approach the problem of collapse assessment with a rigorous probabilistic methodology, the work by Ibarra and Krawinkler (2005) also presents very useful quantitative information about the effect that different input and system parameters may have on the collapse capacity of a structural system. For example, the parametric study developed by the Authors show that among the hysteretic system parameters, the plastic hinge ductility corresponding to the peak strength and the value of the negative post-peak stiffness in the backbone restoring force characteristics are the two parameters having the utmost effect on collapse capacity, while cyclic deterioration is relatively of minor importance, independent of the type of ground motions considered (i.e. nearfault or non near-fault). In addition, it is shown that pinching of the hysteresis cycles has a minor effect on collapse capacities and systems with pinching or peak-oriented loops may have similar collapse capacities, but smaller than bi-linear systems with negative post-peak stiffness. The work also illustrates how P-Delta effects may dominate the response of tall structures, leading to high mean annual frequencies of collapse even in the case of non-deteriorating hysteretic properties. While the reader is referred to the original work by Ibarra and Krawinkler (2005) for more details on the many useful and interesting numerical results, it has to be noted that one main limitation of their work lies in having restricted the collapse mechanism of the investigated generic frames to a beam-sway type. The Authors do also comment on this point, emphasizing that (i) column hinging can significantly reduce collapse capacity and (ii) current US codified rules may not be able to avoid such column plastic hinging. Analytical studies both in support and for interpretation of experimental collapse tests are presented by Lignos et al. (2009). Basically, the phenomenological model developed by Ibarra and Krawinkler (2005) and used for parametric analysis, was here used both for pre-test predictions and post-test interpretation of results. Though the pre-test predictions were adequate to capture reasonably well the maximum acceleration bringing the structure to collapse, significant errors occurred in terms of deformation demand near to collapse (Fig. 13a). Re-calibrating the numerical model using post-test data on moment-rotation hysteresis responses, much better displacements/deformations predictions were possible (Fig. 13b). This is consistent with the concept that collapse is a limit state of stability, where sensitivity of displacement response to small system perturbations is large. Based on this experience, the Authors conclude that phenomenological models for numerical analysis of collapse can be accurate but they should be calibrated using loading protocols that consider many small-amplitude followed by few large amplitude deformation cycles, as it typically occurs in near collapse dynamic response time-histories. One parameter essential for reliable collapse prediction is the rotation corresponding to the peak restoring force (mentioned as post-capping rotation capacity). Furthermore, the Authors conclude that pushover analysis is accurate for predicting the elastic response, the yield limit state and collapse mechanism of a first-mode dominated structure. However, even in this case, higher mode effects can significantly affect peak shear forces. The last conclusion agrees with the numerical results recently found about the response of a dual moment-resisting frame and buckling-restrained braced frame (Maley et al., 2010). In summary, looking at the selected publications on analytical studies about the assessment of collapse safety of structures, it clearly appears that (i) numerical models must reproduce degradation 148

of plastic hinges, especially post-peak negative stiffness and associated strength degradation; (ii) collapse safety must necessarily be assessed on a probabilistic basis and both ground acceleration and system model parameter uncertainties should in principle be included in the analysis; (iii) the need exist for a rational collection of component-level experimental data in order to obtain reliable information about the variability of the mechanical inelastic properties of plastic hinges; (iv) most of the studies (all of those mentioned in this short summary) deals with moment-resisting frames for buildings, while scarce information can be found about braced frames; (v) an approach similar to that adopted for buildings could/should be developed for bridges, with due consideration of the relevant structural features.

2.2.2 CONTRIBUTIONS FROM COST MEMBERS 2.2.2.1 Introduction The contribution by COST C26 members summarized in this report are those relevant to the papers published in the Proceedings of the two intermediate Workshops organized in Prague and Malta. However, since some of the papers were presented on the same topic in both Workshops, the contributions are here organized on the basis of the topic. A short summary of each contribution is given hereafter focusing the attention on those aspects relevant to the general focus of this Chapter, i.e. analysis of structural response in case of exceptional earthquake actions.

2.2.2.2 Experimental tests Reinforced concrete structures Precast industrial buildings consisting of an assemblage of cantilever columns tied at the level of the floor diaphragms are frequently built in Europe. However until recently, neither the system as the whole, nor the connections were fully tested. So neither the appropriate level of seismic forces nor the capacity design of the connections was implemented. Therefore (also in line with the COST C26 definition) seismic loading could be a catastrophic action for such systems. Full-scale pseudo-dynamic and cyclic tests of one-story industrial buildings were done (Fischinger et al., 2008d and Kramar et al., 2009) up to the extreme drifts of the columns – up to 8% (Fig. 14). It was observed that the behaviour of such very slender coulmns (shear span ratio was 12.5) is specific and cannot be described by any known numerical model (see section 2.3.1). The strength and deformation capacity of the pinned beam-to-column connection with one dowell located at the centre of the column (Fig. 15) were investigated with monotonic and cyclic tests within the EU SAFECAST project (Kramar et al., 2010). Results are given in the paper for the Napoli 2010 conference (Fischinger et al., 2010). Seismic response of lightly reinforced RC coupled structural wall (Fig. 16) was investigated by bi-axial shake table test at LNEC, Lisbon. (Fischinger et al., 2008c). The coupling beams were much stronger than foreseen leading to large axial forces in piers and subsequent failure of piers in shear. This indicated the potential inadequacy of the existing numerical models and design procedures with possible catastrophic consequences. The confinement of the thin edges and the detailing of thin diagonally reinforced coupling beams were quite efficient. Reinforced concrete structures built in seismic zones before 1960s or even 1970s were designed to resist mainly the gravity loads and wind. The response of such structures can be catastrophic even in the case of the expected design earthquake. At the COST C26 Prague meeting Mazzolani et al. (2007a) presented the results of full-scale tests on existing reinforced concrete buildings (Fig. 17) seismically upgraded by means of several innovative techniques. In particular, the use of externally bonded carbon fiber reinforced polymers, eccentric braces and buckling restrained braces were investigated and compared. Full-scale experimental tests (Fig. 18) on a real masonry-infilled reinforced concrete (RC) twostory frame building were discussed by Mazzolani et al. (2007b). Two cyclic inelastic tests were carried out. The first test was carried out on the building in its original condition, producing large damage in masonry infill walls, RC frame columns and the staircase structure. Then, some heavilydamaged RC frame columns were repaired, while perimeter masonry panels were re-constructed 149

Figure 14. Precast industrial building tested at ELSA in Ispra.

Figure 15.

Failure of the dowel.

Figure 16.

Figure 18. Full-scale experimental test on a real masonry-infilled RC building (Mazzolani et al., 2007b).

Figure 17. Full-scale tests on seismically upgraded existing RC buildings (Mazzolani et al., 2007a).

and strengthened using carbon fiber reinforced polymer (C-FRP) bars. The latter were applied with the near surface mounting technique, placing them in the horizontal mortar joints. Steel structures As previously discussed, the assessment of collapse safety of structures must necessarily be based on the knowledge of the degrading mechanical response of structural component entering the inelastic range of behavior. Therefore, all the component-level experimental tests which may add information on this aspect, could potentially be useful in subsequent development of analytical studies. From this perspective, test results presented by Landolfo et al. (2008) may be useful in order to develop analytical models of moment-resisting steel frames made of European shapes and subjected to ground shaking trough collapse. In fact, the Authors carried out experimental tests on typical European steel beams under constant moment gradient into the inelastic range of response up to severe strength degradation (Fig. 19a). Empirical formulas for characterizing ductility and plastic overstrength were proposed by the Authors. Though the main aim of the experimental tests was to establish the appropriateness of the European cross-section classification in case of seismic actions, results were also exploited in the perspective of the response of steel structures to catastrophic seismic events. In fact, strength and stiffness degradation was measured in the range of large plastic rotation demand (Fig. 19b). D’Aniello et al. (2008) presented results of experimental tests on buckling restrained braces (BRBs) in the perspective of their response to seismic events imposing displacement demand far in excess of the design displacement. Possible design alternatives were presented, looking for maximizing the brace capacity to sustain extremely large deformations. Namely, two basic alternative ultimate failure modes were identified in case of demountable “all-steel” BRBs: 1) global buckling after exceedance of the compression deformation capacity of the yielding core relative to the restraining sleeve (Fig. 20a); 2) opening of the built-up bolted sleeve due to excessive pushing 150

Figure 19. Tests on steel beam-column members: a) experimental results; b) analytical results.

Figure 20. Tests on buckling-restrained braces.

Figure 21. Tests on removable bolted shear links.

forces transmitted by the yielded and locally buckled core (Fig. 20b). The first type of failure mode is initially associated with an increase of lateral strength because of the larger compression strength of the whole BRB system made up of the core and sleeve behaving as a whole, as respect to the compression strength of the yielding core alone. However, such an increase vanishes with global buckling because of the associated strength and stiffness degradation. In addition, other frame components (columns, beams and connections) may not be able to sustain the increased peak forces transmitted by the braces, leading to progressive collapse of the frame. The second type of failure mode is characterized by smaller strength and stiffness but larger deformation capacity, and it avoids transmitting large overloads to other frame components. The inconvenience may be in strong plastic deformation concentration in the core, leading to fracture and complete loss of brace strength. As preliminary evaluation, the second failure mode was conjectured to be preferable. Stratan and Dubina (2008) presented test results on nearly full-scale specimens of eccentrically braced bare steel frames with removable steel shear links (Fig. 21a). Removable steel links are thought to be used in advanced steel structural systems because of the significant advantage to reduce the repair cost thanks to the ease of substitution of the damaged link elements. Tests were carried out up to complete failure of the device, which sometime occurred with fracture of the bolts used for the end-plate connection. A thinner end-plate was associated to plate-yielding, with a more gradual strength and stiffness degradation. (Fig. 21b) Therefore, test results show that appropriate detailing of end connections is essential to obtain a good performance. This is important from the perspective of the response of structures to exceptional earthquakes, because the abrupt reduction 151

Figure 22. Tests on masonry panels.

Figure 23. Tests on a large-scale model of a mosque with a minaret.

of story shear strength due to bolt fracture may induce story sidesway collapse of the building. Similar results were obtained by Mazzolani et al. (2007) in case of shear links used for seismic upgrading of existing RC buildings. Masonry structures Masonry structures suffer strong degradation under cyclic loads and this simple consideration explains their significant seismic vulnerability. One difficulty in the assessment of collapse of masonry structures derives from the multiple number of possible failure modes. Along with global failure modes, local or intermediate failure modes may appear, both in-plane and out-of-plane of the main masonry walls. Besides, some historical buildings are characterized by special geometrical features that make it necessary specific detailed studies to assess seismic collapse capacity and demand. Results of tests on masonry panels under both monotonic and cyclic loads were presented by Dogariu et al. (2007) (Fig. 22). Though the focus of the work was on strengthening techniques, results on bare masonry panels constitutes the baseline to evaluate the effectiveness of the retrofitting intervention and could profitably be exploited in view of seismic collapse assessment. In fact, results give information about degradation under cyclic loads, in addition to that available in other similar research studies. Such information is precious in order to develop analytical models able to trace the response of masonry shear panels up to complete failure under earthquake actions. Results of shaking table tests on a large-scale model of a historical mosque with a minaret were presented by Krstevska et al. (2007). The masonry model was subjected to multiple shaking according to the following three main steps: 1) test on the original bare masonry model, with a small intensity earthquake to produce small, repairable, damage (Fig. 23a); 2) tests on the repaired and strengthened model, until collapse of the minaret (Fig. 23b); 3) tests on the strengthened mosque (without the minaret, Fig. 23c) until collapse. The strengthening systems were based on FRP materials (Fig. 23c). According to the Authors, based on the experimental evidence, the FRP strengthening system was effective in increasing the seismic coolapse capacity of the mosque. Results of shaking table tests were also presented by De Matteis et al. (2008), but on a scaled model of the central part of the Fossanova gothic church (Fig. 24a). Multiple shaking tests were performed on both the bare masonry model and a strengthened one, according to a two-phases criterion: 1) tests on the bare masonry, to produce moderate, repairable, damage; 2) tests on the strengthened model, until strong damage (let say, close to collapse). The strengthening system consisted of transverse FRP tie rods connecting the ends of the central navy and the aisles (Fig. 24b). 152

Figure 24. Tests on a full-scale model of a Gothic church.

The original structure was able to sustain a peak ground acceleration of 0.14 g, which corresponded to the formation of strong damage (Fig. 24c). The strengthened structure was instead able to reach a peak acceleration at the base equal to 0.4 g, thus testifying the effectiveness of the strengthening system.

2.2.2.3 Analytical studies Reinforced concrete structures In most studies presented by COST C26 members relatively simple lumped plasticity models were used. However several modifications and upradings as well as empirical calibrations were needed to account for the sophisticated mechanisms characteristics for the inelastic response at exceptional seismic event. When modelling the shake-table response of a coupled structural wall, Fischinger et al. (2008c) defined the inelastic characteristics of the shear spring in the MVLEM element by the modified compression field theory and calculated the shear strength of the coupling beams by the FEM code ABAQUS. With these modifications the complex inelastic axial-flexural-shear interaction was successfully modeled. Apostolska et al. (2008) used push-over option in SAP 2000 to evaluate response for 3D wall systems with flexible foundations using capacity spectrum method. It was demonstrated that none of the existing models was able to simulate the degrading inelastic response of the very slender RC columns in precast industrial buildings. Efficient modifications of the Ibarra/Haselton model (see Section 1.3.2) were proposed accounting for the much higher yield rotation than foreseen by the original model and related change in the energy dissipation capacity (Fischinger et.al., 2008d). Using this model, both, the in-cycle and the repeated-cycle strength deterioration was very well modeled (Fig. 25). This model was used to evaluate seismic collapse risk of precast industrial buildings with strong connections (Fischinger et al., 2008d, Kramar et al., 2009). The intensity-measure-based PEER methodology was used (Fig. 26). It is obvious that the methodology works only if the degrading mechanisms in the post-critical range are appropriately modeled. Other large scale vulnerability assessment of RC buildings was done by Kappos (2007). RC members were modelled using lumped plasticity beam-column elements, while infill walls were modelled using the diagonal strut element for the inelastic static analyses, and the shear panel isoparametric element for the inelastic dynamic analyses. It is a particular challenge to model the behaviour of the upgraded RC structures. Both the non-ductile mechanisms of the original structure and the behaviour of strengthening materials and devices might be very complex. Nevertheless, several COST C26 studies (see also Section 2.2.1) demonstrated that even relatively simple models, when appropriately calibrated, could yield good results. Bordea et al. (2007) modeled the enhancement of ductility of reinforced concrete columns strengthned with FRP by increasing concrete strength and ultimate strain (for 4 times) according to the FIB recommendations. In the analyses of the full-scale cyclic tests of a real-masonry infilled building Mazzolani et al. (2007b) used relatively simple code (non-linear SAP) but advanced models for RC. The the FEMA 356 (2000) recommendations were used to define the momentrotation characteristics of the RC members and the shear capacity of FRP reinforced masonry was defined by the information provided by Galati et al. (2005). As expected the major challenge was 153

Figure 25. In-cycle and the repeated-cycle strength deterioration modeled by modified Ibarra model (Fischinger et al., 2008).

Figure 26. Illustration of the intensity-measure-based PEER methodology (Kramar et al., 2009).

Figure 27. Tests on removable bolted shear links.

to select modelling parameters for masonry infill panels, where some engineering judgement was needed. Steel structures The incremental dynamic analysis technique is currently one of the most powerful and widespread analytical tool to assess the seismic behavior of a structure through multiple performance levels. The contributions from COST members were based on this method of analysis and they permit an estimation of the collapse capacity of the investigated structural systems, in terms of comparison among different alternatives. In this contest, Dubina et al. (2007) presented a numerical investigation into the dynamic inelastic response of dual moment-resisting-braced steel frames using the incremental dynamic analysis technique. Though the emphasis of the paper was on the advantages of using high-strength steel for non-dissipative members of steel frames and conventional numerical models were used for dissipative members, the numerical results may be found also useful from the view-point of collapse assessment of steel frames. In fact, the numerical results show that large axial force demand typically characterizes columns of braced bays, leading to the potential of loss of column stability and consequent progressive collapse. The incremental dynamic analysis results presented by the Authors (Fig. 27a) also highlight the potential of sidesway collapse because of column and beam yielding, which is more easily controlled by using high-strength steel for columns and beams (Fig. 27b). Iuorio et al. (2008) discussed the behavior of light steel structures made up of cold-formed members and braced with oriented strand-board (OSB) and gypsum wall-board (GWB) panels (Fig. 28). Though this paper was focused on design under “normal” earthquake loading, previous experimental and numerical studies were carried out using incremental dynamic analyses and full-scale specimen cyclic tests carried out through collapse (Landolfo et al., 2006, Della Corte et al. 2006). Numerical results, based on a refined numerical model calibrated with experimental tests and able to capture pinching, were used in support of the experimental tests and with the final objective to develop and validate a design methodology able to guarantee sufficient collapse margin ratios. 154

Figure 28. Tests on light steel buildings braced with plywood and gypsum panels.

Figure 29.

Numerical model of a typical mosque.

Masonry structures Analytical modeling of masonry structures is complex because of highly and distributed non linear phenomena (cracking and plasticity). Usually, the task is accomplished by means of finite element numerical models able to capture the distributed nature of damage, but using experimental tests in support and for validation of models. Experimental tests may either be inelastic tests on substructures and components or full-scale tests in the elastic range of response of the real structure (on-site dynamic identification). For example, Krstevska et al. (2007) developed finite element numerical models of a historical mosque with a minaret (Fig. 29a), which was previously tested until collapse in several phases (Section 2.2.3). A numerical model was created and analyzed for each of the phases used for the experimental tests. The Drucker-Prager plasticity model was implemented for the analyses. The maximum load sustainable by the structure was determined on the basis of a maximum acceptable local plastic strain, the latter calibrated using shear tests on masonry panels. The results of the numerical simulations were then compared with the experimental results discussing the differences. Two sets of numerical models were built-up to simulate the response of the original prototype and the FRP-strengthened model (Section 2.2.3). Figure 29b shows, for example, the “critical” plastic zones for the strengthened structure subjected to the maximum value of the peak-ground acceleration during the shaking table tests. The Authors suggest that the FE models can be used to compare original and strengthened models, obtaining information about the increase of strength. However, it seems that calibration with large-scale models may still be needed to reproduce the experimental behavior. Analogously, De Matteis et al. (2008) presented results of a numerical model of the Fossanova gothic church, which was tested with a scaled model (Section 2.2.2). Pre-test numerical investigations were presented, which were aimed to help for the design of the experimental tests. The pre-test numerical model (Fig. 30a) was calibrated using results of on-site dynamic identification by the ambient vibration and experimental modal analysis procedures. Numerical results (Fig. 30b) 155

Figure 30.

Numerical model of a gothic church.

were found to be very helpful and substantially able to predict the response subsequently observed with the experimental tests. In particular, the model indicated the location of “critical” zones and the peak values of ground accelerations inducing damage of the relevant zones. As for the study mentioned before, it can be argued from the results presented by the Authors that the FE models always need to be calibrated, at least on the basis of elastic dynamic properties experimentally measured by means of on-site dynamic identification techniques.

2.2.3 CONCLUSIONS AND RECOMMENDATIONS FOR FURTHER DEVELOPMENTS Extreme conditions occur when both loading and structural resistance are combined in such a way to reduce the safety level below acceptable values. Such situation may arise either when the loading is larger than expected (foreseen by the codes existing at the time of the design of the structure) or when the provided capacity was lower than anticipated by the modern standards. Building and structural elements subjected to such exceptional seismic loads are expected to enter into highly inelastic range exhibiting strong stiffness and in particular strength degradation as well as failure modes, which up-to-date codes try to prevent. Until a decade ago there were very few experimental and analytical studies of the structural behaviour up to the collapse. State-of-the-art review has showed that recently the interest has shifted towards near-collapse behaviour. This information is urgently needed for several important reasons: (a) to support development of the PBD procedures, (b) to provide the basis for the seismic risk studies, (c) to gain adequate information of innovative structural systems, (d) to provide data for seismic rehabilitation of the systems designed prior to the introduction of the modern seismic design principles and (in line of the COST Action C26) to (e) to foresee the behaviour of structural systems at exceptional events. Impressive (full-scale) experiments were performed near to the collapse providing suitable data for the development of the adequate analytical models. Among these models empirical calibrated macro-models are prevailing. Although rather simple in concept they provide better physical understanding and consequently better control of the highly complex near-collapse mechanisms. Works performed within the COST Action C26 have considerably contributed to these results. Nevertheless, each collapse mechanism of a particular structural element or structural system is so specific that general solutions are not to be expected. The only long term solution is to perform sufficient experiments, which in the future imposes a tremendous work load on the engineering and research community.

REFERENCES Apostolska R., Bonev Z.V., Blagoev D., Vasseva E., Necevska-Cvetanovska G. (2008). Design seismic response evaluation for 2D frames and 3D wall systems with flexible foundation using capacity spectrum method. Proceedings of the International Symposium “Urban habitat constructions under catastrophic seismic events”, Malta, 23–25 October, pp. 152–157.

156

Bazzurro, P. and Cornell, C. A. (1994a). Seismic hazard analysis for non-linear structures. I: Methodology, ASCE Journal of Structural Engineering, 120(11): 3320–3344. Bazzurro, P. and Cornell, C. A. (1994b). Seismic hazard analysis for non-linear structures. II: Applications, ASCE Journal of Structural Engineering, 120(11): 3345–3365. Bordea S., Stratan A., Dogariu A. and Dubina D. (2007). Seismic upgrade of non-seismic r.c. frames using steel dissipative braces. Proceedings of the COST C26 Workshop, 30–31 March, Prague, Czech Republic, pp. 211–220. Chen S. and Kabeyasawa T. (2000). Modeling of reinforced concrete shear wall for nonlinear analysis. Proc. of the 12th WCEE, Aukland, paper No. 1596. Combescure D. and Chaudat T. (2002). ICONS European program: Seismic Tests on R/C Bearing Walls CAMUS 3 Specimen, CEA, Report DMT, SEMT/EMSI/RT/02-067/A. D’Aniello M., Della Corte G., Mazzolani F.M. (2008). Response of buckling restrained braces to catastrophic seismic events. Proceedings of the International Symposium “Urban habitat constructions under catastrophic seismic events”, Malta, 23–25 October. Della Corte G. (2001). Seismic stability of elastoplastic frames. Ph.D. Thesis, Department of Structural Engineering, University of Naples Federico II, available on-line at www.dist.unina.it. Della Corte G., De Matteis G., Landolfo R., Mazzolani F.M. (2002). Seismic analysis of MR steel frames based on refined hysteretic models of connections. Journal of Constructional Steel Research, Elsevier Science Ltd, 58: 1331–1345. Della Corte G., Fiorino L., Landolfo R. (2006). Seismic behaviour of sheathed cold-formed structures: numerical study. Journal of Structural Engineering, ASCE, Vol. 132, No. 4, 558–569. De Matteis G., Mazzolani F.M., Krstevska L., Tashkov L. (2008). Seismic analysis and strengthening intervention of the Fossanova Gothic Church: numerical and experimental activity. Proceedings of the International Symposium “Urban habitat constructions under catastrophic seismic events”, Malta, 23–25 October, pp. 247–254. Dogariu A., Stratan A., Dubina D., Nagy-Gyorgy T., Daescu C., Stoian V. (2007). Strengthening of masonry walls by innovative metal based techniques. Proceedings of the COST C26 Workshop, 30–31 March, Prague, Czech Republic, pp. 201–210. Dubina D., Dinu F., Ungureanu R., Zaharia R., Grecea D. (2007). High strength steel for seismic resistant building frames. Proceedings of the COST C26 Workshop, 30–31 March, Prague, Czech Republic, pp. 193–200. Elwood K.J. and Moehle J.P. (2002). Shake table tests on the gravity load collapse of reinforced concrete frames. Proc. of the 3rd U.S. – Japan Workshops on Performance-Based Earthquake Engineering Methodology for Reinforced Concrete Building Structures, pp. 407–420. Elwood K.J. and Moehle J.P. (2003). Shear and axial load faiure model for reinforced concrete frames subjected to earthquakes. Proc. of the 4th U.S. – Japan Workshops on Performance-Based Earthquake Engineering Methodology for Reinforced Concrete Building Structures, pp. 163–176. FEMA 356 (2000) – Prestandard and commentary for the seismic rehabilitation of buildings, Washington D.C. Fischinger M, Kramar M, Isakoviæ T. (2008a). Cyclic response of slender RC columns typical of precast industrial buildings. Bulletin of Earthquake Engineering; Vol.6(3):512–534. Fischinger M., Kramar M., Isakoviæ T. (2008b). Using macro elements to predict near-collapse performance of two typical RC building structural systems with lightly reinforced walls and slender precast columns. Proc. of the14th WCEE. Beijing, China. Fischinger M. and Isakoviæ T. (2008c). Seismic resistance of thin, lightly reinforced coupled wall evaluated by large-scale shaking table test. Proceedings of the International Symposium “Urban habitat constructions under catastrophic seismic events”, Malta, 23-25 October, pp. 164–169. Fischinger M., Kramar M. and Isakoviæ T. (2008d) Seismic risk of prefabricated RC industrial buildings with strong connections. Proceedings of the International Symposium “Urban habitat constructions under catastrophic seismic events”, Malta, 23–25 October, pp. 170–175. Fischinger M., Kramar M., Isakoviæ T. (2010). Seismic behaviour of dowel beam-column connections in precast industrial buildings. Proceedings of the Final COST C26 conference, Napoli, 2010. Galati N., Garbin E., Nanni A. (2005). Design guidelines for the strengthening of unreinforced masonry structures using fiber reinforced polymers (FRP) systems. University of Missoury Rolla. Haselton C.B. (2006). Assessing seismic collapse safety of modern reinforced concrete moment frame buildings. Dissertation, Stanford University. Ibarra L. F., Medina R., Krawinkler H. (2002). Collapse assessment of deteriorating SDOF systems. Proc. Twelfth European Conference on Earthquake Engineering, Paper Ref. 665, Elsevier Science Ltd. Ibarra L. F., Krawinkler H. (2005a). Global collapse of frame structures under seismic excitations. Report No. 152, The John A. Blume Earthquake Engineering Center, Department of Civil and Environmental Engineering, Stanford University.

157

Ibarra L.F., Medina R.A. and Krawinkler H. (2005b). Hysteretic models that incorporate strength and stiffness deterioration. Earthquake Engineering & Structural Dynamics 34(12):1489–1511. Kante P. (2005). Seismic Vulnerability of Reinforced Concrete Walls (in Slovenian), Dissertation, University of Ljubljana, pp. 353 Kanvinde A.M. (2003). Methods to evaluate the dynamic stability of structures – Shake table tests and nonlinear dynamic analyses. (EERI paper competition winner), Proceedings of EERI Meeting, Portland, February. Kappos A. (2007). Seismic vulnerability and risk assessment of urban habitat in Southern European cities. Proceedings of the COST C26 Workshop, 30–31 March, Prague, Czech Republic, pp. 115–129. Kim Y., Kabeyasawa Toshimi., Matsumori T. and Kabeyasawa Toshikazu (2007). Proc. of the 2nd NEES/EDefense Workshops on Collapse Simulation of Reinforced Concrete Building Structures, pp. 119–128. Kramar M., Isakovi´c T., Fischinger M. (2009). Seismic Collapse Risk of Precast Industrial Buildings with Strong Connections. Earthquake eng. struct. dyn., (38): 1–21, http://mc.manuscriptcentral.com/eqe. Kramar M., Isakovi´c T. and Fischinger M. (2010). Experimental investigation of pinned beam-to-column connections in precast industrial buildings, 14th ECEE, Ohrid, Sept. 2010 (to be published). Krawinkler H. (2002). Seismic demand analysis. PEER Research Digest No. 2002-1, Pacific Earthquake Engineering Research Center. Krstevska L., Taskov L., Gramatikov K., Landolfo R., Mammana O., Portioli F., Mazzolani F.M. (2007). Experimental and numerical investigations on the Mustafa Pasha Mosque large scale model. Proceedings of the COST C26 Workshop, 30–31 March, Prague, Czech Republic, pp. 158–169. Landolfo R., Fiorino L., Della Corte G. (2006). Seismic behaviour of sheathed cold-formed structures: physical tests. Journal of Structural Engineering, ASCE, Vol. 132, No. 4, 570–581. Landolfo R., Piluso V., Brescia M., D’Aniello M., Mammana O., Tortorelli S. (2008). Rotation capacity vs. demand of steel beams under catastrophic events. Proceedings of the International Symposium “Urban habitat constructions under catastrophic seismic events”, Malta, 23–25 October. Lignos D. G., Krawinkler H., Whittaker A.S. (2008). Shaking table collapse tests of two scale models of a 4story moment resisting steel frame. Proceedings of the 14th World Conference on Earthquake Engineering, October 12–17, Beijing, China. Lignos D. G., Krawinkler H., Whittaker A.S. (2009). Collapse assessment of a 4-story steel moment resisting frame. Proceedings of the Thematic Conference on Computational Methods in Structural Dynamics and Earthquake Engineering (COMPDYN-ECCOMAS 2009), Rhodes, Greece, 22–24 June. Luco, N. and Cornell, C. A. (2000). Effects of connection fractures on SMRF seismic drift demands, ASCE Journal of Structural Engineering, 126: 127–136. Mehanny S.S., Deierlein G. G. (2000). Modeling of assessment of seismic performance of composite frames with reinforced concrete columns and steel beams. Report No. 135, The John A. Blume Earthquake Engineering Center, Department of Civil and Environmental Engineering, Stanford University. Marmo R., D’Aniello M., Portioli F., Landolfo R. (2010). Ultimate response of lap-shear riveted connection under exceptional actions. Proceedings of COST C26 Final Conference, Naples, Italy. Matsumori T., Shirai K. and Kabeyasawa T. (2007). Simulated earthquake test on a full-scale six-story reinforced concrete building at E-Defense, Part 1: Outline of the test program. Proc. of the 2nd NEES/E-Defense Workshops on Collapse Simulation of Reinforced Concrete Building Structures, pp. 5–15. Mazzolani F.M., Della Corte G., Barecchia E., D’Aniello M. (2007a). Experimental tests on seismic upgrading techniques for RC buildings. Proceedings of the COST C26 Workshop, 30–31 March, Prague, Czech Republic, pp. 221–228. Mazzolani F.M., Della Corte G., Fiorino L., and Barecchia, E. (2007b). Full-scale cyclic tests on a real masonryinfilled RC building for seismic upgrading. Proceedings of the COST C26 Workshop, 30–31 March, Prague, Czech Republic, pp. 229–236. Moehle J., and Kabeyasawa T., Eds. (2000–2004). Proc. of the U.S. – Japan Workshops on Performance-Based Earthquake Engineering Methodology for Reinforced Concrete Building Structures. 1st – Maui, Sept. 1999, 381pp.; 2nd – Sapporo, Sept. 2000, 417 pp; 3rd – Seattle, Aug. 2001, 439 pp.; 4th – Toba, Oct. 2002, 337 pp.; 5th – Hakone, Sept. 2003, 417 pp., Japan Ministry of Education, PEER and U.S. NSF. Moehle J., Lynn A., Elwood K., and Sezen H. (2000). Assessing gravity load collapse in concrete frames. Proc. of the 1st U.S. – Japan Workshops on Performance-Based Earthquake Engineering Methodology for Reinforced Concrete Building Structures, pp. 175–190. Moehle J., and Kabeyasawa T., Eds. (2005–2007). Proc. of the NEES/E-Defense Workshops on Collapse Simulation of Reinforced Concrete Building Structures, 1st – Berkeley, July 2005, 419 pp.; 2nd – Kobe Oct./Nov. 2006, 299 pp., National Research Institute for Earth Science and Disaster Prevention, PEER, Japan Ministry of Education, U.S. NSF. Nassar, A. A. and Krawinkler, H. (1991). Seismic demands for SDOF and MDOF systems. Report No. 95, The John A. Blume Earthquake Engineering Center, Stanford University, Stanford, CA.

158

Nakashima M., Matsumiya T., Suita K., Liu D. (2006). Test on full-scale three-storey steel moment frame and assessment of ability of numerical simulation to trace cyclic inelastic behavior. Earthquake Engineering and Structural Dynamics, 35: 3–19. Panagiotakos T.B. and Fardis M. N. (2001). Deformations of reinforced concrete members at yielding and ultimate. ACI Structural Journal 98(2):135–148. Panagiotou M., Restrepo J.I., Conte J.P. and Englekirk R.E. (2006). Seismic response of reinforced concrete wall buildings, 8thUS NCEE, San Francisco. Park R and Paulay T. (1975). Reinforced Concrete Structures, Wiley, New York. Paulay T. and Priestley M.J.N. (1992). Seismic Design of Reinforced Concrete and Masonry Buildings, Wiley, New York. PEER Structural Performance Database (2007) Pacific Earthquake Engineering Research Center, University of California, Berkeley. http://nisee.berkeley.edu/spd. Cited 30 August 2007. Toniolo G (coordinator) (2007). Final report of the EU Research Project: Seismic Behaviour of Precast Concrete Structures with respect to EC8 (Contract No. G6RD-CT-2002-00857). Shome, N. and Cornell, C. A. (1999). Probabilistic seismic demand analysis of nonlinear structures. Report No. RMS-35, RMS Program, Stanford University, Stanford, CA, [Feb 12th, 2004]. Suita K., Yamada S., Kasai K., Shimada Y., Tada M., Matsuoka Y. (2009). Full scale shaking table collapse experiment on a 4-story steel moment frame: Part 2 detail of collapse behaviour. Proceedings of the International Conference on Behaviour of Steel Structures in Seismic Areas (STESSA 2009), Yokohama, Japan, pp. 131–136. Vamvatsikos D., Cornell C. A. (2005). Seismic performance, capacity and reliability of structures as seen through incremental dynamic analysis. Report No. 151, The John A. Blume Earthquake Engineering Center, Department of Civil and Environmental Engineering, Stanford University. Vian D. and Bruneau M. (2001). Tests to structural collapse of single degree of freedom frames subjected to earthquake excitations. Journal of Structural Engineeering, Vol. 129, No. 12, pp. 1676–1685. Vian D. and Bruneau M. (2001). Experimental investigation of P-Delta effects to collapse during earthquakes. Technical report MCEER-01-0001, University at Buffalo – Department of Civil, Structural and Environmental Engineering. Wu C.L., Yang Y.S. and Loh C.H. (2006). Dynamic gravity load collapse of non-ductile RC frames I: Experimental approach. Proc. of the 8th United States National Conference on Earthquake Engineering, San Francisco, CD-ROM, EERI.

159

Urban Habitat Constructions under Catastrophic Events (Final Report) – Mazzolani (Ed). © 2010 Taylor & Francis Group, London, ISBN 978-0-415-60686-8

2.3 Analysis of behaviour of constructions under impact and explosions: Approaches for structural analysis, from material modeling to structural response G. De Matteis University G. d’Annunzio of Chieti-Pescara, Pescara, Italy

E. Cadoni University of Applied Sciences of Southern Switzerland, Lugano, Switzerland

D. Asprone University Federico II of Naples, Naples, Italy

2.3.1 INTRODUCTION Recent catastrophic events on urban habitats, including terrorist attacks, have contributed to change the design approach to urban constructions. In fact, nowadays, a correct approach aims at ensuring the satisfactory performances of the structure during its lifetime, considering all the possible critical actions, which the structure could be subjected to, including severe dynamic load conditions. This approach can be considered a sustainable approach for structural design. It aims at guaranteeing not only the structural performances of the urban constructions, but also the social sustainability of the urban habitat, which is requested to be able to withstand catastrophic events, without causing catastrophic consequences. Hence, designers of urban facilities and infrastructures need to be aware of all the possible events that could act on the structures during their life-time, including catastrophic events. Malicious disruptions, blasts, or impacts have unfortunately become part of the possible scenarios that could occur on constructed facilities. A common feature of extreme loads is that they are intense and of short duration. This applies specifically to impact and blast loads. A further feature is that they result in large deformations. It is the ability to absorb energy through ductility that is the primary factor in survivability. This concept was firstly recognized by the automotive industry with the transfer to the “crumple design” method (Figure 1). Thereby ductile failures are specifically designed for, by selecting relatively weak structural members connected together by relatively strong joints. Anyway, a “robust” design has to be achieved, in order to allow global failure mechanisms and to avoid progressive collapse due to the premature loss of a structural component. Robust design of structures may be achieved by means of high redundancy, i.e. in framed structures by incorporating moment beam-to-column connections having large over strength or by using other dissipative sources in the structure (bracings, dampers and so on). A useful alternative, when the structures are designed for gravity loading only and simple connections are of concern, may be based on the employment of very ductile connections, favoring the development of catenary resistant mechanisms at large deformation stages. The former approach is typically adopted in earthquake prone Countries, while the latter in simply supported steelwork frames subjected to exceptional actions. In particular, in earthquake prone Countries ductility is the key concept of seismic design of structures. Such a concept was firstly introduced by the Japanese seismic code in the latter half of sixties and then incorporated in all modern seismic codes. Nowadays, it is a very common and simple concept that structures may survive under earthquake by mixing strength and deformation capacity. The UK robustness requirements date back to the Ronan Point disaster in 1968, in which part of a 22-storey block of flats collapsed as a result of a gas explosion. The resulting structural integrity requirements contained led to structures with an reasonable ability to resist blast by ensuring that 161

Figure 1.

“Crumple zone” design methodology for protecting car frames.

all structural members are adequately tied together. It is widely believed that this method will prevent progressive collapses by catenary action following the removal of a column, although this load path has not been adequately tested, either in practice or in theory. Further to this, structural elements vital to overall stability (key elements) must be designed to resist abnormal (blast) loads of 34 kN/m2 . This load corresponds to the average peak overpressure estimated to have been developed during the Ronan Point explosion. As such it does not provide protection against high-explosive blasts, which will be much higher in magnitude. Recent experience shown that buildings of urban habitats are vulnerable under extreme loads, such as the ones caused by terrorist attacks and by accidental events, e.g. impact or explosions. In particular, it has been emphasized that terroristic organizations may use unexpected weapons to attack urban habitats, such as explosive devices improvised by vehicle borne. Importantly for the structural engineer these weapons have a proven ability of either causing the complete collapse of entire buildings or severe progressive collapses of multiple bays of buildings, thus highlighting the vulnerability of even the most well protected structures to collapse due to exceptional actions. As a matter of the fact, since the events of September 11th concern regarding the danger posed to building occupants from vehicle borne devices has increased, introducing a wider range of structural engineering experts into this field of research. Addressing these loading conditions represents a critical issue for structural engineers. The most critical aspects can be summarized in the following items: • The assessment of the loading acting on the structure in case of blast and impact events is fundamental but represents a critical concern, since uncertainty related to load magnitude is high, especially for blast actions. • The Structural response in case of such severe dynamic actions represents a critical issue, since dynamic behavior of structural elements under severe dynamic loads can be very different from that exhibited under static actions. • The mechanical behavior of materials can be totally different from that exhibited in quasi-static conditions; thus a specific mechanical characterization is necessary. Hence specific investigations become necessary for all these concerns. The following sessions address these main issues and try to provide information about the state of art related to these topics, emphasizing the contribution to the research development produced within the COST Action C26.

2.3.2 BASIS FOR BLAST AND IMPACT LOAD CHARACTERIZATION An explosion induces mainly a quick and significant increase of pressure in the medium where it occurs, i.e. air or water. Such overpressure propagates as a wave, the so called “blast wave”, and is characterized by its speed, duration and intensity. These parameters are fundamental in order to evaluate the actions that an explosion can induce in the structural elements in its vicinity. The numerical values of these parameters depend on several aspects, such as, type and the amount of the exploding mass, distance of the target of interest from the explosion, geometry of the target, 162

Figure 2.

Blast overpressure in air generated by 1 kg of TNT at 2 m from the charge.

Figure 3. Two masses model for the impact loading.

type of reflecting surfaces (e.g., the ground in case of external explosions or walls or slabs in case of closed-in explosions). In the past decades, several investigations have been performed on such aspects and they have provided reliable numerical procedures for quantification of the overpressure time-histories. In the case of blast explosion, the induced overpressure follows a trend over time similar to that shown in Figure 2, where a positive decaying phase is followed by a weaker negative phase; it can be evaluated according to empirical formulations providing the pressure history p(t) depending on the type and the amount of the exploding charge and on its distance from the target (TM 5-1300, 1990; Henrych, 1979; Kinney and Graham, 1985). The most of such formulations convert the amount of the charge into equivalent weight of TNT, considered as a reference explosive. Hence, the pressure history curve depends on the distance from the target, generally through a cubic decaying rule and on the amount of equivalent TNT, generally through a cubic root rule. As it regards impact loading on structures, it can be divided into soft impact and hard impact. In the first case a deformable body strikes a rigid structure and the kinetic energy of the striking body is partially transferred into deformation energy of the striking body. Impact characterization consists in the evaluation of the contact force. This is the case of a vehicle impact on a wall and of a liquid impact on a containing structure. In the second case, the striking body is rigid and the kinetic energy is transferred into deformation energy of the impacted body. In this case, impact characterization consists in the definition of the displacement and deformation fields induced in the impacted structure. This is the case of hard masses like bullets. In both cases, simplified dynamic models have been proposed and can be used to characterize the impact loading on structural elements (CEB 1988). Such models are essentially based on two masses m1 and m2 representing the impacting body and the impacted body, respectively; the masses are linked to two springs, k1 and k2 , as depicted in Figure 3, representing the impact mechanism and the structural response of the impacted body, respectively.

2.3.3 STRUCTURAL BEHAVIOR UNDER BLAST AND IMPACT LOADING In case of impact or blast events occurring on civil structures failure mechanisms can be distinguished into local and global failure mechanisms. Local failure regards small parts of the structure and is directly due to the effects of the impact or the explosion onto structural elements. It can occur through punching or spalling mechanisms or through structural failure of frame or wall sections. The effects of a local failure depend on the dynamic characteristics of the structural elements, in terms of elastic properties, constraint configurations, ductility of the constituent materials and of the structural elements. 163

On the other hand, global failure of structures often occurs after local failure, being triggered by local failures of single structural elements, and it is related to the capacity of the structure to withstand the loss of structural elements without activate progressive collapse mechanisms, induced by the gravity loads. Global failure initiation depends on the global ductility of the structure and on the quality and the frequency of connections between elements of the structure. Obviously if local failure is more severe, then global failure becomes more probable. Based on these considerations, structural analysis needs to be conducted accounting for both local and global failure mechanisms. With regards to local failure, it should be distinguished between the failure of small portions of the structural elements, induced by punching or spalling mechanisms, and the failure of the structural elements caused by the internal forces. The first failure mode can be assessed through specific models accounting for the dynamic impulse produced by the explosion or by the impact on the surface of the structural elements. Different simplified models are proposed in literature (CEB 1988; TM 5-1300 1990). The second failure mode can occur when the action of blast or impact induces the failure of one or more sections within the element, due to high internal forces, i.e. shear or bending moment. To evaluate such internal forces the dynamic response on SDOF models can be considered. In fact, the rapid and intense loading induced by blast or impact, hit directly the single elements, which behave as independent structures, and can be modeled as independent structural elements (TM 5-1300, 1990). For both failure modes, the constituent materials can exhibit a strain-rate sensitive behavior, able to modify significantly the response from that exhibited under quasi-static loading conditions. These issues need to be taken into account, introducing specific strain-rate dependent formulations for the parameters of the constituent materials. This can be implemented through visco-elastic or visco-plastic constitutive laws or, through a simplified approach, by updating the elastic properties and the strength of the materials, depending on the occurring strain-rate, as proposed by several procedures (CEB, 1988, TM 5-1300, 1990). Due to the short duration of blast and impact events, for structural design purposes, blast and impact loading can be characterized in terms of peak pressure (or equivalently peak force) pmax and impulse i, being pmax the maximum pressure in the pressure history and i the integration of the positive pressure over time. Hence, for each structural element a pmax − i domain curve can be evaluated, reporting the values of impulse and peak pressure able to induce on the structural element a certain damage level. A typical pmax − i relationship is reported in Figure 4. These curves are employed to assess if a certain combination of peak pressure and impulse induced by a blast or impact event is able to cause a certain damage level to the structural elements. These relationships have been widely investigated in literature and can be defined for each structural element, using to a SDOF model (Bangash 1993; Mays and Smith 1995; Krauthammer 1998), in

Figure 4. Tipical pmax – i curve.

164

which the maximum displacement of a control point is considered as the damage parameter, e.g. the top displacement in a cantilever beam or the top deflection in a supported beam. These relationships have been investigated in particular for blast events but can be extended also to the case of impact events. Several closed-form formulations have been derived for different boundary conditions (e.g. cantilever beams, simulating protection walls or double fixed beams, simulating building columns) and for different constitutive laws, including elastic, elastic-plastic, rigid plastic, plastic with hardening and softening (Li and Meng, 2002; Soleiman and Louca, 2007; Ma et al., 2007; Smith and Hetherington, 1994). Generally, it can be observed that two asymptotes are obtained for high values of impulse and peak pressure. In particular, for high values of the impulse, the curve depends only on the peak pressure, corresponding to a quasi-static response of the structure. On the other side, for high values of the peak pressure, the curve depends only on the impulse, corresponding to an impulsive response of the structure. The former and the latter case occur when the duration of the pressure history is much higher or lower than the vibration period of the structural element, respectively. For intermediate values of the duration of the pressure history, the relationship depends on both the peak pressure and the impulse values. In other words, if the loading history is particularly short, the structural response depends only on the total impulse transferred to the structural element, whereas, if the loading history becomes long the structural response depends only on the peak value of the loading, as in a quasi-static regime. It is underlined that this approach implies that the structural response is always independent from the shape of the loading history; this implication is not true in the intermediate regime, but, for the seek of simplicity, it can be considered an acceptable approximation. Once the vulnerability of the structure related to local failure mechanisms has been addressed, structural design needs to consider the possibility that global failure mechanisms could be activated. Global failure occurs after a severe damage of one or more structural elements; once these elements have lost their carrying capacity a progressive collapse mechanism of the structure could be triggered. The progressive collapse can be defined as a mechanism which involves a large part of a structure, activated by local damages to the structural elements. Having lost some elements, the whole structure can become unstable, failing under the gravity loads. That is, the structure can eventually develop a global mechanism, which is widely referred to as the progressive collapse mechanism (Allen and Schriever 1972; SEI 2005; GSA 2003). Such failure mechanism can be addressed through a direct approach or an indirect approach (Ellingwood and Leyendecker 1978). In the direct approach progressive collapse scenarios are directly analyzed, whereas in the indirect approach, resistance to progressive collapse is pursued guaranteeing minimum levels of strength, continuity and ductility. Actually, the progressive collapse mechanism is the predominant mode of failure after a blast event (NRC 2001) and it is the subject of wide research (SEI 2005; GSA 2003; NRC 2001; Agarwal et al. 2003; Bennett 1988). Figure 5 depicts the Murrah building in Oklahoma City which was attacked by a blast event in 1995; the progressive collapse of a large part of the structure occurred. The possibility of this failure mechanism is linked to the capacity of the structure to redistribute loads on other structural elements. It depends on the redundancy of elements and the ductility of connections. The first aspect implies that the static scheme of the structure is much far from

Figure 5.

Progressive collapse of the Murrah building in Oklahoma City.

165

Figure 6.

Basic collapse mechanisms for frame structures.

a statically determinate configuration; the loss of some elements is balanced by the presence of other elements able to carry acting loads. The second aspect is also important, since after the loss of some elements, the new equilibrium configurations are reached with high local deformations, which must be tolerated by the structure. High ductility of connections is thus necessary to allow these static configurations. These considerations make progressive collapse issue quite similar to earthquake issue, since in both situations, structural retrofitting can be addressed by improving the ductility and the redundancy of the structures (Hayes et al., 2005). In the direct approach, the progressive collapse assessment can be conducted through non-linear static analysis of the damaged structure under the gravity loads (Marjanishvili and Agnew, 2006; Yi et al., 2008). In this case, plastic hinge formulations can be used to account for non-linearities; catenary actions can be also introduced to account for the tension forces in the horizontal elements (Yi et al., 2008). As an alternative, the progressive collapse analysis can be considered as a global stability analysis of the damaged structure. A possible approach to performing such analysis would be to conduct a plastic limit analysis. A plastic limit analysis (Corotis and Nafday, 1990; Ellingwood 2006) involves finding the load factor lincreasing the applied loads for which (a) equilibrium conditions are satisfied and (b) a sufficient number of plastic hinges are formed in the structure in order to activate a collapse mechanism in the whole structure or in a part of it. In static loading problems, a load factor l less than or equal to unity indicates that the structure is already unstable under the applied loads. On the other hand, in instantaneous dynamic loading problems, the threshold for l is equal to 2. In case of progressive collapse, it has been shown that a value 2 is probably conservative and the actual value of causing instability in the structure is between 1 and 2 (Ruth et al., 2006). The plastic limit analysis can be conducted by finding the smallest kinematically admissible load for which the above conditions (a and b) are satisfied, employing the principle of virtual work. A kinematically admissible load corresponds to a mechanism in which both the external work done by the applied forces (gravity loads in this case) on virtual deformations and the internal work done by the ultimate moments on the virtual rotations are positive. It can be shown that the collapse mechanism can be described as a linear combination of the independent mechanisms that can be activated in the structure. The number of these independent mechanisms is equal to the difference between the number of possible plastic hinge locations and the degree of indeterminateness in the structure. Figure 6 depicts the types of these basic mechanisms. Asprone et al., (2010) proposed a procedure to conduct the blast assessment of strategic infrastructures in which, through a probabilistic approach, different blast scenarios are analyzed, performing different plastic limit analysis of different damaged configurations of the structure. Furthermore, the procedure for finding the smallest kinematically admissible load is defined as a linear optimization programming with the objective of minimizing the load factor and is implemented through a simplex algorithm. 2.3.4 DYNAMIC BEHAVIOR OF MATERIALS Under dynamic loading conditions, many construction materials present a different mechanical behavior, with respect to that exhibited in quasi-static regime. Generally, both compressive and 166

tensile strength increase; stiffness can also present higher values whereas failure strains can both increase and decrease. The dynamic behavior of materials is due to different phenomena, influencing the mechanical properties. In example, inertia effects in fracture propagations govern the dynamic behavior of ceramic materials, whereas viscous phenomena influence the mechanical response of polymeric materials. To quantify these effects the strain-rate (measured in s−1 ) is used as main parameter to describe the dynamic regime. The low strain-rate range varies from 10−6 s−1 to 10−4 s−1 . It is experienced during the quasi-static tests on materials, service loads on structures and vehicle transits. Higher values, up to 1 s−1 correspond to the medium strain-rate range and are experienced in the structures in case of soft impacts and earthquakes. Strain-rate values up to 102 s−1 correspond the high strain-rate regime, occurring on structures in case of hard impacts and blast events. Several experimental research activities related to dynamic properties of construction materials have been developed. In particular, experimental activities are conducted using different testing procedures. The most used equipments are the Drop-Weight Impact Machine, the Split Hopkinson Pressure Bar (SHPB) and its modifications, and the Hydro-pneumatic machine. In the Drop-Weight Impact Machine a certain mass is left to drop onto the specimen at a certain height, in order to have controlled impact energy (Banthia et al. 1989). By instrumenting the specimens with load cells, strain gauges and displacement transducers, the dynamic mechanical response can be acquired and interpreted. A critical issue in such test is represented by the inertia effects which need to be filtered. This test can be conducted also on structural elements (e.g. three or four point bending tests) to investigate the mechanical response under dynamic loading conditions at a higher scale. In order to design or to assess a structure or component subjected to dynamic loading an accurate knowledge of the elastic and inelastic strength properties of the materials involved is requested. In particular is demanded the knowledge of the complete stress-strain curve. Not all dynamic device are able to measure this information, for example pendulum impact machines such as Charpy or Izod can produce strain rates of up to about 100 s−1 , yielding only energy absorbed to fracture, but not a complete stress-strain curve. In the last fifty years many efforts were addressed to the development of systems designed to fill the strain rate range from 102 s−1 to 103 s−1 , the time duration of many explosive, ballistic impact, crashes and other accident scenarios of interest for both military and civilian applications. A basic technique to measure the shape of a stress pulse in a long elastic bar was first described at the beginning of XX century by Hopkinson (1914). More than thirty years later, Kolsky (Kolsky 1949) improved on Hopkinson’s device, adding displacement gages and oscillographic recording techniques to obtain complete pulse amplitude wave forms in similar elastic bars. Kolsky used a two-bar system, sandwiching a short compression specimen between them. Both the stress and strain could be derived. This modification became know as the Split-Hopkinson Pressure Bar (SHPB). Hopkinson and Kolsky used explosive pellets to generate the stress pulse that propagated along the bar. In the early 1960s, Lindholm (1964) modified the Kolsky technique primarily by altering the bar lengths and the placement of the strain gages, and by using modern strain gage technology to record transient pulse shapes on both bars with electronic circuitry that allowed direct generation of the complete stress-strain curve for a single impact. Additionally, he used a mechanical spring device to launch a striker bar instead of using explosive pellets. The impact of the striker bar against the incident bar generated the stress pulse. The impact speed determines the magnitude of the stress pulse, and the velocity of the striker bar was controlled quite accurately by adjusting the compressed gas pressure. The duration of the stress pulse is controlled by the length of the striker bar. The SHPB became an integral element of a materials research program that studied and characterized the behavior of metals under various loading conditions. Since then, the device has undergone a number of modifications and improvements, and has been applied to a wide range of other materials including geologic materials (Cadoni 2010; Asprone et al., 2009), ceramics, polymers and other soft materials. The propagation of the stress waves without dispersion and uncontrolled reflections and the deformation of the specimens in a state of stress homogeneity are the basic conditions to be satisfied for a correct implementation of the uniaxial elastic stress wave propagation theory in the analysis of the Hopkinson bar experimental records and therefore for the accurate measurement of the dynamic mechanical properties of materials. The basic principles of the Hopkinson bar method are normally not respected in the impact rigs based on the use of drop weight or missiles directly impinging on the specimen for impact load generation and on the use of load cells as load transducers in contact with the specimen. 167

Figure 7.

Modified Hopkinson Bar scheme.

The material properties under impact loading measured with drop weight impact rigs are of low accuracy (and therefore not reliable for the development and calibration of material constitutive laws) because affected by many noisy phenomena like resonant vibrations, rebounds, superposition of waves reflections (Birch et al., 1988). In the seventies Montagnani and Albertini (1974) developed a modification of Hopkinson bar where the stricker was substituted with a bar loaded elastically. The modified Hopkinson bar consists of two half-bars, the incident and transmitter bar respectively, with the specimen introduced in between, as shown in Fig. 7. Elastic energy is stored in a pre-tensioned bar, which is the solid continuation of the incident bar. By releasing this energy (rupturing the brittle intermediate piece in the blocking device), a rectangular wave with small rise-time is generated and transmitted along the incident bar loading the specimen to failure. This is a uniaxial elastic plane stress wave, as the wave-length of the pulse is long compared to the bar transverse dimensions, and the pulse amplitude does not exceed the yield strength of the bar. This modification can produce both direct tension and compression test as well as shear or bending. On the basis of the incident (εI ), reflected (εR ) and transmitted (εT ) records, of the consideration of the basic constitutive equation of the input and output elastic bar material, of the one-dimensional wave propagation theory it is possible to calculate the stress, strain and strain-rate curves by the following equations (Lindholm 1971):

where: σ(t) represents the longitudinal stress in the specimen ε(t) represents the longitudinal strain in the specimen εT (t) represents the longitudinal strain transmitted in the output bar εR (t) represents the longitudinal strain reflected in the input bar ε˙ (t) represents the strain rate in the specimen E0 is the elastic modulus of the bars; A0 their cross-sectional area; A is the specimen cross-sectional area; L is the specimen length; C0 is the sound velocity of the bar material. The Modified Hopkinson Bar allows investigating the dynamic response at the material scale and, with respect to the Drop-Weight Machine tests permit to control better the strain rate level of the test. Given the importance of this topic for structural engineering, several experimental research activities related to dynamic properties of construction materials are available in literature, investigating steel, ceramic materials (i.e. concrete and natural stones) and polymeric materials. A short review of the available literature is here presented. 168

Dynamic mechanical behavior of steel, as experienced with many other metals, presents significant differences if compared with the mechanical behavior exhibited under static load conditions. This is due to several phenomena involved in steel strain-rate sensitiveness, but mainly the reason of such differences stands in the dynamic dislocations evolution, affecting the microscopic scale (Lee and Liu 2006; Mainston 1975; Uenishi and Teodosiu 2004). Available scientific data outline that, as strain rate increases, the following changes in mechanical properties of reinforcing steel can be noticed (Lee and Liu 2006; Mainston 1975; Uenishi and Teodosiu 2004; Malvar 1998): • an increase of yielding stress fy ; • an increase of ultimate tensile stress ft ; • an increase of the ultimate tensile strain εt ; On the contrary no changes are experienced in terms of Young modulus. Unfortunately, available literature focuses the attention on dynamic properties of several steel alloys, mainly for industrial applications (Lee et al., 2005; Couque et al., 1994; Drar 1993; Wang and Wang 2007), whereas few data are available on reinforcing steel properties (Malvar 1998; CEB 1988; Filiatrault and Holleran 2001; Asprone et al., 2009). The differences between dynamic and static properties of concrete are amply illustrated in the literature (Fu et al., 1991; Harris et al., 2000; Cadoni et al., 2000; Malvar and Ross 1998; Cadoni et al., 2006; Asprone et al., 2009). It is accepted that, to evaluate correctly the behavior of concrete under extreme dynamic loads, dynamic mechanical properties have to be necessarily investigated. The scientific data available show that, under high strain rates, concrete can exhibit, both in compression and in tension: • an increase in failure strength (Fu et al., 1991); • an increase in Young’s modulus (Cadoni et al., 2000; van Doormal et al., 1994); • a different evolution of cracks, that do not develop through a local mechanism, like in a static range, but start and grow at the same time in several locations (Cadoni et al., 2000). • an increase in flexural strength (Banthia et al., 1989). The dynamic behavior of both concrete and steel is addressed into several technical codes and instructions, so as to guide the engineer to predict properly the real behavior of a structure under extreme loads. CEB information bulletin no. 187 (CEB 1988) gives formulations to evaluate the dynamic properties of concrete and steel by updating static properties. Defining the Dynamic Increase Factor (DIF) as the ratio between the dynamic and the static value of a certain parameter, DIF–strain rate relationships are suggested for compressive and tensile failure stresses, compressive and tensile ultimate strains and Young’s modulus. Similarly, in TM 5-1300 (TM 5-1300 1990) the issue of the dynamic properties of concrete and steel is addressed by suggesting different DIF for failure strength in several load conditions. In this case, the DIF is not a function of strain rate value, as expressed by CEB formulations, but it just depends on the distance from the blast source, distinguishing between far and close-in explosions. Within rock mechanics literature, a number of studies are available investigating influence of strain-rate on mechanical properties of different types of rocks. Dependence on strain-rate of natural stones was experienced in different loading conditions, such as uniaxial compression (Ray et al., 1999; Ma and Daemen 2004), uniaxial tensile (Kubota et al., 2008; Asprone et al., 2009) and triaxial compression (Lia et al., 1999). Regarding tensile loading conditions, it was accepted that natural stones exhibit higher strength values as strain-rate increases; this behavior is mainly due to microscopic non homogeneity, affecting principally sensitivity to medium strain-rates, and microcrack formations, interfering with failure surfaces development, which influence primarily high strain-rates regime (Cho et al., 2003). Mechanical properties of composite materials are generally affected by strain rate dependence, as shown in (Cho et al., 2003; Hsiao and Daniel 1998; Harding 1993; Welsh and Harding 1985; Sierakowsky 1997). In particular, the initial Young modulus and failure stress increase as the strain rate becomes higher (Sierakowsky 1997; Okoli 2001), under both compression and tensile loading conditions. Focusing on glass fiber reinforced polymers (GFRP), the available literature describes in detail the effects of strain rate under compressive loading (Gary and Zhao 2000; El-Habak 1993; Tay et al., 1995; Huang et al., 2004). In particular, it is shown that mechanical behavior of composites in compression is strongly controlled by resin properties, which appear to be highly 169

Figure 8. Typical moment – curvature relationships for different strain rates.

strain rate dependent. Consequently, a significant increase is generally experienced in initial Young modulus, failure stress and failure strain, under compression. On the other hand, under tensile loading, GFRP also exhibits sensitiveness to the rate of loading, as described in Hardig and Welsh 1983, Makarov et al., 2004, Newill and Vinson 1993 and Asprone et al., 2009. In particular, also in this case, a significant increase in failure stress and initial Young modulus is exhibited, whereas failure strain decreases as strain rate increases (Majzoobi et al., 2005). The variations of the values of the mechanical parameters in the constitutive laws of the construction materials can be implemented to update the stress-strain relationships used as input data for the structural analysis. Hence, the peak pressure-impulse relationships above described can be evaluated through local structural analysis, using the updated formulations of the constitutive laws of the materials, depending on the strain-rate level occurring. As an example, in case of a reinforced concrete column invested by a blast induced overpressure, the fixed end beam scheme can be considered. Hence, to account for the flexural failure mechanism, the ultimate bending moment needs to be evaluated using the updated stress-strain relationships for both concrete and steel, at the strain rate level induced by a blast event (i.e. about 102 s−1 ). By implementing this procedure, significant variations are obtained. Figure 8 reports the moment – curvature relationship of a 250 mm × 250 mm reinforced concrete square column, reinforced with 4 steel bar of 12 mm of diameter. It can be observed that a significant increase of the bending moment values occurs for a strain rate of 102 s−1 . This approach can be also used to conduct more refined 3D Finite Element analysis, using specific dynamics codes, e.g. LS-Dyna or Abaqus, by implementing strain-rate dependent constitutive laws. Through this approach more refined results can be obtained, also accounting for large deformations and geometric non linearity. Furthermore, more sophisticated integration procedures can be also used, as an alternative to the classical Finite Element formulations, such as meshless and particle methods. These formulations, allow simulating large displacements and fracture propagations, without numerical instabilities due to mesh distortions. These methods are widely investigated in computational mechanics community and will be probably even more used in the future to address high dynamic mechanical problems.

2.3.5 COST C26 CONTRIBUTION TO RESEARCH DEVELOPMENT In order to improve the knowledge on the above field the WG3 within the COST Action C26 addressed the protection of structures to survive the effects of impact and explosion loading by developing adequate structural analysis approaches (Mazzolani et al., 2008). These include for example the threat from accidental explosions due to natural gases, as well as impact loading on structures due to other accidental events. The main contributions within the above research areas are considered in the following, providing an overview of the developed research issues. The vulnerability of multi-storey buildings with pre-cast load bearing walls to collapse following natural gas explosions has been considered by Langone, De Matteis and Mazzolani (2008), based on the application of the key element strategy to control the connection requirements (both strength and ductility) between the wall elements (Figure 9a). In fact, it has been stated that the adopted 170

Figure 9. Precast walls with ductile connections (a) and propagation of front flame (Study Case I).

Figure 10.

Numerical analysis with a stand-off distance of 14 m and related pressure on the building façade.

connecting system for precast RC walls is very important to provide stability and robustness to the whole structure. To this purpose a specific design methodology has been developed, which is based on the definition of a specific Pressure-Impulse diagram as a practical design tool for determining the required connection capacity according to the estimated peak pressure value into the compartment. In order to assess the pressure due to gas explosion and compare the corresponding values to the ones determined by applying the above empirical relationships, three different common situations were analysed, namely regular room with window as venting (Case I), irregular room with window as venting (Case II) regular room linked to another with window as venting (Case III). The results obtained shown that regular compartment geometry and combustion laminar flow limit significantly the value of the peak pressure. Such a value can be estimated in the range of 154–167 mbar by means of empirical relationships, and appears to be rather higher than the actual value determined by applying refined numerical methods (Figure 9b). In more realistic compartments (Case II and Case II), s turbulences and domino effects may be of concern. In such a case, the peak pressure increases significantly and can be evaluated by applying more sophisticated empirical relationships. (De Matteis et al., 2006). Kilic and Smith (2010a) developed an interesting numerical work for predicting pressure developed due to explosions behind rigid blast walls (sometime called ‘blast barriers’). Simulations are done by using the Arbitrary Lagrangian Eulerian (ALE) technique of the commercial finite element code LS-Dyna. The effects of the blast wall height and the stand-off distance of the building behind the blast wall in reducing the overpressure of the blast wave were investigated. A model building of 15 m height is placed behind the rigid blast wall, and pressures along the front façade were computed (Figure 10). It was concluded that the medium blast wall height of 3.00 m provides reasonable protection for the pressures on the front façade of the building with peak pressures of 171

Figure 11. scenario.

Considered column loss scenarios a) and evaluated dynamic increase factors for different loss

300 kPa, 125 kPa, and 75 kPa, respectively. As the stand-off distance between the blast wall and the building is increased, the attenuation of the blast wave provides a higher level of protection for the building behind. Although this paper has concentrated on the reduction in overpressure developed on the building façade, this overpressure reduction was accompanied by a reduction in the level of blast impulse delivered. This is an important point to note given that building damage is often determined by impulse. Also, the design of the deformable wall is addressed by developing numerical analysis with increasing amount of explosive material. Significant wall failure occurs when the amount of explosive material is increased well beyond the 20 kg design level (Kilic and Smith, 2010b). Dinu and Dubina (2008) investigated the response of high rise steel buildings as a result of column loss. The main objective of the study was to evaluate the redundancy of buildings designed for seismic loading due to the loss of a structural members caused by terrorist action. A case study, for which different hazard scenarios are taken into account, has been developed and the work confirmed the great benefit of installing truss systems in the uppermost parts of frame which can be capable of redistributing column loads following localized frame damage. Successive research activity shown that buildings designed to resist seismic loads have a good ability to avoid global collapse in case of column loss and the strategies employed to resist seismic actions generally aim to provide ductility and redundancy. Alternate load path analysis on two types of frames, designed for low and high seismic hazard, shown that rotation capacity of beam-to-column connections is critical in assuring force redistribution after the loss of columns. High resistance materials (i.e. high strength steel S460) may prove suitable for critical members (i.e. columns) that should not fail prematurely and therefore preventing global collapse. Static non-linear analysis may reproduce the behavior of the structure with sufficient accuracy, if the dynamic amplification of the gravity loads above the damaged area is properly accounted for. The case study considered the blast as cause of the column loss, but other types of extreme actions may cause similar effects (fire, for instance) (Figure 11a). If fire after blast scenario is also considered, the allowable level of damage from the columns loss scenario need to be adjusted, in order to take into account more damages are expected in the aftermath of the fire. If static nonlinear analysis is employed, this can be done by considering larger values of DIF (Figure 11b). By considering the two loading events, i.e. blast and fire after blast, it is possible to optimize the performance criteria of the structure (Dinu et al., 2010). Demonceau and Jaspart (2008) analyzed the behaviour of steel and composite building frames further to a column loss, when significant membrane effects developed within the structure, by analytical and experimental methods. The main objective of the research was the evaluation of the influential parameters and the validation of simple analytical procedures for predicting the response of a frame due to column loss. The difficulty to simulate the actual behaviour of joints subjected to combined bending moments and axial forces by simple static analysis has been pointed out, suggesting the application of complex FEM model for accurate prediction of the structural response of the system. The above study is combined with experimental investigations by Kuhlmann and Rolle (2008), which highlighted that the ductility demand to beam-to-column connections due to progressive collapse assessment of steel and composite frames could be faced by designing partial strength joints with sufficient ductility (Figure 12a), but the existing analytical criteria for predicting strength and ductility of joints should be extended in order to account correctly for the effects due to 172

Figure 12. Simulation of column loss in a composite frame (a) and experimental behaviour of a bolted end-plate partial strength connection (b).

large deformations, as for example in case of progressive collapse assessment of framed steel or composite structures. To avoid progressive collapse initiated by local damage a redistribution of force from the damaged part of the structure has to be enabled by alternate load path. Activation of alternate load paths by change of the bearing mechanism from pure bending state to more or less pure membrane state is a measure but only possible by allowing large global deformations resulting in high deformation requirements for the joints. Therefore highly ductile and partial-strength joint solutions were investigated in order to achieve a development of the plastic hinges in the beams. The effect on the global behavior was investigated analyzing the collapse resistance of the whole structure (Figure 12a). First results showed that composite beam-column structure is able to resist the event of a column loss under the accidental load combination for about 70–80% utilization of ULS loading. The identified requirements for the partial-strength joints concerning ductility and M-N-resistance are also feasible and within the range of the available rotation capacity and strength determined by the experimental investigation (Rolle and Kuhlmann, 2010). Contrarily, the studies conducted by Byfield and Paramasivam (2008) emphasized that many steel beam-to-column joints (simple and semi-rigid) may have insufficient ductility to successfully bridge damaged columns to avoid progressive collapse of the structure. In particular, the result of the prying action was shown to cause early joint fracture in simple connections, while semi-rigid connections due to low reserve of strength do not allow large redistribution of the accidental load. De Matteis et al., (2008) addressed the structural model of submerged floating tunnels under explosion. Detonation of a high explosive and consequent effect on the structure was analysed. The aim of the study was to evaluate the structural robustness of the prototype of the submerged tunnel proposed within the Sino-Italian joint laboratory of Archimedes Bridge (SIJLAB), when it is subjected to unexpected events like internal explosions or impact with external boats. The scope of the analysis was to carry out the stress due to the accidental detonation explosion (for instance, due to the terrorist attacks). Firstly, the equivalent static action due to the detonation explosion is determined. Secondly, the stress due to the explosion is evaluated by using detailed FEM model (Figure 13b). Finally, the most critical parts of the structures are determined in order to allow the evaluation of the structural robustness. The activity is still in progress since the applied numerical models have to be improved to account for other important phenomena, such as the actual interaction between the behaviour of the tube and the dynamic effect of the blast. Kilic and Altay (2007) studied the impact of narrow-body commercial airliners into structures with spirally-confined reinforced concrete columns. The numerical modelling was carried out using LS-Dyna and included the impact of fuel mass on spirally-confined reinforced concrete columns (Figure 13a). The fuel mass was found to be capable of destroying the columns at high velocity. Ongoing activities include (1) an investigation into the behaviour of reinforced concrete and steel structures under extreme loads such as blast and explosion; 2) development of advanced finite element techniques to simulate the response of structures to blast loadings; 3) improvement of blast-resistant design of structures through the use of blast shields and structural detailing for 173

Figure 13.

Numerical simulations: a) aircraft impact and b) analysis of submerged tunnel.

Figure 14.

Deformed guardrail after impact: a) testing campaign and b) numerical simulation.

critical members and 4) prevention of failure of reinforced concrete slabs under reverse curvature caused by blast loading/explosions. Gresnigt’s impact related activities concerned the impact of pressurised steel pipelines. Static tests were carried out simulating denting and scratching due to rough excavator teeth. Analytical models for load deformation have been developed, backed up with the development of accurate finite element analyses [Gresnigt, 2007]. Seiler (2008, 2010) studied the effects from impacts on guardrails from vehicles. Work involved tests for impact of vehicle (small size, bus and truck) on guardrails. Accurate simulations were carried out considering a simplified model, which incorporates the most influential parameters such as the dynamic interaction between vehicle and the road retaining structure (Figure 14). In fact, for economical reasons plastic behavior and geometrical nonlinearities of the construction should be considered for achieving correct results. As computation of the highly nonlinear behavior could lead to divergent results special emphasis should be focused on numerical stability by introducing numerical damping and appropriate modeling. The scope of the work was the substitution of testing campaign by refined numerical simulations. Also, a comparison between impact forces due to German standard and due to numerical simulation was drawn. The simulations carried out show that impact forces provided by German standard are lying on the safe side. 174

Figure 15. VTT compressed air-driven impact apparatus.

Some specific experimental testing campaigns have been also carried out. Tyas (2008) developed accurate studies to describe the purposes, possibilities and limitations of small-scale blast testing of structural components. Issues relating to the scaling of results to larger scenarios, and the use of experimental work to validate numerical modelling were addressed. Also, the features of commonly used instrumentation devices were highlighted, in addition to a presentation of state-of-the-art micro-scale fibre-optic blast measurement devices. In order to investigate the impact of aircrafts into the reinforced concrete shells of nuclear power plants Lastunen et al. (2008) developed a specific test apparatus at the VTT of Finland (Figure 15). Projectiles are accelerated to high velocities using an acceleration tube and pressure accumulator and the results are used for designing new facilities against aircraft impact. The investigation also includes FE modeling validated using the results from the experimental testing. Based on such results a numerical model was validated to design a new Finnish nuclear power plant to resist aircraft impact. REFERENCES Agarwal J., Blockley D., Woodman N., 2003. Vulnerability of structural systems, Structural Safety, No. 25, pp: 263–286. Allen, D. E., Schriever, W. R., 1972. Progressive Collapse, abnormal loads and building codes, Division of Building Research Council, Québec. Asprone D., Cadoni E., Prota A., 2009. Tensile High Strain-Rate Behavior of Reinforcing Steel from an Existing Bridge, ACI Structural Journal, Vol. 106, n. 4, pp. 523–529. Asprone D., Cadoni E., Prota A., Manfredi G., 2009. Dynamic behavior of a Mediterranean natural stone under tensile loading, International Journal of Rock Mechanics and Mining, Vol. 46, pp. 514–520. Asprone D., Cadoni E., Prota A., Manfredi G., 2009. Strain-rate sensitiveness of a pultruded E-glass/polyester composite, ASCE Journal of Composites for Construction,Vol. 13, 6, 558–564. Asprone, D. Jalayer, F., Prota, A. and Manfredi, G., 2010. Proposal of a probabilistic model for multi-hazard risk assessment of structures in seismic zones subjected to blast for the limit state of collapse Structural Safety, Volume 32, Issue 1, pp. 25–34. Asprone, D., Cadoni, E. and Prota, A., 2009. Experimental Analysis on Tensile Dynamic Behavior of Existing Concrete under High Strain Rates, ACI Structural Journal, Vol. 106, n. 1, pp.106–113. Bangash, M. Y. H. 1993. Impact and explosion-Analysis and design. Blackwell Scientific Publication, Oxford. Banthia, N., Mindess, S., Bentur, A. and Pigeon, M. 1989. Impact testing of concrete using a Drop-weight Impact Machine, Experimental Mechanics, Vol. 29, pp. 63–69. Bennett, R.M., 1988. Formulations for probability of progressive collapse, Structural Safety, No. 5, pp: 67:77, 1988. Birch, R.S., Jones, N., Jouri, W.S., 1988. Performance assessment of an impact rig, Proceedings of the Institute of Mechanical Engineers, Vol. 2, C4, pp 275–285. Byfield M. P., Paramasivam, S., 2008. The prevention of disproportionate collapse using catenary action. In Urban Habitat Construction Under Catastrophic Events (Eds. F.M. Mazzolani, E. Mistakidis, R.P. Borg, M. Byfield, G. De Matteis, D. Dubina, M. Indirli, A. Mandara, J.P. Muzeau, F. Wald, Y. Wang), Malta University Publishing, Malta, ISBN 978-99909-44-42-6, 336–340. Cadoni, E. 2010. Dynamic Characterization of an Orthogneiss Rock Subjected to Intermediate and High Strain Rate in Tension, Rock Mechanics and Rock Engineering, DOI 10.1007/s00603-010-0101-x. Cadoni, E., Albertini, C., Solomos, G. 2006. Analysis of the concrete behavior in tension at high strain-rate by a modified Hopkinson bar in support of impact resistant structural design, Journal de Physique, Vol. 3, pp. 647–652. Cadoni, E., Labibes, K., Berra, M., Giangrasso, M. and Albertini, C., 2000. High strain-rate tensile behaviour of concrete, Magazine of Concrete Research, Vol. 52, No.5, pp. 365–370.

175

CEB, 1988. Comité Euro-International du Béton Concrete structures under impact and impulsive loading. CEB Bulletin 187, Lausanne, Switzerland. Cho, SH, Ogata, Y and Kaneko, K, 2003. Strain-rate dependency of the dynamic tensile strength of rock, International Journal of Rock Mechanics & Mining Sciences 40, pp. 763–777. Corotis R.B., Nafday A.M. 1990. Application of mathematical programming to system reliability, Structural Safety, No. 7, pp: 149–154. Couque, H., Leung, C.P., Hudak Jr., S.J., 1994. Effect of planar size and dynamic loading rate on initiation and propagation toughness of a moderate-toughness steel Engineering Fracture Mechanics, Vol. 47, pp. 249–267. De Matteis, G., Eboli, A., Mazzolani, F.M., 2008. Blast effects on the Archimede Bridge. In Urban Habitat Construction Under Catastrophic Events (eds. F.M. Mazzolani, E. Mistakidis, R.P. Borg, M. Byfield, G. De Matteis, D. Dubina, M. Indirli, A. Mandara, J.P. Muzeau, F. Wald, Y. Wang), Malta University Publishing, Malta, ISBN 978-99909-44-42-6, 339–344. De Matteis, G., Langone, I., Mazzolani, F.M., Rebecchi, V., 2006. Effect of Gas Explosion on RC Buildings with Structural Key Elements. In Proc. of IABSE Annual Meeting and Symposium on Responding to Tomorrow’s Challenges in Structural Engineering, Budapest (Hungary), September 13-15, IABSE Report Volume 92, Paper No. A-0369 (CD-ROM). Demonceau, J.F., Jaspart, J.P., 2008. Experimental and analytical investigations on the behaviour of building frames further to a column loss. In Urban Habitat Construction Under Catastrophic Events (Eds. F.M. Mazzolani, E. Mistakidis, R.P. Borg, M. Byfield, G. De Matteis, D. Dubina, M. Indirli, A. Mandara, J.P. Muzeau, F. Wald, Y. Wang), Malta University Publishing, Malta, ISBN 978-99909-44-42-6, 300–305. Dinu, F., Dubina, D., 2008. Response of high rise steel buildings as a result of column loss. In Urban Habitat Construction Under Catastrophic Events (Eds. F.M. Mazzolani, E. Mistakidis, R.P. Borg, M. Byfield, G. De Matteis, D. Dubina, M. Indirli, A. Mandara, J.P. Muzeau, F. Wald, Y. Wang), Malta University Publishing, Malta, ISBN 978-99909-44-42-6, 277–282. Dinu, F., Dubina, D., De Matteis, G., 2010. Direct design approach for seismic resistant steel frame buildings under extreme loading. Proc. of. COST C26 Final Conference “Urban Habitat Construction Under Catastrophic Events”, Naples, 16–18 September 2010. Drar, H., 1993. On predicting the temperature and strain rate dependences of the fracture toughness of plain carbon steels Materials Characterization, Vol. 31, pp. 91–97. El-Habak, A. M. A. 1993. Compressive resistance of unidirectional GFRP under high rate of loading, Journal of Composites Technology and Research, Vol. 15, No. 4, pp. 311–317. Ellingwood, B. R. and Leyendecker, E. V., 1978. Approaches for Design against Progressive Collapse, Journal of Structural Division, Vol. 104, No. 3, pp. 413–423. Ellingwood, B. R., 2006. Mitigating Risk from Abnormal Loads and Progressive Collapse, Journal of Performance of Constructed Facilities, Vol. 20, No 4, pp. 315–323. Filiatrault, A, Holleran, M., 2001. Stress-strain behavior of reinforcing steel and concrete under seismic strain rates and low temperatures, Materials and Structures, Vol. 34, pp 235–239. Fu, H.C., Erki, M.A. and Seckin, M. 1991. Review of effects of loading rate on concrete in compression, Journal of Structural Engineering Vol 117, pp. 3645–3659. Gary, G, Zhao, H. 2000. Dynamic testing of fibre polymer matrix composites plates under in-plane compression, Composites Part A, Vol. 31, pp. 835–840. Gresnight, A.M. 2007. Impact of pressurised steel pipelines. Proceedings of COST Action C26 Workshop “Urban Habitat Constructions under Catastrophic Events”, Prague, March 30–31, 347–357. GSA 2003. General Services Administration, Progressive collapse analysis and design guidelines for new federal office buildings and major modernization projects, GSA, Washington, D.C., 2003. Harding, J. and Welsh, L.M., 1983. A tensile testing technique for fiber reinforced composites at impact rates of strain, Journal of Materials Science, Vol. 18, pp. 1810–1826. Harding, J., 1993. Effect of strain rate and specimen geometry on the compressive strength of woven glassreinforced epoxy laminates, Composites, Vol. 24, No. 4, pp. 323–332. Harris, D.W., Mohorovic, C.E., Dolen, T.P. 2000. Dynamic Properties of Mass Concrete Obtained from DamCores ACI Materials Journal, Vol.97 pp 290–296. Hayes JR, Woodson SC, Pekelnicky RG, Poland CD, Corley WG and Sozen M, 2005, Can Strengthening for Earthquake Improve Blast and Progressive Collapse Resistance? Journal of Structural Engineering, Vol. 131, No. 8. Henrych, J 1979. The Dynamics of Explosion and its Use, Elsevier. Hopkinson, B., 1914. A Method of Measuring the Pressure Produced in the Detonation of High Explosives or by the Impact of Bullets, Philos. Trans. R. Soc. (London) A, 213, pp. 437–456. Hsiao, H. M. and Daniel, I. M., 1998. Strain rate behavior of composite materials, Composites Part B, Vol 29B, 521–533.

176

Huang, Z., Nie, X. and Xia, Y., 2004. Effect of strain rate and temperature on the dynamic tensile properties of GFRP, Journal of Materials Science, Vol. 39, pp. 3479–3482. Kilic, S and Altay, G., 2007. Aircraft Impact on Reinforced Concrete Structures, Proceedings of COST Action C26 Workshop “Urban Habitat Constructions under Catastrophic Events”, Prague March, 30–31, 302–306. Kilic, S.A., Smith, P.D., 2010a. Simulation of pressures behind rigid blast walls. Proc. of. COST C26 Final Conference “Urban Habitat Construction Under Catastrophic Events”, Naples, 16-18 September 2010. Kilic, S.A., Smith, P.D., 2010b. Behaviour of deformable blast walls for protective structural design. Proc. of. COST C26 Final Conference “Urban Habitat Construction Under Catastrophic Events”, Naples, 16–18 September 2010. Kinney, G. F. & Graham, K. J. 1985. Explosive shocks in air Berlin and New York, Springer-Verlag. Kolsky, H., 1949. An Investigation of the Mechanical Properties of Materials at Very High Rates of Loading,” Proc. Phys. Soc. (London) B, 62, pp. 676–700. Krauthammer, T. 1998. Blast mitigation technologies: developments and numerical considerations for behaviour assessment and design. Proc. Int. Conf. on Structures Under Shock and Impact, SUSI, Computational Mechanics Inc., 3–12. Kubota, S, Ogata, Y, Wada, Y, Simangunsong, G, Shimada, H and Matsui, K, 2008. Estimation of dynamic tensile strength of sandstone, International Journal of Rock Mechanics & Mining Sciences 45, pp. 397–406. Kuhlmann, U., Rölle, L., 2008. Ductile and partial-strength steel & composite joints as basis for redundant and robust frame structures. In Urban Habitat Construction Under Catastrophic Events (Eds. F.M. Mazzolani, E. Mistakidis, R.P. Borg, M. Byfield, G. De Matteis, D. Dubina, M. Indirli, A. Mandara, J.P. Muzeau, F. Wald, Y. Wang), Malta University Publishing, Malta, ISBN 978-99909-44-42-6, 288–293. Langone, I., De Matteis, G., Mazzolani, F.M., 2008. Progressive collapse of buildings with key elements subjected to gas explosion”, In Urban Habitat Construction Under Catastrophic Events (Eds. F.M. Mazzolani, E. Mistakidis, R.P. Borg, M. Byfield, G. De Matteis, D. Dubina, M. Indirli, A. Mandara, J.P. Muzeau, F. Wald, Y. Wang), Malta University Publishing, Malta, ISBN 978-99909-44-42-6, pp. 294–299. Lastunen, A., Kuutti, J., Kolari, K., 2008. Aircraft impact testing. In Urban Habitat Construction Under Catastrophic Events (Eds. F.M. Mazzolani, E. Mistakidis, R.P. Borg, M. Byfield, G. De Matteis, D. Dubina, M. Indirli, A. Mandara, J.P. Muzeau, F. Wald, Y. Wang), Malta University Publishing, Malta, ISBN 97899909-44-42-6, 267–270. Lee WS, Liu CY, 2006. The effect of temperature and strain-rate on the dynamic flow behaviour of different steels, Materials Science and Engineering, A 426, pp. 101–113. Lee, W.-S., Lin, C.-F., Chen, B.-T., 2005. Tensile properties and microstructural aspects of 304L stainless steel weldments as a function of strain rate and temperature, Proceedings of the Institution of Mechanical Engineers, Part C: Journal of Mechanical Engineering Science, Vol. 219, pp. 439–451. Li Q.M. & Meng H., 2002. Pressure-Impulse Diagram for Blast Loads Based on Dimensional Analysis and Single-Degree-of Freedom Model Journal of Engineering Mechanics, Vol. 128, No. 1, pp. 87–92. Li, H.B., Zhao, J. and Li, T.J., 1999. Triaxial compression tests on a granite at different strain rates and confining pressures, International Journal of Rock Mechanics and Mining Sciences 36, pp. 1057–1063. Lindholm, U.S. 1971. High strain-rate tests, in measurement of mechanical properties, Techniques of Metals Research Vol. 5, pp. 199–271. Lindholm, U.S., 1964. Some Experiments with the Split-Hopkinson Pressure Bar,” Journal of the Mechanics and Physics of Solids, 12, p. 317. Ma G.W., Shi H.J., Shu D.W., 2007. P-I diagram method for combined failure modes of rigid-plastic beams, International Journal of Impact Engineering, Volume 34, Issue 6, pp. 1081–1094. Ma, L. and Daemen, J.J.K. 2004. Strain rate dependence of mechanical properties of welded Topopah Spring tuff SINOROCK2004 Symposium, International Journal of Rock Mechanics and Mining Sciences Vol. 41, CD-ROM. Mainstone, R.J. 1975. Properties of materials at high rates of straining or loading, part 4, State-of-the-Art report on Impact loading of Structures, Materials and Structures, Vol. 8, No. 44, pp. 102–116. Majzoobi, G.H., Saniee, F.F. and Bahrami, M., 2005. A tensile impact apparatus for characterization of fibrous composites at high strain rates, Journal of Materials Processing Technology Vol. 162–163, pp. 76–82. Makarov, G., Wang, W. and Shenoi, R.A., 2004. Deformation and fracture of unidirectional GFRP composites at high strain rate tension, Proc. of 2nd International Conference on Composites Testing and Model Identification, Bristol, UK, Paper No. 52. Malvar L.J. and Ross C.A. 1998. Review of strain-rate effects for concrete in tension, ACI Materials Journal Vol. 95, pp. 735–739. Malvar, L.J., 1998. Review of static and dynamic properties of steel reinforcing bars, ACI Materials Journal, Vol. 95, pp. 609–616.

177

Marjanishvili, S and Agnew, E, 2006. Comparison of Various Procedures for Progressive Collapse Analysis Journal of Performance of Constructed Facilities, Vol. 20, No 4, pp. 349–364. Mays, G. C., & Smith, P. D. 1995. Blast effects on buildings-Design of buildings to optimize resistance to blast loading. Thomas Telford, London. Mazzolani, F.M., Byfield, M., De Matteis G., 2009. Recent Developments in Blast and Impact Research in Europe, Proc. of PROTECT 2009 International Workshop “Performance, protection & Strengthening of Structures under Extreme Loading” (Sonoda Y., Banthia N. And Fujikake eds), Hayama, Japan, 19-21 August, 2009, published by National Defence Academy Protect Workshop, Japan, ISBN 978-4-9904806-0-8 (CD-ROM). Newill, J. F., and Vinson, R., 1993. Some high strain rate effects on composite materials Proc. Of ICCM9, Vol. 5, pp 269. NRC 2001. National Research Council, Protecting people and buildings from terrorism, Committee for Oversight and Assessment of Blast-effects and Related Research, National Academy Press, Washington, D.C. Okoli, I., 2001. The effect of strain rate and failure modes on the failure energy of fiber reinforced composites, Composite Structures, Vol. 54, pp. 299–303. Ray, S.K., Sarkar, M. and Singh, T.N., 1999. Effect of cyclic loading and strain rate on the mechanical behaviour of sandstone, International Journal of Rock Mechanics and Mining Sciences, vol. 36, pp. 543–549. Roelle, L., Kuhlmann, U., 2010. Robust design – Alternate Load Path Method as design strategy, Proc. of COST C26 Final Conference “Urban Habitat Construction Under Catastrophic Events”, Naples, 16–18 September 2010. Ruth, P., Marchand, K. A. and Williamson, E. B. 2006. Static Equivalency in Progressive Collapse Alternate Path Analysis: Reducing Conservatism While Retaining Structural Integrity, Journal of Performance of Constructed Facilities, Vol. 20, No 4, pp. 349–364. SEI 2005. ASCE/Structural Engineering Institute, “Minimum design loads for buildings and other structures”, ASCE/SEI 7, Reston, Va. Seiler, C., 2010. Tests and simplified numerical simulation of vehicle impact on guardrails with respect to normative regulations. Proc. of. COST C26 Final Conference “Urban Habitat Construction Under Catastrophic Events”, Naples, 16–18 September 2010. Sierakowsky, R. L., 1997. Strain rate effects in composites Appl. Mech. Rev., Vol. 50, No. 11, 741–761. Smith P.D. & Hetherington J.G., 1994. Blast and ballistic loading of structures, Butterworth-Heinemann, Oxford. Soleiman Fallah A. & Louca L.A., 2007. Pressure-impulse diagrams for elastic-plastic-hardening and softening single-degree-of-freedom models subjected to blast loading, International Journal of Impact Engineering, Volume 34, Issue 4, pp. 823–842. Tay T. E., Ang H. G. and Shim V. P. W., 1995. An empirical strain rate-dependent constitutive relationship for glass–fibre reinforced epoxy and pure epoxy, Composite Structures, Vol. 33, pp. 201–210. TM5-1300 1990. Structures to resist the effects of accidental explosions Departments of the Army, the Navy and the Air Force – USA, TM 5-1300. Tyas, A., 2008. Small-scale blast load and structural response measurements. In Urban Habitat Construction Under Catastrophic Events (Eds. F.M. Mazzolani, E. Mistakidis, R.P. Borg, M. Byfield, G. De Matteis, D. Dubina, M. Indirli, A. Mandara, J.P. Muzeau, F. Wald, Y. Wang), Malta University Publishing, Malta, ISBN 978-99909-44-42-6, 333–338. Uenishi, A. and Teodosiu, A., 2004 “Constitutive modelling of the high strain-rate behaviour of interstitial-free steel”, International Journal of Plasticity, Vol. 20, pp. 915–936. van Doormal, J.C.A.M., Weerheijn, J. and Sluys, L.J. 1994 Experimental and numerical determination of the dynamic fracture energy of concrete Journal de physique Vol. 4, pp. 501–506. Wang, G.Z., Wang, Y.L., 2007. Effects of loading rate, notch geometry and loading mode on the local cleavage fracture stress of a C-Mn steel, International Journal of Fracture, Vol. 146, pp. 105–121. Welsh, L.M. and Harding, J., 1985. Effect of strain rate on the tensile failure of woven reinforced polyester resin composites, Journal de Physique, Vol. C5, 1985, pp. 405–414. Yi, WJ, He, QF, Xiao, Y, and Kunnath, SK 2008. Experimental Study on Progressive Collapse-Resistant Behavior of Reinforced Concrete Frame Structures, ACI Structural Journal, Volume 105, Issue 4, pp. 433–439.

178

Urban Habitat Constructions under Catastrophic Events (Final Report) – Mazzolani (Ed). © 2010 Taylor & Francis Group, London, ISBN 978-0-415-60686-8

2.4 Consequences of natural disasters on constructions Carlos Coelho Civil Engineering Department, University of Aveiro, Portugal

Ruben Paul Borg Faculty for the Built Environment, University of Malta, Malta

Vlatko Sesov University of Skopye, Macedonia

Maurizio Indirli ENEA, Bologna, Italy

2.4.1 INTRODUCTION Most of the highly populated cities all over the world, (but also a large amount of little towns and villages) are prone to natural hazards. Natural hazards are defined as processes, occurring in the biosphere that may constitute a damaging event. The main hazardous catastrophes are earthquakes, volcanic eruptions, landslides, tsunamis, coastal erosions, floods, hurricanes, drought, etc. With regards urban areas, both wild and man-induced fires can be also considered. Consequently, the disaster risk DR (probability of harmful consequences, expected loss of lives, people injured, property, livelihoods, economic activity disrupted, environment damaged, etc., see Figure 1) results from the combination between hazard H, vulnerability V (human condition or process resulting from physical, social, economic and environmental factors, determining probability and scale of damage from the impact of a given hazard) and physical exposure E (elements at risk, an inventory of those people or artifacts exposed to hazards), divided by the RM factor (Risk Management). The Risk Assessment (RA) is an estimate of the social and economic impact that hazards can have on people, buildings, services, facilities and infrastructures. It is worth noting that, in absolute terms (UNPD 2002, Munich Re Group, 2004) the economic cost of disasters has been increasing over the decades (Figure 2). In addition, the urban building aggregates enshrine notable cores (like urban and social tissues, historical and architectonical constructions, precious monuments, museums and archaeological evidences) of invaluable significance; this kind of patrimony, which must be handed down intact to posterity as far as possible, is often protected by international and national cultural heritage boards. Nevertheless, a huge amount of such treasures was lost for ever, due to past natural catastrophes; some important examples can be highlighted: the 79 A.D. Vesuvius eruption, Italy (when Pompeii, Ercolano and Stabiae were completely covered by pyroclastic flows); the disruption of San Francisco (California, USA) and Valparaiso (Chile) during the 1906 earthquakes; the 1963 Vajont landslide, Italy (which swept away small towns); the Firenze’s flood (1966); the Lisbon great fire (1988); the Indian Ocean tsunami (2004); the hurricane Katrina (2005). Therefore, the accomplishment of effective pre- and post-disaster risk management is crucial, in order to minimize disaster impacts and implement potent policies and coping capacities of society or individuals. This is achieved by managing the multifaceted nature of risk, realizing integrated hazard models and adopting appropriate governance for development and reconstruction planning. In this regard, different strategy levels have to be foreseen; during the emergency phase, it is necessary to understand, well and quickly, the dynamic development of each environmental process, provide a detailed damage assessment and address prompt civil defense interventions; furthermore, prevention policies are also mandatory: hazard mapping, vulnerability studies, building inventory, mitigation programmes and citizenship preparedness. 179

Figure 1.

Risk definition.

Figure 2. World natural catastrophes (increasing) and economic losses (decreasing) from 1950 to 2005 (source Munich Re).

In the last decade, Geomatics (emerging technology playing a vital role in natural disasters mitigation) has been developing; it is a conglomerate of measuring, mapping, geodesy, satellite positioning (GPS), photogrammetry, computer systems and computer graphics, remote sensing RS (Shinozuka, 2005), geographic information systems GIS (Valpreda, 2003) and environmental visualization. The earth observation satellites provide comprehensive, synoptic and multi-temporal coverage and monitoring of large areas for a wide range of scales, from entire continents to details of a few meters in real time and at frequent intervals. This approach, which started primarily with earthquake applications, has broadened rapidly to tsunami, hurricanes, storms, wildfires, landslides and other matters (Indirli, 2007).

2.4.2 A BRIEF SUMMARY ON RISK ASSESSMENT Nowadays, a proper risk assessment must include a multidisciplinary approach, through the implementation of interactive digitized databases, collection of a huge amount of input data and the organization of a user-friendly instrument with strong import-export capabilities. The integrated use of several tools (hazard models, building classification and inventory database resources, RS and GIS mapping, etc.), the identification of analyses procedures and algorithms, the elaboration of reliable outputs have to be foreseen. The basic steps for a correct risk assessment (Table 1) are 180

Table 1. Definition of the main risk assessment procedures. Process

Outputs

Step 1 • definition of the study region • creation of a region base map • hazard of interest identification

Identify hazards • study region • base map • hazard of interest list

Step 2 • hazards database construction • performing a data gap analysis • profile and priority of hazards

Profile hazards • updated and completed hazard profiles • map of hazard areas • hazards prioritized list

Step 3 • inventory database construction • performing a data gap analysis

Inventory assets • inventory data tables and maps • inventory data • data sources list

Step 4 • construction of estimate losses scenarios and risk assessment tools • evaluation of the results

Estimate losses and risk assessment tools • loss and exposure for the study region

Step 5 • mitigation options identification • mitigation options verification

Consider mitigation options • mitigation options list

• risk assessment outputs • tables • maps

Table 2. List of possible natural hazards. • • • • • • • • • • • • • •

Natural hazard type

earthquake tsunami landslide and mudslide subsidence hurricane tornado flood coastal storm and erosion volcanic eruption drought wildfire winter storm (ice and snow) avalanche ...

discussed in the following paragraphs (FEMA 386-2, 2001). Moreover, the database should be flexible, freely available for use by any country and organization through Internet access, open-source, capable to be multi-hazard and international in scope, encouraging the worldwide community to participate to its development and validation (Indirli, 2007). 2.4.2.1 Hazard identification The first step regards the identification of natural hazards that might affect the community or the territory (Table 2). Accidental actions such as fires, internal explosions and seismic events are, more or less, covered by the Eurocode program. Avalanches, erosion, extreme snow, extreme winds, floods, landslides, rockfalls, tsunamis, volcanic eruptions, etc., represent infrequent natural actions 181

Table 3. Natural hazard profile. • • • • • • •

Hazard event Frequency Probability Duration Magnitude Intensity Hazard areas

specific occurrence how often likelihood (statistical measure) how long an event lasts severity (technical measure) effect of an event at a particular place geographic areas within study region

that are not covered by the Eurocode program. It has to be noted that most of these phenomena are characterized by large fluid masses moving with a different degree of velocity according to their density and viscosity. Volcanic eruptions or tsunamis do not depend on climatic phenomena. However most of the other disasters have natural origins and due to climate change, they seem to become more frequent, and not as infrequent as in the past. The main problem is that, by definition, exceptional and infrequent events are associated with a very low probability of occurrence. Therefore, databases concerning these events are rather limited. One of the goals of the COST Action C26 (COST, 2006) is to define a suitable methodology to predict the structural behaviour of constructions under extreme conditions. One of the first steps is the identification and the classification of the relevant infrequent events and, when possible, the description of the associated scenarios. It is important to avoid omissions, in considering the full range of potential hazards (including new and unexpected ones), and in assessing whether they may affect the area considered. Several approaches can be considered in the investigation: historical information, newspapers and Internet websites, modern event experience, technical information and experts’ opinion, review of existing plans and reports. The results can be summarized in models and maps. After the preliminary research described above, it is indispensable to focus on the most prevalent hazards in the community or territory, through an accurate consultation of specific hazard websites (if any). Hazard grouping is a crucial point: a list may give the impression that hazards are independent of one another; on the contrary, they are often related (primary and secondary hazards). The identification of global hazard factors for a given area (or a building), is another crucial step to be carried out, due to the difficult definition of combination methods and algorithms (Indirli, 2007). 2.4.2.2 Hazard profile In order to prepare the profile of hazards and their potential consequences, various aspects are of particular interest; the frequency of occurrence, magnitude and intensity, location and/or spatial extent, duration, seasonal pattern, speed of onset, and availability of warning (Table 3, Indirli, 2007). Some hazards (such as floods, coastal storms, wildfires, tsunamis, and landslides) occur in predictable areas and can be easily mapped. Other hazards (such as tornadoes, which can occur anywhere) may be profiled simply by recording the maximum potential wind speed. This type of information will be used to evaluate the potential impact on individual structures or elements in the selected area. Several kinds of consequences can be investigated: effects on people, critical facilities and community functions, property, and sites of potential secondary hazards. The use of both historical information and modeling is recommended. The aim is to determine which hazards merit special attention, and therefore great attention must be given to compare and prioritize risks, create and apply scenarios, investigate data sources. The definition of complete, accurate and detailed base maps of different scales (from global to local) is fundamental (also creating GIS platforms and elaborating satellite imagery), where objects (like buildings, roads, rivers, coastlines, etc.) must be well distinguishable. Hazard level areas (Low, Moderate, High, Extreme) have to be clearly identified. Focusing on seismic hazard maps, nowadays several studies can provide more accurate tools for the urban environment protection, like neo-deterministic models and earthquake scenarios (to be preferred especially when dealing with objects that should be 100% safe, as strategic facilities and cultural heritage). In fact, Seismic hazard assessment, necessary to design earthquake-resistant 182

structures, can be performed in various ways, following a probabilistic or a neo-deterministic approach. Case studies indicate the limitations of the PSHA (Probabilistic Seismic HazardAnalysis) currently used methodologies, deeply rooted in engineering practice, providing indications that can be useful but not sufficiently reliable (Decanini et al., 2001, Klügel et al., 2006), as shown in recent examples (earthquakes of: Michoacan 1985, Kobe 1995, Bhuj 2001, Boumerdes 2003, Bam 2003, E-Sichuan 2008, L’Aquila 2009 events). In fact, the PSHA could not be sufficiently reliable to completely characterize the seismic hazard, because of the difficulty to define the seismogenetic zones and evaluate correctly the occurrence of the earthquakes (frequency-magnitude relations), and the propagation of their effects (attenuation laws). A more adequate description of the seismic input can be done following a neo-deterministic approach (Neo-Deterministic Seismic Hazard Assessment NDSHA), which allows for a realistic description of the seismic ground motion due to an earthquake of given distance and magnitude (Panza et al., 2001). The approach, that is feasible to apply at urban scales, is based on modelling techniques that have been developed from the knowledge of the seismic source generation and propagation processes. It is very useful because it permits the definition of a set of earthquake scenarios and the computation of the associated synthetic signals, without having to wait for a strong event to occur. A complete description of the neo-deterministic methodology, from the definition of the hazard to the seismic input calculation for the design of a building, is given in Zuccolo et al. (2008). This methodology provides algorithms for the space-time medium terms forecasting and the realistic simulation of ground shaking, including seismic input synthetic time series, calibrated against observations, whenever possible; thus, the use of methods that showed many shortcomings (based on acceleration peak values and probabilistic earthquake return periods) can be overcame. An important example is the UNESCO-IUGS-IGCP project 414 (Panza et al., 2002) for the seismic microzonation of several towns (e.g. Delhi, Beijing, Rome, Naples, Santiago de Cuba, Bucharest, Sofia). Furthermore, this approach has been extended to Chile (Indirli et al., 2006; MAR VASTO, 2007). A similar approach is advisable to evaluate the occurrence of tsunamis, together with the development of reliable analytical models of sea waves propagation and accurate recording systems (Panza et al., 2000). With regards to flood, concepts as flood elevation and flood hazard areas must be very well defined; in the USA, BFE (Base Flood Elevation), is the elevation of the water surface resulting from a flood that has a 1% chance of occurring in any given year; SFHA (Special Flood HazardArea) is the shaded area that identifies an area that has a 1% chance of being flooded in any given year. Usually, tornadoes strike random and in a wide portion of territory; they are classified by the Fujita Measurement Scale (from category F0 for light tornadoes to F5 for incredible tornadoes); to profile the hazard, it is necessary to choose the design wind speed with accuracy, (that is provided in USA by ASCE, the American Society of Civil Engineers). Coastal storms, due to typhoons or hurricanes, can cause tidal elevation increase (called storm surge), inland flooding and water force, wind speed, and erosion. Storm surge water levels depend on wind speed and are measured by the five categories (from minimal to catastrophic) of the Saffir-Simpson Scale. The best predictor of future landslides is past landslides, because they tend to occur in the same places. This hazard is very complex and requires geologic expertise. Landslide inventories identify areas that appear to have failed in the past; landslide susceptibility maps depict areas that have the potential for landslides; landslides hazard maps show the real extent of the threat: where landslides have occurred in the past, where they are likely to occur now, and where they can occur in the future.

2.4.2.3 Hazard independent effects 2.4.2.3.1 The effect of snow Exceptional snow loads For altitudes smaller than 1500 m, exceptional snow loads are specified in EN1991-1-3:2003, and the code is based on the assumption of a return period equal to 50 years. Exceptional snow loads are considered as accidental loads (EN, 1991a–b). Avalanches Avalanches are one of the infrequent actions not taken into account in the Eurocodes. It is possible to identify two main types of avalanches (Givry and Perfettini, 2004) depending on the state of 183

the snow: dry snow or powder avalanche and wet snow avalanche. The difference between these two kinds of phenomena defines their mode of failure, their way of displacement down the slope and their relative power. Nevertheless, it is to be noted that there is a transition between these two extreme modes that are qualified as dry, damp, moist, wet and saturated. Compared to a dry snow avalanche, a wet snow one is slower and the runout distance is shorter. However, the impact on obstacles (trees or constructions for instance) is very important due to the high density of wet snow. Being slower, wet snow avalanches appear to be less dangerous for humans than dry avalanches, but regarding construction it is the opposite. Nevertheless, if the robustness of the construction can be strongly affected by wet snow avalanches, the openings are affected by dry snow avalanches due to its related high pressure, similar to a blasting effect. A dry snow avalanche is not really affected by the site topography as the wet snow avalanche which reduces with the slope. The avalanche path corresponds to a terrain feature where an avalanche occurs. It is composed of a starting zone, a track and a deposition or runout zone. 2.4.2.3.2 The effect of wind Extreme winds Cyclones, hurricanes, tornados or typhoons are extreme winds whose dynamic action leads generally to severe damages on constructions. The name of the extreme winds depends on their geographical location and on their maximum speed. To be initiated, tropical cyclones need certain thermodynamic conditions to be respected above a large mass of warm water. Therefore, they form above seas or oceans. They are named hurricanes in the Atlantic Ocean and typhoons in the Pacific Ocean. Tornados are initiated above the earth during a severe storm when special thermodynamic conditions are found between huge cloudy masses and winds. Cyclones Three types of cyclonic perturbations are commonly defined: tropical depressions, tropical storms and tropical cyclones (Chaboud, 2003). A tropical cyclone is constituted by an eye at its centre, which is a relatively warm and calm zone, surrounded by an area about 16–80 km wide in which the strongest thunderstorms and winds circulate around it. Up to now, the extreme wind speed due to a tropical cyclone is estimated to be equal to 305 km/h. To be initiated and sustained, tropical cyclones need large unstable volumes of warm water (more than 26◦ C over 60 m in depth) so, their strength decreases over land because of the lack of water. This is why the coastal regions are much more affected by cyclones than inland regions, since the wind speed decreases as much as the depression progress on earth. Tornados Much smaller than a tropical cyclone regarding its influence diameter, a tornado is a violently rotating column of air starting from the lower part of a cumulonimbus cloud and in contact with the earth. Presenting different shapes, the form of a tornado is typically a visible condensation funnel, whose narrow part moves on the earth. On the path of the displacement, the damages on constructions are generally localised but very important due to the high speed of the rotation. In most cases, a cloud of debris collected on the way, moves around the funnel at its lower part and it contributes to increase in damage. If the radial action of the vortex is important, it is generally combined with a vertical suction opposite to the gravity which amplifies the damages. A powerful tornado may extract light constructions from their foundations. Most of the tornadoes create a very localised strong wind whose speed may reach 175 km/h. Their lower part is generally about 100 m and they travel only on a small distance (about 10 km) before they dissipate. Nevertheless, much more powerful tornadoes have been observed: with a wind speed close to 500 km/h, with a base diameter close to 1.5 km and whose path on the ground may be longer than 100 kilometres. 2.4.2.3.3 The effect of water River floods River floods are one of the main hazards encountered by people living in the whole European Union. Since these floods can take numerous forms, such as flash floods, estuarine floods, mud floods etc, almost all kinds of landscapes can be impacted. In the past decades, an increase in terms of frequency and of importance of such catastrophic events has been noticed, and can be related 184

to numerous causes: modification of land occupation (increase of impervious areas, changes in agricultural practices), erratic river banks management (artificialized hydrosystems) and the effects of climate change on storm events frequency. Coastal floods Highly energetic wave regimes, some negative effects from coastal interventions, littoral occupation, exterior interventions in harbors, the weakening of river sediment supply and the generalized sea level rise and other effects of climate change (frequency of storm events, rotation of waves provenience) can be pointed out as the main causes of the increasing number of constructions exposed to waves. Overtopping and flooding are being more frequent events on the coastal zones, jeopardizing buildings and infrastructures. Erosion All European coastal states are to some extent affected by coastal erosion. About 20 000 km of coasts, corresponding to 20%, faced serious impacts in 2004. Most of the impact zones (15 100 km) are actively retreating; some of them in spite of the coastal protection works done (2 900 km). In addition, another 4 700 km have become artificially stabilized (European Commission, 2004). The dynamical variability of sandy beaches, where the alongshore sediment transport is controlled by waves, currents, wind, water level, sediments sources and sinks and sediments properties, can represent erosion situations with exposure of constructions. The main causes for coastal erosion are the generalized sea level rise, caused by climate change, some negative effects from coastal interventions, littoral occupation, exterior interventions in harbours, which cause serious perturbations in the littoral drift system and river sediment supply reduction (Coelho et al., 2006). The river sediment supply has been weakening due to sand mining, for construction and navigation, and dam construction with consequent sand retention and hydrological regime regularization (Santos et al., 2002). One of the usually preferred solutions to solve erosion problems is beach artificial nourishment. This can represent a very expensive solution, when the sedimentary deficit is very high, and there is no large sand deposit availability. However, it is essential to try to mitigate coastal erosion processes in specific locations. At the moment, the so-called hard coastal defences are indispensable to protect some of the existing settlements, but an adequate plan of monitoring of the existent coastal defence structures, should be foreseen, keeping the maintenance costs of the structures at a low level. Regarding the political point of view, it is crucial to regulate urban seafront extension. In some cases, the policy options of managed realignment – identify a new line of defence and re-settle the populations in the hinterland – have to be considered. The solution for coastal erosion problems must be a compromise between the passive acceptance of erosion, some beach nourishment and coastal intervention for urban front protection. Tsunamis Tsunami are series of waves created by the fast displacement of huge volumes of water (an ocean for instance) strongly and rapidly affected by a natural phenomenon at a huge scale. Generally, they can be initiated by earthquakes, submarine volcanic eruptions or landslides (seabed slides) for example, but not by strong winds whose impulsion is not short and strong enough. The effects of a tsunami can be classified as insignificant to catastrophic, affecting coastal population and constructions. The waves move as growing circles from the initial location to all the surrounding coats at high speeds (700 km/h is a mean value in the Pacific Ocean) with a large wavelength (hundreds of kilometres) and they can travel great transoceanic distances with small overall energy loss. Far from coasts, most classical tsunamis have wave heights smaller than one meter but in their motion, they mobilise the whole water column from the surface to the sea bed. That explains why, when it approach the coast where the sea bottom become less deeper, the wave front becomes higher and can reach 20 m or more, and the residual energy creates a violent displacement. Most of the damages are due to the enormous mass of water accompanying the initial wave front. On the one hand, they are originated by the wave impact on obstacles and, on the other hand, by the flood resulting from the sea level rising. The energy of the phenomenon is sufficient to project any kind of object found on its path (ships for instance but also any kind of debris) and sometimes far from the coast. Their combination is powerful enough to shear weak brittle houses at their base or to submerge and to create bending actions on rather high constructions depending of the wave height. 185

The influence area depends also on the relief but it can be measured several kilometres far from the coast itself whose position can be strongly affected. Other details can be found in Rossetto et al. (2010a). 2.4.2.3.4 The effect of volcanoes Volcanic eruptions This topic is largely discussed in another part of the Final Report (Section 2.5). 2.4.2.3.5 The effect of landslides and rockfall Landslide describes a wide variety of processes that result in the downward and outward movement of slope-forming materials including rock, soil, artificial fill, or a combination of these. The materials may move by falling, toppling, sliding, spreading, or flowing. The various types of landslides can be differentiated by the kinds of material involved and the mode of movement. Other classification systems incorporate additional variables, such as the rate of movement and the water, air, or ice content of the landslide material. Although landslides are primarily associated with mountainous regions, they can also occur in areas of generally low relief. In low-relief areas, landslides occur as cut-and fill failures (roadway and building excavations), river bluff failures, lateral spreading landslides, collapse of mine-waste piles (especially coal), and a wide variety of slope failures associated with quarries and open-pit mines. The conventional stability analysis of slopes where sliding is possible along some definable surface is usually preformed by calculating the factor of safety, i.e. by comparing the shearing resistance available along the failure surface; with the shearing stresses imposed on the failure surface. Most analytical methods are based on limit equilibrium, with typical failure forms such as infinite slope or finite slope with planar or curved failure surface considered. However, in the recently introduced performance-based approach, emphasis is placed not on whether the slope is stable or unstable, but on the magnitude of deformation after failure. Several techniques are currently available to asses the post-failure velocity and travel distance of the moving mass. The basic model assumes that during the shaking a slope will suffer displacement only when the ground acceleration exceeds a threshold value, the critical acceleration, which can be derived from the static factor of safety of the slope in question. The sliding mass will continue to move until the shaking drops below the critical acceleration. If the cumulative displacement caused by shaking, known as Newmark displacement is sufficient to cause a reduction in the shear strength of the soil or rock mass then a re-calculation of the slope stability is carried out using residual shear strength parameters to establish whether failure occurs. Thus the analysis is bi-linear, allowing for a change in the strength parameters of the slope forming materials based on the deformation of the slope. 2.4.2.4 Combination of hazards scenarios. In this section, some combinations of actions of exceptional events are presented, evaluating realistic scenarios. 2.4.2.4.1 Snow and avalanches Considering the combination of exceptional events, wet snow or dry snow avalanches are never combined together due to their different origin. An avalanche is never combined with an earthquake even if this last may be the creating factor because they do not occur in the same time interval. Winds and avalanches are never combined because the wind action being smaller, it can be expected to be included in the avalanche load case. Snows and avalanches are obviously combined because heavy snow is generally the creating factor. 2.4.2.4.2 Extreme winds and rains In most cases, extreme winds are associated to torrential rains creating floods which amplify the damages effects. The extreme winds due to tropical cyclones are often combined with torrential rains, high waves, and storm surges: strong wind creates a high pressure able to damage civil engineering structures themselves and, on the other hand, it transforms debris into flying objects able to damage covering and cladding; heavy rains can create river and stream floods but also landslides; storm surges, by increasing the sea level, can produce extensive coastal flooding up to about 50 km inland depending on the relief. 186

2.4.2.4.3 River and coastal floods Combination of heavy rainfall with severe marine conditions (storms and/or high tides) can highly increase the damaging potential of flood events. Usually, storm events, with important winds and rainfalls, low atmospheric pressure, high tides and significant wave heights represent natural conditions for floods combination in coastal and estuarine areas. In fact, fluvial and coastal floods are frequently related and present common characteristics. Coastal shores and river banks management is very important to reduce flood damaging costs and control the consequences of flood events. Land use is very important on soil erosion and thus, on river sediment transport. On the other hand, fluvial floods are usually associated with important sediment transport rates. These sediments represent the natural nourishment of the coastal areas, preventing coastal erosion. In estuarine regions, the extreme rainfall events can present more negative impacts when associated to high tide periods, mainly during spring tides. Storms and low atmospheric pressure can amplify the consequences. In spite of all this, the subject of combined flood effects of river and sea actions is still poorly referred in the literature. 2.4.2.4.4 Landslides and other actions Although there are many actions that cause landslides, there are three that cause most of the damaging landslides around the world. These are described as follows, with more than one action in combination. Landslides and water Slope saturation by water is a primary cause of landslides. This effect can occur in the form of intense rainfall, snowmelt, changes in ground-water levels, and water level changes along coastlines, earth dams, and the banks of lakes, reservoirs, canals, and rivers. Landsliding and flooding are closely allied because both are related to precipitation, runoff, and the saturation of ground by water. In addition, debris flows and mudflows usually occur in small, steep stream channels and often are mistaken for floods; in fact, these two events often occur simultaneously in the same area. Landslides can cause flooding by forming landslide dams that block valleys and stream channels, allowing large amounts of water to back up. This causes backwater flooding and, if the dam fails, subsequent downstream flooding. Also, solid landslide debris can “bulk” or add volume and density to otherwise normal streamflow or cause channel blockages and diversions creating flood conditions or localized erosion. Landslides can also cause overtopping of reservoirs and/or reduced capacity of reservoirs to store water. Landslides and seismic activity This topic is also discussed in other parts of the Final Report. Many mountainous areas that are vulnerable to landslides have also experienced at least moderate rates of earthquake occurrence in recorded times. The occurrence of earthquakes in steep landslideprone areas greatly increases the likelihood that landslides will occur, due to ground shaking alone or shaking – caused dilation of soil materials, which allows rapid infiltration of water. Widespread rockfalls also are caused by loosening of rocks as a result of ground shaking. Saturated soil condition due to rainy days and very steep inclination of the natural and manmade slopes makes soil very vulnerable to earthquake shaking. Landslides and volcanic activity This topic is also discussed in another part of the Final Report. Landslides due to volcanic activity are some of the most devastating types of landslides. Volcanic lava may melt snow at a rapid rate, causing a deluge of rock, soil, ash, and water that accelerates rapidly on the steep slopes of volcanoes, devastating anything in its path. These volcanic debris flows (also known as lahars) reach great distances, once they leave the flanks of the volcano, and can damage structures in flat areas surrounding the volcanoes. 2.4.2.5 Inventory assets Inventory assets organize a huge amount of data, about patterns that can be affected by hazardous events, which are better stored on a GIS platform (Table 4). An example of potentially vulnerable 187

Table 4. Inventory assets. Demographics Building stock Essential facilities Transportation systems Lifeline utility systems High potential loss facilities Hazardous material facilities Cultural heritage

Figure 3. Organization HAZUS-MH.

of

building

• population, employment, housing • residential, commercial, industrial • emergency operations centers, hospitals, schools, shelters, police and fire stations • airways, highways, railways, waterways • potable water, waste water, oil, gas, electric power, communication systems • dams and levees, nuclear facilities, military installations • facilities housing industrial/hazardous materials • historical centers, archaeological remains, monuments, museums

data

in Figure 4. Building HAZUS-MH.

classification

system

in

X X X X

X

X X X

X X X X X X X X

wildfire

X X

landslide

X X

coastal storm

X X

tornado

tsunami

Type/type of foundation Code design level/construction date Roof material Roof construction Vegetation Topography Distance from the hazard zone Lowest floor elevation Base floor elevation

earthquake

Building characteristics

flood

Table 5. Building data requirements by hazard.

X X

X X X X X X

assets is shown in Figures 3–4, taken from HAZUS-MH (HAZUS, 1992), briefly described in paragraph 4. HAZUS can summarize the number and value of structures in the area considered, according to the types of structure or the occupancy class. Critical buildings and facilities must be classified separately (Indirli, 2007). In order to gather building-specific information, the following data must be provided: building size; replacement value to its pre-damaged conditions; content value; function use or value; displacement cost due to hazard; occupancy or capacity. Other data, summarized in Table 5, is hazard-specific. 188

Figure 5.

Sana’a GIS database.

Figure 6. 3D Digital mapping joining hazard and vulnerability for earthquakes (source Midorikawa).

Table 6. Estimate losses. Building types

• concrete, pre-cast concrete, reinforced and unreinforced masonry, steel, wood, etc.

Building occupancies

• residential, commercial, industrial, education, agriculture, religion, government, etc.

Essential facilities

• Emergency Operations Centers, hospitals, schools, police and fire stations, shelters, etc.

Transportation/lifelines

• highways, railways, light rail, bus station, ports, ferries, airports, bridges, tunnels, etc.

Utilities

• waste water, potable water, oil, gas, electric power, communication facilities, etc.

It is worth noting that the system can be implemented depending on the features of the countries considered; for Europe, specific categories can be defined for cultural heritage assets and historical centers, which are widespread, critical and precious. Vulnerability factors can be calculated with particular regard to masonry buildings, by including specific algorithms already developed by the scientific community (Bernardini et al. 1990, D’Ayala et al., 2003, Valluzzi et al. 2004). A remarkable study on vulnerability evaluation and building inventory is the Sana’a GIS implementation (Figure 5) after a detailed in-field survey, provided by the Ferrara University to the Yemeni authorities, in the framework of the Conservation and Rehabilitation Plan for the Old City and other historic neighbouring settlements. A digitized database has been completed, classifying all the buildings in various categories, depending on their architectural relevance, according to the ICOMOS Washington Charter (ICOMOS 1987). Finally, a GIS-based application can join hazard and vulnerability data, merging together inputs coming from updated cadastral maps, RS satellites images and in-field DGPS surveys (diagnostics investigations and damage assessment included). Thus, geo-referred risk maps (Figure 6), in which single building structural features are linked to the surrounding environmental and social context, identify each house separately, giving a sharp classification of the danger level. (Midorikawa 2005). 2.4.2.6 Estimate losses In this analysis, the information must be provided together with the data of the previous steps; a true “risk assessment” takes into account all the possible hazards and not just a single event. Table 6 shows a brief summary of estimate losses. Building damage (structural, content, use and function) is a reliable indicator of risk. The level of building damage can be used to rank risks from various natural hazards and estimate risk in absolute terms. Human losses can be calculated in a credible way by using HAZUS procedures obtained for earthquakes. 189

Figure 7.

Satellite images of Bam (Iran) before (left) and after (right) the 2003 earthquake (source QuickBird).

Table 7. Mitigation options, general overview. Regulatory measures

• legislation which organizes and distributes responsibilities to protect a community from hazards • regulations that reduce financial and social impact of hazards through measures (insurances • new/updated design and construction codes • new/modified land use and zoning regulations • incentives that provide inducements for implementing mitigation measures

Repair and rehabilitation of existing structures

• removal or relocation of structures in high hazard areas • repair and strengthening of essential and high-potential-loss facilities

Protective and control structures

• deflect destructive forces from vulnerable structures and people • erect protective barriers (safe rooms, shelters, protective vegetation belts, etc.)

The loss estimate analysis can be concluded calculating the loss to each asset, the damage for each hazard event, and finally creating a composite map showing the most affected areas. It is important to note that the risks to existing structures (i.e. great part of European historical centers), built before the introduction of updated standards, must be accurately evaluated in the risk assessment procedures. Risk assessment must take into account all the data coming from post-earthquake in-situ damage investigations. In addition, RS image processing can also provide a damage prompt evaluation for large areas, comparing the situation “before” and “after” the event (Figure 7, Indirli, 2007). 2.4.2.7 Mitigation options Tables 7 and 8 show the principal mitigation options (general overview and hazard-targeted respectively). The adoption of updated building and safety codes is mandatory. With regards to earthquake, the adoption of revised set of rules by several GovernmentAuthorities is a step already achieved in many earthquake-prone countries, especially after the Northridge (1994) and Kobe (1995) seismic events, but also following the primary school collapse of San Giuliano di Puglia (Figure 8), Molise Region 2002, Italy (Presidenza del Consiglio dei Ministri, 2003; Line Guida, 2006; Indirli et al., 2004a-b). Another significant case is given by the hurricane Katrina (FEMA 2006a); the main recommendations (both for flood and wind) include the adoption of updated building codes (IBC 2006, IRC 2006, NFPA 5000 2006), incorporating flood load (ASCE 7-05, 2006) and flood-resistant construction standards (ASCE 24-05, 2006), with particular regard to foundations (Figure 9); design wind speeds are provided by ASCE 7 (ASCE 7, 2006); the use of FEMA 550 (FEMA 2006b) is 190

Table 8. Mitigation options, hazard-targeted. Earthquake regulatory measures

Flood regulatory measures

• building codes • master planning regulations

• guide development outside flood-prone areas • new development to address flood hazards • codes to address rehabilitation of older buildings

Repair and rehabilitation of existing structures • raise earthquake resistance • retrofitting/hardening • strengthen and repair of structural and non-structural elements

Repair and rehabilitation of existing structures • rehabilitation of older buildings • acquisition/demolition • relocating intact buildings out of floodplain • retrofit of infrastructures

Protective and control structures • securing around buildings and critical infrastructures • stabilizing soils and securing hazardous sites before new construction

Protective and control structures • decreasing run-off • increasing discharge capacity • containing, diverting or storing flood water

Figure 8. Rescue for children at the school site at San Giuliano di Puglia after the 31st October, 2002 earthquake.

Figure 9. Well-elevated and embedded pile foundation (left) and nearly failed house due to insufficient pile embedment (right) after the hurricane Katrina (USA).

also mandatory. Thus, the comparison of codes and standards regarding the mitigation of natural hazards, inside and outside the European Community, should be a fundamental step in future activity. Other details are given in Section 4.4.

2.4.3 VULNERABILITY ASSESSMENT Due to the different characteristics of the events, some examples are presented in this section. However, general procedure can be referred. Usually, the vulnerability assessment depends of 191

several parameters related to local conditions. The representation on maps of the vulnerability will allow the identification of the best location of constructions to reduce danger, by ranking priorities. Different scales of risk analysis may be considered; the study scale will determinate the studied consequences, the associated responsibilities and the action possibilities. The risk is the combination of hazard occurrence and subject damage. The action is the hazard itself and the subject can be persons, structures, infrastructures, communications, environment and economy. The next sections represent examples of analysis with respect to snow, water and landslide actions, allowing a better comprehension of the described procedures. 2.4.3.1 Snow In the Alps, the hazard is classified into 3 categories: the red zone where it is strictly forbidden to build any kind of new construction because of the high probability of danger; the blue zone where it is possible to build some constructions but where some special specifications are required; the white zone which is expected to be without danger. An approximation of the reference dynamic pressure Pd may be evaluated using the Bernoulli relationship: Pd = 1/2ρ V2 if ρ is the average snow unit weight (kg/m3 ) and V the displacement speed of the avalanche (m/s). So, a dry snow avalanche with a unit weight equal to 10 kg/m3 and a speed of displacement equal to 77.5 m/s (≈280 km/h) gives the same pressure (≈30 kPa) as a wet snow avalanche whose unit weight is equal to 400 kg/m3 and speed equal to 12.25 m/s (≈44 km/h). The avalanche loading on constructions may present very high values. For the blue zone, 30 kPa (corresponding to the previous examples) is a reference value used in many European countries in the case of a wet snow avalanche; this value comes from Switzerland which is considered as the reference country in Europe regarding this phenomenon. It is to be noted that trees, stones or ice blocks can amplify the effect of a wet snow avalanche by the addition of an impact load. Its value depends on the reference dynamic pressure. For instance, in Switzerland, a value of 100 kN is used with Pd = 30 kPa as 33 kN is used with Pd = 10 kPa; this load is expected to be applied to a surface equal to 500 cm2 at any level of the avalanche. Avalanches risk mitigation efficiency depends on the accurate knowledge of the studied system. The avalanche risk analysis may be described in several main phases: definition of the study scale and limits; determination of the constitutive elements and their function; identification of the risk scenarios; quantification of those scenarios in terms of occurrence and consequences; proposing mitigation actions. The definition of risk analysis scale is very important as it conditions the risk analysis result and the efficiency of the study process. Indeed, a risk analysis at the mountain scale will generate an important loss of time if it is expected to know the damage risk of a resort building regarding a potential avalanche hazard that may progress in a particular slope; in this case, the risk analysis should be made at the slope scale. To simplify, it is possible to relate the object and the scales of this kind of studies: at the mountain scale: global environmental impacts; at the massif scale: local environmental impacts; at the slope scale: economical and sociological impacts on persons, structures, infrastructures, communications, etc.; at the snowy coat: behavioural knowledge of the avalanche departure, flow and deposit. Considering a risk analysis from the civil engineering point of view, the scales of main interests are the slope scale and the snowy coat scale. Risk scenarios are basically the chains of events starting from an avalanche departure that lead to catastrophic damage on subjects: persons, structures, infrastructures, communications networks, etc. When risk scenarios are identified, quantified (departure occurrence and consequences effects) and classified by order of importance, it becomes possible to reduce the risk probability (prevention action) or the risk gravity (protection action) in order to put a snowy slope into secure conditions. Two categories of construction arrangements are available to protect buildings against avalanche risk: overall arrangements and specific arrangements for each construction. The construction principles that correspond to overall arrangements are: building grouping, orientation and shape of buildings, not-increasing the risk for the neighbourhood, provision for an avalanche outlet. The building principles that correspond to specific arrangements for each construction are: foresee an access and an entrance on the non-exposed facades, design facades without hold-in corner when these are facing the avalanche-prone slopes, no storage of polluting or dangerous products in poorly resistant constructions, foresee an appropriate distribution of the buildings: the more vulnerable buildings have to be located upstream in the avalanche direction. The protection works are classified into two main categories, depending on their location in the departure zone (active protection) 192

or in the flow or deposit zones (passive protection). In both cases, these actions can be provided permanently (without human intervention) or temporarily (with decision taken). The main permanent active protection refers to: reforestation on seat, wind barrier, snow barrier, buzzard roof, wind transfer, tire racks, wicker racks, and fillets. The main permanent active protection refers to: galleries, stems, deflectors, stopping dike, stakes, and road detector of avalanche. The major temporary protections consist in the application of several rules and ensure these are observed by the population in order to prevent a direct exposure to the avalanche risk: traffic restrictions and regulation. This may be done in order to proceed to an evacuation, or on the contrary to maintain people in a safe location until the end of the critical period.

2.4.3.2 Water Flood vulnerability analysis and risk assessment are important for the mitigation of the effects of hazard. Erosion and floods are a common problem in Europe which can be observed on a mediumterm scale. Several causes are, to a varying degree contributing to erosion and floods. On the one hand, one finds the dynamic nature of coastal and river zones and climate change, and on the other the anthropogenic influences. It is increasingly clear that people and assets in some littoral and fluvial urban fronts are endangered and that serious damage and high costs should be expected. To implement sustainable actions, plans should be conceived based on a medium to long-term evolution assessment. Due to inherent uncertainty the assessment of future conditions can only be done on the basis of scenario evaluations, for which numerical models comprising the present state of knowledge may be used as tools. To help in ranking action priorities, vulnerability and risk must also be assessed. The risk can be accomplished by crossing vulnerability and degree of exposure information in risk maps depicting spatial analysis (as a recommendation from the European Union, Directive 2007/60/CE, see Directive, 2007). Parameters, such as wave energy, tides, bathymetry and topography, shoreline morphology, sediment budget and meteorological conditions are important for a vulnerability analysis. However, zones which are very vulnerable to floods may not necessarily be considered as being at risk. An approach similar to the vulnerability analysis needs to be established for evaluating the degree of exposure. Storm surges in combination with river floods can be highly damaging to infrastructure or property, and cause substantial human and economic losses. In terms of exposure levels the parameters considered important are population density, economic activities potentially affected by floods, and ecological, cultural and historical values exposed to devastation by sea actions or floods. Spatial classification facilitates mapping the degree of exposure. Risk maps consist of a classification obtained from crossing vulnerability with degree of exposure. Analysis based on the spatial classification and on the weighing of the parameters, is important for the evaluation of the vulnerability and the risk of floods (see also Coelho et al., 2009).

2.4.3.3 Landslide Vulnerability to landslide hazards is a function of location, type of human activity, use, and frequency of landslide events. The effects of landslides on people and structures can be lessened by total avoidance of landslide hazard areas or by restricting, prohibiting, or imposing conditions on hazard-zone activity. Local governments can reduce landslide effects through land-use policies and regulations. Individuals can reduce their exposure to hazards by educating themselves on the past hazard history of a site and by making inquiries to planning and engineering departments of local governments. They can also obtain the professional services of an engineering geologist, a geotechnical engineer, or a civil engineer, who can properly evaluate the hazard potential of a site, built or unbuilt. The hazard from landslides can be reduced by avoiding construction on steep slopes and existing landslides, or by stabilizing the slopes. Stability increases when ground water is prevented from rising in the landslide mass by: • • • •

covering the landslide with an impermeable membrane; directing surface water away from the landslide; draining ground water away from the landslide; minimizing surface irrigation. 193

Table 9. Possible damages to constructions as function of the sustained wind speed. Wind speed (km/h)

Damages

<150 150–180 180–210

Negligible damages to constructions. Some coastal flooding. Minor damages to roofs and openings. Significant flooding damages. Some structural damages to small constructions (mainly curtain wall failures). More important flooding damages near the coast: small structures destroyed and larger structures damaged by floating debris. Significant structural damages. More important curtain wall failures with some complete roof structure failures on small constructions. Important erosion of coastal areas. Complete roof failures on most constructions and industrial buildings. Some complete building failures. Flooding causes major damage to lower floors of all concerned structures.

210–250

>250

Slope stability is also increased when a retaining structure and/or the weight of soil/rock are placed at the toe of the landslide or when mass is removed from the top of the slope. 2.4.4 DAMAGE ASSESSMENT METHODOLOGIES Following a natural disaster, engineers undertake structural assessments for many different purposes; for example, for the assessment of structural safety, quantification of the severity of the event effects or for insurance loss calculation. These purposes are common irrespective of the hazard that may have caused the structural damage. A critical review and comparison of existing methods for the post-event damage assessment of structures under different natural hazards is presented. These hazards have different levels of development in terms of structural assessment methods and universal acceptance of these methods. Structural damage assessments are an integral and essential part of the recovery process from a natural disaster, and occur independently of the nature of the hazard causing the disaster. Immediately after the event engineers must assess all buildings within the affected area to assess damage, safety, and usability, to identify buildings requiring emergency strengthening (e.g. to avoid collapse during aftershocks or further volcanic ash fall), to provide reliable data to the authorities, and to plan further relief and rehabilitation measures. A systematic collection of damage data reduces the time required to complete the work, ensures that no valuable information is lost, and leads to a realistic assessment of building capacity. This first stage of structural assessment is often carried out through rapid screening. In the next phase, structures deemed unsafe are assessed in more detail to determine the extent of required repair or need for demolition. In addition to their use for recovery, structural damage assessments often provide data for future research studies on the revision of existing urban plans. The methodology to be adopted for the structural assessment must therefore strike a balance between the need for a rapid and efficient procedure, and the need for detailed data collection for future studies. A comparison of different damage assessment methodologies is presented in Rossetto et al. (2010b). Again, some examples of hazards damage assessment are presented here. 2.4.4.1 Cyclones A simple classification of potential damages due to cyclones, depending on the wind speeds, is described in Table 9. 2.4.4.2 Landslides Three kinds of landslide impacts and damages in constructed areas can be outlined: submersions, logjams and dams or bridges failure have caused collapses and destruction of buildings; landslides and massive erosion highly damaged the communication networks (roads, electricity . . .); sediments carried out by the flow have invaded the building, while water imbibed the masonry, causing a long term weakening of the structures. 194

Table 10. Criteria for assigning degree of damage (DD) in Japan. DD

Damage state of structural member

I II III IV V

Visible but narrow cracks on surface of concrete (crack width w < 0.2 mm) Visible cracks on surface of concrete (0.2 mm < w < 1.0 mm) Local spalling of cover concrete, major cracks (1 mm < w < 2 mm) Full spalling and crushing of concrete, exposed reinforcing bars Buckling of bars, crushing of concrete core, visible vertical deformation of column/wall

2.4.4.3 Earthquakes This topic is widely discussed in other Sections of the Final Report. Several methods for postearthquake inspection and rapid assessment of buildings have been developed in a number of countries. Among these, procedures used in Japan (JBDPA, 1990), USA (ATC 20, 1989; ATC 20-1, 1989; ATC20-2, 1995), New Zealand (NZSEE, 2009), the Balkans (UNDP/UNIDO, 1985) and Italy (Protezione Civile 2010a-c; MEDEA, 2005; GNDT_INGV, 2010a,b) deserve particular attention (Kappos, 2003). An example of the Japan methodology is presented in Table 10. 2.4.4.4 Tsunami In the case of tsunami, very few guidance documents have been developed for use in post-event damage assessments. The Intergovernmental Oceanographic Commission of UNESCO (IOC, 1998) has published a post tsunami field guide developed from existing earthquake and tsunami field guides and more recent tsunami surveys (Farreras, 2000). While concentrating on collecting scientific data such as tidal levels, run-up elevations and bathymetric data, it indicates that data on structural damage should be collected where possible, noting the possible cause of the damage and distinguishing tsunami damage from earthquake damage in a near source event. The guidance for building damage assessment is brief and recommends rough (non-specialized) classification of damage, estimating the nature and category of the damage and its apparent cause. Several approaches exist for identifying tsunami intensity (e.g. Ambraseys, 1962; Papadopoulos and Imamura, 2001). However, these methods do not provide techniques for identifying structural damage. Most rapid field investigations presented in literature, refer to damage assessments based on earthquake assessment methodologies directly. Rigorous, multi-stage building assessments using forms such as those of ATC-20 (ATC 20, 1989) have not been carried out, or at least have not been published. Instead, the damage scales in EMS-98 are the most commonly used (e.g. in Miura et al., 2006). A few studies have attempted to modify earthquake damage assessment methods and scales to take into account damage relating to fast-flowing water, such as foundation failure due to scour or floating debris impact damage. A modified version of the EMS-98 damage scales for use in tsunami damage assessment in Thailand and Sri Lanka following the Indian Ocean Tsunami was proposed by Rossetto et al. (2007) and EEFIT (2006). In these studies damage attributed to different building types was also adopted to assign intensity values to the surveyed locations, using a modified version of the Tsunami Intensity scale of Papadopoulos and Imamura (2001). An example of the damage scale descriptions for masonry buildings proposed by EEFIT (2006) is shown in Table 11. Taking into account damage to different structural types allows the intensity values to be compared in countries with different building stock, to obtain a comparative intensity for tsunami impact assessment. The results of these surveys do not provide sufficient information however to improve knowledge on the structural response of buildings under tsunami loading and therefore are not useful for the re-evaluation of codes of practice, assessment of existing structures etc. 2.4.4.5 Volcanic eruption This topic is largely discussed in another part of the Final Report (Section 2.5). 195

Table 11. Tsunami damage scale descriptions for masonry structures typical of Sri Lanka proposed by EEFIT (2006). Damage State

Damage description for structure

None (DM0)

No visible structural damage to the structure observed

Light (DM1)

Damage limited to chipping of plaster on walls, minor cracking visible. Damage to windows, doors. Damage is minor and repairable. Immediate occupancy

Moderate (DM2)

Out-of-plane failure or collapse of parts of or whole sections of masonry wall panels without compromising structural integrity. Masonry wall can be repaired or rebuilt to restore integrity. Most parts of the structure intact with some parts suffering heavy damage. Scouring at corners of the structures leaving foundations partly exposed but repairable by backfilling. Cracks caused by undermined foundations are clearly visible on walls but not critical. Unsuitable for immediate occupancy but suitable after repair

Heavy (DM3)

Out-of-plane failure or collapse of masonry wall panels beyond repair, structural integrity compromised. Most parts of the structure suffered collapse. Excessive foundation settlement and tilting beyond repair. Collapse of wall sections due to scouring and damage non-repairable. Structure requires demolition since unsuitable for occupancy

Collapse (DM4)

Complete structural damage or collapse, foundations and floor slabs visible and exposed, collapse of large sections of foundations and structures due to heavy scouring

2.4.5 AN EXAMPLE OF FIELD INVESTIGATION: L’AQUILA EARTHQUAKE The topic regarding earthquakes is largely discussed in other parts of the Final Report. A field investigation was conducted in the city of L’Aquila (Abruzzo Region, Central Italy) after the April 6th, 2009 seismic event. (A general description of the seismic event and its consequences is given in Indirli, 2010). After the earthquake, the PLINIVS Centre (Naples), conducted an extensive damage survey of the whole historic centre of L’Aquila, with the contribution of experts of COST Action C26 (COST 2006). The COST researchers, in addition to an overall review of the consequences of the earthquake in L’Aquila and its surroundings, performed a detailed damage assessment of structures of three areas of the historic city centre. The AeDES (Protezione Civile, 2010a) and the MEDEA (MEDEA, 2005) survey methodologies were utilised, with respect to masonry and reinforced concrete structures (see details in Borg et al., 2010; Kouris et al., 2010). REFERENCES ASCE 7, 2006.American Society of Civil Engineers, Minimum Design Loads for Buildings and Other Structures. ASCE 7-05, 2006. American Society of Civil Engineers, Flood load. ASCE 24-05, 2006. American Society of Civil Engineers, Flood -resistant construction standards. Ambraseys N.N., 1962. Data for the investigation of the seismic sea waves in the eastern Mediterranean. Bull. Seism. Soc. of Am. 52, 895–913. ATC-20, 1989. Procedures for post-earthquake safety evaluation of buildings, 1989. ATC20-1, 1989. Field Manual: post-earthquake safety evaluation of buildings, 1989. ATC-20-2, 1995. Revised placards and forms, 1995. Bernardini, A., Gori R., & Modena C. 1990. Application of coupled analytical models and experimental knowledge to seismic vulnerability analyses of masonry buildings, Earthquake Damage Evaluation and Vulnerability Analysis of Buildings Structures. A. Kortize (ed.). INEEC, Omega Scientific. Borg, R.P., Indirli, M., Rossetto T., Kouris, L. 2010. L’Aquila earthquake The April 6th, 2009: the damage assessment methodologies. Proceedings of COST Action C26 Final International Conference on Urban habitat construction under catastrophic events, Naples, 16–18 September 2010.

196

Chaboud, R. 2003. Cyclones. Encyclopédia Universalis. DVD-ROM, France. Coelho, C., Silva, R., Veloso-Gomes, F. 2006. Shoreline Evolution Model: A Tool to Compare Coastal Protection Measures”. Coastal Dynamics, Geomorphology and Protection. 8th International Conference LITTORAL 2006 – Coastal Innovations and Initiatives, Gdansk, Poland, 115–123, ISBN 83-88617-82-6. Coelho C., Granjo M., Segurado-Silva C., 2009. Map of Costal Zone Vulnerabilities to Wave Actions – Application to Aveiro District (Portugal). International Symposium on Integrated Coastal Zone Management. Arendal, Norway, 11th to 14th June 2007, chapter 24, pp. 318–330, ISBN 978-1-4051-3950-2. COST, 2006. COST, European COoperation in the field of Scientific and Technical research, Transport and Urban Development, COSTAction C26: “Urban Habitat Constructions Under Catastrophic Events”. 2006– 2010. D’Ayala D. & Speranza E., 2003. Definition of Collapse Mechanisms and Seismic Vulnerability of Historic Masonry Buildings, Earthquake Spectra, 19 March 2003. Decanini L., Mollaioli F., Panza G.F., Romanelli F. & Vaccari F. 2001. Probabilistic vs deterministic evaluation of seismic hazard and damage earthquake scenarios: a general problem, particularly relevant for seismic isolation. Proc. 7th International Post-Smirt Seminar on Seismic Isolation, Passive Energy Dissipation and Active Control of Vibration of Structures, Assisi, Italy, 2–5 October, 2001. Directive, 2007. Directive 2007/60/EC of the European Parliament and of the Council of the 23rd October 2007 on the assessment and management of flood risks. Official Journal of the European Union, L 288/27. EEFIT, 2006. The Indian Ocean tsunami of 26 December 2004: mission findings in Sri Lanka & Thailand. Pomonis A, Rossetto T, Peiris N, Wilkinson SM, Del Re D, Koo R, Gallocher S, Earthquake Engineering Field Investigation Team (EEFIT) Report, UK. http://www.eefit.org.uk. EN, 1991a. EN 1991-1-3:2003. Eurocode 1 – Actions on structures – Part 1-3: General actions – Snow loads. European standard, CEN, Brussels. EN, 1991b. EN 1991-1-7:2003. Eurocode 1 – Actions on structures, Part 1-7: General Actions – Accidental actions. European standard, CEN, Brussels. European Commission 2004. Living with Coastal Erosion in Europe – Sediment and Space for Sustainability. Luxembourg: Office for Publications of the European Communities, 40 pp., ISBN 92-894-7496-3. Farreras S,F., 2000. Post-Tsunami Field Survey Procedures: An Outline. Natural Hazards, 21, 207–214. FEMA 386-2, 2001. Understanding your risks. Identifying hazards and estimating losses, FEMA (Federal Emergency Management Agency) Report, 2001. FEMA 549, 2006a. Hurricane Katrina in the Gulf Coast, Mitigation Assessment Team Report, Building Performance Observations, Recommendations, and Technical Guidance, FEMA (Federal Emergency Management Agency) Report, 2006. FEMA 550, 2006b. Recommended Residential Construction for the Gulf Coast: Building on Strong and Safe Foundations, FEMA (Federal Emergency Management Agency) Report, 2006. Givry, M., Perfettini, P. 2004. Construire en montagne – La prise en compte du risque d’avalanche. Ministère de l’écologie et du développement durable. Ministère de l’équipement, des transports, du logement, du tourisme et de la mer. Available on http://www.avalanches.fr. GNDT-INGV 2010a. Scheda di II livello per la valutazione della vulnerabilità in edifici in cemento armato/II Level Form for the evaluation of vulnerability in r.c. buildings. http://gndt.ingv.it/Strumenti/ Schede/Schede_vulnerabilita/ scheda_secondo_livello_ca.pdf. GNDT-INGV 2010b. Scheda di II livello per la valutazione della vulnerabilità in edifici in muratura/II Level Form for the evaluation of vulnerability in masonry buildings. http://gndt.ingv.it/Strumenti/ Schede/Schede_vulnerabilita/ scheda_secondo_livello_mur.pdf. HAZUS, 1992. www.fema.gov/hazus. IBC, 2006. International Building Code, USA. ICOMOS 1987. Charter for the conservation of the historic towns and urban areas, Washington Charter, adopted by ICOMOS, October 1987. Indirli M. et al., 2004a. The reconstruction of San Giuliano di Puglia after the October 31st 2002 earthquake. Proc. 13th World Conference on Earthquake Engineering, August 2004, Vancouver, B.C., Canada. Indirli, M. et al. 2004b. Seismic protection of historical centers using innovative techniques, with focus on San Giuliano di Puglia after the 2002 Molise earthquake. In Proc. IV International Conference on Structural Analysis on Historical Constructions (SAHC), 10-13 November 2004, Padua, Italy. Indirli M., Valpreda E., Panza G., Romanelli F., Lanzoni L., Teston S., Berti M., Di Bennardo S. & Rossi G., 2006. Natural multi-hazard and building vulnerability assessment in the historical centers: the examples of San Giuliano di Puglia (Italy) and Valparaiso (Chile). Proc. 7th European Conference “SAUVEUR” Safeguarded Cultural Heritage, Understanding & Viability for the Enlarged Europe, Prague 31st May – 3rd June 2006. Indirli M., 2007. Overview on risk assessment approaches for natural hazards. Proc. of COST Action C26 Seminar on Urban Habitat Constructions Under Catastrophic Events, Prague, 30–31 March 2007.

197

Indirli, M. 2010. The 6th April 2009 L’Aquila earthquake: from ruins to reconstruction. Seismicity and Earthquake Engineering: L’Aquila Earthquake of April 2009. Eds. R.P. Borg, ISBN-978-99932-0-879-2, Kamra Tal-Periti, April 2010, La Valletta, Malta. IOC, 1998. Intergovernmental Oceanographic Commission (of UNESCO): 1998, Post-Tsunami Survey Field Guide (1st edn), Manuals and Guides #37, Paris, France. IRC, 2006. International Residential Code, USA. JPDPA, 1990. Japan Building Disaster Prevention Assessment, 1990. Kappos, A.J., 2003. Methods for post-earthquake inspection and rating of buildings: a world view, Marmara earthquake rehabilitation programme ActionC3: Scheme for post-earthquake inspection and rating of buildings (course notes). Klügel, J.U., Mualchin L. and Panza G.F. 2006. A scenario-based procedure for seismic risk analysis. In Engineering Geology, 88, 1-22. Kouris, L., Borg, R.P., Indirli M. 2010. The L’Aquila earthquake April 6th, 2009: a review of seismic damage mechanisms. Proceedings of COST Action C26 Final International Conference on Urban habitat construction under catastrophic events, Naples, 16-18 September 2010. Linee guida, 2006. Linee Guida per la valutazione e riduzione del rischio sismico del patrimonio culturale con riferimento alle norme tecniche per le costruzioni (Guidelines for evaluation and mitigation of seismic risk to cultural heritage). Italian Ministry for Cultural heritage, July 2006. MAR VASTO, 2007. Risk Management in Valparaíso/Manejo de Riesgos en Valparaíso, Servicios Técnicos (acronym MAR VASTO), funded by BID/IDB (Banco Inter-Americano de Desarrollo/Inter-American Development Bank). Project ATN/II-9816-CH, BID/IDB-ENEA Contract PRM.7.035.00-C, March 2007 – October 2008, www.marvasto.bologna.enea.it. MEDEA, 2005. Manuale di Esercitazioni sul Danno Ed Agibilità per edifici ordinari/User’s manual on damage and safety for ordinary buildings. http://gndt.ingv.it/Att_scient/Molise2002/San_Giuliano/ Strumenti%20di%20rilievo.pdf. Midorikawa S., 2005. 3D Seismic Hazard and Risk Maps for Earthquake Awareness of Citizens with Aids of GIS and Remote Sensing Technologies. Proc. of the 3rd International Workshop on Remote Sensing for Post-Disaster Response, September 12-13, 2005, Chiba, Japan. Miura, H., Wijeyewickrema, A. & Inoue, S. 2006. Evaluation of tsunami damage in the eastern part of Sri Lanka due to the 2004 Sumatra earthquake using remote sensing technique. Proc. 8th National Conference on Earthquake Engineering, Paper No. 8, NCEE-856. Munich Re Group, 2004. Topics Geo Annual review, Natural catastrophes, knowledge series. NFPA 5000, 2006. Building Construction and Safety Code, published by the National Fire Protection Association. NZSEE, 2009. Guidelines for Building Safety Evaluation During a State of Emergency. August 2009. Prepared by the New Zealand Society for Earthquake Engineering (NZSEE). Panza, G.F., Romanelli F. &Yanovskaya, T., 2000. Synthetic Tsunami mareograms for realistic oceanic models, Geophysical Journal International, 141, 498–508. Panza G.F., Romanelli F. & Vaccari F., 2001. Seismic wave propagation in laterally heterogeneous anelastic media: Theory and applications to seismic zonation, Advances in Geophysics, Vol. 43, pp. 1–95, Academic Press, San Diego. Panza G.F. et al., 2002. Realistic modeling of seismic input for megacities and large urban areas (the UNESCO/IUGS /IGCP project 414). Episodes, 25–3, 160–184. Papadopoulos, G. A., and Imamura, F. 2001. A proposal for a new tsunami intensity scale. ITS Proceedings, Session 5, No.5-1. pp569–577. Presidenza del Consiglio dei Ministri, 2003. OPCM 3274, 3316, 3431, Gazzetta Ufficiale: May 8th, 2003, Ser. Gen. n. 105, Suppl. Ord. n. 72; October 10th, 2003, Ser. Gen. n. 236; May 10th, 2005, Ser. Gen. n. 107, Suppl. Ord. n. 85. Protezione Civile 2010a. AeDES, Scheda Agibilità e Danno nell’Emergenza Sismica/First Level form for safety assessment, damage investigation, prompt intervention for ordinary buildings in the post-earthquake emergency. http://www.protezionecivile.it/cms/attach/editor/schedadanni.pdf. Protezione Civile 2010b. Scheda per il rilievo del danno ai Beni Culturali: chiese/Form for the damage survey to cultural heritage: churches. http://www.protezionecivile.it/ cms/attach/adc.pdf. Protezione Civile 2010c. Scheda per il rilievo del danno ai Beni Culturali: palazzi/Form for the damage survey to cultural heritage: buildings. http://www.protezionecivile.it/ cms/attach/bdp.pdf. Rossetto T, Peiris LMN, Pomonis A, Wilkinson SM, Del Re D, Koo R, Gallocher S. 2007. The Indian Ocean tsunami of December 26, 2004: observations in Sri Lanka and Thailand, Natural Hazards 42, 105–124. Rossetto T., Coelho C., Carlier J.P., Allsop W., Lloyd T.O., 2010a. Tsunami impact evaluation for coastal areas. Proceedings of COST Action C26 Final International Conference onUrban habitat construction under catastrophic events, Naples, 16-18 September 2010.

198

Rossetto T., Kappos, A.J., Kouris L.A., Indirli M., Borg R.P., Lloyd T.O., Sword-Daniels V., 2010b. Comparison of damage assessment methodologies for different natural hazards. Proceedings of COST Action C26 Final International Conference onUrban habitat construction under catastrophic events, Naples, 16–18 September 2010. Santos, F.D., Forbes, K., Moita, R. 2002. Climate change in Portugal. Scenarios, Impacts, and adaptation measures. SIAM Project, Gradiva, Lisboa, Portugal. Shinozuka M., 2005. Application of Remote Sensing Technology in Natural Hazard Assessment -Examples from 2004 and 2005. Proc. of the 3rd InternationalWorkshop on Remote Sensing for Post-Disaster Response, September 12–13, 2005, Chiba, Japan. UNPD, 2002. United Nations Development Programme, Bureau for Crisis Prevention and Recovery, Reducing disaster risk, a challenge for development, www.undp.org/bcpr. UNPD-UNIDO, 1985. United Nations 1985. Post earthquake damage evaluation and strength assessment of buildings, UNIDO/UNDP-PER/79/015, vol. 4. Valluzzi MR., Cardani G., Binda L., & Modena C. 2004. Analysis of the seismic vulnerability of masonry buildings in historical centres and intervention proposals, Proc. 6th International Symposium on the Conservation of Monuments in the Mediterranean Basin. Lisbon, Portugal, 7–10 April 2004, (on CD-ROM). 2004: 561–565. Valpreda E., 2003. GIS and natural Hazards, chapter 22, in: Natural Disaster and Sustainable Development, C. Margottini-Springer-Verlag Eds, 2003. Zuccolo E., Vaccari F., Peresan A., Dusi A., Martelli A., Panza G.F. 2008. Neo-deterministic definition of seismic input for residential seismically isolated buildings. Engineering Geology, Vol. 101, 89–95.

199

Urban Habitat Constructions under Catastrophic Events (Final Report) – Mazzolani (Ed). © 2010 Taylor & Francis Group, London, ISBN 978-0-415-60686-8

2.5 Consequences of volcanic eruptions on constructions Ruben Paul Borg Faculty for the Built Environment, University of Malta, Malta

Maurizio Indirli ENEA, Bologna, Italy

2.5.1 INTRODUCTION About 500 million people are at risk in view of volcanic hazards. In the past 500 years, over 200,000 people have lost their lives due to volcanic eruptions. An average of 845 people died each year, between 1900 and 1986 as a result of volcanic hazards. In the coming years, these numbers are predicted to rise. The reason is not due to an increase in volcanic activity, but due to an increase in the population in the areas surrounding the active volcanoes. In Europe, this is the case of Naples (Italy), where Vesuvius and Phlegrean Fields, threaten the safety of about one million people. In other parts of the World, metropolitan areas, like Tokyo (Mt. Fuji), Mexico City (Popocatépetl) and Auckland (Auckland Field) are affected by eruptive risk. The peculiar importance of this aspect has led the EU Network COST Action C26 “Urban habitat constructions under catastrophic events” (2006–2010; COST, 2006) to introduce the analysis of the volcanic risk in the urban areas in its research activities, substantially referred to the following topics: to identify the volcanic actions on constructions; to evaluate the volcanic vulnerability of the urban environment in view of an eruption; and to propose simple and economical mitigation interventions. Furthermore, Vesuvius (more precisely Somma-Vesuvius, located about 15 km away from Naples, Italy, close to a densely populated area) has been selected as the case study within the COST Action C26 Working Group 4 ‘Risk Assessment for Catastrophic Scenarios in Urban Area’ (Alterio et al., 2010; De Gregorio et al., 2010a-b; Faggiano et al., 2010; Florio et al., 2010; Mazzolani et al., 2008, 2009a-c and 2010a-b; Sword-Daniels et al., 2010; Zuccaro et al. 2010a-d). The part regarding the volcanic hazard is largely discussed in Section 1.5, the part on volcanic actions also in Section 1.5, the part on mitigation measures for the reduction of volcanic risk both in Sections 1.5 and 4.4. This Section is dedicated in particular to volcanic damage and the field investigations carried out in the Vesuvius area. The aim of the investigation was to collect the data and parameters influencing the volcanic vulnerability for different constructions. 2.5.2 VOLCANIC DAMAGE 2.5.2.1 Structural damage due to volcanic action A volcanic eruption is characterized by a series of subsequent physical phenomena, including volcanic earthquakes, ash-fall, pyroclastic flows, lahars, landslides, volcanic missiles and tsunami (see other Sections of this Report). As a consequence, the damage impact due to a volcanic eruption depends upon several disastrous events, different from each other, but tightly connected, where each event contributes to the final scenario. The evaluation of the possible effects due to a volcanic eruption in an urban area is therefore very complex, depending on the type of eruption, and also on the development over time of the different phenomena characterizing it. Other important questions are the location considered and the typological-structural characteristics of buildings/infrastructures in the area. Therefore, the classification of structures and the identification of the different building characteristics is an important step in the assessment of the vulnerability of structures to volcanic events. 201

The collapse of the substained column and the conseguent pyroclastic flow (PF) are very frequent phenomena in explosive eruptions. In this case the magmatic material erupted is composed by a mixture of molten and solid pyroclasts in a continuous gas phase. This mixture is the product of gas exsolution and magma fragmentation processes that occur during the magma ascent from the deep magma reservoir to the ground surfaces. When this mixture of gas and pyroclasts penetrates into the cooler atmosphere it mixes up with the surrounding air forming a volcanic jet. Crater geometry, outlet pressure and velocity, temperature and gas content control the effectiveness of the mixing and, therefore, the global evolution of the explosive event. If the mixture at the top of the jet is reduced below the atmospheric density, then the eruption forms a convective buoyant plume called “Plinian“ column. Otherwise, the eruptive mixture collapses forming a ground PF that can propagate to great distances from the vent. The action on the vertical surfaces of buildings affected by the flow is a combination of impact and thermal stress, proportional to its mass and velocity. Pressure and temperature values vary depending on the characteristics of the eruption column and on the morphology of the invaded areas. In case of a Sub-Plinan I eruption the pressure can reach 10 kPa and the temperature can reach up to 400◦ C. However, considering the way of propagation of the flow within the territory, velocity and temperature values are not uniform, but weaker in lateral areas of the cloud, generally decreasing depending on distance from the vent. This means that not all buildings struck by the flow are destroyed, but it is possible to identify levels of damage as a function of impact characteristics and building vulnerability. Damage resulting from the impact of pyroclastic flows on buildings depends on the combination of several factors: the duration of the phenomenon, the temperature of the flow and pressure produced by the impact. In general, the impact of pyroclastic flows can be classified into three main categories: a) the values of pressure and temperature are likely to damage the structure, until partial or total collapse; b) the values of pressure and temperature are not likely to damage the structure, but there is a breakthrough of non-structural parts (window frames or infill panels) that allows the penetration of the flow into the building; c) for lower values of pressure and temperature none of the building technical elements is expected to collapse, but the difference between the external and internal pressure causes the infiltration of the flow inside the building. While it is clear that in the first case the damage is very serious, it should be noted that, in the other two cases, the flow infiltration can lead to the destruction of the building, mainly in case of breakage, as the strong internal pressure caused by flow infiltration can “inflate” the building, causing the failure of the roof or windows to the outside. It is also important to consider the possibility that the high temperature of the flow entering the building could trigger fires, which could destroy the building even in the absence of mechanical damage. For these reasons, three different aspects of vulnerability should be identified, with respect to the three different types of expected damage: a) the vulnerability of the major elements (masonry walls, frame), b) the vulnerability of non-bearing (coverage, cladding), c) understanding vulnerability as “permeability” to infiltration, and expressed as ACH (Air Change for Hour). The most significant parameter in this case is the dynamic pressure, whereas the temperature is less decisive. The evaluation of the building structure vulnerability to pyroclastic flow actions requires the estimation of the limit horizontal pressure at the collapse state of the standard buildings. A vulnerability analysis has been carried out by means of fundamental theorems of limit state analysis applied to R.C. frames and to the masonry walls (Zuccaro et al., 2004, 2008 and 2010a–d; Spence et al., 2004). It should be considered that the lateral pressure caused by the flow is quite different from seismic action. Pyroclastic flow action is not cyclical, therefore the ductility, as energy dissipation capacity is less important. Also, unlike the case of earthquake, mass is not directly proportional to lateral action, but plays a stabilizing function. The slenderness of the building is a factor strongly conditioning the level of vulnerability. An investigation was carried out, on the behavior of several 202

Table 1. Pyroclastic Flows – Structural Classification. Type

Description

Ap

Weak Masonry Buildings of 3–4 storeys with deformable floor. Weak or strong Masonry Buildings with more then 4 storeys. Medium Masonry Buildings of 1–2 storeys with deformable floor. Strong Masonry Buildings of 3 or more storeys with rigid floor. Strong Masonry Buildings of 1–2 storeys with rigid floor. Non aseismic r. c. buildings of more than 6 storeys (High). Non aseismic r. c. buildings of 4–6 storeys (Medium). Non aseismic r.c. buildings of 1–3 storeys (Low).

Bp Cp Dp Ep Fp

sample buildings loaded by lateral increasing pressure, exchanging the typological and geometrical characteristics of the buildings, and computing the collapse values. The results of numerical analysis show a significantly different behavior between masonry and reinforced concrete structures, thus suggesting the definition of two separate vulnerability scales). Three classes of vulnerability where identified (A, B, C) for masonry structures and three for R.C. structures (D, E, F), defining the buildings assignment criteria and the collapse probability as a function of lateral pressure by flow (Table 1). With regards to seismic impact scenarios, the PLINIVS Study Centre of the University of Naples “Federico II” (Centre of competence of the Italian Civil Protection) has developed a computerized tool to build up the seismic impact in the volcanic areas around Vesuvius and Campi Flegrei in Campania Region (Italy). The final result of the assessment is the damage distribution for the buildings and the damage to population in each cell, which can be represented in the form of damage maps using a GIS tool. The assessment provided also a sub map identifying the roads selected in the evacuation plan and the interruption probability of the road-links along the escape routes as a result of an earthquake of expected magnitude (Zuccaro et al., 2004, 2008 and 2010a–d). The PLINIVS Study Centre also developed tools intended to evaluate a reasonable estimate of the cumulative damage impact scenario as a consequence of a Vesuvius eruption of assigned intensity. The model is integrated in a Geographic Information System (GIS), and refers to Vulnerability Functions and on cumulative damage on the buildings. The tool is intended to control the progressive impact up to the final impact scenario Zuccaro et al. (2004, 2008 and 2010a–d). 2.5.2.2 The structural damage assessment The damage assessment of buildings was carried out by Spence et al. (1996), Blong (2003) and Spence et al. (2005), using a damage scale based on the MSK earthquake intensity scale for buildings, and by applying earthquake engineering principles and survey techniques. Spence et al. (1996) developed a 6-point damage scale for use in assessing the damage from Mount Pinatubo, to summarise the damage and to provide damage distribution data. Blong (2003) carried out a damage assessment for Rabaul, Papua New Guinea, and used the same damage scale to quantify the impacts. The study noted that additional volcanic hazards should be included in such assessments, including: mudfills, lahars and the secondary impact of corrosion. Spence et al. (2005) again used the 6-point damage scale to classify the damage from volcanic ashfall. The study generated vulnerability curves for European buildings based on empirical and analytical data. The data was gathered from the area around Vesuvius and from other areas in Europe. These surveys provide little information on the building characteristics, type, age and condition, and many surveys do not provide adequate detail for the purpose of understanding building vulnerability to volcanic hazards. The COST Action C26 (COST, 2006) undertook an assessment of buildings in the Vesuvius area, in order to classify the building typology and to predict the possible modes of failure when impacted by volcanic hazards. While the C26 preliminary activity is reported in several publications (Dobran, 2007; Mazzolani et al., 2008, 2009a-c), the complete results of the Vesuvius field investigation are reported in Alterio et al. (2010), De Gregorio et al. (2010a–b), Faggiano et al. (2010), Florio et al. (2010), Mazzolani et al. (2010a–b), Zuccaro et al. (2010a–d). The study involved two field surveys undertaken during 2009 and 2010, to identify structural typologies in the area around Vesuvius, and 203

record their characteristics. The surveys and subsequent vulnerability analyses were carried out as a collaboration between various universities and institutes across Europe. The surveys were carried out for different building types (historic centre, residential area, and schools of Torre del Greco, the most populated municipality around Vesuvius), and cultural heritage (the Vesuvian Golden Mile Villas). The detailed surveys recorded information including: regularity of building in plan and height, the number and height of storeys, number and size of openings, frescos, mouldings, number of statues and pieces of original furniture. Data was also collected on construction materials, construction methods, building age, existing strengthening or improvement, general state of repair, site morphology and the existence of cornices, lintels, stringcourses, tie-beams, connection of walls to roof and of floors to walls (where seen). In addition, the failure mechanisms were classified using the MEDEA form (MEDEA, 2005) and vulnerability was assessed to the three hazards of: earthquake, pyroclastic flow and ash fall. The results showed the prevalence of buildings designed to resist ordinary vertical loads, which showed insufficient safety against the volcanic actions. The adopted survey approach should be optimised for others to use successfully and unambiguously, with the aim of increasing the quantity of acceptable field-based surveys that can then be added to the collective database. This will improve understanding of vulnerability to volcanic hazards.

2.5.3 THE VOLCANIC VULNERABILITY OF STRUCTURES 2.5.3.1 The Vesuvius structural vulnerability assessment Vesuvius, or more precisely Somma-Vesuvius, is an explosive volcanic complex, located about 15 km away from Naples (Italy). Moreover, the area is densely populated; and a possible Plinian eruption implies disastrous consequence. This situation has induced the European project COST Action C26 ‘Urban habitat constructions under catastrophic events’ to introduce the ‘Vesuvius case’ as a case study within its research activities, in the program of the Working Group 4 ‘Risk Assessment for Catastrophic Scenarios in Urban Area’. The main purpose was to examine the effects on the constructions, due to a probable Vesuvian eruption (Mazzolani et al., 2010). In the COST C26 research, typical constructions in the area have been assessed, through a detailed in situ survey, carried out with the contribution of the PLINVS Centre (Hydrological, Volcanic and Seismic Engineering Centre, Director prof. Giulio Zuccaro) and several experts belonging to various European Universities (members of the C26 Action). The town of Torre del Greco (the most populated city in the Vesuvius area with about 90,600 inhabitants), was selected for the investigation, out of the 18 municipalities surrounding the Vesuvius crater. Therefore, most of the pilot study areas are located in Torre del Greco, while the monumental buildings considered in the investigation are widespread along a line parallel to the coast, in different Vesuvian cities (Alterio et al., 2010; De Gregorio et al., 2010a; Florio et al., 2010). The building stock considered in the investigation is the following (Figures 1–4): • an important portion of the historic centre of Torre del Greco (281 buildings); • a decentralized residential area in Torre del Greco, located 4 km away from the vent (20 buildings); • 15 schools, distributed on the whole city territory, Torre del Greco; • 9 monumental buildings (located in Torre del Greco, Ercolano, Portici and San Giorgio a Cremano city), were also included in the investigation. These consist of historic Villas, and represent a small sample of an important portion of the Vesuvian cultural-architectural heritage, constituted by the 122 Golden Mile Villas (or Vesuvian Villas). The areas and buildings selected are representative of the Vesuvian urban environment in terms of exposure, comprising both ordinary buildings and strategic constructions. The identification of the construction types has been performed through a visual examination, accompanied by the compilation of an ad hoc form, taking into account the factors which affect the building vulnerability to volcanic effects. Afterwards, the collected data was used in the application of the specific methodology for the assessment of the volcanic vulnerability, which has been developed by the PLINIVS Centre within the EXPLORIS European project (2002–2005; EXPLORIS, 2006). Examples of classification are given in Figures 5–8. 204

Figure 1.

Pilot areas in Torre del Greco city.

Figure 2. Torre del Greco historical centre.

Figure 3. The primary school Mazza (Torre del Greco).

Figure 4a. The localization of the Golden Mile villas in the Vesuvian Area.

Figure 4b. Villa delle Ginestre, near Torre del Greco.

2.5.3.2 The vulnerability assessment methodology The ‘quick’ methodology for the volcanic vulnerability assessment and the survey form described, are proposed within EXPLORIS (EXPLORIS, 2006). A dynamic model was developed with reference to the Vesuvius case, with the scope of simulating the whole eruptive process. Therefore, an Event Tree framework can be determined, summarizing the potential eruption scenarios for the 205

Figure 5. Main vertical structures in the historic centre of Torre del Greco.

Figure 7. Percentage of openings on the building façades in the historic centre of Torre del Greco.

Figure 6. Roofs structure in the historic centre of Torre del Greco.

Figure 8. Conditions of windows in the historic centre of Torre del Greco.

next volcanic crisis of Vesuvius and the possible associated hazards which may develop (Neri et al., 2008). At the moment, EXPLORIS considers only three volcanic phenomena, earthquakes (EQ), ashfalls (AF) and pyroclastic flows (PF). The eruptive event is studied from the first precursory seismic event up to the final pyroclastic flow, by evaluating the damage accumulated on the buildings and the distribution of damage on the territory at each step of the process. The evaluation of the volcanic impact on the constructions is very complex and depends on the possible eruptive scenario, which has been assumed. The combination of the three volcanic phenomena can increase the damage on buildings, in comparison with the effects of each phenomenon acting separately. In fact, the sequence of events during the eruption causes a progressive reduction of the resistance properties of the buildings, in function of the temporal evolution of the damaging process. Therefore, some basic considerations and simple preliminary assumptions have been necessary in order to simplify such a complex task, also considering the great uncertainty in the definition of the load history derived from different eruptive scenarios, which can be identified by the Event Tree sets in EXPLORIS. First: in order to evaluate the possibility of simultaneous events, a preliminary distinction is necessary; they are either continuous (as AF which tends to occur during all the eruption) or discrete (as EQs which last for seconds or as PFs each pulse of which can last few minutes). Second: the events have a different probability distribution in space. For impact purposes, AF and EQ can be assumed to have an almost uniform distribution in the most-affected zones, so that a large number of buildings can be contemporary affected. On the contrary, the PFs are less uniformly distributed in space; nevertheless, they can be numerous, so that a considerable number of buildings would be involved. In addition, some simple assumptions can be made about the time distributions of the expected events, based on the assumed eruptive scenario. Hence, impossible or highly improbable combinations can be eliminated. 206

The final impact scenario can be examined by trying to parameterize the cumulative damage which the structure experiences from the possible sequence of events. The problem can be treated as a sort of progressive deterioration of the building’s resistance characteristics, which is essentially represented by the damage level. This requires the assumption of one damage scale as descriptor of the global structural damage for the different building classes. The seismic damage scale has been then assumed, intended to describe also the damage level caused by the other phenomena (AF and PF). In this regard, the vulnerability under the action of PF of non structural elements (NSE), like windows, doors, infill panels etc., has been taken into account separately, being compatible with the results obtained by Baxter et al. (2005a) and Spence et al. (2004). However, the consequences on the structure of the NSE failure, fire, roof explosion, casualty etc., have been considered in term of seismic equivalent global damage and they have affected the vulnerability curves evaluation (Zuccaro et. al., 2008). The total damage suffered by the buildings during the whole eruptive process is evaluated by using a computerised model prepared for this purpose. Therefore the building damage distribution caused by the sequence of seismic events in the area, by the accumulation of vertical load due to ash-fall out and by the lateral pressure consequent to the pyroclastic flows, can be simulated. The study area has been subdivided into cells, through the use of a circular grid centred on the crater which is adopted in the model developed in EXPLORIS. By using an automatic GIS script, the building inventory of the vulnerability classes in the single cell of the mesh is updated by shifting the class according to the damage, which has been caused by the previous event, and so on. 2.5.3.3 Field investigation and data collection The aim of the investigations was to collect the data and parameters influencing the volcanic vulnerability for each construction. For this purpose, the survey form, developed by PLINIVS and illustrated in Table 2, has been adopted. It is divided into the following sections: • the Identification section is intended to locate the building with reference to the geographical parameters of the region; • the General Information Section refers to the building type (ordinary building, warehouse, electrical station, etc.), destination (residence, hospital, school, etc.), use (fully used, partially used, not used and abandoned) and exposure (ordinary, strategic, exposed to special risk) of the construction; • the Condition Section refers to age and state of conservation of the structure (poor, mediocre, good and excellent) and typology of the finishes (economic, ordinary, luxury); • the Descriptive Characteristics Section refers to the number of total storeys starting from the lowest ground level, the number of floors above the ground, including the penthouse, the number of residential apartments, the presence of occupied or not basement, the height of the first storey, minimum and maximum heights up to the roof, the presence of barriers with height >2 m, the orientation (angle between the longest or the main façade and the North) and the position of the unit in the block; • the Structural Characteristics Section refers to the principal typology (reinforced concrete, masonry, wood, steel and mixed), primary vertical structures (sack masonry with or without reinforcements, hewn stones masonry, masonry or tuff blocks, RC frames with weak or resistant cladding, etc.), primary horizontal structures (timber floor, floor with steel beams, concretetile structures, vaults, etc.), geometry of the roofing (plane, single pitched, multi pitched and vaults), thickness of the walls and the curtain walls and typology of the curtain walls (tuff blocks or squared stones, concrete blocks, etc); • the Openings Section refers to the percentage of openings on the façade, the number of small, typical and large windows, their material (timber, PVC, aluminium or timber-aluminium, light steel and steel of security anti-intrusion type), their protection and their conditions (perfect, efficient, poor, bad or lack of windows); • the Interventions Section refers to the age and type of repairs (extraordinary maintenance, upgrading and retrofitting); • the Regularity Section refers to the regularity and distribution of curtain walls in plan and along the height, the type of the structure (single or two-directional frames, single or two-directional 207

Figure 9. Position in the aggregate: a) isolated, b) internal, c) external, d) internal corner.

walls and walls with frames), soft floor (pilotis on a part of the ground floor, totally open ground floor and intermediate soft storey) and possible presence of stocky beams or columns. These parameters define each building in terms of geometry, typology and importance and mainly measure the volcanic vulnerability of the construction itself. In particular, these parameters can be divided into two main sections. The first section provides information on the main vertical and horizontal structures, the regularity in plan and in elevation, the age and conservation of the construction, the number of storeys. These aspects are associated with the evaluation of the seismic vulnerability of buildings. The second section is specific to the building behaviour under the effect of an explosive eruption, referring to the roof structure typology, and the openings. The information on the type of the roof structure is associated with the collapse due to ash-fall deposits during an eruption. The information on openings, including opening shape, the size and the protection of the openings, is associated with the pyroclastic flows. This was reported in the case of the Montserrat eruption (Baxter et al., 2005b). The form in Table 2 is suitable for the collection of data related to the volcanic vulnerability assessment of ordinary constructions. However, in the case of the survey activity related to the historic monuments (Vesuvian Villas), updates to the form were necessary, since cultural heritage is of great importance and the information collected must be more detailed. In view of this, the updates indicated in Table 3, have been proposed. Reference was made to the Italian survey form for the cultural heritage damage (DPCM, 2006). The following sections are included. • The Location Section refers to the location of the building with reference to the site characteristics (flat land, peak, filling or inclined soil, depression), the urban context (urban centre, urban periphery, industrial or commercial area, historical centre), the surrounding infrastructures (pedestrian or vehicle access, access for heavy facilities, neighbourhood parking) and the presence of other risks (landslide, inundation, industrial or natural threats). • The Descriptive Characteristics section refers to the type of artistic heritage, as frescoes, mosaics, mouldings, tapestries, altars, statues, books, prints, paintings on different bases, furniture, furnishings and archaeological finds. • The Structural Characteristics section refers to the state of general conservation of vertical and horizontal structures and roofs, the presence of steel or RC tie beams, internal elements (arcades, lodges, inner courts) and material and constructive discontinuities. • The Interventions Section identifies the possible type of structural interventions, like extraordinary maintenance, upgrading, retrofitting, enlargement or raising. • The Regularity section indicated the plan layout, which can be rectangular, extended rectangular, L-shaped, C-shaped or with court yards.

2.5.3.4 Vulnerability classes The information obtained by the survey activity has been used for the evaluation of the vulnerability of the constructions with respect to the volcanic actions, through the methodology, proposed within EXPLORIS (EXPLORIS, 2006). In particular, the aim of the analysis was to analyse representative samples. The vulnerability analysis has been conducted considering two groups of buildings: the constructions of the residential area (281 independent structural unities) and the buildings of the residential area and the schools (54 independent structural unities). The methodology applied is based on the assignment of specific vulnerability classes with respect to each considered exceptional action, in function of the structural elements typology. In particular, under the effect of the seism, the classes are four, As , Bs , Cs , Ds with vulnerability decreasing respectively, according to the combinations of the horizontal and vertical structures indicated in Table 4. Under the effect of 208

Table 2. PLINIVS Centre survey form for the assessment of the volcanic vulnerability. Section

Datum

Identification

Progressive number of the block Number of the building in the block Type Destination Use Exposure

General information

Condition

Age State of conservation of the structure Typology of the finishes

Descriptive characteristics

Number of storeys Number of apartments Basement Occupied basement Height of first storey Minimum height Maximum height Fence Orientation Position

Structural characteristics

Principal typology Primary vertical structures Primary horizontal structures Geometry of the roofing Structure of the roofing Thickness of the walls Thickness of the curtain walls Typology of the curtain walls

Openings

Percentage of openings on the façade Number of small windows Number of typical windows Number of large windows Material of small windows Material of typical windows Material of large windows Protection of small windows Protection of typical windows Protection of large windows Conditions of the windows

Interventions

Type of repairs Age of repairs

Regularity

Regularity in plan Regularity along the height Distribution of curtain walls in plan Distribution of curtain walls along the height Distribution of the structure Arcade (soft floor) Squat element

the ash fall, the classes are five, Ar , Br , C1r , C2r , D with vulnerability decreasing respectively, in function of the roof typologies indicated in Table 5. Finally, under the effect of the pyroclastic flows, the classes are three for the masonry buildings, Ap , Bp , Cp , and three for the RC ones, Dp , Ep , Fp , with vulnerability decreasing respectively in each group, in function of the vertical and horizontal structures in Table 6. 209

Table 3. Survey form integration for the monumental building. Section

Datum

Location

Site characteristics Urban context Infrastructures Presence of risks

Descriptive characteristics

Type of artistic heritage

Structural characteristics

State of general conservation of vertical structures State of general conservation of horizontal structures State of general conservation of roofs Presence of steel or RC tie Internal elements Material and constructive discontinuities

Interventions

Type of interventions

Regularity

Plan layout

Table 4. Vulnerability classes under effect of the seism.

Vertical structure

Horizontal structure Poor Rigidity Vaults and/or wooden floor without ties

Poor Technology “SAP” floor

Medium Rigidity Vaults and/or wooden floor with ties

Medium High Rigidity Iron beam floor

High Rigidity Reinforced Concrete floor

Weak masonry Rubble masonry neglected (lavic stone, nor squared tuff, etc.)

As

As

As

As

As

Medium quality Rubble masonry maintained (lavic stone, nor squared tuff, etc.)

As

As

Bs

Bs

Bs

Good masonry Squared masonry (lavic stone, tuff, etc.)

As

As

Bs

Bs

Cs

Framed Structures (RC or steel)

–

Bs

–

–

Ds

With reference to the results of the survey activity, the two groups of the examined buildings present the distributions of the vulnerability classes indicated in the Tables 7a–c with respect the seism, the ash fall and the pyroclastic flows respectively. Under the effect of each considered action, an analysis of the results indicates that the buildings located in the historical centre have a higher vulnerability when compared to those in the residential area and the schools. In fact, in the historical centre in general, masonry buildings are more than 3 storeys high, with poor or medium rigidity horizontal structure and roofs consisting of timber, 210

Table 5. Vulnerability classes under effect of the ash fall. Class

Description

Ar

Weak pitched wooden roof

Br

Flat standard wooden roof RC flat roof–SAP type Weak steel with little vaults flat roof

C1r

Flat RC roof older than 20 years

C2r

Flat RC roof younger than 20 years Recent flat RC flat roof

Dr

Recent pitched RC roof Recent pitched steel roof

Table 6. Vulnerability classes under effect of the pyroclastic flows.

M

RC

Class

Description

Ap

Weak masonry buildings of 3-4 storeys with deformable floor Weak or strong masonry buildings with more than 4 storeys

Bp

Medium masonry buildings of 1-2 storeys with deformable floor Strong masonry buildings of 3 or more storeys with rigid floor

Cp

Strong masonry of 1-2 storeys with rigid floor

Dp Ep Fp

Non aseismic RC b. of more than 6 storeys Non aseismic RC buildings of 4-6 storeys Non aseimic RC buildings of 1-3 storeys

Table 7. Percentage distributions of the vulnerability classes of the historical centre and the residential area plus the schools, under effect of the seism (a), ash fall (b) and pyroclastic flows (c) Seism Historical centre Residential area and schools

As 36 2

Bs 48 11

Cs 16 63

Ds 0 24

(a)

Ash fall Historical centre Residential area and schools

Ar 1 2

Br 81 7

C1r 18 48

C2r 0 43

(b)

Pyroclastic flows Historical centre Residential area and schools

Ap 25 0

Bp 65 35

Cp 0 0

Dp 0 0

(c)

steel and SAP flat floor (Bs, Br, Bp). However the buildings in the residential area and the schools in general consist of good masonry construction or RC frame structures of 1-2 storeys with high rigidity RC floor (Cs, Ds, C1r, C2r, Fp). 2.5.3.5 Expected damage In this section, the expected damage obtained, is presented with reference to the specific adopted scenarios. For the historical centre, the considered volcanic event is going to develop according to the following phases. Before the eruption, three discrete seismic events (EQ) occur with an intensity of VI-VII-VIII degree of the European Macro-seismic Scale (EMS ‘98), respectively; while, during the eruption, continuous ash-falls (AF), whose deposits are increasing in time, are accompanied by one EQ event of moderate intensity (VI EMS) plus continuous low tremors. Approaching the end 211

Figure 10.

Radial sectors with 16 different wind profiles.

Table 8. Historical centre eruptive damage scenarios. Lost buildings n◦

Broken windows n◦

Sector

Actions

8

EQ+AF+PF EQ+AF

11 5

142 0

9

EQ+AF+PF EQ+AF

50 46

111 0

10

EQ+AF

188

0

of the AF phase and immediately afterwards, a series of three PF events occur, having durations of two minutes each, randomly distributed in time during the last eruption phase. The impact of ash-fall is strongly dependent by the wind direction during the eruption, so the model considers 16 possible ‘sectors’ for the direction of the prevalent wind (Macedonio et al., 2008), as illustrated in Figure 10. In particular, for the volcanic vulnerability of the considered pilot areas, according to Civil Protection Plan Simulation on the Vesuvian area, three impacts are evaluated, using three different sectors for the wind direction. They are the sector 10, which can cause more damage in the study area than other sectors, and the two sectors 8 and 9, which can cause very little damage. The whole eruption process has been simulated for each of these sectors. The results of impact model simulation are illustrated in Table 8. They show that the PF impact seems most relevant when the ash impact is less important with wind in sectors 8 and 9. This happen because the PF action is almost applied to the whole building set. On the contrary, when the wind direction is in sector 10, most parts of buildings are considered already ‘lost’ by ash-fall, so the PF action is applied just to few buildings. For the group of 54 buildings, in the case of the constructions of the residential area and the schools, the assumed exceptional actions are two earthquakes of VII and VIII degree (EMS ’98) and ash fall with wind in the direction of the sectors 8, 9 and 10 (Figure 10). These actions are considered to act independently. Under the effect of the two earthquakes, the damage levels are indicated in Table 9, with reference to the EMS’98 (D0 = no damage, D1 = light damage, D2 = moderate damage, D3 = severe damage, D4 = partial collapse, D5 = total collapse). According to the low seismic vulnerability of this group of constructions (As, Bs), the expected damage results to be low (D0–D3). Instead, the ash fall action can lead to failure of many constructions, especially with the main direction of the wind in Sector 9, where 53 buildings out of 54 can collapse (Table 10). 2.5.3.6 Air fall deposits model De Gregorio et al (2010b) carried out a specific analysis on air fall deposits due to explosive eruptions, with respect to the action model and robustness of the Vesuvian roofs. In particular, the 212

Table 9. Residential area and schools. Damage levels under effect of earthquakes (EQ) of VII and VIII degree of the buildings. Building with damage level Di [n◦ ] Actions

D0

D1

D2

D3

D4

D5

EQ (VII) EQ (VIII)

28 23

19 20

6 9

1 2

0 0

0 0

Table 10. Residential area and schools. Lost buildings under effect of ash fall (AF) with main direction in the sectors 8, 9 and 10. Sector

Actions

Lost buildings n◦

8 9 10

AF AF AF

38 3 53

study (focused on the analysis of a specific volcanic event constituted by the pyroclastic deposits, falling on the roofs due to gravity) refers to two important aspects. The first aspect concerns the proposal of a model of the action on the basis of a similitude between the air fall deposits and the snow load. In fact, as for the snow, the action produced by the air fall is a gravitational load, which depends on the slope and the exposure of roofs. Besides this, the thermal degradation of mechanical properties of materials also occurs, being produced by the high temperature of the clasts, with a range between 150 and 400◦ C. The second aspect considered, is the robustness evaluation against the air fall deposits of the most common roof types in the Vesuvian area, which are made of timber, steel and reinforced concrete. Accordingly, some mitigation systems have been identified.

2.5.4 THE VOLCANIC RISK ASSESSMENT The risk management of natural hazards including Volcanic action, is a complex issue often due to very significant potential consequences and substantial uncertainties. A framework for risk based decision making in the field of engineering is defined (Narasimhan et al., 2010) and reported in detail in Section 4.4 and 5.4.

REFERENCES Alterio L., De Gregorio D., Faggiano B., Di Feo P., Florio G., Formisano A., Mazzolani F.M., Cacace F., Zuccaro G., Borg R., Coelho C., Indirli M., Kouris L., Sword-Daniels V., 2010. Survey activity for the seismic and volcanic vulnerability assessment in the vesuvian area: the golden mile villas. Proceedings of COST Action C26 Final International Conference onUrban habitat construction under catastrophic events, Naples, 16–18 September 2010. Baxter, P.J., Cole, P.D., Spence, R., Zuccaro G., Boyd R. and Neri, A. 2005a. The impacts of pyroclastic density currents on buildings during the eruption of the Soufrière Hills volcano, Montserrat. Bulletin of Volcanology. 67: 292–313. Baxter, P.J., Cole, P.D., Spence, R., Zuccaro G., Boyd R. and Neri, A. 2005b. The impacts of pyroclastic density currents on buildings during the eruption of the Soufrière Hills volcano, Montserrat, Bulletin of Volcanology. 67: 292–313. Blong, R., 2003. Building damage in Rabaul, Papua New Guinea, 1994. Bulletin of Volcanology, vol. 65.

213

COST, 2006. COST, European COoperation in the field of Scientific and Technical research, Transport and Urban Development, COST Action C26: “Urban Habitat Constructions Under Catastrophic Events”, 2006–2010. De Gregorio D., Faggiano B., Florio G., Formisano A., De Lucia T., Terracciano G., Mazzolani F.M., Cacace F., Conti G., De Luca G., Fiorentino G., Pennone C., Zuccaro G., Borg R., Coelho C., Gerasimidis S., Indirli M., 2010a. Survey activity for the seismic and volcanic vulnerability assessment in Vesuvian area: the historical centre and the residential area in Torre del Greco. Proceedings of COST Action C26 Final International Conference onUrban habitat construction under catastrophic events, Naples, 16–18 September 2010. De Gregorio D., Faggiano B., Formisano A. & Mazzolani F.M., 2010b. Air fall deposits due to explosive eruptions: action model and robustness assessment of the Vesuvian roofs.Proceedings of COST Action C26 Final International Conference onUrban habitat construction under catastrophic events, Naples, 16–18 September 2010. Dobran F. 2007. Urban Habitat Constructions Around Vesuvius. Environmental Risk and Engineering Challenges. Proc. of COST Action C26 Seminar on Urban Habitat Constructions Under Catastrophic Events, Prague, 30–31 March 2007. DPCM, 2006. Survey of the cultural heritage damage. Model B. Official Journal 7.3.2006, n.55 (in Italian). EXPLORIS, 2006. Explosive Eruption Risk and Decision Support for EU Populations Threatened by Volcanoes (EXPLORIS). EU Contract n˚ EVR1-CT-2002-40026, 2001-2006. Faggiano B., Nigro E., De Gregorio D., Zuccaro G., Cacace F., 2010, Volcanic actions and their consequences on structures, Proceeding of the International Conference COST Action C26 Urban habitat constructions under catastrophic events, Naples, Italy, 16–18 September 2010. Florio G., De Gregorio D., Formisano A., Faggiano B., De Lucia T., Terracciano G., Mazzolani F.M., Cacace F., Conti G., De Luca G., Fiorentino G., Pennone C., Zuccaro G., Borg R.P., Coelho C., Gerasimidis S., Indirli M., 2010. Survey activity for the seismic and volcanic vulnerability assessment in the Vesuvian area: relevant masonry and r.c. school buildings in Torre del Greco. Proceedings of COST Action C26 Final International Conference onUrban habitat construction under catastrophic events, Naples, 16–18 September 2010. Macedonio, G., Costa, A. and Folch, A. 2008. Ash fallout scenarios at Vesuvius: Numerical simulations and implications for hazard assessment, Journal of Volcanology and Geothermal Research. 178: 366–377. Mazzolani, F.M., Faggiano, B. and De Gregorio, D. 2008. Actions in the catastrophic scenarios of a volcanic eruption. Proceeding of COST Action C26 Symposium on” Urban habitat construction under catastrophic events”, Malta, 23–25 October 2008. Datasheet n◦ 5.1: 449–467. ISBN 978-99909-44-40-2. Mazzolani F.M., Faggiano B., De Gregorio D., 2009a. The catastrophic scenario in explosive volcanic eruptions in urban areas. Proceeding of Protection of Historical Buildings, PROHITECH 09, Rome, Italy, 21–24 June. Vol. 2: 1529–1534. ISBN: 978-0-415-55805-1. Mazzolani F.M., Faggiano B., Formisano A., De Gregorio D., 2009b. Vulnerability evaluation of RC structures in the Vesuvian area. Proceeding of Protection of Historical Buildings, PROHITECH 09, Rome, Italy, 21–24 June. Vol. 2: 1523–1528. ISBN: 978-0-415-55805-1. Mazzolani F.M., Indirli M., Zuccaro G., Faggiano B., Formisano A. and De Gregorio D. 2009c. Catastrophic effects of a Vesuvian eruption on the built environment. Proc. PROTECT 2009, 2nd International Workshop on Performance, Protection & Strengthening of Structures under Extreme Loading, Shonan Village Center, Hayama, Japan, 19–21 August 2009. Mazzolani F.M., Faggiano B., Formisano A., De Gregorio D., Nunziata C., Mandara A., 2010a. Volcanic and tectonic earthquakes effects in the Vesuvian urban habitat. Proceeding of the International Conference 14th ECEE, European Conference on Earthquake Engineering, Ohrid, Republic of Macedonia, August 30–September 03. Paper n. 1179. (in press). Mazzolani F.M., Faggiano B., Formisano A., De Gregorio D., Indirli M. and Zuccaro, G. 2010b. Survey activity for the volcanic vulnerability assessment in the Vesuvian area: the ‘quick’ methodology and the survey.Proceeding of the International Conference COST Action C26 Urban habitat constructions under catastrophic events, Naples, Italy, 16–18 September 2010. MEDEA, 2005. Manuale di Esercitazioni sul Danno Ed Agibilità per edifici ordinari (User’s manual on damage and safety for ordinary buildings). http://gndt.ingv.it/Att_scient/ Molise2002/ San_Giuliano/ Strumenti%20di%20rilievo.pdf. Narasimhan H., Borg R.P., Cacace F., Zuccaro G., Faber M.H., De Gregorio D., Faggiano B., Formisano A., Mazzolani F., Indirli M., 2010, A framework and guidelines for volcanic risk assessment. Proceeding of the International Conference COST Action C26 Urban habitat constructions under catastrophic events, Naples, Italy, 16–18 September 2010. Neri A., Esposti Ongaro T., Macedonio G., De’ Vitturi M., Cavazzoni C., Erbacci G., Baxter P. 2007. 4D simulation of explosive eruption dynamics at Vesuvius. Geophysical Research Letters, vol. 34, L04309, doi: 10.1029/ 2006GL028597.

214

Rossetto T., Kappos A.J., Kouris L.A., Indirli M., Borg R.P., Lloyd T.O., Sword-Daniels V., 2010, Comparison of damage assessment methodologies for different natural hazards. Proceeding of the International Conference COST Action C26 Urban habitat constructions under catastrophic events, Naples, Italy, 16–18 September 2010. Spence, R. J. S., Antonios, P., Baxter, P. J., Coburn, A. W., White, M., Dayrit, M., & Field Epidemiology Training Team, 1996. Building Damage Caused by the Mount Pinatubo Eruption of June 15, 1991. Philippine Institute of Volcanology and Seismology & University of Washington Press, 1996. Spence R. J. S., Baxter P. J., Zuccaro G. 2004. Building vulnerability and human casualty estimation for a pyroclastic flow: a model and its application to Vesuvius. Journal of Volcan. and Geothermal Research 133 (2004) 321–343. Spence, R. J. S., Kelman, I., Baxter, P. J., Zuccaro, G., & Petrazzuoli, S. 2005. Residential building and occupant vulnerability to tephra fall. Natural Hazards and Earth System Sciences, vol. 5. Sword-Daniels V., Rossetto T., Twigg J., Johnston D., Wilson T., Cole J., Loughlin S., & Sargeant S., 2010. Review of the impacts of volcanic ash fall on urban environments. Proceedings of COST Action C26 Final International Conference on Urban habitat construction under catastrophic events, Naples, 16–18 September 2010. Zuccaro G., Ianniello D. 2004. Interaction of pyroclastic flows with building structures in an urban settlement: a fluid-dynamic simulation impact model. Journal of volcanology and geothermal research 133, 345–352. Zuccaro G., Cacace F., Spence R.J.S. and Baxter P.J. 2008. Impact of explosive eruption scenarios at Vesuvius, Journal of Volcanology and Geothermal Research. 178: 416–453. Zuccaro G. & Cacace F., 2010a. Seismic impact scenarios in the volcanic areas in Campania. Proceedings of COST Action C26 Final International Conference onUrban habitat construction under catastrophic events, Naples, 16–18 September 2010. Zuccaro G., & Leone M.F. 2010b. Building technologies for the mitigation of volcanic risk. Proceedings of COST Action C26 Final International Conference onUrban habitat construction under catastrophic events, Naples, 16–18 September 2010. Zuccaro G., Cacace F. & Nardone S., 2010c. Human and structural damage consequent to a Sub-Plinian like eruption at Mount Vesuvius. Proceedings of COST Action C26 Final International Conference onUrban habitat construction under catastrophic events, Naples, 16–18 September 2010. ZuccaroG., Cacace F. & Rauci M., 2010d. Vulnerability functions for building structures under pyroclastic flow actions. Proceedings of COST Action C26 Final International Conference on Urban habitat construction under catastrophic events, Naples, 16–18 September 2010.

215

Chapter 3: Evaluation of vulnerability of constructions

Urban Habitat Constructions under Catastrophic Events (Final Report) – Mazzolani (Ed). © 2010 Taylor & Francis Group, London, ISBN 978-0-415-60686-8

3.1 Vulnerability of existing buildings under fire E. Nigro & G. Cefarelli University of Naples “Federico II”, Naples, Italy

F. Wald Czech Technical University, Prague, Czech Republic

M. Hajpál ÉMI, Budapest, Hungary

R. Zaharia The “Politehnica” University of Timisoara, Romania

N. Lopes & P. Vila Real Universidade de Aveiro, Portugal

L. Kwasniewski & Z. Drabowicz Warsaw University of Technology, Warsaw, Poland

D. Pantousa Volos University, Greece

Edgaras Geda, Darius Bacinskas & Viktor Gribniak Vilnius Gediminas Technical University, Lithuania

M. Heinisuo Tampere University f Technology, Finland

3.1.1 STATE OF ART 3.1.1.1 Introduction The safety of existing buildings is a very topical issue. In general, in countries of the world, the national construction building codes are periodically updated in response to changing technology, new materials and products, and the changing needs of building occupants and the community at large. Besides they can provide changes in required level of performance and modifications of administrative provision. However, a building constructed according to a past building code probably will not be able to satisfy all the provisions of the new rules. This paper deals with the safety of the existing buildings with particular reference to the fire vulnerability. Moreover, the procedures that can be used in European Countries are deal with based on the contribution of the WG1 members. The safety of existing buildings is a very topical issue. In general, in countries of the world, the national construction building codes are periodically updated in order to improve the level of safety, health, welfare and property protection. However, a building constructed according to a past building code probably will not be able to satisfy all the provisions of the new rules. Therefore it is necessary to individuate the approach which designers and contractors should take for the modification, repair and addition of existing buildings. An example is provided by the International Existing Building Code (2006) of International Council Code. The aim of this code is to maximize the safety and health of the workers involved with the building modifications, repairs or additions as well as building users and those whose proximity to the building could be affected by its failure. 219

The International Existing Building Code classifies works on existing buildings into alteration levels 1, 2 and 3; change of occupancy; additions; historic buildings and relocated buildings. Specific requirements for each class of works is outlined under the headings of Special Use & Occupancy; Building Elements and Materials; Fire Protection; Means of Egress; Accessibility; Structural, Electrical, Mechanical, Plumbing and Other Requirements. Moreover, useful information on how existing buildings and building rehabilitation are regulated are provided by David B. Hattis, that in 1981 wrote about this subject on the Bulletin of the Association for Preservation Technology. The safety in case of fire is a significant aspect of safety of existing buildings. This aspect affects particularly the historic buildings. Fire has always been a threat to culturally valuable historic buildings and surroundings. Building construction works, day to day activities, events and exhibitions all create different degrees of risks. Moreover, historic buildings are often built from easily-ignited materials, see Built Heritage: Fire Loss to Historic Buildings of COST Action C17, 2006. 3.1.1.2 Existing building regulations The States currently regulate new and existing building by one or more of the following types of regulations: – – – –

Construction codes; Building maintenance codes (property maintenance, health, fire prevention); Past construction codes; Retroactive laws and regulations.

Construction codes are generally referred to as “building codes” and they regulate structural, fire, accident and health safety and everything is connected to provide a certain level of safety, health, welfare and property protection for building occupants and for general public. In general, in countries of the world, the national construction building code are periodically updated. The updating of codes represents a constant increase in the implied levels of safety, health, welfare, and property protection. The building code change periodically in response to changing technology, new materials and products, and the changing needs of building occupants and the community at large. The updating of codes, in general, consists of the substitution of references to materials and methods of construction no longer used in modern construction with the references to new materials and methods of construction. Besides they can provide changes in required level of performance and modification of administrative provision. However, a building constructed according to a past building code probably is not be able to satisfy all the provisions of the new rules. In many cases, it is impossible to make an existing building completely satisfying the provisions in force. A building code traditionally permits the continued use and occupancy of existing buildings at the time the code is adopted. Most building codes and property maintenance codes imply that an existing building and any required safety equipment and devices must be maintained at the level required by the code under which the building had been constructed. In fact, these past codes establish levels of health, safety, welfare, and property protection which are usually different from new codes, and are often lower than those of current new construction codes. However, in some cases states or local governments have declared certain building features to be unsafe or otherwise undesirable and have required that all buildings of a certain occupancy or class be altered to remove the unsafe or undesirable condition. In other cases they have required the installation of some specific features that contribute to a building’s increased safety (e.g. sprinkler, smoke detectors). All existing buildings covered by a retroactive regulation are required to be modified to conform to the new minimum provisions. The levels of health, safety, welfare, and/or property protection required by such retroactive regulations may be the same as, or lower than, the respective levels required by codes for new construction. The need to increase the levels of safety of existing buildings may be dictated by the need of repairs (for damaged buildings), change of use or occupancy and/or structural, by the need of applications of laws retroactive. A change of use or occupancy may introduce new or greater hazards in the existing buildings. A careful re-examination is required to determine that the building 220

will be safe for the new occupancy. Generally, the building codes require that the entire building comply with the new construction requirements for the new occupancy, although this is often stated in various ways, depending on which model code is being followed. 3.1.1.3 Fire vulnerability For the existing building a fire event may represent a severe condition because the existing buildings are generally designed without any structural or non-structural fire safety concept. Principal points of weakness in existing buildings have been repeatedly demonstrated in actual fires. The presence of open stairways and elevator shafts; the improper disposition or inadequate protection of combustible contents, construction, or interior finish; the lack of adequate means for restricting spread of fire; and the omission of devices that will give prompt notice of excessive temperatures are all familiar features of older types of buildings. Many others could be cited. Building construction works, day to day activities, events and exhibitions all create different fire scenarios and, therefore, different degrees of risks. Belong to the existing buildings also the historic buildings. They are often built from easilyignited materials. Therefore, fire has always been a threat to culturally valuable historic buildings and surroundings. In addition to the current serious levels of loss to life and contents, the number, authenticity and quality of European historic buildings is now recognized as being steadily eroded through the effects of fire but the full extent of this is unknown. Human factors, lit candles, open fires and chimneys in poor condition are also responsible for starting many incidents, as are lightning strikes. 3.1.1.4 Fire vulnerability reduction Each building owner is on constructive notice as to his obligation under a building code, and many owners will no doubt be willing to make necessary changes in existing buildings, once the necessity for doing so is appreciated. However, no general improvement can be counted upon without enforcement of minimum measures by public authorities. Such measures may take the form of adequate enclosure of elevator shafts; separating banks of elevators into not more than three in the same enclosure; adequate enclosure of stairways; a requirement that doors opening into an exit way shall be self-closing; adding stairways or other means of exit where provision of means of escape is deficient; subdividing excessively large areas; elimination of grills in exit ways; closing of movable transoms and substituting wired glass for plain glass in them; avoidance of ventilating systems (natural or mechanical) that exhaust air from assembly or sleeping rooms into exit ways; and providing suitable alarm and extinguishment devices. The extent to which such changes will be required will vary with the occupancy, greatest emphasis being placed on improving the protection to those occupancies where people are infirm or are confined, or where sleeping quarters are provided. Especially in need of attention are old hotels in small communities, buildings converted to multi-family occupancy, and farm residences and resort hotels, particularly where they are outside the fire fighting zone and are of substandard construction. 3.1.2 COMMON RULES FOR EXISTING BUILDING The building needs to be repaired or brought up to some minimum level of safety if it is dangerous, in accordance with safety code, or for enforce a property maintenance code, or for enforce a retroactive law or provision. The rehabilitation works, which do not change the building use or occupancy, cannot in any case reduce the existing fire safety level. In general, the building codes address two categories of building rehabilitation: a) maintenance, alteration and repair of existing buildings not involving a change of use or a change of occupancy; b) change of use or occupancy in existing buildings. The 25–50% Rule is commonly used by building codes as a means to control rehabilitation with no change of use or occupancy. The specific wording of the 25–50% Rule varies from code to code. Typically it requires the upgrading of existing buildings to the performance levels required for new construction if repair or alteration work exceeds 50% of the value of them building, and allows various lower performance levels to continue to exist in buildings when lesser work is involved. 221

Varying degrees of compliance with new construction requirements are specified for work between 25 and 50% of the value of the building, and for work below 25% of building value restoration with original materials is typically permitted. In case of replacement or modification of floor distribution and/or equipment of active protection fire-fighting, the partial modification of the construction characteristics and/or of the system of ways of exit, and/or enlargements, the provisions of code applies only to the installations and/or parts of the construction subject of amendments. The codes may provide two distinct approaches: prescriptive and performance-based. In the prescriptive approach the acceptable materials, sizes and methods of construction are prescribed in the code. In the performance-based approach any material, design or method of construction meeting the specified level of performance is acceptable. Codes today for existing building, if fire safety interventions are necessary, provided primarily non structural intervention (e.g., sprinklers, compartimentation, evacuation plan, …). The performance-based design (performance approach) and fire safety engineering concepts applied to existing buildings requires in general the consideration of several aspects: – assessment of the vulnerability of existing buildings to fire: e.g., fire resistance of the global structure and of more vulnerable structural members, as timber floors and roofs; – risk assessment methodologies; – protection of fabric and content; – prevention of fire and fire spread; – insurance considerations. The strange and criticized category “existing building being a risk for human life” refers mostly to the buildings which have insufficient capabilities for evacuation of occupants or have improper covering materials. Considering the consequences of the loss of a historic building, the risk analysis should include: – loss of economic value (in terms of providing a modern replacement of premises of the same quality as the building which has been lost); – loss of historic cultural and emotional value; – loss of a positive image for the local community; – loss of economic impact on the tourist industry; – additional costs for reconstruction. The special characteristics of historic buildings should be described and analyzed in the risk analysis to recognize the: – – – – – –

particular vulnerability of the building; activities taking place in the building; fabric of the building and its structural features; surroundings of the building and the activities that take place there; probability of fire ignition; length of time required for the fire brigade to arrive.

3.1.3 QUESTIONNARIE 3.1.3.1 Q&A A questionnaire on the existing building and fire design were sent to the European members of COST-C26 WG1. The questionnaire was composed of 22 questions. The questions were on the procedure, that the national code provide for existing building, in case of the change of the fire safety rules or of the purpose of the building. Today 8 countries answered to the questions. In Table 1 are reported the each countries answers to each questions. The results are presented at Action WG1 web page www.fsv.cvut.cz/ www/wald/COST_C26_Prague. 222

3.1.3.2 Comments on questionnaire answers In general, in much European Countries the continuous modifications of the law on the existing building currently created a complex and messy system, far away from the unification. However, Czech Republic have a CSN 73 0834 “Fire protection of buildings – Changes of buildings”, where are specified the condition of changing the purpose of the existing building including structural aspects. In general, codes today for existing building, if fire safety interventions are necessary, provided primarily non structural intervention (e.g., sprinklers, compartimentation, evacuation plan, …). The performance-based approach for structural fire safety is already adopted by several International Codes, as well as in Eurocodes. National fire code includes performance based design in Czech Republic, UK, Finland, Hungary and Italy. It is possible in Belgium if a derogation to the Fire Regulation is agreed on by decision of the Minister of Interior. In France it is possible to apply it partially for fire resistance and smoke propagation. Recent modifications of Polish law regarding building infrastructure and fire protection have introduced a category of “existing buildings being a risk for human life”. Only such existing buildings are subjected to the same requirements as those for newly constructed buildings. The regular design requirements are regulated by the National Building Code and Eurocodes, while requirements regarding applied materials, active fire protection, localization, evacuation are controlled by numerous national standards and regulations. The prescriptive method dominates. 3.1.4 FIRE OF BUDAPEST SPORT HALL SERVED AS MARKED Budapest Sport Hall was the largest covered arena dedicated to sport activities and cultural events in Hungary, see Figure 1. This steel reinforced concrete structure was constructed between 1978– 1982 and opened in 1982 It had a seating capacity of 7000 and an overall capacity of 12500 persons.

Figure 1. The structure of the Budapest sport stadium.

223

Figure 2. The plan view with the change to the Christmas market.

The main dimensions were: diameter at the footing 117 m and at the roof 127 m, high 26.5 m, place 12000 m2 . It were 5 floors: – – – –

under the arena were the cellar depots, permanent rink and exercise room; on the arena floor were the serving rooms for the competition; on the 1. and 2. floor were entrances for the public, toilets, buffet; on the 3. floor were press boxes, enclosure, place for pressmen, room for interpreters, bureau, monitoring rooms for the building; – on the 4. floor were mechanical equipments, 700 m2 exhibition room and here connected the steel structure to reinforced concrete ring. The primary use was temporarily changed to host markets and other trade events, see Figure 2. The fire initiated on the 18th of December 1999, when a Christmas Market took place within the covered hall. Kiosks and small pavilions made from combustible materials were installed in the auditorium, on the ground floor and on the first floor corridors and even between the stands. The estimated fire load of goods and kiosks was more than 3000 MJ/m2 . After the catastrophic event a damage survey was performed and it has proved that most of part of the structure suffered long-term heat flux and exposure to high temperature fire (>800–900◦ C), which exceeded the fire resistance of the structures. The highest fire exposure was detected at the structures of vault of the arena, resulting in the complete collapse of the roof, see Figure 3. The structural units of the 3rd floor “ring structure” of the Sport Hall have been damaged significantly since there was only a glass dividing structure with very limited fire resistance.

3.1.5 PROPOSED FURTHER DEVELOPMENTS In the European countries the continuous modifications of the laws on the existing building currently created a complex and messy system, far away from the unification. Therefore, it is necessary provide the unified European approach which designers and contractors should take to the modification, repair and addition of existing buildings as the International Existing Building Code. 224

Figure 3. The structure of the hall exposed to fire.

Figure 4. The structure of the hall after collapse.

REFERENCES International Council Code 2006. International Existing Building Code. COST Action C17, 2006. Built Heritage: Fire Loss to Historic Buildings. David B. Hattis, 1981. How Existing Buildings and Building Rehabilitation Are Regulated, Bulletin of the Association for Preservation Technology, Vol. 13, No. 2, Regulating Existing Buildings, pp. 9–12. Rules for fire safety in case of change of use of building: www.fsv.cvut.cz/www/wald/COST_C26_Prague.

225

Urban Habitat Constructions under Catastrophic Events (Final Report) – Mazzolani (Ed). © 2010 Taylor & Francis Group, London, ISBN 978-0-415-60686-8

3.2 Performance based evaluation and risk analysis E. Mistakidis Department of Civil Engineering, University of Thessaly, Volos, Greece

R. Vacareanu Department of Reinforced Concrete, Technical University of Civil Engineering, Bucharest, Romania

A.J. Kappos Department of Civil Engineering, Aristotle University of Thessaloniki, Thessaloniki, Greece

3.2.1 INTRODUCTION This chapter of the general report of the activities of the action COST-C26 concerns the topic of performance based evaluation of structures and the topic of risk analysis. Moreover, it contains information on performance related aspects in seismic vulnerability assessment. This is part of the activity developed within the Working Group 2 (WG2) of the Action, with the title “Earthquake resistance”. In this respect, first a short state of the art is given for topics covered and afterwards the activity of the action members is presented. This activity is generally documented by the proceedings of the main events of the Action, i.e. the workshop that took place in Prague in 2007, the seminar organized in Malta in 2008, and the final conference that was organized in Naples in 2010. Moreover, some interesting contributions that were given in the working group meetings of the Action are presented. For the interested reader, all the presentations of the WG2 members are available on the web (http://www.civ.uth.gr/cost-c26/index.files/WG2.htm). 3.2.2 PERFORMANCE BASED STRUCTURAL EVALUATION AND DESIGN 3.2.2.1 A short state of the art Major seismic events during the past years, such as those that occured in Northridge, California (1994), Kobe, Japan (1995), Turkey (1999), Taiwan (1999), Central-Western India (2001), L’Aquila (2009) and Chile (2010) have continued to demonstrate the destructive power of earthquakes. Every seismic occasion reminds the need to improve the seismic performance of the built environment through the development of advanced procedures and guidelines. Such procedures have been termed Performance Based Seismic Engineering (PBSE) in the literature and can be applied for the design of new structures but also for the evaluation of the seismic adequacy of the existing buildings stock. This matter of rehabilitation of existing buildings in regions of high seismicity has attracted the attention of researchers from various scientific areas for more than 30 years. Engineers practiced analytical techniques for a quick estimation of the performance of structures such as the capacity spectrum method (Freeman 1975,1987, Deierlein and Hsieh 1990). Later, Mahaney (Mahaney et al., 1993) introduced the acceleration-displacement response spectrum format which had the advantage of representing structural capacity and demand in terms of force and displacement on the same plot. These methods have application to low or midrise structures, in which the response is characterized by the fundamental mode of vibration. The reliability of these methods for structures in which higher modes of vibration are significant may not be adequate. During the past 10 years a number of methods have been presented, which are used today both for the seismic design of new structures and for the seismic evaluation of existing structures. As a result, various advanced methods have till now been proposed such as: – the adaptive pushover method (Bracci et al., 1997, Elnashai 2000) – the “N2” method (Fajfar and Fischinger, 1988, Fajfar and Gašperšiè, 1996) 227

– the incremental response spectrum analysis (Aydinoglu 2003) – modal and multimodal methods (Chopra and Goel 2002, Chopra et al., 2004, Sasaki et al., 1998) – upper bound pushover analysis (Jan et al., 2004) A very detailed state of the art is contained in FEMA 440 (FEMA, 2005). The following sections present the activities of the members of COST-C26 in this direction. 3.2.2.2 Advances in performance based evaluation and design within the context of the COST Action C26 The activity of COST-C26 within the four years period 2006-2010 is demonstrated by the papers and worksheets included in the proceedings of the three major events organized by the action, i.e. the Prague workshop, the Malta seminar and the Naples final conference. Moreover, in the various working group meetings organized by the action, useful presentations were given by the working group members. In the following, the output of the action in the field of performance based structural evaluation and design is presented. The report Typology of seismic motion and seismic engineering design (Mistakidis et al., 2007), summarizes the activity of the action in the corresponding topics. It contains two contributions related to performance based design. The first concerns a study on the magnification of seismic action on short period structures. The study was performed using a nonlinear SDOF oscillator subjected to various ground motions recorded in Greece. In order to cover various structural typologies, different force-displacement models were used. The study compared the results of the various nonlinear analyses performed with the formulas given in FEMA356 for the estimation of the target displacement using the Displacement Coefficient Method (DCM). In the second contribution, an interesting extension of the performance based design procedure was presented that considered uncertainties through the notion of fuzzy analysis. Data and models are uncertain in general, which has a significant influence on the results of the analysis. Therefore the uncertainty has to be described with suitable models and considered within the analysis. In the paper, informal and lexical uncertainties are described and quantified on the basis of fuzzy set theory with the aid of assessed intervals. In the field of seismic structural analysis, this framework gives the ability to treat deficits of information describing input variables, human mistakes and mistakes in fabrication, utilization and maintenance of structures, etc. After the presentation of the elements of the underlying theory, a simple application if presented in which the uncertainty in the input parameters affects the structural response. A fuzzy capacity curve is obtained, which is afterwards combined with the seismic demand to obtain a fuzzy target point with which all the relevant performance requirements should be checked. In the paper Performance-based seismic retrofit of masonry and R.C. buildings (Mandara et al., 2007), applications of performance based seismic retrofit for reinforced concrete and masonry buildings were presented. The R.C. frames were reinforced by eccentric steel braces and the masonry walls were strengthened by additional ties. All the analyses were carried out complying with the basic assumptions of the Performance Based Design. A damage-controlled nonlinear static procedure was defined to estimate maximum lateral displacement and plastic dissipated energy of RC frames, in order to keep damage indices in structural elements within tolerable limits at each performance level. These multi-level objectives were pursued with a procedure based on the Capacity Spectrum Method, the Inelastic Demand Response Spectra and the estimation of the duration-related damage, that is a function of the energy absorbed in the structure. The procedure was applied to regular RC frames strengthened with an eccentric bracing system composed of steel braces and vertically placed shear links, showing a great increase of stiffness with a minimal added weight. In the second part, dealing with masonry walls, the results of a study on the seismic upgrading of in-plane loaded masonry walls strengthened by metal tying techniques were presented. The calculations presented in the paper, show in quantitative way the significant improvement that can be attained with the use of steel ties, in terms of both strength and ductility. The keynote lecture of the seminar that took place in Malta with the title The performance based approach in seismic rehabilitation of Buildings (Pascu, 2008) presented a brief review and state of the art of Performance Based Seismic Design (PBSD) methods, focusing on the developments in the U.S. Moreover, the recent European Code EN 1998-3 and the Romanian Code P100-3 228

were presented and discussed. Some particular features of the PBSD were also discussed, as the demand assessment through incremental dynamic analysis, nonlinear response history analysis and nonlinear static procedures. Finally, an example of the rehabilitation design of a residential building in Bucharest was given. The paper Bucharest soil conditions and input ground motion for the structural performance analysis (Lungu et al., 2008) addressed the main problem of input ground motion for structural performance analysis. The study concerned the Bucharest soil conditions and was based on the available data obtained from more than 400 boreholes. The GIS techniques were applied in order to map significant soil parameters for the territory of Bucharest. The geological results, permit seismic microzonation of Bucharest to be used as a tool for urban planning and earthquake risk reduction. For this task, the results were correlated with shear wave velocity measurements in several locations having depth between 30 m and 200 m and with analysis of recorded strong earthquakes. The paper Design seismic response evaluation for 2D and 3D frames with flexible foundation using capacity spectrum method (Apostolska et al., 2008), addressed the important issue of generalizing the capacity spectrum method so that it incorporates the behaviour of the foundation, especially for the case of flexible ones. For this task, the influence of foundation flexibility on the capacity curve and on the capacity spectrum method as a whole was studied for 2D RC frames and 3D wall systems. For the 2D RC frames, the objects of investigations were 3-storey and 10-storey single frames. The frames were loaded with gravity load and the seismic effect was represented through the design response spectrum according to Eurocode 8. For the 3D structural wall systems, the model consisted of RC walls, columns and slabs. The results from the investigations showed that the largest values for the behaviour factor could be achieved if the fixed base is considered (Figure 1). The smallest target displacements are observed in the same case. If the foundations are flexible (Figure 2), the target displacement is increased but the behaviour factor decreases. When the soil is soft and the structure reaches the target displacement, the global behaviour of the structure may remain completely elastic. This mode of deformation implies that the soil fails before yielding happens in structure. Safe design solutions could be provided if soil deformations are taken into account in capacity curves.

Figure 1.

Capacity spectrum method applied to fixed-base structure: (a) in X-direction; (b) in Y-direction.

Figure 2. Capacity spectrum method applied to structure with flexible foundations (UFM 20 000 kN/m3 ) (a) in X-direction; (b) in Y-direction.

229

The paper Seismic performance of RC building structures with masonry infill (Apostolska et al., 2010) dealt with another interesting aspect of seismic evaluation and design which concerns the role of wall infills in structural response. Actually, the frames with infill are composite structures consist of bare frames (RC or steel) and infill which very often are not homogenous and have significant impact on the seismic performance of the structure. Experience from past earthquakes and results from experimental-analytical investigations have shown that the interaction between unreinforced infill and frame can cause either positive or negative effects. In case that the bare structural system has little seismic resistance, its effects are positive. On the other hand, if the contribution of masonry infill to lateral strength and/or stiffness is large relative to that of the frame itself, its effect is negative. In such a case the infill may override the seismic design and render ineffective the efforts of the designer to control the inelastic response by spreading inelastic deformation demands throughout the structure. In this study, 5 and 7 story RC buildings were studied, with and without infill. Both linear and nonlinear analysis were applied and the results were given in the form of shear forces and relative storey displacements and stiffnesses. The results showed that the effect of infill on the seismic performance of the structures is significant. In the elastic range and in the beginning of the nonlinear range, masonry infill increases the seismic resistance of RC systems and has a significant influence on the nonlinear dynamic response. Further research is needed for the definition of appropriate mathematical models for the nonlinear behaviour of infill. The paper Development and validation of a procedure of seismic performance evaluation of structures (Ayala et al., 2010) dealt with the approximate methods for the construction of the capacity curve that is being used for the evaluation of seismic performance of structures. A number of methods are reviewed as the capacity spectrum method, the displacement coefficient method and the N2 method. In the sequel, methods for the construction of the capacity curve are evaluated such as the incremental dynamic analysis, the static pushover analysis and the evolutive modal spectral analysis. The application of these methods is demonstrated through two numerical examples treating an 8-storey and a 17-storey regular buildings. The results verify that the evolutive modal spectral method gives results that compare well with those of the incremental dynamic analysis. Moreover, it indirectly considers the dissipation of energy due to hysteresis. This contribution entitled Performance based evaluation of seismic retrofitting techniques by Dogariu and Dubina, given during the meeting that took place in Aveiro (see http://www. civ.uth.gr/cost-c26/index.files/WG2.htm), presented a comparison between the Performance Based Seismic Assessment methods contained in the FEMA documents, Eurocode 8 and the recent Romanian code P100-3. Then, it shortly commented the numerical methods used for the evaluation of the structural response. Finally, it introduced the problem of masonry elements which are related to small deformability and lack of ductility. In these elements, the increase of resistance leads to an increase of rigidity and as a result to larger forces that can cause a sudden failure. On the other hand, an increase of the ultimate deformation and ductility may result to a stable post-cracking behaviour, to modification of the failure mode from shear to bending, or to a combination of them and a distribution of cracking to larger areas. In such structures, the reversibility and the minimum impact of the interventions is usually required, leading to much more complex decision making processes. The paper presented a decisional matrix that takes into account structural, technical and economical aspects. A number of methods were proposed, based on shear steel and aluminium panels and stainless steel and zinc coated wire meshes. Due to the large in-plane stiffness of masonry walls, the suggested metal based systems may not eliminate completely the damage to masonry. Therefore, it is reasonable to allow for limited and controlled damage to the masonry structure. Experimental results are presented from monotonic tests on masonry specimens, which establish the beneficial effect of the proposed reinforcing techniques that enhance the ductility, ensure the strength demand in the range between Life Safety and Collapse Prevention performance levels and allow a better prediction of the structural behaviour during earthquake. The authors discussed a case study which concerned an existing masonry building for which various retrofitting schemes were applied. The research study concluded to complete performance matrices for the considered schemes. Then, the decisional matrix presented earlier was used in order to consider together structural, technical and economical aspects and arrive to a certain decision regarding the optimum retrofitting scheme. The contribution A robustness based method for the validation of seismic retrofit by F. Mazzolani, A. Mandara, B. Faggiano, A. Formisano, A. Marzo given during the meeting 230

that took place in Aveiro (see http://www.civ.uth.gr/cost-c26/index.files/WG2.htm), introduced a robustness based method for the validation of seismic retrofit schemes. The method has a deterministic background and is based on the structural performance curve, expressed in terms of action and damage, obtained by means of any well established static nonlinear procedure. Given the damage level induced by a nominal performance demand, the robustness index is evaluated as a function of the action and the damage that occurs when the resistance of the structure is maximized. The method is then integrated in the typical multi-performance procedure that is used for the evaluation of existing structures. The notion of “Robustness capacity” is introduced to fill in the gap between Life Safety (LS) and Collapse Prevention (CP) performance levels, especially for the cases of very rare or catastrophic events. In this framework, various retrofitting schemes that were applied for the increase of the seismic capacity of existing structures were evaluated. The robustness index was calculated for the strengthening schemes of the ILVA-IDEM research project, the case of the Deutsche Bank in Naples and for the case of the Mustafa Pasha mosque in Skopje. The considered cases covered reinforced concrete, steel and masonry buildings. It was concluded that all the interventions were effective, providing robust structures under exceptional loading conditions.

3.2.3 PERFORMANCE-RELATED ASPECTS IN SEISMIC VULNERABILITY ASSESSMENT 3.2.3.1 Measures of performance in vulnerability analysis A realistic and usable definition of limit states (LS), usually referred to in the US as performance levels (PL), and a viable procedure for identifying them by visual inspection and/or by analytical methods, is at the heart of all assessment procedures and is also the basis for deriving fragility curves. As a general rule, LS are defined in terms of (acceptable) degree of damage and associated implications on the functionality of the structures and possible disruption of their use, with economic implications that can exceed the cost of repairing the structures. The proposed number of LS to be verified varies significantly in the various documents (whether code-type or not), depending mainly on the objectives and limitations of each work. One should also point out the difficulties in classifying damage in a consistent way using visual inspection, as opposed to the relative easiness in specifying a large number of ranges of analytical damage parameters and/or indices, each corresponding to a (conveniently named) LS. Due mainly to space limitations, the remainder of this section focusses on the definition of performance levels, both in analytical and in economic terms, within the framework of the so-called ‘hybrid’ method for seismic vulnerability assessment. The methods combines the use of statistical data from previous earthquakes with series of inelastic analyses of structures representative of particualr structural configurations. In this repsect, one can claim that the definitions of performance in this method is fairly rperesentative of what ona finds in other methods, i.e. purely empirical ones or purely analytical ones. More details on the various components of the hybrid procedure are given in another paper in this conference (Vamvatsikos et al.,), as well as in a number of studies cited later in this section. The basic idea is that a damage index, typically the ratio of repair cost to replacemnt cost for a damaged structure is estimated both from actual damage statistics (‘empirical’ data) and from the aforementioned series of inelastic analyses of representative structural systems. The ‘primary’ vulnerability curves (plots of degree of damage as a function of the earthquake intensity) are then obtained by appropriately weighting the empirical and the analytical data (Kappos and Panagopoulos 2010). Then this economic damage index is used to define a number of damage states (performance levels) for which fragility curves are derived using assumptions regarding the form of the type of probability distribution and the variability in the response. The performance-related aspects of this procedure are presented in the remainder of this section. 3.2.3.2 Performance-based estimation of economic loss To obtain the analytical prediction of the economic damage index (see §3.2.3.1), a series of inelastic dynamic analyses are run for the models representing the ‘generic’ buildings in each category of the 231

classification scheme used for vulnerability purposes (Kappos et al., 2006). From each analysis, the cost of repair (which is less than or equal to the replacement cost) is estimated for the building type analysed, using the models for member damage indices proposed by Kappos et al. (1998). The total loss (L) for the entire building is derived from empirical equations (calibrated against cost of damage data from Greece)

where Dc and Dp are the global damage indices (≤ 1) for the R/C members and the masonry infills of the building, respectively. Due to the fact that the cost of the R/C structural system and the infills totals less than 40% of the cost of a (new) building, the above relationships give values up to 38% for the loss index L, wherein replacement cost refers to the entire building. In the absence of a more exact model, situations leading to the need for replacement (rather than repair/strengthening) of the building are identified using failure criteria for members and/or storeys, as follows: – In R/C frame structures, failure is assumed to occur (and then L = 1) whenever either 50% or more of the columns in a storey ‘fail’ (i.e. their plastic rotation capacity is less than the corresponding demand calculated from the inelastic analysis), or the interstorey drift exceeds a value of 4% at any storey (Dymiotis et al., 1999). – In R/C dual (wall+frame) structures, failure is assumed to occur (L = 1) whenever either 50% or more of the columns in a storey ‘fail’, or the walls (which carry most of the lateral load) in a storey fail, or the interstorey drift exceeds a value of 2% at any storey (drifts at failure are substantially lower in systems with R/C walls). This set of failure criteria proposed by Kappos et al. (2006) resulted after evaluating a large number of inelastic time-history analyses. Although they represent the writer’s best judgement (for an analysis of the type considered herein), it must be kept in mind that situations close to failure are particularly difficult to model, and all available procedures have some limitations. For instance, although in most cases the earthquake intensity estimated to correspond to failure (damage state 5 in Table 1) is of a reasonable magnitude, in some cases (in particular wall/dual structures, especially if designed to modern codes) PGAs associated with failure are unrealistically high and should be revised in future studies. Having said this, their influence in a risk analysis is typically limited, since the scenario earthquakes do not lead to accelerations more than about 1 g. 3.2.3.3 Definition of damage for the derivation of fragility curves Assuming a lognormal distribution (common assumption in seismic fragility studies), the conditional probability of being in or exceeding, a particular damage state dsi , given the peak ground acceleration (PGA) is defined by the relationship

where: – PGA,dsi is the median value of peak ground acceleration at which the building reaches the threshold of damage state, dsi , see Table 1. – βdsi is the standard deviation of the natural logarithm of peak ground acceleration for damage state, dsi , and –  is the standard normal cumulative distribution function. Each fragility curve is defined by a median value of peak ground acceleration that corresponds to the threshold of that damage state and by the variability associated with that damage state; these two quantities are derived as described in the following. Median values for each damage state in the fragility curves are estimated for each of the building systems analysed. The starting point for estimating these values is the plot of the damage index (calculated from inelastic time history analysis as described in Kappos et al., 2006) as a function 232

Table 1. Damage grading and loss indices (% of replacement cost) for R/C buildings. Damage State

Damage state label

Range of loss index -R/C

Central index (%)

DS0 DS1 DS2 DS3 DS4 DS5

None Slight Moderate Substantial to heavy Very heavy Collapse

0 0–1 1–10 10–30 30–60 60–100

0 0.5 5 20 45 80

Figure 3. Evolution of economic damage (loss) index for medium-rise (left) and high-rise (right) buildings with R/C frame system designed to moderate codes.

of the earthquake intensity (PGA); some plots of this type are given in Figure 3 and they refer to buildings with dual system designed to moderate codes. Several trends can be identified in the figure, for instance that the least vulnerable building is the fully infilled one, with the exception of very low PGA values, for which the loss is higher than in the other two types; this is mostly due to damage in the masonry infills, which is accounted for in the loss model used (Kappos et al., 1998). Having established analytically the loss index L, the final value to be used for each PGA in the fragility analysis depends on whether an empirical value is available for that PGA or not. 3.2.3.4 Displacement-based definition of damage in hybrid fragility curves The hybrid methodology can be used to calculate vulnerability (fragility) curves for URM buildings in terms of spectral displacement. When appropriate capacity curves are available, the straightforward procedure (used in HAZUS) to derive fragility curves consists in defining damage states in terms of structure displacements (typically top-storey drift) and transforming these into displacements of the equivalent SDOF system, i.e. spectral displacements; these are then used as the mean values of the lognormal distribution defined for each damage state. The corresponding variabilities (β values) can be estimated in a way similar to that used for R/C structures. Instead of using semiempirical interstorey drift values (the HAZUS approach), the Thessaloniki group (Kappos 2001, Kappos et al., 2006) has suggested expressing the damage state thresholds in terms of the basic parameters of the capacity curve (yield displacement and ultimate displacement, both referring to a bilinearised capacity curve); this proposal is shown in Table 2. It should be clear that, depending on the height of the building and the failure mechanism, Sdy and Sdu values vary for each building type. Although straightforward, the aforementioned procedure cannot be directly integrated within the hybrid approach. For the latter to be materialised, one possible way is to define damage states in terms of the loss index, already employed in the case of R/C structures. Four damage states (plus the no-damage state) are proposed for URM buildings, defined according to the loss index (L) shown in Table 2; note that the range of L for each state is different from that used for R/C buildings (Table 1). To correlate these damage states to an analytical expression of damage, the loss index is 233

Table 2. Damage states in terms of displacements, and associated loss indices (%), for URM buildings.

Figure 4.

Damage State

Damage state label

Spectral displacement

Range of loss index

DS0 DS1 DS2 DS3 DS4 DS5

None Slight Moderate Substantial to heavy Very heavy Collapse

<0.7Sdy 0.7Sdy ≤ Sd < Sdy Sdy ≤ Sd < 2Sdy 2Sdy ≤ Sd < 0.7Sdu 0.7Sdu ≤ Sd < Sdu >Sdu

0 0–4 4–20 20–50 50–100

Economic loss index in URM buildings, as a function of roof displacement.

expressed as a function of yield and ultimate displacement of each building as shown in Figure 4; this model is based on the definitions of damage in terms of spectral displacement shown in the third column of Table 2, but recognising that for > 0.9 u , a URM building should be replaced (L = 100%) rather than repaired. 3.2.4 PROBABILISTIC SEISMIC RISK ASSESSMENT 3.2.4.1 A short state of the art The probabilistic seismic risk assessment explicitly takes into account the uncertainties in the basic variables involved in the analysis. Both aleatory and random uncertainties can be considered in the probabilistic risk analysis. The outcome of the analysis describes the mean annual frequency of a certain amount of losses for a given structural system (or an extended built system). Based on this result, rational decisions on seismic risk reduction can be made. The probabilistic seismic risk analysis integrates the results of the probabilistic seismic hazard analysis and of the seismic fragility/vulnerability analysis. The seismic hazard is a characteristic of the earthquake that might produce structural damage or losses. The outcome of a probabilistic seismic hazard analysis is the mean annual frequency with which a seismic hazard will occur. The seismic fragility describes the probability of reaching or exceeding a considered level of seismic damage for a structural system given the level of the seismic hazard. The seismic fragility might be regarded as a cumulative distribution function of the seismic capacity of the building. The seismic vulnerability describes the probability of reaching a considered level of seismic losses given the level of the seismic hazard. Risk is the expectancy of damage and/or losses or of other negative future happenings derived on the basis of present knowledge. The outcome of a probabilistic seismic risk analysis is the mean annual frequency with which a certain level of damage and/or loss will occur for a given structural system (or for an 234

extended built system). The risk analysis recognizes basically the impossibility of deterministic prediction of events of interest, like future earthquakes, exposure of elements at risk, or chain effects occurring as a consequence of the earthquake-induced damage. Since the expectancy of losses represents the outcome of a more or less explicit and accurate predictive analysis, a prediction must be made somehow in probabilistic terms, by extrapolating or projecting into the future the present experience. A probability-based prediction must be grounded in the conceptual and methodological framework of the theory of probabilities. Probabilistic seismic risk is the outcome of the convolution of seismic hazard, exposure of elements at risk and vulnerability of the elements at risk, using the total probability theorem. In the most general format, the general relation for the determination of the total risk can be expressed as (Whitman & Cornell, 1976):

in which P[ ] signifies the probability of the event indicated within the brackets, Ri denotes the event that the state of system is i, Sj means that the seismic input experienced is level j, and P[Ri /Sj ] states the probability that the state of the system will be Ri given that the seismic input Sj takes place. A more specialized format for the evaluation of the probability of failure of a structural system due to seismic ground motions is obtainable, given the following assumptions: – the ground motion parameter, a is a continuous variable; – the structural system condition is either safe or failed; – both the fragility curve and the seismic hazard curve are conditioned on the same ground motion parameter, a. The ground motion parameter might be the peak ground acceleration, PGA or the spectral response at a frequency of interest in the analysis. The fragility curve describes the probability of being in or exceeding an adverse state (i.e. structural failure) given the intensity of the ground motion parameter. The probability of failure of a structural system due to seismic ground motion is (McGuire, 2004), Figure 5:

where: – HA (a) – mean annual frequency of exceedance of ground motion amplitude, a; – PF|a – the fragility characteristic describing the probability of failure given the ground motion amplitude, a. Relation (5) can be reshaped in the following form (Reed & Kennedy, 1994), Figure 6:

Relations (5) and (6) resemble the basic structural reliability/safety problem in which the structural resistance, R (the capacity) and the structural effect of the loads, S (the demand) are convoluted in order to get the probability of a limit state violation. For the seismic case, the structural resistance is replaced by the structural fragility and the structural effect of the loads is replaced by the seismic hazard. Sometimes the solution is not that straightforward, because the structural damage correlates better with other seismic hazards (i.e. spectral response at a frequency of interest in the analysis). If this is the case, the probabilistic seismic risk analysis of a given structural system can be obtained as follows (again, by using the total probability formula) (Cornell & Krawinkler, 2000):

235

Figure 5. Convolution of seismic hazard (mean annual frequency of exceedance of PGA) and seismic fragility (conditional probability density function, PDF).

Figure 6. Convolution of seismic hazard (mean annual frequency of reaching PGA) and seismic fragility (cumulative distribution function, CDF).

where: – P(≥ ds ) is the annual probability of exceedance of damage state ds ; – (≥d s | Sd) is the standard normal cumulative distribution function of damage state ds conditional upon spectral displacement Sd; – f(Sd | PGA) is the probability density function of spectral displacement Sd given the occurrence of peak ground acceleration PGA, and – f(PGA) is the probability density function of PGA. One can change relation (7) for obtaining the mean annual rate of exceedance of various damage states for a given structural system:

236

Figure 7. Variation of spectral displacement conditional upon PGA (Vacareanu et al., 2003).

where: – λ(≥ds ) is the mean annual rate of exceedance of damage state ds ; – λ(PGA) is the mean annual rate of occurrence of PGA; – P(Sd |PGA) is the probability of reaching spectral displacement Sd given the occurrence of PGA (Figure 7), and – P(≥d s | Sd ) is the probability of exceedance of damage state ds conditional upon spectral displacement Sd . Thus, in this case, the probabilistic assessment of seismic risk involves three steps: (i) probabilistic seismic hazard assessment, λ(PGA); (ii) probabilistic assessment of seismic structural response, P(Sd|PGA), and (iii) probabilistic assessment of seismic structural vulnerability, P(≥d s |Sd). A closed-form analytic expression for the mean annual frequency of exceeding a damage measure and/or a structural limit state can be derived based on certain simplifying assumptions. The expression for mean annual frequency of exceedance of a limit state is derived by taking into account the random uncertainty in three main elements: seismic hazard, structural response (as a function of ground motion intensity) and capacity (Jalayer and Cornell, 2003). The hazard corresponding to a specific value of the seismic ground motion intensity measure (here, peak horizontal ground acceleration) is defined as the mean annual frequency that the intensity of future ground motion events are greater than or equal to that specific value x and is denoted by HPGA (x). It is advantageous to approximate such a curve in the region of interest for the analysis by a power-law relationship:

For a given level of peak horizontal ground acceleration, there will be variability in the displacement-based demand results over any suite of ground motion records applied to the structure. It is assumed here that this variability is a result of randomness in the seismic phenomena. Based on the results of inelastic dynamic analysis, IDA, it is convenient to introduce a functional relationship between the peak horizontal ground acceleration and the median value of maximum interstory drift (Jalayer and Cornell, 2003):

The mean annual frequency, MAF of exceedance of a limit state, LS is derived as (Jalayer and Cornell, 2003) (relation is adapted for peak ground acceleration):

237

where: – k0 and k – parameters defining the shape of the hazard curve, given in relation (9); – a, b – parameters defining the relation of between maximum interstory drift angle and peak horizontal ground acceleration, given in relation (10); – βD|PGA – conditional standard deviation of the natural logarithm for the displacement-based demand given peak horizontal ground acceleration; – ηC – median value of the limit state threshold (capacity); – βC – standard deviation of the natural logarithm for the limit state threshold (capacity). Consequently, the limit state frequency HLS is equal to the hazard function HPGA evaluated at the peak horizontal ground acceleration corresponding to the median drift capacity times two magnifying coefficients accounting for the aleatory uncertainties (randomness) in drift demand for a given peak ground acceleration and the aleatory uncertainties (randomness) in drift capacity itself. The seismic fragility functions can be obtained using the methodology described in (FEMA – HAZUS, 2003). Methods for determining the probability of Slight, Moderate, Extensive and Complete damage to general building stock designed to earthquake resistant seismic codes or not seismically designed are developed and presented in HAZUS ®MH MR4 Technical Manual. The probability distribution of the previously mentioned damage states is defining the building fragility/vulnerability function (curves). The fragility curves describe the probability of reaching or exceeding different states of damage given peak building response. The probability of being in or exceeding a given damage state is modeled as a lognormal cumulative distribution function. For structural damage, given the spectral displacement, Sd, the probability of being in or exceeding a damage state, ds, is modeled as (FEMA – HAZUS, 2003):

where: – Sd,ds is the median value of spectral displacement at which the building reaches the threshold of the damage state, ds, – β ds is the standard deviation of the natural logarithm of spectral displacement of damage state, ds, and –  is the standard normal cumulative distribution function. The median values of spectral displacement at which the building reaches the threshold of the damage state, ds, as well as the standard deviation of the natural logarithm of spectral displacement of damage state, ds, are provided in HAZUS ®MH MR4 Technical Manual for 36 model building types considered representative for the building stock in USA. 3.2.4.2 Earthquake losses Human and economic losses are the most important features of earthquake-induced phenomena. Sometimes the economic burden and pressure induced by the consequences of an earthquake disaster caused irreparable economic crisis for poor countries. Table 3 presents a combination of human and economic losses for earthquakes where monetary evaluations were available (Coburn & Spence, 2002, www.usgs.gov). It is therefore stringent need to make decisions towards the goal of seismic risk reduction. Different strategies may be taken to mitigate earthquake disasters, based on appropriate risk assessment. The following mitigation strategies might be considered: – Correlating land use planning with seismic hazard zonation; – Regulating the earthquake resistant design of buildings and structures and enforcing the seismic codes; – Strengthening, or removing unsafe buildings and structures; – Enhancing critical lifelines and facilities; 238

Table 3. Human and economic losses produced by earthquakes in 20th century. Date UTC

Location

Deaths

Losses ($bn)

Magnitude

1 2 3 4 5 6 7 8 9 10

1963 July 26 1972 Dec 23 1976 Feb 4 1976 Jul 27 1977 Mar 4 1979 Apr 15 1980 Nov 23 1985 Sep 19 1986 Oct 10 1988 Dec 7

1,070 5,000 23,000 255,000 1,500 101 4,680 9,500 1,000 25,000

0.98 2 1.1 6 2.0 4.5 45 5 1.5 17

6.2 6.2 7.5 8 7.2 7 7.2 8.1 5.5 7

11 12 13

1989 Oct 17 1990 Jun 21 1990 Jul 16

63 40,000 1,621

8 7.2 1.5

6.9 7.7 7.8

14 15 16 17 18

1994 Jan 17 1995 Jan 16 1999 Jan 25 1999 Aug 17 1999 Sep 20

FYROM, Skopje Nicaragua, Managua Guatemala China, Tangshan Romania, Vrancea Montenegro Italy, southern Campania Mexico, Michoacan El Salvador Turkey-USSR border region Spitak, Armenia Loma Prieta Western Iran, Gilan Luzon, Philippine Islands Northridge Japan, Kobe Colombia Turkey Taiwan

57 5,502 1,185 17,118 2,297

30 82.4 1.5 20 0.8

6.8 6.9 6.3 7.6 7.6

Human losses in Table 3 are represented as a function of magnitude in Figure 8. In Figure 9 human losses are represented versus economic losses, also based on data in Table 3. Based on data in Table 3, the number of deaths from an earthquake can be related to the magnitude of the earthquake by the following relations (Vacareanu et. al., 2004):

where – D is the number of deaths, and – M is the magnitude of the earthquake. The economic losses can be related to the number of deaths from an earthquake by the following relations (Vacareanu et al., 2004):

where – L are the economic losses expressed in billion US$, and – D is the number of deaths.

3.2.4.3 Case studies The application of the methodologies for performing probabilistic risk assessment shown in 3.2.3.2 are given in the datasheet and paper Seismic collapse risk of precast industrial buildings with strong connections (Fischinger et al., 2009), the datasheet of R. Vacareanu, D. Lungu, A. Aldea, C. Arion, Risk Analysis and paper Seismic Fragility of High-Rise RC Moment-Resisting Frames. Estimation of Drift Hazard (Vacareanu et al., 2006). 239

Figure 8.

Human losses as a function of magnitude.

Figure 9.

Human versus economic losses caused by earthquakes.

The paper Seismic collapse risk of precast industrial buildings with strong connections by M. Fischinger, M. Kramar, T. Isakoviæ (Fischinger et al., 2009) presents a systematic seismic risk study that has been performed on some typical precast industrial buildings that consists of assemblages of cantilever columns with high shear-span ratios connected to an essentially rigid roof system with strong pinned connections. These buildings were designed according to the requirements of Eurocode 8. The numerical models and procedures were modified in order to address the particular characteristics of the analyzed system. They were also verified by pseudodynamic and cyclic tests of full-scale large buildings. The intensity measure (IM )-based solution strategy described in the PEER methodology was used to estimate the seismic collapse risk in terms of peak ground acceleration capacity and the probability of exceeding the global collapse 240

limit state. The effect of the uncertainty in the model parameters on the dispersion of collapse capacity was investigated in depth. Reasonable seismic safety (as proposed by the Joint Committee on Structural Safety) was demonstrated for all the regular single-storey precast industrial buildings addressed in this study. However, if the flexural strength required by EC8 was exactly matched, and the additional strength, which results from minimum longitudinal reinforcement, was disregarded as well as large dispersion in records was considered, the seismic risk might in some cases exceed the acceptable limits. Detailed information on the procedure for computing MAF of exceedance of a limit state, specialized for drift hazard, as well as numerical results, are given in the paper Seismic Fragility of High-Rise RC Moment-Resisting Frames. Estimation of Drift Hazard by R. Vacareanu, P. Olteanu and A. B. Chesca (Vacareanu et al., 2006). The case study refers to a high-rise reinforced concrete moment-resisting frame structure designed according to the earthquake resistant design code in force in Romania, P100-1/2006, that is in line with the provisions of EN 1998-1. The seismic motion Intensity Measure is peak horizontal ground acceleration, PGA. The seismic motions used in the Inelastic Dynamic Analysis, IDA consist of nine suites (classes) of random processes comprising ten samples each. Target elastic acceleration spectra are used to generate acceleration time-history samples. For parametric analysis purpose, the accelerograms are artificially generated at predefined values of PGA. The computer program IDARC 2D (Valles et al., 1996) is used for performing nonlinear dynamic analyses. The expression for mean annual frequency of exceedance of Damage Measure is derived by taking into account the uncertainty in three main elements: seismic hazard, structural response (as a function of ground motion intensity) and capacity. Based on the results of IDA one can derive the mean annual frequency that the displacement-based demand (Damage Measure) (e.g., maximum peak interstory drift θmax) exceeds a given value d, also referred to as the “drift hazard”. The hazard corresponding to a specific value of the peak horizontal ground acceleration is defined as the mean annual frequency that the intensity of future ground motion events are greater than or equal to that specific value x and denoted by HPGA (x). It is advantageous to approximate such a curve in the region of interest by a power-law relationship. The hazard curve for Bucharest site due to Vrancea subcrustal seismic source can be expressed in approximate analytical form as (Vacareanu et al., 2002) (for PGA expressed in fractions of g):

For parametric IDA the accelerograms are generated at predefined values of PGA, as follows: 0.064 g, 0.08 g, 0.10 g, 0.125 g, 0.16 g, 0.20 g, 0.25 g, 0.32 g, 0.40 g (g – acceleration of gravity). The acceleration time-history samples are generated such as to fit target response spectra compatible with earthquake resistant design regulations in force in Romania, Figure 10. The structural system of the building analysed is a high-rise reinforced concrete moment resisting frame type designed according to the provisions of the earthquake resistant design codes in force in Romania. It is a thirteen-storey building; the height of the first two storeys (groundfloor and first story) is of 3.60 m each and all the remaining stories of 2.75 m height each. The total building height above the grade is 37.45 m. The building has two spans of 6.00 m each in the transversal direction and five spans of 6.00 m each in the longitudinal direction. The structural model is implemented in SAP2000 computer program (CSI Berkeley). The lateral load-resistant structural system plan at the groundfloor level is given in Figure 11. The concrete is of class C32/40 and the steel is of quality S355. Details on the structural system of the building can be found elsewhere (Vacareanu et al., 2006). The total weight of the building is 63070 kN. The first periods of vibration are 1.3 s for longitudinal translation, 1.2s for transversal direction and 1.07 s for general torsion. The IDA is performed on the structural model in the transversal direction consisting of six – twospan - reinforced concrete moment resisting frames acting together due to the action of horizontal diaphragms located at each storey. Nonlinear dynamic analyses are performed for 10 acceleration time-histories generated at each value of PGA (overall, 90 nonlinear dynamic analyses). Figure 12 shows the results of the IDA, e.g. maximum interstory drift, D versus PGA. For a given level of peak horizontal ground acceleration, there will be variability in the displacement-based demand results over any suite of ground motion records applied to the structure. It is assumed here that this variability is a result of randomness in the seismic phenomena. It is convenient to introduce a functional relationship between the peak horizontal ground acceleration and the median value 241

Figure 10. Target spectrum and response spectra of generated accelerograms at 0.24 g.

Figure 11.

Structural system – groundfloor level.

of maximum interstory drift, as it is stated by relation (10). The coefficients in relation (10) are: a = 20.929 and b = 1.746. The mean annual frequency, MAF that the displacement-based demand exceeds a given value d, also referred to as the drift hazard, is computed. The hazard curve for the drift demand HD (d) is equal to the hazard function HPGA evaluated at the peak horizontal ground acceleration corresponding to this drift demand times a (magnifying) factor related to the dispersion in the drift demand for a given peak horizontal ground acceleration, as it is given by relation (10) specialized for drift hazard. The drift hazard curve values derived from Equation (11) for given maximum interstory drift values are given in Figure 13. The application of IDA has been demonstrated for a high-rise RC moment-resisting frame structure. The approach proposed by (Jalayer and Cornell, 2003) to derive the mean annual frequency that the displacement-based demand exceeds a given value d has been followed and drift hazard values have been obtained. Only aleatory (random) uncertainties are considered in this paper. The additional epistemic (knowledge) uncertainties, not considered here, will increase the drift hazard. 242

Figure 12. IDA results: Maximum interstory drift (median values) versus peak horizontal ground acceleration.

Figure 13.

Hazard curve derived for maximum interstory drift values.

REFERENCES Apostolska, R., Bonev, Z. P., Blagoev, D., Vasseva, E. & Necevska-Cvetanovska G. 2008. Design seismic response evaluation for 2D and 3D frames with flexible foundation using capacity spectrum method, in: Mazzolani F.M., et al. (eds.) (2008). Urban Habitat Constructions under Catastrophic Events; Proceedings of the International Symposium, Malta, 23–25 October 2008. University of Malta. Apostolska, R., Necevska-Cvetanovska, G. & Cvetanovska, J. 2010. Seismic performance of RC building structures with masonry infill, in: Mazzolani F.M. et al. (eds.) (2010). Urban Habitat Constructions under Catastrophic Events; Proceedings of the Final Conference in Naples, 16–18 September 2010. Ayala, A.G., Mendoza, M. & Apostolska, R. (2010). Development and validation of a procedure of seismic performance evaluation of structures, in: Mazzolani F.M. et al. (eds.) (2010). Urban Habitat Constructions under Catastrophic Events; Proceedings of the Final Conference in Naples, 16–18 September 2010. Aydino˘glu, N.M. 2003. An incremental response spectrum analysis procedure based on inelastic spectral displacements for multi-mode seismic performance evaluation. Bulletin of Earthquake Engineering 1:3–36. Bracci, J.M., Kunnath, S.K. & Reinhorn, A.M. 1997. Seismic performance and retrofit evaluation of reinforced concrete structures, J. Struct. Engineering, ASCE 123 (1): 3–10.

243

Chopra, A.K., Goel, R.K., & Chintanapakdee, C. 2004. Evaluation of a modified MPA procedure assuming higher modes as elastic to estimate seismic demands. Earthquake Spectra, Earthquake Engineering Research Institute 20 (3): 757–778. Oakland, California. Chopra, A.K., & Goel, R.K. 2002. A modal pushover analysis procedure for estimating seismic demands for buildings. Earthquake Engineering and Structural Dynamics 31: 561–582. Coburn, A. & Spence, R. 2002. Earthquake Protection. England: John Wiley & Sons Ltd. Cornell C.A. & Krawinkler H. 2000. Progress and Challenges in Seismic Performance Assessment. PEER Center News (4)1:1–3. Deierlein, G.C. & Hsieh, S.-H. 1990. Seismic response of steel frames with semirigid connections using the capacity spectrum metnod. Proc. 4th U.S. Nat. Conf. on Earthquake Engineering 2: 863–872. Dymiotis, C., Kappos A.J. & Chryssanthopoulos, M.C. 1999. Seismic reliability of R/C frames with uncertain drift and member capacity. J. Str. Engng, ASCE, 125(9): 1038-1047. Elnashai, A.S. 2000. Advanced inelastic static (pushover) analysis for seismic design and assessment. G. Penelis International Symposium on Concrete and Masonry Structures. Aristotle University of Thessaloniki, Thessaloniki, Greece 23–34. Fajfar, P. & Fischinger, M. (1988), N2 - A Method for Non-Linear Seismic Analysis of Regular Buildings, Proceedings of the 9th World Conference on Earthquake Engineering 1988, Paper No. 7-3-2, Tokyo-Kyoto, Japan, August 2–9, Vol. V: 111–116. Fajfar, P. & Gaspersic, P. 1996. The N2 method for the seismic damage analysis of RC buildings. Earthquake Engineering and Structural Dynamics 25 (1): 31–46. FEMA, 2003. Multihazard Loss Estimation Methodology. Earthquake Model – HAZUS®MH MR4 Technical Manual. Fischinger, M., Kramar, M. & Isakoviæ, T. 2009. Seismic collapse risk of precast industrial buildings with strong connections, Earthquake Engng Struct. Dyn. (2009). Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/eqe.970, John Wiley & Sons, Ltd. Freeman, S.A. 1987. Code designed steel frame performance characteristics. Dyn. of Struct. Proc. Struct. Congr. ’87, Struct. Div. ASCE, New York, N.Y., 383–396. Freeman, S.A., Nicoletti, J.P. & Tyrell, J.V. 1975. Evaluations of existing buildings for seismic risk -a case study of Puget Sound Naval Shipyard, Bremerton, Washington. Proceedings of U.S. National Conference on Earthquake Engineering, Berkeley, California 113–122. Jalayer, F., Cornell, C.A. 2003. A Technical Framework for Probability-Based Demand and Capacity Factor Design (DCFD) Seismic Formats, PEER Report 2003/08, Pacific Earthquake Engineering Center, College of Engineering, University of California Berkeley Jan, T.S., Liu, M.W. & Kao, Y.C. 2004. An upper bound pushover analysis procedure for estimating the seismic demands of high rise buildings. Engineering Structures 26: 117–128. Kappos, A.J. 2001. Seismic vulnerability assessment of existing buildings in Southern Europe, Keynote lecture, Convegno Nazionale ‘L’Ingegneria Sismica in Italia’ (Potenza/Matera, Italy), CD ROM Proceedings Kappos, A.J. & Panagopoulos, G. 2010. Fragility curves for reinforced concrete buildings in Greece. Structure and Infrastructure Engineering. 6(1–2): 39–53. Kappos, A. J., Panagopoulos, G., Panagiotopoulos, Ch. & Penelis, Gr. 2006. A hybrid method for the vulnerability assessment of R/C and URM buildings. Bull. of Earthquake Engineering 4 (4): 391–413. Kappos, A., Pitilakis, K., Morfidis, K. & Hatzinikolaou, N. 2002. Vulnerability and risk study of Volos (Greece) metropolitan area. 12th European Conference on Earthquake Engineering (London, UK), CD ROM Proceedings (Balkema), Paper 074. Kappos, A.J., Stylianidis, K.C., & Pitilakis, K. 1998. Development of seismic risk scenarios based on a hybrid method of vulnerability assessment. Nat. Hazards 17(2): 177–192. Lomnitz, C. & Rosenblueth, E. (eds.). 1976. Seismic Risk and Engineering Decisions. R. Whitman and C. A. Cornell - Design, Chapter 9. Elsevier, Amsterdam - Oxford - New York. Lungu D., Arion C. & Calarasu E. 2008. Bucharest soil conditions and input ground motion for the structural performance analysis, in: Mazzolani F.M., et al. (eds.) (2008). Urban Habitat Constructions under Catastrophic Events; Proceedings of the International Symposium in Malta, 23–25 October 2008. University of Malta. Mahaney, J.A., Paret, T.F, Kehoe, B.E. & Freeman, S.A. 1993. The capacity spectrum method for evaluation structural response during the Loma Prieta Earthquake; Proc. 1993 Nat. Earthquake Conf., Central U.S. Earthquake Consortuim, Memphis, Tenn. 2. 501–510. Mandara A., Avossa A.M., Ferraioli M., Ramundo F. & Spina G. 2007. Performance-based seismic retrofit of masonry and R.C. buildings, in: Wald F., Mazzolani F.M., et al. (eds.) (2007). Urban Habitat Constructions under Catastrophic Events; Proceedings of Workshop in Prague, 30–31 March 2007. McGuire, R., 2004. Seismic Hazard and Risk Analysis, EERI Publication No. MNO-10.

244

Mistakidis E., Apostolska-Petrusevska R., Dubina D., Graf W., Necevska-Cvetanovska G., Nogueiro P., Pannier S., Sickert J.-U., Simões da Silva L., Stratan A. & Terzic U. 2007. Typology of seismic motion and seismic engineering design, in: Wald F., Mazzolani F.M., et al. (eds.) (2007). Urban Habitat Constructions under Catastrophic Events; Proceedings of Workshop in Prague, 30–31 March 2007. Pascu R. 2008. The performance based approach in seismic rehabilitation of Buildings, in: Mazzolani F.M., et al. (eds.) (2008). Urban Habitat Constructions under Catastrophic Events. Proceedings of the International Symposium in Malta, 23–25 October. Reed, J. W. & Kennedy, R. P. 1994. Methodology for Development Seismic Fragilities, TR-103959 Research Project. Vacareanu, R., Aldea, A. & Arion, C. 2003. Probabilistic risk assessment of current buildings VIe Colloque National AFPS, Genie Parasismique-Aspects Dynamiques et Vibratoires en Genie Civil 3: 153–160. Vacareanu, R., Lungu, D., Aldea, A. & Arion, C. 2004. Seismic Risk Scenarios Handbook, WP7 Report – RISK UE Project An advanced approach to earthquake risk scenarios with applications to different European towns, Contract: EVK4-CT-2000-00014. Vacareanu, R., Olteanu, P. & Chesca, A. B. 2006. Seismic Fragility of High-Rise RC Moment-Resisting Frames. Estimation of Drift Hazard, Proceedings of First European Conference on Earthquake Engineering and Seismology, paper no. 1000, Geneva, Switzerland. Vamvamvatsikos, D., Gerasimidis, S., Nigro, E., Muzeau, J.P., Kouris, L., Kappos, A., Rossetto, T., Stathopoulos, T., Faggiano, B. & Talon, A. 2010. Structural Vulnerability Assessment under Natural Hazards: A review, Proceedings of COST-C26 Final Conference, Naples 16–18 September 2010. Web resources: www.usgs.gov http://www.civ.uth.gr/cost-c26/index.files/WG2.htm

245

Urban Habitat Constructions under Catastrophic Events (Final Report) – Mazzolani (Ed). © 2010 Taylor & Francis Group, London, ISBN 978-0-415-60686-8

3.3 Vulnerability and damageability of constructions under impact and explosion F. Dinu The “Politehnica” University of Timisoara, Romania

3.3.1 INTRODUCTION Most buildings are vulnerable when exposed to abnormal loading, like explosion or impact. The loads associated to such events are much larger than the design loads. Their effect on structures is aggravated by the nature of the events, e.g. high strain rate, dynamic amplification. As demonstrated in case of WTC collapse, if fire spreads in the building, the risk of collapse increases. Risk associated to such events can be reduced but cannot be eliminated. Local failure is likely and the main target is to mitigate the progressive collapse (Fig. 1). Some structures, for example those provided with structural redundancy and multiple load paths can survive these events. It is the case of the seismic resistant structures, which are designed to sustain significant lateral loads, and are equipped with continuity of members, effective tying, ductile details, and so on. Most existing ordinary buildings, do not meet the requirements for progressive collapse mitigation. In these cases, the reduction of the hazard and exposure (e.g. physical barriers) play an important role. The vulnerability and robustness of buildings and other constructions under explosion and impact are addressed in the next section. General methodology for vulnerability assessment and factors for reducing the risk of global failure (or progressive collapse) are presented. 3.3.2 RISK ASSESSMENT FOR MAN-MADE HAZARDS Risk avoidance as public policy is uneconomical and discourages innovation and new technology. Individuals and society as a whole must accept risk in order to achieve objectives, and this willingness to accept reasonable risks is a sign of progress and health (NISTIR 2007). However, the aftermath of recent natural and man-made disasters has made it clear that judgmental approaches to risk management are not acceptable. Rational approaches to progressive collapse mitigation require risk-informed assessment and decision making. Assessing and managing risks should be seen relative to the occurrence of hazards, i.e. risk management in the situations before, during and after the events of hazards (Faber 2008): – Before a hazard occurs, the issue of concern is to optimize investments into so-called preventive measures such as e.g. protecting societal assets, adequately designing and strengthening societal infrastructure as well as developing preparedness and emergency strategies. – During the event of a hazard, the issue is to limit consequences by containing damages and by means of rescue, evacuation and aid actions. After a hazard event, the situation is to some degree comparable to the situation before the event; however, the issue here is to decide on the rehabilitation of the losses and functionalities and to reconsider strategies for prevention measures. A range of different terms to characterize the effect of hazards are applied across different disciplines. Among these, vulnerability, resilience, robustness and adaptive capacity are used most frequently (Faber 2008): – Vulnerability is commonly related to risk over time in terms of expected potential future losses considering all possible events which may lead to such. 247

Figure 1. Progressive collapse: a) Ronan Point collapse: a gas explosion on the 18th floor resulted in a progressive collapse; b) Alfred P. Murrah Building, after the terrorist attack with a truck bomb.

Figure 2. Vehicle barriers.

Figure 3.

Management of blast risk.

The vulnerability of a system is defined as the ratio between the risks due to direct consequences and the total value of the considered asset or portfolio of assets considering all relevant exposures and a specified time frame. Conditional vulnerabilities may be defined through the vulnerability conditional on given exposures. – Robustness is often applied to characterize the response of a system to given changes in the system state variables. Different interpretations of robustness are available in the different disciplines. In structural engineering, a robust structural system is understood as a structure which will not loose functionality at a rate or extent disproportional to the cause of the change in the state variables. 248

Figure 4. Typical pressure-time record from detonation of explosives (adapted from Mainstone 1974).

The robustness of a system is defined as the ratio between the direct risks and the total risks, (total risks is equal to the sum of direct and indirect risks), for a specified time frame and considering all relevant exposure events and all relevant damage states for the constituents of the system. A conditional robustness may be defined through the robustness conditional on a given exposure and or a given damage state. If we consider the event H that causes abnormal loads, LD the event that causes local damage and C the event that causes global collapse from the local damage LD, then the probability of structural collapse due to H can be expressed as (NISTIR 2007):

where: P(H) = probability of abnormal load or hazard P(LD|H) = probability of local damage given that the abnormal load occurs, and P(C|LD) = probability of collapse given that local damage occurs. If we consider that P[H] ≈ λH , where λH = rate of occurrence of the abnormal load or hazard, the eq. 1 can be rewritten as:

Probability of collapse of buildings and other constructions depends on the hazard, but also on the capacity of the structures to withstand the abnormal loads with local damage and to arrest the propagation of collapse following a local failure. If we analyze the eq. 2, three basic strategies for reducing the probability of progressive collapse are identified: – event control: – local strengthening – alternate load path Event control – Reduction of the collapse probability can be done by reducing the exposure to hazard. The rate of occurrence for bomb explosion, the mean rate of occurrence is approximately 2 × 10−6 /yr. Gas explosions have larger rate of occurrence, approximately 1.8 × 10−5 /yr, while for vehicular collisions, the probability is even lower, around approximately 6 × 10−4 /yr. In case of possible bombing threats, increasing standoff distance reduces the explosion pressure and therefore the associated risk (Fig. 4). Local strengthening – refers to the design of key members, e.g. columns, to resist the abnormal load without collapse. Special load combinations and overpressure need to be considered in design, depending on the type of threat. 249

Table 1. Reliability class vs. annual failure probability. Reliability Class

Annual failure probability

RC3 RC2 RC1

10−7 10−6 10−5

Alternate load path – refers to the redistribution of forces after the loss of primary members. Once the key member fails, there is a potential for developing alternate load paths, and thus preventing the failure propagation (progressive collapse). The later two methods are more under the control of the structural engineer and refer to direct or indirect design. Evaluation and characterization of exposure is very important for the quantification of risk and identification of measures for reduction of losses. For example, in case of blast events, the exceedance probability of the pressure at a given location would be of interest. In many risk assessments, the joint representation of several exposures is required. This is e.g. the case when considering blast and fire after blast, which can constitute an important scenario for the assessment of the risks. While the vulnerability is not determined by the inherent characteristics of the building structural system, certain systems clearly are better suited to confront it. A question that needs further study is to what extent structural details designed to resist earthquakes also help resist progressive collapse. Ferahian (1972) showed that structural elements designed to withstand an El-Centro earthquake should be capable of resisting a gas explosion also. The GSA design guidelines against progressive collapse rely heavily on seismic criteria. For example, in case of Alfred P. Murrah Federal Building in Oklahoma City, the report suggested that if more recently developed detailing, such as those present in special moment frames used in seismic regions had been in place, the collapsed area would have been reduced at least by 50% and at most by 80% (FEMA 277). While current building codes and standards keep failure rates at a very low level, it is difficult to define a socially acceptable failure rate. For example, EN 1990 (2002) expresses the safety of structure by means of reliability classes, RC. Each RC is characterized by a reliability index, β, which depends on the probability of failure Pf . The choice of a level of reliability should take into account the possible consequences of failure and the exposure of the construction works to hazard. Reliability Class 1 (RC1) has low consequences for loss of human life and economic, social or environmental consequences small or negligible. Class (RC2) is medium and considerable. Class 3 (RC3) is high and very great. The reliability index can be converted into annual failures of probability as follows (Table 1). In the frame of the Cost Action C26, aspects of the vulnerability and robustness of buildings and other urban habitat infrastructure components have been investigated in the working group 3 (WG3 – Impact and explosion). Numerical simulations and experimental investigations have been performed, in order to predict the response of the structures under abnormal loading. Different hazard scenarios, like accidental explosions (terrorist attacks, gas explosion) and impact (vehicular collision, aircraft impact) have been studied. The results of the common research are presented in the following section.

3.3.3 VULNERABILITY AND ROBUSTNESS ASSESSMENT This section provides a review of the main contributions of the members to the goals of the WG3, task 3. It includes the following topics: – vulnerability assessment in case of blast effects: • C. Pérez-Jiménez, M. Minguez-Fica & J. De la Quintana, Labein Tecnalia-saix, Bilbao, Spain • G. Solomos, F. Casadei, G. Giannopoulos, European Commission-Joint Research Centre, Via E. Fermi 2749, 21027 Ispra (VA), Italy; M. Larcher Institut für Mechanik und Statik, Universität der Bundeswehr München, 85577 Neubiberg, Germany 250

– robustness analysis of buildings to impact: • I. Björnsson & S. Thelandersson, Lund University, Division of Structural Engineering, Lund, Sweden • J. Vaiciunas & V. Dorosevas, Faculty of Civil Engineering and Architecture, Kaunas University of Technology, Lithuania – robustness analysis of buildings to blast: • J. Mediavilla, TNO Defence, Security and Safety, Rijswijk, The Netherlands; F. Soetens & J.W.P.M. Brekelmans, TNO Built Environment and Geosciences, Delft, The Netherlands. • M.P. Byfield & S. Paramasivam, School of Civil Engineering and the Environment, University of Southampton, Southampton, United Kingdom • F. Dinu, Romanian Academy, Timisoara Branch, Centre for Advanced and Fundamental Technical Sciences & D. Dubina, The “Politehnica” University of Timisoara, Romania. 3.3.3.1 Vulnerability assessment in case of blast effects Historical records indicate that the majority of terrorist incidents have occurred in an urban environment in presence of nearby buildings forming the street geometries (e.g. WTC 2001). The recent terrorist attacks (Madrid 2004, London 2005, Moscow 2010) have shown that rail transport constitutes also a high-impact target, and have exposed its vulnerability. While security measures will always be taken for preventing and foiling such attacks, improved urban area configuration and architectural design may also significantly contribute towards mitigating the effects of explosions. Thus, reliable simulation tools for assessing structural vulnerabilities would be necessary. Vulnerability of urban area configurations to blast effects In order to make the buildings less vulnerable, it is necessary to identify and analyze the geometrical parameters that determine the vulnerability of urban area configuration to blast effect. The behavior of different geometrical configuration of single buildings as well as different urban layout against blast actions by means of validated numerical methods was investigated. Evaluation of the human damage in urban layout scenarios will be presented by means of maps of damage based on probabilistic methods. Even the number of possible urban area configuration is limitless; the set of scenarios was limited to the following cases: – Single buildings: the objective is to determine how the geometry of two common typologies of building (rectangular, RB, and cylindrical, CB) submitted to a blast explosion is affected and affects the nearby regions due to the reflection of the shock wave. Dimensions of the scenarios are shown in Figure 1a where the height of both buildings, H, is assumed to be 24 m. The width, a, and long, b, of the rectangular building is 20 m as well as the diameter, D, of the cylindrical one. A 100 kg and 1000 kg TNT charges, are chosen for each building typology. The distance between the charge and the centre of analyzed building is 25 m, given a scaled distance of 5.4 y 2.5 m/kg1/3 , respectively. – Rectangular urban layout: the objective is to determine the influence of different typologies of buildings, position and orientation placed on a rectangular layout. Maps of human damage at 1 m above the ground are analyzed. Dimensions of the reference urban layout are shown in Figure 1b. The height of both buildings, H, is assumed to be 24 m. Both longitudinal, WL, and transversal, WT, street widths are 10 m. A spherical 1000 kg TNT charge is placed in the middle of the longitudinal street at 1 m above the ground and 10 m far from the building wall. In all simulations, buildings are assumed to be rigid; air and high explosive (TNT) are modeled by an Euler processor with equations of state being ideal gas and Jones–Wilkins–Lee (JWL), respectively (AUTODYN 2009). Each simulation has been divided in three stages: in the first stage, the initial detonation and propagation of the spherical blast wave is modeled in 1D, using 100 × 10 mm cells. Then, in the second stage the results from the 1D analysis are remapped to a 2D simulation, using approximately 150,000 × 100 mm cells. Finally, the results obtained from the 2D simulation are remapped to a 3D simulation; using approximately 1,040,000 × 500 mm cells and 800,000 × 500 mm cells for single building and urban layout simulations, respectively. The 3D numerical models were extended sufficiently far in each direction to ensure that the presence of the 251

Figure 5.

a) Schematic single buildings: rectangular and cylindrical b) schematic rectangular urban layout. Table 2. PROBIT equations for different human damage to be caused by explosions. Type of damage

PROBIT equation

Eardrum rupture Death for whole body impact

Y1 = −15.6 + 1.93ln(ps ) (1) Y2 = 5 − 2.44 ln((7.38 × 103 /ps ) (2) + (1.3 × 109 /(ps × i)

where ps is the maximum peak overpressure, kPa; i is the maximum peak impulse, Pa × s.

boundaries did not affect the results at the measuring locations. Blast parameters were measured over the front and rear part of the rectangular and cylindrical building at three different planes above the ground (1 m, 10 m and 20 m). For evaluating the damage suffered by the population submitted to the effects of urban blast, PROBIT equation, Y, are used. Table 2 represents the selected PROBIT equation found in literature (Fernando Diaz 2007). For single building simulation, the maximum peak pressure percentage with respect to a scenario in which no building is considered, WB, is represented for the rectangular and cylindrical building scenario. Both 100 kg and 1000 kg TNT charge are represented (Fig. 6). The presence of buildings creates an enhancing pressure region and shadowing pressure region in the front and rear part, respectively. The enhancing region, in both scenarios, increases with height. However, the maximum reflected pressure value reduces. This effect is represented in Figure 7. The mentioned tendency of maximum peak pressure and influenced region with height is appreciated. Figure 8 shows pressure–time curve for a point placed at the front and the rear part of the rectangular and cylindrical building for both 100 kg and 1000 kg TNT charge. In case of rectangular urban layout scenario, the results are discussed separately, for longitudinal lines and transversal lines, respectively. For both lines, the global tendency of maximum peak pressure and impulse is to decrease with distance. Building orientation and geometry are important parameters when planning an urban area against blast effects. Four cases were analyzed, two with a rectangular building rotated 45◦ (one in first line and one in second line) and two with a cylindrical building (one in first line and one in second line) (Fig. 9). Figure 10 shows the perceptual maximum peak pressure and relative maximum peak impulse differences with the reference rectangular urban layout scenario. Positive numbers represent higher values in the analyzed scenario than in the reference scenario. For the building in the upper part, No differences in the maximum peak pressure and impulse are appreciated in the lower part of the urban layout when cylindrical or the rotated rectangular building is placed in the upper part. For buildings in the lower part, both attenuation and enhancing effects, with respect to the reference scenario are found. Enhancing effects appear in the transversal street, while attenuation effects appear in the rest of the urban layout. 252

Figure 6.

a) Rectangular building scenario; b) cylindrical building scenario.

Figure 7. Pressure and influenced distance as a function of the building height. Data taken in central plane of the front wall of RB scenario (100 kg TNT charge scenario).

Maps of human damage have been also obtained in these scenarios. The following human damages are considered: eardrum rupture and death to whole body impact. Thus, taking into account the PROBIT equation, Table 2, the following maps of human damage is obtained (Fig. 11). Numbers indicate the probabilistic damage, in %. It can be appreciated that for distances close to the explosive charge, the percentage of eardrum rupture is over 90% and it is slightly affected by the influence of building geometry or orientation. The incident shock wave has enough energy to provoke this human damage and the amplified energy of the reflected wave has not really influence. Building orientation and geometry are important parameters also for human damage distribution. In case of comparison with reference scenario, the behaviour of the rotated rectangular building and cylindrical scenario create more safety urban layout. 253

Figure 8. building.

Pressure-time curve at different height (1 m, 10 m, 20 m); (a) front part and (b) rear part of the

Figure 9. Rectangular urban layout scenarios; a) – b) lower and upper rotated rectangular building scenario, respectively; c) – d) lower and upper cylindrical building scenario.

Figure 10. a) Perceptual maximum peak pressure differences with the reference rectangular urban layout scenario; b) relative maximum peak impulse differences with the reference rectangular urban layout scenario.

254

Figure 11. Human damage in reference urban layout scenario; a) eardrum rupture and b) death by whole body impact.

Figure 12.

Idealised form of a pressure-time function.

Assessment of explosion effects in railway stations In the study, air blast waves resulted from the detonation of a solid high-explosive charge (a TNT equivalent). The magnitude of the pressure of an air blast wave that arrives at a certain point depends on the distance and on the size of the charge. An idealized form of a pressure-time function at a certain distance from the explosive is shown in Figure 12, where ta (it includes the detonation time itself) is the arrival time, pmax is the peak overpressure over the reference pressure p0 (atmospheric pressure), td is the duration of the positive phase, and tn is the duration of the negative phase. The positive impulse I is the integral of the overpressure curve over the positive phase td . Modified Friedlander equation was used to describe the positive phase of the free-field air blast wave, where the pressure p at time t can be expressed as:

Numerical simulations are performed within Europlexus (2010), an explicit finite element code for non-linear dynamic analysis. This finite element tool has been jointly developed by the French Commissariat a l’Energie Atomique (CEA) and the Joint Research Centre (JRC). Among the main advantages of Europlexus over similar software is its ability to handle complex fluid-structure interaction problems. The bursting balloon model (Larcher et al., 2010) was used for modelling explosion and blast loading. The advantages of this approach lie in the fact that for big structures and spaces (e.g. a rail station) larger dimension elements can be used, and thus the computation time becomes reasonable, while at the same time both the structure and the fluid are modelled. A relatively old typical train station has been selected for evaluation. The geometry of its structures has been acquired using a 3D laser scanning technique. It is composed of two principal domains which are connected through a short, narrow passage. The dimensions of the waiting hall area are about 50 × 30 × 12 m and those of the long corridor about 100 × 10 × 8 m (Fig. 13). The full simulation of the explosion is performed using an Eulerian formulation for the explosive and for the fluid representing the air. Apart from the nature of the problem, this choice is justified by the fact that the subsequent risk analysis requires the calculation of pressure and impulse of the air inside the volume of the structure. 255

Figure 13.

Geometrical finite element model of structure consisting of a main hall and a long corridor.

The simulation of terrorist attacks up to possible complete failure and fragmentation of some structural components introduces a new issue, for which a novel dedicated fluid-structure interaction (FSI) model has been developed (Casadei 2008, Giannopoulos 2010). The formulation of human injuries risk is based on the work of González Ferradás et al. (2008), Yet-Pole (2008) and Mannan (2005). It uses the peak overpressure pmax (Pa) and the positive impulse I (Pa · s) calculated inside each fluid finite element in order to determine the probability of eardrum rupture and the probability of death. Three different causes of death are considered i.e. head impact, whole body impact and lung haemorrhage, using the following PROBIT equations, Y:
Y1 = 5 − 8.49 ln

2430

 (4)

pmax +(4×108 ) pmax I


Y2 = 5 − 2.44 ln

7380

 (5)

pmax +(1.3×109 ) pmax I

 Y3 = 5 − 5.74 ln

4.2 × 105



pmax +1694 I

(6)

Y1 is the death PROBIT function due to head impact, Y2 is the one for whole body impact and Y3 is the one for lung haemorrhage. A PROBIT function for body impact by flying debris has not yet been implemented. The PROBIT function of eardrum rupture Y4 is described through the equation presented in Table 2. The probability of occurrence R (or the percentage of the affected population) of the corresponding injury is next determined for each of the above PROBIT functions using Equation 7 (Ferradás et al. 2008), which is a very good approximation of the relevant cumulative normal distribution, i = 1,4: Ri = −3.25Yi3 + 48.76Yi2 − 206.6Yi + 270.35

(7)

Several scenarios with different quantities of explosive have been run. Two cases for the same explosive charge of approximately 250 kg TNT equivalent, placed on the floor at the centre of the main hall, are presented in Figure 14 and Figure 15. It is observed that both approaches manage to reproduce the main characteristics of the expected structural behaviour, such as large deformations, failure of the roof, fragmentation, projectile formation and motion. The deformation and damage pattern of the structure at the selected four time instants is quite similar. This is also true for the whole response period and for the picture of the final damage. Thus it can be concluded that for structures directly exposed to an air blast the pressure-time functions can provide a good and inexpensive way for calculating their behaviour. However, this approach is not sufficient if the blast waves arrive indirectly to the structure (reflections, channelling), or if information about the pressure field is required. The risk analysis performed for this structure reveals the areas for which 256

Figure 14.

Structural response at four time instants using the pressure-time functions approach.

Figure 15.

Structural response at four time instants using the full fluid-structure interaction approach.

257

Figure 16. Top view of death risk contours for the main hall and corridor.

the human injury risk is high. As an example, Figure 16 shows the results in three-dimensional space for the death risk. For the main hall it is also worth noticing spots at corners where enhanced values of death risk are encountered due to wave reflection phenomena.

3.3.3.2 Robustness analysis of buildings to impact Vehicular collisions, including trucks, trains and barges, can have a major impact of the vulnerability of constructions (buildings, bridges, and so on). A common robustness requirement given by structural codes states that a structure shall be designed and executed in such a way that it will not be damaged to an extent disproportionate to the original cause (EN 1990 2002). In the following, the robustness of a bridge in case of train collision and the robustness a concrete frame building in case of car impact will be detailed. Bridge response to train collision The bridge structure being considered is a multi span post-tensioned reinforced concrete bridge located in Malmö, Sweden. The bridge crosses multiple rail tracks as well as a highway. The structure has been designed according to Swedish bridge, road and railway standards; e.g. BRO 2004 (Vägverket 2004). The longitudinal section of the bridge is shown in Figure 17. The bridge has a total span of 172.5 m and consists of a cast in-situ concrete cross section with post tensioning cables running the length of the bridge. The longitudinal geometry of the cross section varies from span to support sections. The traffic running along the bridge consists of road, cycle and pedestrian lanes and the total deck width is around 20 m. In addition to the bridge girder there are four internal supports and two land abutments all of which are composed of reinforced concrete cast in-situ. The inner supports are constructed as wall elements approximately 12m in length. The bridge deck and supporting sub-structure are connected using a pot bearing system. There are a total of four bearings per support the majority of which are uni – or multidirectional with the exception of the central support which contains two fixed bearings. There is a four lane road highway running under the bridge and a double lane road running over it. The probability of collapse can be expressed by the eq. 1. This equation serves as a good indicator of the various factors that can bring about the collapse of the bridge. In order to increase the system’s collapse resistance, countermeasures aimed at minimizing P(C|LD) and P(LD|H) terms should be considered. The latter is included in standard structural component based design whereas the former is not directly reflected in the verification procedures of current design codes. If also the term P[H], which refers to exposure, can be modelled, the evaluation of the probability of collapse can be done directly. The exposure event examined here is the occurrence of train impact to a bridge support as a result of derailment. The probability of the event occurring is determined with the aid of railway accident statistics. 258

Figure 17.

Longitudinal section of multi-span bridge crossing rail tracks and highway.

The hazard scenario which was considered is a train collision to supports 2, 3 & 4 as a result of derailment. The examination of this scenario starts with a check of the probabilities associated with train derailment near the bridge, on track adjacent to a support and in the direction of that support. Then the corresponding impact force to the bridge support and the probability of support failure can be determined. Finally, the behaviour of the remaining bridge structure given support failure can be analyzed. The expected number of derailments per year is based on the Swedish railways accident statistics between the years 1985 and 1995 (Sparre 1995):

where W is the associated exposure variable and ξ is the intensity factor. Once the annual derailment rate has been determined for a length of track in the region near the bridge, the probability of the train derailing on tracks adjacent to a support and within the critical region for which collision is possible is evaluated. This critical region was defined for a set of limiting derailment angles; i.e. a maximum and minimum angle of departure from the original direction of travel towards a bridge support. The post derailment model used for this analysis is based on a simplified derailment mechanism in which the derailment angle and lateral movements of a derailed train are considered (Östlund et al., 1995). The train is represented by a single mass travelling along a constant path governed by the derailment angle, which is limited to a degree where rolling of the derailed vehicle is possible. The ground force is modelled as a Coulomb friction force acting in the reverse direction of the velocity vector. Events occurring just after the point of derailment, such as interactions between the train wheels, rail and concrete sleepers, were not considered (refer to Brabie & Andersson 2008). The determination of the collision forces used for this analysis will be based on the so called hard impact model, in which the support wall is assumed rigid and the impact energy is mainly dissipated by the train. The train is then modelled as a single DOF system and a linear deformation during impact is assumed. The impact force is then determined (EN 1991 2006):

where m is the mass of the train, vr is the velocity of the train at impact and k is the equivalent spring stiffness of the train. The equivalent spring constant was estimated to 10 MN/m. The impact force was determined using Monte-Carlo simulations. Figure 18 shows the cumulative probability distributions of this force for supports 2, 3 & 4. The probability of support failure is conditioned on the impact force and normal force in the support wall. The failure modes that were considered depended on the support being checked but in most cases it was modelled as a fixed column where combined normal force and bending about the weaker axis of the wall was decisive. The probability of support failure given train impact was determined. The resulting annual probability of failure for supports 2, 3 and 4 given train impact was determined as 6.8 × 10−3 , 2.4 × 10−5 and 3.0 × 10−3 respectively. Figure 19 shows a sensitivity analysis of the probability of failure for support 2 given baseline variations of some of the variables used in the analysis. 259

Figure 18. Longitudinal section of multi-span bridge crossing rail tracks and highway.

Figure 19. Sensitivity analysis of probability of support failure given train impact for support 2.

To account for the dynamic effects of sudden failure of a support, a force based dynamic increase factor (DIF) may be utilized. In the study, a dynamic factor of 1.5 was assumed. In all cases the resistance was between 25% and 60% less than the applied load. In terms of a probabilistic analysis, the expected probability of failure is likely to settle in the vicinity of 1. The evaluation of train collision due to derailment for the bridge case determined an annual marginal probability of global system failure at 6.7 × 10−7 . In terms of structural reliability, this corresponds to a reliability index of β = 4.82 for a reference period of one year. This constitutes a consequence class CC2 according to Eurocode 0 (EN1990 2002). Thus in terms of total system failure given one rare exposure type, the bridge is within realm of acceptable limits for code based design. It was also found that for supports 2 and 4 the failure probabilities given train impact were significantly high, corresponding to reliability indices less than the recommended absolute minimum according to Eurocode 0 (EN 1990 2002). However, coupled with the probability of the train impact occurring, the support failure probability decreased significantly. The direct consequence involves the crashworthiness of the train itself, in other words, the safety of possible passengers onboard. Given that the support withstands the collision, there may only be minor indirect consequences for the rail network. On the other hand, if the support were to fail, it was shown that global collapse was almost certain. In which case, the consequences, direct and indirect, material and immaterial, would be expected to increase drastically. The direct consequences expected to occur due to impact include rebuild and repair cost of the bridge structure, and possible human casualties such as those on the train or in vehicles on the bridge deck or running under it. The indirect consequences include user costs as a result of road and railway closure. The latter may bring with it economic losses considering the volume of rail traffic on the rails running under the bridge. Response of RC frames to impact loading The study aims to calculate the qualitative changes of the RC framed structures, under impact loading and to apply a point of risk level D for such abnormal situations. The studied structure is presented in Figure 20. Three main parameters influence the response of the structure: – impact load characteristics; – structural characteristics; – contact between the structure and the colliding object. In order to evaluate the response, the displacements of the RC slab from impactor with initial velocity interaction in Figure 21. The following three states of RC slab were used in the analysis: (A) normal state, concrete and steel materials are in good condition; (B) damage of the reinforcements; (C) damage of the concrete. The impactor was modelled as a rigid 197 kg weight of cylindrical body with the density of 7830 kg/m3 , with an elastic modulus 159 GPa and Poisson’s ratio of 0.2 260

Figure 20. slab.

RC framed structure: 1) column; 2) column bracket; 3) RC slab connection; 4) beam; 5) floor

Figure 21.

RC slab interaction with impactor in state A.

Figure 22. Fragments of directional deformation results in RC slab of state of reinforcement A (a), and of state of reinforcement B (b), (z axis).

Adopting very small standard displacement st, for example 0.5 mm, points of risk level of various conditions of RC slabs during different moments of shock impact can be calculated. Comparison of the results is given in Figure 23. 3.3.3.3 Robustness analysis of buildings to blast The design methods used to mitigate the potential for progressive collapse include: – tying force method – the specific local resistance method (designing key elements to withstand abnormal loads) – the alternate load path method (allowing for redistribution of load in the event of the loss of a key member). First method is an indirect method while the other two are direct methods. Each method has advantages and disadvantages, but here they are ordered by increasing level of complexity. Tying 261

Figure 23.

Comparison the point of risk level. Indicated to parentheses in RC slab state.

Figure 24.

Catenary action.

Figure 25. Typical steel frame, considered to study the feasibility of steel frame a) and details of beam-column (end plate) connection b).

force method is recommended for facilities that require low level of protection, while the direct methods are recommended for important facilities. In the following section, the application of direct and indirect approaches to the robustness evaluation of multi-story buildings is presented. Tying force method The tying force method relies on catenary action to redistribute loads in the event of loss of support at lower level (Fig. 24). The method is a threat-specific design approach. The mechanics of this mechanism are investigated by way of a case study of a steel framed building (Fig. 25). Thus the example would correspond most closely to a glass clad building with open-plan internal architecture. The example frame resembles a typical medium rise office development, using a steel frame with simple connections and incorporating composite metal decking. The weak link in the catenary action mechanism is the tensile capacity of the primary beam to column connections. The tying capacity of the end-plate connections joining the perimeter beams to the columns (Fig. 25b) was calculated to be 587 kN, which is the design value determined from industry standard design guides (SCI 2002). In accordance with standard practice for the tying force method this capacity was calculated in the absence of beam rotation. Clearly, rotation will affect the tying capacity of the connection, with the tensile load capacity calculated in the absence of rotation representing the maximum tying capacity. The analysis presented here has included the full tying capacity of the connection, although the beam rotation is limited to the maximum rotation available for this particular connection, which was calculated to be limited to 4.38 degrees, beyond this rotation the connection will fail. Small moments are generated in the connection at this limit of rotation, however 262

Figure 26.

End plate connection and its T stubs.

they have little effect on the redistribution of loads and have therefore been ignored. The composite slab has a tensile (tying) capacity and this was included in the catenary action calculations. Failure was deemed to occur when the tensile capacity of the catenary system was exceeded. The floor comprises a 125 mm thick composite slab, reinforced with A142 fabric reinforcement and formed on 1.2 mm gauge decking. Composite action between the beams and slab is achieved with 19 mm diameter, 100 mm high headed studs, spaced at 300 mm centers, i.e., 40 studs in each half span of the secondary beams. For the main beams the studs are provided at 450 mm centers, i.e., 22 studs in each half span. The floor slab is extended by 0.5 m beyond the edge beams to support a curtain wall system. All bolts were M20 grade 8.8 and the member and connection details of the frame are shown in Figure 25. The tying capacity of the slab was taken as the sum of the axial strength, Tslab = Tdeck + Tmesh , where, Tdeck is the strength of the profiled metal decking and Tmesh is the strength of the mesh reinforcement embedded in the slab. Under tension, the profiled sheet failed either by shear or bearing or insufficient end distance. The minimum centre-to-centre spacing of stud shear connectors should be five times the nominal shank diameter along the beam (BS 5950 – 3.1:1990, clause 5.4.8.4.1). Furthermore, the end distance for profiled sheet measured to the centre line of the studs should not be less than 1.7 times the stud diameter (BS 5950-4:1994, clause 6.4.3). Because of these provisions the critical failure mode is bearing. Hence, the maximum tying force carried by profiled sheet was 42 kN/m. Tmesh was equal to 65 kN/m. These provide a total tensile capacity of the slab, Ts equal to 107 kN/m. The integrity of the catenary action mechanism mainly depends on the response of the connection which includes the interaction of tensile force, bending moment and rotation. The rotation of bolted connections is mainly produced by deformation in the column flange, end plate and bolts. Deformation of the column flange and end plate is estimated using the equivalent T stub method, Yee and Melchers (1986). Whilst rotation capacity was calculated based on the plastic deformation of T-stubs (Faella et al, 2000). In a bolted T-stub, the flanges are connected by means of two bolts only, i.e., with only one bolt row. The deformation and strength of the T stub, due to axial load mainly depends on flexural strength of the flanges and axial strength of the bolts. In this analysis, the end plate connection shown in Figure 25.b is divided into T stubs as shown in Figure 26. The force displacement of each component including bolts, 2 T-stubs and column web, up to failure is modelled based on the non-linear stress-strain curve presented by Kato et al. (1990). As the ultimate strength of the end plate (116.3 kN) is the lowest among the connection components, it is considered the weakest joint component. The contribution in the ultimate plastic deformation from the compression of the column web is much lower, hence it can be ignored. The deformation of other components is obtained corresponding to this axial force. The sum of the deformations (corresponding to 116.3 kN) of the joint components is the ultimate deformation capacity of one bolt row which is equal to 29.86 mm. After the removal of column support, the column starts to drop down and the beams on either side rotate about the joints. The joints are assumed to rotate about the bottom of the end plate. The lateral movement of columns is assumed to be arrested by the composite slab (as the stiff arrangement of secondary and primary beams bonded together by shear studs can resist the tying loads without buckling). The rotation capacity of the connection is 4.38◦ , calculated by dividing the ultimate deformation of top most T-stub (29.9 mm) by its distance from bottom end plate (390 mm). 263

After removal of the intermediate support, the beams on either side of it form an inverted three hinged arch. This is a determinate structure, can be analyzed using force equilibrium. The horizontal thrust at the support:

where W1 is the reaction from the secondary beam, W2 is the reaction from the main beam plus the weight of column for that storey, q is the UDL acting directly onto the beams Resolving vertically provides the vertical reactions:

√ and the tying force is given by T = H 2 + V 2 The load W1 is determined by multiplying the accidental limit state load by half the area supported by beam SS1 . The factor of safety (FoS) against disproportionate collapse is defined herein as the ratio between tying capacity of the joint and tying force acting in the joint at the maximum rotation limit for the joint, which in this example was 4.38◦ . The FoS is calculated with and without considering the catenary action in slab, as the contribution of the reinforcement mesh and profiled sheet is uncertain. A further issue of uncertainty is the dynamic amplification of loads. Considering all the above issues, the following cases are considered to estimate lower and upper bound values of FoS. In this upper bound case the full tensile strength of the slab is included and the dynamic amplification of loads is ignored (DAF = 1.0). This case provided a factor of safety of 0.19 and can be considered as an upper bound estimate. In the lower bound case we do not rely on catenary action in the slab because of the uncertainty as to whether the slab can accommodate large deflections without fracture. In addition, the DAF has been taken as equal to 2.0, in accordance with US practice (GSA 2003). The accidental load of this case provides a factor of safety of 0.08. The specific local resistance method The response of a steel frame building under blast load is studied by means of finite element computer models. A simplified multi-storey steel building has been considered, with typical values of spans and heights presented in Figure 27. Standard steel profiles (HEA, HEB, IPE, UNP) are used for beams and columns, made of S235 steel, dimensioned to withstand typical load combinations (static, live, wind), using appropriate load factors. The building is not dimensioned against blast loading. The dead weight is 5 kN/m2 and the live weight is neglected, since this is the most critical load combination with the blast load. All floors and roofs are made of prefab cellular concrete 0.1 m thick. The façade is made of 0.01 m thick glass curtain and 0.15 m thick reinforced concrete parapet. C30 concrete and FeB500 steel bars are used, 0.56% in both directions. Three blast scenarios are considered. The explosive charge is placed on the ground floor, in front of the building. The combinations of explosive charge m and standoff distance d are studied: 5 kg-1 m; 100 kg −10 m and 1000 kg −25 m. Simulations have been performed with LS-DYNA (Hallquist 2007). The blast load is modelled using ConWep, which is based on TM 5-855-1 (1998). This gives the reflected pressure-time distribution at each point of the structure based on the distance, incidence angle and amount of explosive, assuming that the building is rigid (no fluid-structure interaction). Belytshko-Tsay shell elements are used on the façade, floor and roof. Hughes-Liu beam elements are used for the columns and beams, with fixed joints. The façade glass is modelled as an elastic material. The façade glass is modelled as an elastic material. Reinforced concrete (floors, roof and parapets) is modelled as a homogenized material, based on EN1992. Two failure criteria have been added to account for crushing and rebar traction failure. Beams and columns are modelled as an elastoplastic material, with a strain based failure criterion. Figure 28 shows the final state of the building after the explosion, for the three blast scenarios. In the first case (5 kg −1 m) only the glass of the first floors of the front facade breaks. In the other two cases, all the glass has been blown away. In the second case, there is fracture of the front concrete parapets of the first and second floor. The third blast case (1000 kg −25 m) is the 264

Figure 27.

Geometry and dimensions of steel multi-storey building.

Figure 28. Final state for the 3 blast scenarios. (left) 5 kg −1 m; (middle) 100 kg −10 m; (right) 1000 kg −25 m.

Figure 29.

Plastic strain plots. (left) 5 kg −1 m; (middle) 100 kg −10 m; (right) 1000 kg −25 m.

most destructive, with failure of the concrete parapets of the first five floors, the side facades and collapse of the concrete roof. The damage undergone by the steel frame can be visualized by means of the plastic strain. Figure 29 shows the plastic strain contour plots for the three loading cases. In the first and second blast scenario there is hardly any plasticity (no damage) in any of the steel elements. The third blast scenario, the most severe, there is extensive damage of the front columns and even failure, side columns and roof beams. There is risk of progressive collapse. Figure 30 shows how the extent of damage in the steel structure increases with increase in glass strength. The main conclusion that can be drawn from these simulations concerns the effect of façade toughness on the damage and response of the structure. The tougher the façade, the more energy is absorbed by the structural elements. As a consequence, the damage increases and higher dynamic forces are transmitted to the foundation. The extent of damage of the three blast load scenarios studied is very different. In the first case, 5 kg −1 m, failure is limited to the front glass elements. 265

Figure 30.

Plastic strain plot, 1000 kg −25 m. (left) ft = 25 MPa; (middle) ft = 80 MPa; (right) ft = 135 MPa.

In case 2,100 kg −10 m, all glass is gone and there is localized failure of some of the concrete parapets. In the case 3, the concrete roof fails, there is extensive damage of the concrete elements of the front façade and some of the steel columns fail. The building is about to collapse. The alternate load path method This method is a non threat-specific design approach. If in the “specific local resistance”, the design strategy focuses on limiting P(LD/H) (see eq.1), in an “alternate load path” design strategy the focus is on limiting P(C|LD), namely to avoid the propagation of the failure beyond the point of initial damage. The method is recommended for structures that require Medium or High Levels of Protection. In the following, the application of alternate load path to the robustness evaluation of a multistory building is presented. The structure was designed according to EN1998-1 (2004) and P100/1 (2006) seismic codes. In order to develop a multilevel evaluation, different extension of damage was considered. It included loss of several columns, located at the perimeter and inside of the building. Columns are removed one by one and 3D dynamic nonlinear analyses are employed. In order to develop a multilevel evaluation, different extension of damage was considered. It included loss of several columns, located at the perimeter and inside of the building. Columns are removed one by one and 3D dynamic nonlinear analyses are employed. The building frame system uses steel braced and unbraced frames. The cruciform cross sections columns, made of hot rolled profiles were partially encased in reinforced concrete to increase the strength, stiffness and fire resistance. Beams and braces are made of I hot rolled sections. S355 steel was generally used for frame members, excepting the braces designed as dissipative members, which are of S235 steel. The actions considered in design were: dead load: 6.2 kN/m2 , live load: 2.0 kN/m2 , snow load: 1.50 kN/m2 , wind load: 0.55 kN/m2 , base shear force, ag = 0.24 g (q = 4, TB = 0.16 s; TC = 1.6 s; TD = 2.0 s; β0 = 2.75). The load combination for dynamic analysis is D + 0.5L + 0.2 × W, where D, L and W are dead, live and wind loads. If dynamic analysis is performed, the removal of the member should take place in a fraction of the period of vibration associated with the expected structural response mode. This fraction can be considered one-tenth to one-twentieth of the corresponding period of the structure. In order to activate the vertical vibration, the load was applied for one-twentieth of the corresponding period of the structure and than kept constant (Fig. 32). Depending on the level of protection (LOP) and the importance of building, the UFC 4-023-03 guidelines (2009) recommends considering the dynamic effects due to instantaneous removal of a column in different ways. If a static nonlinear analysis is performed, the dynamic effect of gravity load is simulated through an amplification of the gravity loads above the affected area, only. This dynamic increase factor DIF can vary from 1.15 to 2, depending on the level of allowable plastic deformation in the members. The DIF can be chosen as a function of the level of nonlinear behaviour (i.e., structural performance level) that the designer wishes to employ or, else, the level of nonlinear behaviour can be assigned, resulting in a specific DIF. Based on the results of previous studies, the loss scenarios considered columns from the top and top-right of the structure, as illustrated in Figure 33. Each removal was considered one at a time. The structure remains elastic for all cases. It order to evaluate worse case scenarios, removal of two columns at a time were assessed. For all combinations of two columns removal at a time, the structure remains stable and is not damaged to an extent disproportionate to the original damage. Deflection time history is shown in Figure 34 for worst one column and two columns loss scenarios. Figure 35 266

Figure 31. structure.

Structural system: transversal frame, longitudinal frame, current floor plan and view of the

Figure 32. Application of vertical load on the model with lost column in the dynamic analysis.

shows the plastic hinge formation and axial force diagram for case C4-5. The use of mega-trusses at the mid-height and top of the building improves the global behaviour and thus the robustness.

3.3.4 FURTHER DEVELOPMENTS The present report is the main outcome of the joint research developed in the frame of the Cost Action C26. The list of research topics includes vulnerability assessment in case of blast effects, robustness analysis of buildings to impact and robustness analysis of buildings to blast. As stipulated in the Memorandum of Understanding, the research was motivated by the increasing terrorist threat in urban areas and the need for a coordinated research at a European level. The vulnerability of buildings and infrastructure from man-made hazards is not specifically addressed in construction codes in Europe and therefore urgent action is required to extend design standards. Further developments need to include: – accounting for accidental actions from man-made hazards (blast, impact) – improved requirements for structural robustness (prevention of progressive collapse from terrorist attacks) 267

Figure 33.

Column removal locations.

Figure 34.

Dynamic analysis results.

Figure 35.

Plastic hinge formation and axial force diagram for loss of two columns at a time – case C4-5.

– risk oriented design rules for important facilities – design of new buildings and assessment of existing buildings under blast and impact • direct and indirect methods • methods of analysis • acceptance criteria • structural detailing for structures subjected to blast, impact (principles, type of connections, tying systems) – possible extension of seismic design principles and details – new materials and construction techniques • unobtrusive and aesthetic facades • nonfrangible glass for fenestration – performance based format for multi-threat assessment (eg. blast and fire). 268

REFERENCES AUTODYN user’s manual-version 12 2009. Century Dynamics Inc. Alonso, F. D. et al. 2007. Consequence analysis to determine the damage to humans from vapour cloud explosions using characteristic curves. Journal of Hazardous Materials, vol. 150, issue 1(146–52). Larcher M, Casadei F., Solomos G., 2010. Risk analysis of explosions in trains by fluid–structure calculations. Journal of Transportation Security, 3(1). EN 19900 2002. Basis of Structural Design. Europ. Committee for Standardization. Vägverket 2004. Vägverkets allmänna tekniska beskrivning för nybyggande och förbättring av broar. Bro 2004. Publication 2004:56. Borlänge, Sweden. Sparre, E. 1995. Urspårningar, kollisioner och brander påsvenska järnvägar mellan åren 1985 och 1995. Department of Mathematical Statistics, Lund University. 1995:E9. Östlund, L. et al. 1995. Dubbelspårsutbyggnad Kävlinge-Lund – Konsekvenser och skyddsåtgärder vid urspårning eller collision. Lund Institute of Technology. Department of Structural Engineering. TVBK-7048. Lund, Sweden. EN 1991 2005. Actions on Structures: Part 1-7: General Actions – Accidental Actions. Europ. Committee for Standardization. Steel Construction Institute 2002. Joints in Simple Construction. SCI, Berkshire. British Standards Institution 1990. Structural use of steelwork in building. BSI, London, BS 5950: Part 3. British Standards Institution 1994. Structural use of steelwork in building. BSI, London, BS 5950: Part 4. Yee, Y. L. & Melchers, R. E. 1986. Moment Rotation Curves for Bolted Connections”, Journal of Structural Engineering, ASCE, 112(3), pp. 615–635. Faella, C. et al. 2000. Structural Steel Semi-rigid Connections, CRC Press, London. Kato, B. et al. 1990. Standardized mathematical expression for stress-strain relations of structural steel under monotonic and uniaxial tension loading. Materials and Structures, Springer Netherlands, 23(1), pp. 47–58. GSA 2003. Progressive collapse analysis and design guidelines for new federal office buildings and major modernization projects. Office of Chief Architect, Washington, D.C. TM 5-855-1 1998. Design and Analysis of Hardened Structures to Conventional Weapons Effects. US-Army. Hallquist, J. O. 2007. LS-DYNA keyword’s user manual, version 971, Livermore software technology corporation. NISTIR 7396 2007. Best Practices for Reducing the Potential for Progressive Collapse in Buildings, National Institute of Standards and Technology, Technology Administration, U.S. Dep. of Commerce. Faber, M. H. 2008. Risk Assessment in Engineering: Principles, System Representation & Risk Criteria. JCSS Joint Committee of Structural Safety, Edited by M. H. Faber, June, 2008, ISBN 978-3-909386-78-9. Larcher, M. et al. 2010. Risk analysis of explosions in trains by fluid–structure calculations. J Transp Secur 3: 57–71. González Ferradás, E. et al. 2008. Consequence analysis by means of characteristic curves to determine the damage to humans from bursting spherical vessels. Process Safety Environ Protect 86: 121–129. Mannan S. and Lees, F.P. 2005. Lee’s loss prevention in the process industries; Volume 2: Hazard Identification, Assessment and Control. Amsterdam: Elsevier, ISBN 0750678577. Yet-Pole, I. & Te-Lung, C. 2008. The Development of a 3D Risk Analysis Method, J. Hazard Mater 153(1–2): 600–608. Sparre, E. 1995. Urspårningar, kollisioner och brander påsvenska järnvägar mellan åren 1985 och 1995. Department of Mathematical Statistics, Lund University. 1995:E9. Östlund, L. et al. 1995. Dubbelspårsutbyggnad Kävlinge-Lund – Konsekvenser ochskyddsåtgärder vid urspårning eller collision. Lund Institute of Technology. Department of Structural Engineering. TVBK-7048. Lund, Sweden. Brabie, D. & Andersson, E. 2008. Post-derailment dynamic simulation of rail vehicles – methodology and applications. Vehicle System Dynamics, 46(1): 289–300. Yee, Y. L. & Melchers, R. E. 1986. Moment Rotation Curves for Bolted Connections. Journal of Structural Engineering, ASCE, 112(3), pp. 615–635. Kato, B. et al. 1990. Standardized mathematical expression for stress-strain relations of structural steel under monotonic and uniaxial tension loading. Materials and Structures, Springer Netherlands, 23(1), pp. 47–58.

269

Urban Habitat Constructions under Catastrophic Events (Final Report) – Mazzolani (Ed). © 2010 Taylor & Francis Group, London, ISBN 978-0-415-60686-8

3.4 Performance assessment under multiple hazards D. Vamvatsikos Department of Civil and Environmental Engineering, University of Cyprus, Cyprus

E. Nigro Department of Structural Engineering, University of Naples “Federico II”, Naples, Italy

L.A. Kouris, G. Panagopoulos & A.J. Kappos Department of Civil Engineering, Aristotle University of Thessaloniki, Greece

T. Rossetto & T.O. Lloyd Department of Civil, Environmental and Geomatic Engineering, University College London, UK

T. Stathopoulos Department of Building, Civil and Environmental Engineering, Concordia University, Canada

3.4.1 INTRODUCTION Structural vulnerability and associated methodologies for its assessment have been identified as a key research field in structural engineering. Vulnerability itself can be defined in multiple ways and it can be evaluated using widely different formats that are typically inconsistent with each other, especially when considering different hazards. For example, it can be defined either deterministically or probabilistically, it can be based on the concept of one or more limit-states or performance levels and it can be evaluated using static or dynamic methods including or ignoring aleatory randomness and epistemic uncertainty. Thus, at least for frequent actions from well understood hazards such as wind, fire or snow, there are several methods to estimate it, some complex and other simplified, some of which are deeply entrenched in the professional practice, forming a cornerstone of past, current and forthcoming design codes and guidelines. On the other hand, infrequent actions from extreme natural hazards, such as floods, hurricanes, volcanoes, avalanches, tsunamis and earthquakes are often less well understood, and researching methodologies for their assessment is an ongoing project. With the emergence of multi-hazard assessment concepts, it is now important to collectively discuss such methods, understand their merits and attempt to cast them in a format that is suitable for integration within a single practical assessment framework. Therefore, in the sections to follow we will present a review of existing methodologies for structural vulnerability assessment under earthquake, tsunami and wind actions.

3.4.2 SEISMIC VULNERABILITY 3.4.2.1 General Seismic vulnerability can be defined as the degree of loss to a given element at risk (e.g. buildings) resulting from the occurrence of an earthquake event (Coburn & Spence, 2002). The reliable estimation of the economic, as well as human, losses incurred by an earthquake is a necessity for the development of seismic risk scenarios which are now widely accepted as an essential tool for seismic risk management and for prioritizing the pre-earthquake strengthening of the built environment (e.g. Bal et al., 2008; Kappos et al., 2008; Lang & Bachmann 2004; Strasser et al., 2008). 271

3.4.2.2 Classification of the building stock With regard to the scope of a vulnerability study, an approach using detailed assessment of individual buildings or a coarser one utilizing appropriate classification of building populations may be adopted. The first one is usually tailored to buildings of great significance, such as monumental or important (e.g. hospitals) public buildings, while the latter is more suitable for risk scenarios in a greater urban area. The vulnerability of monumental masonry buildings with unique historical value and a very limited population in every urban area is commonly estimated for every single monumental structure. This is because each monument is unique and is associated with particular modelling difficulties, associated with its long history (involves.g. multiple phases of construction, repairs, alterations and degradation of the materials), which are exacerbated by limitations in the experimental investigation of the strength of the materials etc. Methods involving statistical data with structural damage from past earthquakes are insufficient in this case. Thus, the vulnerability is individually estimated using detailed (Rota et al., 2005) or more simplified (Augusti et al., 2001; Lagomarsino 2006; Lagomarsino & Podestà 2004a,b; Valluzzi 2007) models. For the vulnerability assessment of ordinary building populations several classification methods have been proposed, taking into account characteristics that affect their seismic performance, such as the construction material (i.e. concrete, steel, brick or stone masonry, etc.), the level of seismic design and detailing, the building height, the configuration of infill panels etc. Various sets of building classifications have been proposed in the literature, as a result of the different construction practices applied in each country. An effort to introduce a classification scheme that establishes a common basis for vulnerability studies in Europe has been made within the framework of the project Risk-UE (Kappos et al., 2006, 2008, Lagomarsino & Giovinazzi 2006), in a similar fashion that HAZUS (FEMA-NIBS 2003) classification is currently considered as a reference for North America.

3.4.2.3 Damage definition The choice of a damage scale for the assessment of buildings is fundamental to the definition of vulnerability functions. From the simple Green-Yellow-Red characterization to more refined damage state definitions, a wide variety of damage state sets has been proposed in order to describe damage levels from negligible damage up to collapse of the structure (ATC-13; FEMA-NIBS 2003; FEMA 273/356; SEAOC 1995; EMS-98, etc.). Comparisons between several approaches have been presented by Hill & Rossetto (2008), along with the proposal of a homogenized scale for R/C buildings based on experimental data (Rossetto & Elnashai 2003). Each damage state can be defined in terms of structural and non-structural damage, as well as in economic or loss terms such as the ratio of repair cost to replacement cost (Kappos et al., 2006). Damage state descriptions can be different for various building classes since damage evolves at varying rates in structures with different characteristics (i.e. R/C and masonry buildings). Furthermore, economic approaches introduce a time and location dependency that can limit the scale application and lead to erroneous physical damage predictions (Miyakoshi et al., 1997), especially if used in absolute, instead of normalised terms.

3.4.2.4 Ground motion characterisation The choice of a ground motion parameter that represents the seismic demand is crucial for the vulnerability assessment of buildings. Mercalli-type intensity based approaches (e.g. ATC-13) can be misleading since it is a rather subjective quantity, associated with a great amount of uncertainty, and is also dependant on the performance of the building stock. Nevertheless, the fact that the limited available damage data (see next section) is usually associated with intensity levels leads to the need of incorporating them into many vulnerability assessment procedures. Direct ground motion quantities such as PGA or PGV can be utilized (Kappos et al. 2006, 2010; Boatwright et al., 2001) or even spectral quantities like Sd (HAZUS) or Sa (Singhal & Kiremidjian, 1996) to account for the frequency content of seismic motion. An extensive investigation on the correlation of building performance with recorded ground motion and the subsequent development 272

of empirical motion-damage relationships in the form of lognormal fragility curves has been carried out by King et al. (2005). 3.4.2.5 Vulnerability functions While for individual buildings the capacity curves (derived from inelastic pushover analyses) seem to be convenient for the description of their seismic performance, for populations of buildings a probabilistic approach is usually adopted. The building stock of an urban area is classified in a limited number of categories (classes) with, approximately, equal vulnerability (see section 2.2). Each class is related to a cluster of vulnerability (fragility) curves, or equivalently to a damage probability matrix (DPM). Vulnerability curves relate, for a predefined damage degree, the severity of the seismic motion with the probability that the damage suffered by the structure will exceed this specific damage degree. Similarly, each term of a DPM represents the probability that a building class suffers a certain degree of damage (e.g. light, moderate, severe, collapse), when struck by an earthquake of a predefined severity level (the macroseismic intensity is usually utilized herein). Existing vulnerability curves can be classified into the four generic groups of empirical, judgmental, analytical, and hybrid, according to whether the damage data used in their generation stems mainly from observed post-earthquake surveys, expert opinion, analytical simulations, or combinations of these (Rossetto & Elnashai 2003). 3.4.2.5.1 Empirical approach The construction of empirical vulnerability curves (or the corresponding DPM’s) requires available statistical damage data reported in post-earthquake surveys from previous seismic events (Rota et al., 2008; Spence et al., 2008). The observational source, when available, is the most realistic as all practical details of the exposed stock are taken into consideration alongside soil–structure interaction effects, topography, site, path and source characteristics (Rossetto & Elnashai 2003). The most common problem when applying a purely empirical approach is the unavailability of (sufficient and reliable) statistical data for several intensities. By definition, Modified Mercalli Intensities up to five lead to negligible damage, particularly cost-wise, therefore gathering of damage data is not feasible, while, on the other hand, events with intensities greater than nine are rare, especially in Europe, so there are not enough data available. This unavailability leads to a relative abundance of statistical data in the intensity range from 6 to 8 and a lack of data for the other intensities (or associated ground motions), making difficult the selection of an appropriate cumulative distribution, since the curve fit error is significant and the curve shape not as expected. The absence of available data necessitates recourse to other procedures such as expert judgement (Fäh et al. 2001; Oliveira 2008). A convenient but sometimes misleading approach of adopting data from different regions with similar construction practices should be treated with caution. A good example of the problems involved in adopting models from another country is outlined in the paper by Barbat et al. (1996) who had to adapt the vulnerability models developed for Italian masonry buildings to study the ones in Barcelona. 3.4.2.5.2 Judgement-based and rating methods The concept of judgement-based methods involves use of the opinions of expert panels of engineers with experience in earthquake engineering who are asked to make estimates of the likely damage distribution within building populations when subjected to earthquakes of different intensities. Although the reliability of such methods can be questionable due to the subjectivity of each expert engineer, these methods were used as the predominant source in the United States for the generation of damage probability matrices and vulnerability curves (ATC-13). Rating methods adopt the idea that experienced engineers fill in special questionnaires, which include structural characteristics of the buildings that affect their seismic vulnerability. The final outcome of these methods is typically a vulnerability index. The magnitude of this index represents the building capacity against earthquake. Sometimes, for the derivation of the rating, analytical procedures are also needed, hence the entire procedure can be classified as hybrid. A very detailed method for masonry buildings was developed by the GNDT (1989) (see also Benedetti et al., 1988; Casciati et al., 1994). The GNDT method includes the filling of two forms: at level 1 collection, building by building, of some informative elements; at level 2 qualitative and quantitative aspects referring to the configuration, foundation type, material quality etc. are 273

scored (in four classes) and lead to the vulnerability index. Modification of this methodology has been developed for the application in other types of masonry buildings (Gent Franch et al., 2008). These methodologies have been applied for the estimation of the expected building damage for a deterministic hazard in several urban centres (Cole, Xu, & Burton 2008; Grant et al., 2007; Faccioli et al., 1999). 3.4.2.5.3 Analytical approach Analytical vulnerability curves adopt damage distributions simulated from the analyses of structural models under increasing earthquake loads. The analytical procedure followed ranges in complexity from the elastic analysis of equivalent single-degree-of-freedom systems (Mosalam et al., 1997) to inelastic pushover analysis (HAZUS, Rossetto & Elnashai, 2005), or non-linear time history analyses, of realistic models of reinforced concrete (R/C) structures, mostly in 2D (Singhal & Kiremidjian 1997, Masi 1998, 2006), to reduce the cost of analysis. Several analysis-based curves have also been proposed for the vulnerability assessment of masonry buildings (Barbat et al., 2008; Borzi et al., 2008; Erberik 2008a,b; Lang 2002; Lang & Bachmann 2003; Park et al., 2009). Significant work has recently appeared in codified documents on the analytical estimation of vulnerability or fragility of buildings and bridges, based on nonlinear analysis methods. For example, ASCE/SEI 41 (2007) and Eurocode 8 (2004) offer a comprehensive methodology based on the static pushover method. On the other hand, the SAC/ FEMA-350 (2000) guidelines propose the use of nonlinear timehistory analyses, encompassing the use of incremental dynamic analysis (IDA, Vamvatsikos & Cornell 2002) for the assessment of seismic demand and capacity. Purely analytical approaches should, in principle, be avoided, since they might seriously diverge from reality, typically (but not consistently) overestimating the cost of damage (Kappos 2001). Analytical methods should be supported by experimental results in order to increase their reliability (Ruiz-Garci’a & Negrete 2009). 3.4.2.5.4 Hybrid approach Hybrid vulnerability curves attempt to compensate for the scarcity of observational data, subjectivity of judgemental data and modelling deficiencies of analytical procedures, by combining data from different sources. Kappos and his co-workers have developed over the previous years a hybrid methodology that combines statistical data with appropriately processed results (utilising repair-cost models) from nonlinear dynamic or static analyses, which permit interpolation and (under certain conditions) extrapolation of statistical data to PGAs and/or spectral displacements for which no data is available (Kappos et al., 1998, 2006, 2010). An extensive set of 54 building classes for R/C and 4 unreinforced masonry (URM) building classes has been analysed, representing most of the common typologies in S. Europe. All statistical data are from earthquakes that struck Greece in the past few decades. The analytical part of the procedure differs with regard to the structural material, since for URM buildings only pushover analyses have been utilized, while for R/C buildings both incremental inelastic dynamic (for 16 dully selected accelerograms) and static analyses have been tackled. Vulnerability curves are derived in terms of peak ground acceleration or spectral displacement. A lognormal distribution was assumed for constructing fragility curves for each class (common assumption in seismic fragility studies). Median values for each damage state in the R/C fragility curves were estimated based on the plot of the damage index (defined as the ratio of repair cost to replacement cost) against a function of the earthquake intensity (PGA) through incremental dynamic analysis, until collapse. These plots are then corrected using the corresponding available statistical data and appropriate empirical weighting factors based on the reliability of the statistical data (Kappos & Panagopoulos 2010). Fig. 1 shows a complete set of fragility curves (for 5 damage states) for old, medium-rise, R/C buildings with dual (wall+frame) system, without significant discontinuities in the arrangement of masonry infills. The hybrid approach for vulnerability assessment of masonry buildings combines statistical data with appropriately processed results from nonlinear static analyses. The statistical data used for masonry buildings were from Greek earthquakes, i.e. the Thessaloniki 1978 and the Aegion 1995 events, with some additional data from the Pyrgos 1993 earthquake used for comparison (Penelis, Kappos & Stylianidis 2003). Non-linear analysis of masonry buildings is generally more cumbersome than that of R/C ones. A simplified equivalent frame model with concentrated nonlinearity at the ends of the structural elements can be used for the non-linear static analysis of 274

Figure 1. Fragility curves in terms of PGA for low code, medium rise, regularly infilled R/C buildings with dual system.

Figure 2.

Fragility curves for single-storey stone-masonry buildings.

masonry buildings (Penelis 2006; Borzi, Crowley, & Pinho 2008). The damage (limit) states can be defined according to a drift-based damage index (Calvi 1999). An alternative definition, more suitable when pushover curves have been derived for the building classes studied, is to express the damage states as a function of the yield and the ultimate displacement of each building (Penelis et al., 2003). However, statistical data are not available in such terms; as said previously, in Greece statistics were available in terms of the economic damage index (ratio of repair to replacement cost). Hence, a correlation between the two sets of definitions is necessary for applying the hybrid approach, as proposed in Penelis et al. (2003) and Kappos (2007). Fragility curves for URM buildings can be derived either in terms of PGA (as in Fig. 2) or of spectral displacement (Sd ); in the hybrid procedure these values are inevitably based on the spectra of the specific ground motions recorded in the (broader) areas wherein damage statistics are available. It is noted that the Sd -based procedure is more sensitive to the type of ‘representative’ spectra selected for each earthquake intensity. An earthquake loss scenario for contemporary and historical buildings in Thessaloniki has been developed by the AUTh group (Kappos et al., 2007; Kappos, Panagopoulos & Penelis 2008) using the aforementioned methodology. 3.4.2.6 Epistemic uncertainty Epistemic uncertainties stem from the incomplete knowledge of the actual problem and its parameters, or simply from the, often unavoidable, modelling and methodology errors. The estimation 275

Figure 3. 200 realizations of static pushover capacity curves for a two story masonry building, caused by epistemic uncertainty (Vamvatsikos & Pantazopoulou 2010). The black solid line represents the base case assumed when neglecting uncertainty.

of the seismic vulnerability under the influence of such uncertainties has been recognized as an important constituent of the structural design and analysis process, as exemplified, at least qualitatively, by the SAC/ FEMA-350 guidelines (FEMA, 2000). Nevertheless, only recently have we seen actual attempts to quantify this effect for realistic structural models in a way that is consistent with current performance-based earthquake engineering frameworks. Such studies include mainly the work of Dolsek (2008), Liel et al. (2009) and Vamvatsikos & Fragiadakis (2010) who propose methods to account for the uncertainty in modelling parameters and its effect on the estimated structural fragility using Monte Carlo techniques on incremental dynamic analysis. However, such methodologies remain computationally intensive and often difficult to apply for practical purposes. As a partial remedy, Fragiadakis & Vamvatsikos (2010) have offered a simplified process based on the static pushover and the SPO2IDA tool (Vamvatsikos & Cornell 2005) that manages at least two orders of magnitude reduction in the processing load at an insignificant loss of accuracy for first-mode dominated buildings. Finally, on the same track, Vamvatsikos & Pantazopoulou (2010) have recently applied such efficient techniques for the simplified evaluation of the seismic vulnerability of groups of masonry structures, typical of historical city cores. A representative example appears in Fig. 3 where the multiple uncertain pushover capacity curves of a two story building appear versus the single curve that would typically be used in a deterministic assessment. Nevertheless, there is still considerable room for refinement in this area, and future developments will play a major role in the new generation of seismic guidelines.

3.4.3 PERFORMANCE ASSESSMENT OF STRUCTURES SUBJECTED TO LANDSLIDES AND FLOWSLIDES 3.4.3.1 Introduction Flowslides and debris flows can be considered as one of the most dangerous slope movements for their capability to produce casualties and remarkable economic damage. Such phenomena are widespread in many countries and involve different kind of soils, generally in a loose state, which in the post failure stage collapse and rapidly reach the toe of the slope; the initial mobilised mass often increases during its path downslope either by inducing additional slope failure and/or by eroding the stable in place soils (Cascini et al., 2003). Significant examples of this type of slope movements have occurred in several areas of the world. For example, those periodically occurring in the Campania Region (South Italy) triggered by critical rainfall events. They involve unsaturated 276

Figure 4.

a) Height of debris flow; b) Breaking of brick or tuff external walls in R.C. framed structure.

pyroclastic soils – originated by the explosive phases of the Somma-Vesuvius volcano – which mantle the limestone and tuffaceous slopes over an area of about 3000 km2 . In this area, there are more than 200 towns that frequently suffer from flowslides, as pointed out by historical data acquired over a period from the 16th century up to the present (Cascini & Ferlisi, 2003). One of the worst events occurred on May 5–6, 1998, when 159 casualties and serious damages were recorded in four towns (Bracigliano, Quindici, Sarno and Siano) located at the toe of the Pizzo d’Alvano relief. During the quoted hydrogeological disaster of 1998 in the Campania Region, numerous flowslides due to the detachment of the pyroclastic deposits from the calcareous massif of “Pizzo d’Alvano” impacted the constructions which were located near incisions and valleys, determining wide-spread damage. The following paragraphs contain the description and the assessment models concerning the effects of the dynamic impact of the flowslide on constructions, generally buildings, with special care taken of structural resistance and/or vulnerability. 3.4.3.2 Damage and collapse mechanisms The surveys and the analysis of building damage during the quoted hydrogeological disaster of 1998 in the Campania Region, realised immediately after the event, allow us to better understand the impact of debris flows on constructions and their collapse mechanismsand allow an evaluation of the impact velocity of the flows on the constructions (Faella & Nigro, 2003a,b). The effects of the debris flow impact on the constructions are significantly different depending on the following parameters: – position of the construction with reference to the impact direction; – level of kinetic energy of the debris flow, related to its velocity; – structural typology (reinforced-concrete or masonry buildings, structural or non-structural members). On the basis of the analysis of the structural and non-structural damages in the buildings impacted by the debris flows, it is possible to derive the following synthesis (see Faella & Nigro, 2003a), referring to the main collapse mechanisms (Figures 4–7): – Reinforced concrete framed buildings: A) Collapse of the ground floor external walls, directly impacted by the flows, without significant damage to the structural parts (columns and beams); B) Serious damage or collapse of single structural elements, generally columns, without collapse of the whole structure, but with formation of plastic hinges at the ends and/or in the midspan of the columns; C) Serious damage and/or collapse of the structure, with formation of floor mechanism (plastic hinges at the top and bottom of the column); D) Translation of part of the building as a consequence of the collapse of the ground floor bearing structures. 277

Figure 5.

R.C. structures: a) Breaking of brick or tuff external walls; b) Failure of corner column.

Figure 6.

Reinforced concrete structures: Plastic collapse mechanism of columns.

Figure 7.

a) Masonry building impacted by debris flows; b) Residual parts of masonry buildings.

– Masonry buildings: E) Serious damage and/or collapse of bearings walls at ground floor, directly impacted by flows, without collapse of the overall building; F) Serious damage and/or collapse of the overall building. The described types of damages may be interpreted by means of appropriate collapse mechanisms in order to assess the bearing capacity at the ultimate limit state of elements or of the 278

overall structure and, with furthermore considerations of hydrostatic and hydrodynamic, also an approximate evaluation of the impact velocity of the debris flows on the constructions.

3.4.3.3 Performance assessment of structures subjected to flowslides The analysis and the interpretation of the structural and non-structural damages in the buildings impacted by the debris flows point out some types of collapse mechanisms for reinforced concrete and masonry buildings, described in the previous paragraphs. The described damage types can be interpreted by means of appropriate collapse mechanisms, which allow us to assess the ultimate bearing capacity of members or of the overall structure. The comparison between the ultimate bearing capacity and hydrostatic and hydrodynamic thrusts due to the flow impact on the structures allows assessment of the impact velocity which determines the collapse of the member or the structure. In the hydrodynamic models the hypothesis of a fluid stream of constant density is assumed, neglecting the possible presence of mass concentration (for instance trees, rocks and other transported material). In Table 1 and Table 2 the mechanical models related to the main collapse mechanisms and the analytical formulations to evaluate the corresponding impact velocities are summarised with reference to respectively reinforced concrete and masonry structures. More details can be found in Faella & Nigro (2003b). It is important, to point out the uncertainties in both of the hydrodynamic and structural models. In the hydrodynamic models, the direction of the debris flow has to be assumed on the basis of the position of the construction, considering with approximate formulations the influence of the impacted member shape; moreover, the height of the debris flows is generally assumed equal to the first floor height on the basis of the surveys of the real cases. In the structural models, the approximations refer to the material strength and the evaluation of the internal forces due to vertical loads. Nevertheless, the whole approximations don’t invalidate in significant way the results of the assessing models, due to the moderate influence of the different parameters of uncertainty.

3.4.3.4 Application of the models for the assessment of structures subjected to flowslides The models described in the previous paragraph are now applied to some significant buildings, selected between those surveyed in the post-event of the quoted hydrogeological disaster of 1998 in the Campania Region, with the purpose to assess the debris flow impact velocity on the basis of the surveyed damages. In some cases it is possible only to deduce a lower bound of the velocity, as for instance in the case of masonry walls destroyed by the debris flow and in the case of global collapse of the building. In other cases, instead, the range which contains the probable impact velocity can be evaluated: this is possible, for instance, when the debris flow has destroyed the external walls of a reinforced concrete building without the failure of the ground-floor columns, or when some impacted columns have collapsed and others have withstood the impact due to their greater bearing capacity. (the last one is the case of Figure 5b). In the application of the interpretative models described in the previous paragraph it is assumed that the specific weight of the fluid is equal to γ = 14.00 kN/m3 (density ρ = 1427.1 kg/m3 ). The complete results in terms of debris flow impact velocity are reported in Faella & Nigro (2003b). The analysis of the results allows some interesting remarks: – The collapse of masonry buildings impacted by debris flows occurs in the presence of relatively low velocities (approximately lower than 5÷6 m/s); in some cases, moreover, only hydrostatic thrust is enough to determine the collapse. – The collapse of external walls in reinforced concrete buildings occur for very low velocities (about 3 m/s). – Reinforced concrete buildings completely impacted by debris flows exhibit intermediate value of collapse velocity (about 10 m/s); in this case the collapse model is interpreted by the twoplastic-hinges mechanism (see type-C mechanism – Table 1), related to the formation of storeyfailure-mechanism at the ground floor. – In the case of reinforced concrete buildings only partially impacted by debris flow, instead, the failure of single columns may occur; the corresponding velocities are greater than the previous 279

Table 1. Collapse resistant models for assessment of reinforced concrete buildings.

cases (about within the range 15÷20 m/s), due to the most favourable three-plastic-hinges failure mechanism (see type-B mechanism – Table 1). – Obviously, the obtained results are related to the examined building types, characterised by two or three-floors buildings; if the floors number increases, also the collapse velocities increase 280

Table 2. Collapse resistant model for assessment of masonry buildings.

both for masonry and reinforced concrete buildings: in the first case, the resistance capacity of the ground-floor walls increases thanks to the increments of the acting vertical load and the wall thickness; in the second case, the geometric dimensions of the ground floor columns and the corresponding axial forces increase, determining the increment of the ultimate bending moments. 3.4.3.5 Final remarks The main topic of this subsection is to investigate the possible effects of flowslides on the urban areas exposed to such risks, e.g., the urban areas around the Vesuvius. With this aim, mechanical models deduced utilizing also hydrodynamic concepts are introduced; the models are capable to interpret the effects of the landslide impact on the constructions and the collapse mechanisms of various types of structures. The application of these models to building types representative of the urban areas around the Vesuvius allow us to estimate their vulnerability against the expected landslide events, providing some useful information concerning the risk mitigation.

3.4.4 TSUNAMI VULNERABILITY 3.4.4.1 Introduction Vulnerability analysis is a concept still in its infancy for tsunami risk assessment. It is fraught with issues due to scarcity of events, which result in lack of knowledge on the behaviour of tsunami waves in the near and onshore regions. Furthermore, the rarity of events also means a lack of good information on tsunami impacts, in terms of damage to structures and infrastructure.Also, until the 2004 Indian Ocean Tsunami few structural engineers have taken an interest in looking at the vulnerability of structures to tsunami. The following paragraphs outline some of the current state of art, its limitations and the issues still to be solved for the generation of accepted tools for the vulnerability assessment of structures impacted by tsunami. 3.4.4.2 Tsunami damage to structures When observing damage to structures along coastal regions hit by tsunami it is common to see a large variation in the degree of damage to buildings (from total devastation where the majority of the buildings have collapsed, to light damage where only windows and shutters are damaged). This variation is partly due to structural vulnerability to lateral forces and partly due to differences in water forces resulting from local variations in shoreline topography, bathymetry, and possibly the presence of coastal vegetation and coral formations (Rossetto et al., 2007). Quantification of the influence of the latter factors requires further research, however a close correlation is commonly observed between locations of highest sustained damage and highest measured run-up (EEFIT 281

2006, 2009). Four main causes of damage to buildings and infrastructure can usually be identified (Rossetto et al., 2007): – – – –

lateral water flow, wave loading, ground scour (causing subsidence or foundation failure of buildings), debris impact.

Large horizontal components of water movement and turbulence are associated with tsunami waves entering shallow waters, which result in the entrainment of sediment and coastal erosion. Also, in the case of tsunami, large objects such as water tanks, cars and boats can be carried by the water flows and impact buildings. Typically the degree of damage decreased with increasing distance from the coastline and increasing obstacles between the building and the sea. Furthermore, the amount of damage was influenced by the construction type, with older masonry buldings being particularly vulnerable, whilst modern R/C frame structures perform better (EEFIT, 2006). In the case of tsunami, very few guidance documents have been developed for use in post-event damage assessments. The Intergovernmental Oceanographic Commission, IOC (of UNESCO, 1998) has published a post tsunami field guide developed from existing earthquake and tsunami field guides and more recent tsunami surveys (Farreras, 2000). While concentrating on collecting scientific data such as tidal levels, run-up elevations and bathymetric data, it indicates that structural damage should be collected where possible, noting the possible cause of the damage and distinguishing tsunami damage from earthquake damage in a near source event. The guidance for building damage assessment is brief and recommends rough (non-specialized) classification of damage, estimating the nature and category of the damage and its apparent cause. Several approaches exist for identifying tsunami intensity (e.g. Ambraseys 1962 and Papadopoulos and Imamura 2001). However, these methods do not provide techniques for identifying structural damage. Most of the literature presenting rapid field investigations largely bases their damage assessments on earthquake assessment methodologies directly. Rigorous, multi-stage building assessments using forms do not exist, or at least have not been published. Instead, the damage scales in EMS-98 are the most commonly used (e.g. in Miura et al., 2006). A few studies have attempted to modify earthquake damage assessment methods and scales to take into account damage relating to fast-flowing water, such as foundation failure due to scour or floating debris impact damage. A modified version of the EMS-98 damage scales for use in tsunami damage assessment in Thailand and Sri Lanka following the Indian Ocean Tsunami was proposed by Rossetto et al. (2007) and EEFIT (2006). In these studies damage attributed to different building types was also adopted to assign Intensity values to the surveyed locations, using a modified version of theTsunami Intensity scale of Papadopoulos and Imamura (2001). 3.4.4.3 Vulnerability relationships Due to their rarity, observational damage data required for the generation of empirical vulnerability curves (see Section 3.4.2.5.1) is insufficient. Nevertheless, the potential value of such vulnerability curves has meant several researchers have tried to derive vulnerability functions based on particular events using data from damage surveys (e.g. Peiris 2006, Ruangrassamee et al., 2006 and Reese et al. 2007). These empirical vulnerability functions are based on few data points and due to the nature of recent tsunami events encompass only certain types of non-engineered buildings (generally low-rise masonry). Koshimura et al. (2009) have attempted to improve their sample of damage data through the interpretation of building damage from pre- and post- tsunami satellite imagery. Although this does produce a larger sample size, only the damage state of collapse can really be identified from the satellite imagery and so curves for lesser damage states cannot be derived. Also, structure type cannot be discerned from the roof type and hence all buildings are considered together. This is incorrect in view of the observations made above in 3.4.4.2. Assets that play a key role in the response to disaster like tsunami are often elements of the transport infrastructure. A functional road network is essential for rapid evacuation, the deployment of medical supplies and movement of injured persons. Serviceable routes for transportation continue to be vital during the recovery stage for the management of reconstruction. However, though there are limited studies regarding vulnerability of structures, even fewer exists for the assessment of the often critical bridges in transport infrastructure. Shoji & Moriyama (2007) examined the vulnerability of bridge structures in Indonesia and Sri Lanka following the Boxing Day 2004 Indian 282

Ocean tsunami. 60 data from Sri Lanka and 27 from Indonesia (collected by the JSCE) were used to derive vulnerability curves using inundation height only as the severity parameter. Differences in vulnerability were found between the two locations and also between bridge construction types, though the data was limited in quantity. Similar studies and damage data collection for future events would be advantageous to this type of work. 3.4.4.4 Characterising tsunami loads on buildings All existing empirical vulnerability functions for tsunami adopt inundation depth as the parameter describing tsunami flow intensity, as this is one of the only measurable parameters of tsunami onshore flow that can be obtained in the field following an event (e.g. through observation of water level marks on the sides of buildings). Most tsunami design codes, where they exist (FEMA 2008, Okada et al., 2005), predominantly use inundation heights to derive maximum forces for design, so this is a reasonable parameter to link to vulnerability. However, it should be noted that design formulae for pressures and forces are also dependant on velocity, so height of water alone may not be the sole parameter that should be considered. Unfortunately readings of tsunami velocities are almost always not available. Discrepancies in the determination of these forces also exist between various design guides from around the world. Koshimura et al. (2009) adopt a numerical model to simulate the onshore flow of the Indian Ocean Tsunami in Banda Aceh. The numerical model is based on non-linear shallow water wave equations and the presence of structures is accounted for as an additional roughness term. This numerical model is shown to provide reasonable inundation depths but unrealistic flow velocities. This observation is common to most commercial numerical models for onshore flow estimation, with none being able to account for the complex interaction between the water, buildings, sediments etc. Hence, the derivation of vulnerability functions is hindered significantly by a lack of appropriate numerical models, and the development of the latter is hampered by the scarcity of field data for their calibration and validation. Furthermore, as in the case of flow slides, modelling of transported debris in the water is difficult and is commony omitted from analyse, despite the potentially high induced damage levels from debris impact on structures. All these issues pose a significant problem for the development of methods for the assessment of individual structures for tsunami actions. Codes and guidance is an ongoing area of research. Where they exist, the prescriptive steps to assessing a structure in terms of its tsunami vulnerability is not usually dealt with, rather pointers as to what analysis is needed are given. Some codes give example calculations for certain types of force (FEMA 2005), but significant engineering judgment is required for all such designs. An issue with the force calculations present in codes is the data they are based upon. Actual measurements from tsunami in the ocean have been limited to tide gauge data (elevations) which often get drowned out and are subject alteration of the wave due to interactions with the continental shelf. Tsunami buoys and bottom pressure sensors have been deployed in an attempt to acquire readings in this area for the validation of numerical models. This data while extremely useful for offshore modeling bears very little correlation with what actually happens in the near shore region, and inundation zone. Unfortunately there is no such data recording forces on structures and what information is available is due to physical scale models. Wave modeling until recently has been entirely conducted using piston-type wave generators which have limited stroke length, so the wave length of the generated wave is a limiting factor to the generation of realistic tsunami waves. To address this problem a novel pneumatic system of wave generation has been developed (Lloyd et al., 2009, Rossetto et al., 2010). In 2008–9 large-scale tests were carried out in a flume in HR Wallingford in the UK specifically to look at near shore and onshore processes (Figure 8). In these experiments, velocity, pressure and force measurements of waves on model building structures have been determined and a better understanding of tsunami forces will be gained enabling better vulnerability analysis in the future. Physical modeling undoubtedly still has a large role in better understanding tsunami and developing better design codes to deal with them. 3.4.5 VULNERABILITY TO STRONG WIND EVENTS – HURRICANES Existing studies on the vulnerability of structures to extreme wind events are generally classified as dealing with (a) damage assessment, (b) field examinations of wind-structure interaction 283

Figure 8. (a) Tsunami generator in flume at HR Wallingford; (b) Model instrumented structure subjected to “tsunami” wave.

and (c) hurricane risk assessment from the insurance perspective. We will discuss each of these categories in the pages to follow. 3.4.5.1 Damage assessment Research in this area mostly deals with observed damage from extreme wind events. The main objectives are to correlate building damage intensity to measured wind speeds and to examine building performance for those cases where wind speeds were close to building code values. A good example is the work of Mehta et al. (1983). Therein, buildings were grouped in various categories: – Fully engineered buildings, which performed well, even for wind speeds above the codespecified values. Limited damage was observed on roofing material and façade. – Pre-engineered buildings suffered from structural framing damage for wind speeds close to, or over, the code-suggested values. Weak links (e.g. overhead door) were identified in such structures and held responsible for progressive damage. – Marginally engineered buildings, which were affected significantly at all wind speed regimes. – Non-engineered buildings, which were severely damaged when wind speed reached the codespecified values. Furthermore, wind-induced damage can be classified to structural (lack of uplift load path, roof sheathing loss at corners and gable end wall loss) and non-structural (loss of roof shingles and vinyl siding, vulnerability of soffits, breach through attic vents and better performance of hip roofs over gable roofs) – see Van de Lindt et al. (2007). The concept of wind load path has been also used and discussed by Stathopoulos et al. (2008). 3.4.5.2 Field studies of wind-structure interaction This category deals with the effect of wind-structure interaction, which can be very important for groups of buildings, where the wind force acting on each building is heavily influenced by the nearby structures that may be shielding it or channeling the wind. Typical studies of this type include the following: – – – – – –

National Bureau of Standards (Marshall 1975) Aylesbury Experiment (Eaton & Mayne 1975) Texas Tech University Project – WERFL (Levitan et al., 1990) Silsoe Structure – BRE (Robertson & Glass, 1988) Southern Shores Project (Caracoglia & Jones 2004) Florida Coastal Monitoring Project – FCMP (Datin et al., 2006) 284

– Load Paths on Wood Buildings and Engineering Design of Low-Rise Wood Buildings Projects (Doudak 2005 and Zisis 2007) – International Hurricane Research Center and Florida International University (Leatherman et al. 2007) – Insurance Research Lab for Better Housing (Bartlett et al., 2007). 3.4.5.3 Hurricane risk assessment Within the context of insurance and risk modeling, it has become important to estimate the risk faced by structures, especially in the catastrophic event of a hurricane. Friedman (1984) has discussed the Risk related information related to hurricane events in the context of defining risk assessment models to be used for insurance purposes. Furthermore, he has dealt with the loss-producing potential of a structure influenced by various factors, among which is vulnerability, and also with wind speed-damage models used by insurance companies Berz & Smolka (1988) attempted to carry out risk assessment and rating, which involve not only detailed weather data but information regarding local design and construction practices. Khanduri & Morrow (2002) discussed vulnerability assessment of buildings to strong wind events. Their effort was mainly focused on refining the “general” vulnerability models by specific and detailed models related to factors such as different building types, occupancy, construction material, height etc. Finally, Stewart (2003) offers a discussion about building vulnerability models for hurricane/cyclone events, especially on the effect of vulnerability variations (e.g. retrofitting, enhancement of building standards etc) on existing models.

3.4.6 CONCLUSIONS Although we have only briefly touched upon a number of difficult problems, it appears that there is a common structure underlying the estimation of seismic, landslide, tsunami and wind vulnerability of structures. There is a significant focus on the use of empirical data from historical events while there are many efforts underway to provide powerful analytical tools to assess vulnerability at various scales, ranging from a single building up to an entire city. This is especially true for the more mature fields of seismic and wind research were abundant data and methodologies are available. On the other hand, comparative research efforts in hazard modeling are still under way but on the right path to catch up eventually. In summary, it seems that the proper groundwork already exists for preparing a common framework for vulnerability assessment in all three areas, and it is expected that future research will lead towards this promising direction. REFERENCES Ambraseys, NN. 1962. Data for the investigation of the seismic sea-waves in the eastern Mediterranean. Bull. Seismol. Soc. Am. 52, 895–913. ASCE/SEI 41, 2007. ASCE Standard ASCE/SEI 41-06, Seismic Rehabilitation of Existing Buildings, ASCE, Reston, VA. ATC. 1985. Earthquake damage evaluation data for California. Applied Technology Council Report ATC-13, Redwood City, California, USA. ATC. 1996 ATC-40: Seismic evaluation and retrofit of concrete buildings, Rep. SSC 96-01, CSSC-ATC, Redwood City, California, USA. Augusti, G., M. Ciampoli & P. Giovenale. 2001. Seismic vulnerability of monumental buildings. Structural Safety 23, (3): 253–74. Bal, I. E., H. Crowley & R. Pinho. 2008. Displacement-based earthquake loss assessment for an earthquake scenario in istanbul. J.Earthqu.Eng. 12, (SUPPL. 2): 12–22. Barbat, A.H., Moya, F.Y., Canas, J.A., et al. 1996. Damage Scenarios Simulation for Seismic Risk Assessment in Urban Zones. Earthquake Spectra 12(3): 371–394. Barbat, A. H., L. G. Pujades & N. Lantada. 2008. Seismic damage evaluation in urban areas using the capacity spectrum method: Application to Barcelona. Soil Dynamics and Earthquake Engineering 28, (10–11): 851–65.

285

Bartlett, F. M., Galsworthy, J. K., Henderson, D., Hong, H. P., Iizumi, E., Inculet, D. R., et al., 2007. The three little pigs project: A new test facility for full-scale small buildings. 12th International Conference on Wind Engineering, Cairns, Australia. 1623–1630. Berz, G. & Smolka, A., 1988. Windstorm hazard and insurance. Journal of Wind Engineering and Industrial Aerodynamics, 30, 191–204. Benedetti, D., G. Benzoni & M. Parisi. 1988. Seismic vulnerability and risk evaluation for old urban nuclei. Earthquake Engineering & Structural Dynamics 16 (2): 183–201. Boatwright, J., Thywissen, K. & Seekins, L.C. 2001. Correlation of Ground Motion and Intensity for the 17 January 1994 Northridge, California, Earthquake. Bulletin of the Seismological Society of America, 91(4): 739–752. Borzi, B., H. Crowley & R. Pinho. 2008. Simplified pushover-based earthquake loss assessment (sp-bela) method for masonry buildings. International Journal of Architectural Heritage 2, (4): 353–76. Calvi, G. M. 1999. A displacement-based approach for vulnerability evaluation of classes of buildings. J.Earthqu.Eng. 3, (3): 411–38. Caracoglia, L., & Jones, N., 2004. Full-scale pressure measurement during hurricanes. Proceedings of 5th Intern. Colloquium on Bluff Body Aerodynamics & Applications, Ottawa, ON, Canada, 485–488. Casciati, F., L. Faravelli & X. D. Liu. 1994. Lessons of rehabilitation design learned from statistical analyses of masonry seismic-vulnerability data. Structural Safety, 16, (1–2) (10): 73–89. Cascini L., Ferlisi S., 2003. Occurrence and Consequences of Flowslides: a Case Study, in L. Picarelli (Ed.), Proceedings of Int. Conf. on Fast Slope Movements-Prediction and Prevention for Risk Mitigation (FSM2003), Napoli (Italy), 11–13 May 2003. Patron Ed.: Bologna. Cascini L., Sorbino G., Cuomo S., 2003. Modelling of flowslides triggering in pyroclastic solis, in L. Picarelli (Ed.), Proceedings of Int. Conf. on Fast Slope Movements-Prediction and Prevention for Risk Mitigation (FSM2003), Napoli (Italy), 11–13 May 2003. Patron Ed.: Bologna. CEN 2004, Design of Structures for Earthquake Resistance, Part 3, Assessment and Retrofitting of Buildings, EN 1998–3, Brussels. Coburn, A. & Spence R. 2002. Earthquake Protection, Second Edition, John Wiley & Sons, Ltd. Cole, S. W., Y. Xu & P. W. Burton. 2008. Seismic hazard and risk in Shanghai and estimation of expected building damage. Soil Dynamics and Earthquake Engineering 28, (10–11): 778–94. Datin, P. L., Liu, Z., Prevatt, D. O., Masters, F. J., Gurley, K., & Reinhold, T. A., 2006. Wind loads on single-family dwellings in suburban terrain: Comparing field data and wind tunnel simulation. Structures Congress 2006, St. Lois, Missouri, USA., 201(40889) 13. Dolce, M., A. Kappos, A. Masi, Gr. Penelis & M. Vona. 2006. Vulnerability assessment and earthquake damage scenarios of the building stock of potenza (southern italy) using italian and greek methodologies. Engineering Structures, 28, (3) (2): 357–71. Dolsek M., 2008. Incremental dynamic analysis with consideration of modeling uncertainties. Earthquake Engineering and Structural Dynamics, 38: 805–825. Doudak, G., 2005. Field determination and modeling of load paths in wood light-frame structures. PhD Thesis, McGill University, Montreal, QC, Canada. Eaton, K. J., & Mayne, J. R., 1975. The measurement of wind pressures on two-storey houses at Aylesbury. Journal of Wind Engineering and Industrial Aerodynamics, 1, 67–109. EEFIT, 2006. The Indian Ocean Tsunami of 26 December 2004: Mission Findings in Sri Lanka & Thailand. Pomonis, A., Rossetto, T., Peiris, N., Wilkinson, S.M., Del Re, D., Koo, R., and Gallocher S., Earthquake Engineering Field Investigation Team (EEFIT) Report, The Institution of Structural Engineers, London, UK. http://www.eefit.org.uk. EEFIT, 2009. The South Pacific islands earthquake and tsunami of 29 september 2009. So, E., Hunt-Raby, A., Robinson, T., Fraser, S. and Lloyd, T. Earthquake Engineering Field Investigation Team (EEFIT) Report, The Institution of Structural Engineers, London, UK. http://www.eefit.org.uk. Erberik, M. A. 2008a. Fragility-based assessment of typical mid-rise and low-rise RC buildings in turkey. Engineering Structures 30, (5): 1360–74. Erberik, M. A. 2008b. Generation of fragility curves for turkish masonry buildings considering in-plane failure modes. Earthquake Engineering & Structural Dynamics 37, (3): 387–405. Faccioli, E., V. Pessina, G. M. Calvi & B. Borzi. 1999. A study on damage scenarios for residential buildings in catania city. Journal of Seismology 3, (3): 327–43. Faella C., Nigro E., 2003a. Dynamic Impact of the Debris Flows on the Constructions during the Hydrogeological Disaster in Campania-1998: Description and Analysis of the Damages, in L. Picarelli (Ed.), Proceedings of Int. Conf. on Fast Slope Movements-Prediction and Prevention for Risk Mitigation (FSM2003), Napoli (Italy), 11–13 May 2003. Patron Ed.: Bologna.

286

Faella C., Nigro E., 2003b. Dynamic Impact of the Debris Flows on the Constructions during the Hydrogeological Disaster in Campania-1998: Failure Mechanical Models and Evaluation of the Impact Velocity, in L. Picarelli (Ed.), Proceedings of Int. Conf. on Fast Slope Movements-Prediction and Prevention for Risk Mitigation (FSM2003), Napoli (Italy), 11–13 May 2003. Patron Ed.: Bologna. Faella C., Nigro E., 2003c. A Hypothesis of Technical Code on the Structural Design and Repairing in Urban Areas with High Debris Flow Risk, in L. Picarelli (Ed.), Proceedings of Int. Conf. on Fast Slope MovementsPrediction and Prevention for Risk Mitigation (FSM2003), Napoli (Italy), 11–13 May 2003. Patron Ed.: Bologna. Fäh, D., F. Kind, K. Lang & D. Giardini. 2001. Earthquake scenarios for the city of Basel. Soil Dynamics and Earthquake Engineering 21, (5) (7): 405–13. Farreras SF. 2000. Post-Tsunami Field Survey Procedures: An Outline. Natural Hazards 21, 207–214FEMA 1997. NEHRP Guidelines for the Seismic Rehabilitation of Buildings, FEMA-273, Federal Emergency Management Agency,Washington D.C. FEMA 2000. Prestandard and commentary for the seismic rehabilitation of buildings, FEMA-356, Federal Emergency Management Agency, Washington D.C. FEMA 2000. Recommended seismic design criteria for new steel moment-frame buildings. Report No. FEMA350, SAC Joint Venture, Federal Emergency Management Agency, Washington DC. FEMA & NIBS. 2003. Multi-hazard loss estimation methodology – earthquake model: HAZUS® MH technical manual, Federal Emergency Management Agency, Washington D.C. FEMA. 2005. Coastal Construction Manual (3 vols.). 3rd ed. (FEMA 55). Federal Emergency Management Agency, Washington DC. FEMA. 2008. Guidelines for design of structures for vertical evacuation from tsunami (FEMA P646). Federal Emergency Management Agency, Washington DC. Fragiadakis M. & Vamvatsikos D. 2010. Fast performance uncertainty estimation via pushover and approximate IDA. Earthquake Engineering and Structural Dynamics, 39(6): 683–703. Friedman, D.G., 1984. Natural hazard risk assessment for an insurance program. The Geneva Papers on Risk and Insurance, 9, No. 30, 57–128. Gent Franch, K. A., G. M. Giuliano Morbelli, M. A. Astroza Inostroza & R. E. Gori. 2008. A seismic vulnerability index for confined masonry shear wall buildings and a relationship with the damage. Engineering Structures 30, (10): 2605–12. Grant, D. N., J.J. Bommer, R. Pinho, G. M. Calvi, A. Goretti & F. Meroni. 2007. A prioritization scheme for seismic intervention in school buildings in italy. Earthquake Spectra 23, (2) (May 2007): 291–314. Grünthal G. ed. European macroseismic scale 1998 (EMS-98). Cahier du Centre Européen de Géodynamique et de Séismologie, Conseil de L’Europe, Luxemburg, Vol.15. Hill, M., Rossetto, R. 2008. Comparison of building damage scales and damage descriptions for use in earthquake loss modelling in Europe, Bull. Earthq. Eng., 6(2): 335–365. Imamura, A. 1942. History of Japanese tsunamis. Kayo-No-Kagaku (Oceanography), 2, 74–80 (in Japanese) Kappos, A.J. 2007. Seismic vulnerability and risk assessment of urban habitat in Southern European cities. Keynote lecture, Urban Habitat Constructions under Catastrophic Events Workshop (COST C26), Prague, Mar., 115–129. Kappos, A.J., Panagopoulos, G., Panagiotopoulos, C. & Penelis, Gr. 2006. A hybrid method for the vulnerability assessment of R/C and URM buildings. Bull. Earthq. Eng., 4(4): 391–413. Kappos, A. J., V. Lekidis, G. Panagopoulos, I. Sous, N. Theodulidis, Ch Karakostas, T. Anastasiadis, T. Salonikios & B. Margaris. 2007. Analytical estimation of economic loss for buildings in the area struck by the 1999 athens earthquake and comparison with statistical repair costs. Earthquake Spectra 23, (2): 333–55. Kappos, A. J., K. C. Stylianidis & K. Pitilakis. 1998. Development of seismic risk scenarios based on a hybrid method of vulnerability assessment. Natural Hazards 17, (2): 177–92. Kappos, A. J., G. Panagopoulos & Gr. G. Penelis. 2008. Development of a seismic damage and loss scenario for contemporary and historical buildings in thessaloniki, greece. Soil Dynamics and Earthquake Engineering 28, (10–11) (0): 836–50. Kappos A.J. & Panagopoulos G. 2010. Fragility curves for reinforced concrete buildings in Greece. Structure and Infrastructure Engineering. 6(1–2): 39–53. Khanduri, A.C. & Morrow, G.C., 2003. Vulnerability of buildings to windstorms and insurance loss estimation. Journal of Wind Engineering and Industrial Aerodynamics, 91, 455–467. King, S.A.,. Kiremidjian, A.S, Sarabandi, P. & Pachakis, D. 2005. Correlation of Observed Building Performance with Measured Ground Motion. Report TR148. The John A. Blume Earthquake Engineering Center. Department of Civil and Environmental Engineering Stanford University.

287

Koshimura, S., Oie, T., Yanagisawa, H., Imamura, F. 2009. Developing Fragility Functions For Tsunami Damage Estimation Using Numerical Model and Post-Tsunami Data From Banda Aceh, Indonesia. Coastal Engineering Journal, Vol. 51, No. 3 (2009) 243–273. Lagomarsino, S. & S. Podestà. 2004a. Seismic vulnerability of ancient churches: I. damage assessment and emergency planning. Earthquake Spectra 20(2): 377-394. Lagomarsino, S. & S. Podestà. 2004b. Seismic vulnerability of ancient churches: II. statistical analysis of surveyed data and methods for risk analysis. Earthquake Spectra 20(2): 395–412. Lagomarsino, S. 2006. On the vulnerability assessment of monumental buildings. Bulletin of Earthquake Engineering 4, (4): 445–63. Lagomarsino, S. & Giovinazzi, S. 2006. Macroseismic and mechanical models for the vulnerability and damage assessment of current buildings. Bull. Earthq. Eng. 4(4): 415–443. Lang, K. 2002. Seismic vulnerability of existing buildings. PhD thesis, ETH, Zurich, Hochschulverlag AG. Lang, K. & H. Bachmann. 2004. On the seismic vulnerability of existing buildings: A case study of the city of Basel. Earthquake Spectra 20 (1): 43–66. Lang, K. 2003. On the seismic vulnerability of existing unreinforced masonry buildings. Journal of Earthquake Engineering 7 (3): 407. Lantada, N., L. G. Pujades & A. H. Barbat. 2008. Vulnerability index and capacity spectrum based methods for urban seismic risk evaluation. A comparison. Natural Hazards: 1–24. Leatherman, S. P., Gan Chowdhury, A., & Robertson, C. J., 2007. Wall of wind full-scale destructive testing of coastal houses and hurricane damage mitigation. Journal of Coastal Research, 23(5), 1211–1217. Levitan, M. L., Mehta, K. C., Chok, C. V., & Millsaps, D. L., 1990. An overview of Texas Tech’s wind engineering field research laboratory. Journal of Wind Engineering and Industrial Aerodynamics, 36(Part 2), 1037–1046. Liel A., Haselton C., Deierlein G. & Baker J. 2009. Incorporating modeling uncertainties in the assessment of seismic collapse risk of buildings. Structural Safety 31(2):197–211. Lloyd TO, Allsop, W, Charvet I, Robinson D, Robinson T, Rossetto T. Physical modelling of violent flows and their impact using a new tsunami generator. 9th ICE Coasts, Marine Structures and Breakwaters Conference 2009. Marshall, R. D., 1975. A study of wind pressures on a single-family dwelling in model and full scale. Journal of Wind Engineering and Industrial Aerodynamics, 1, 177–199. Masi, A. 2003. Seismic Vulnerability Assessment of Gravity Load Designed R/C Frames. Bulletin of Earthquake Engineering 1(2): 371–395. Mehta, K.C., Minor J.E. & Reinhold T.A., 1983. Speed-damage correlation in Hurricane Frederic. Journal of the Structural Division Vol. 109, ASCE, No. 1. Miura, H., Wijeyewickrema, A. & Inoue, S. 2006. Evaluation of tsunami damage in the eastern part of Sri Lanka due to the 2004 Sumatra earthquake using remote sensing technique. Proc. 8th National Conference on Earthquake Engineering, Paper No. 8, NCEE-856Miyakoshi J., Hayashi Y., Tamura K. & Fukuwa N. 1997. Damage ratio functions of buildings using damage data of the 1995 Hyogoken-Nanbu Earthquake. ICOSSAR 1997. pp. A.19–A.24 Mosalam K.M., Ayala G., White R.N. & Roth C. 1997. Seismic fragility of LRC frames with and without masonry infill walls. Journal of Earthquake Engineering 1(4): 693–719. Oliveira, C. S. 2008. Lisbon earthquake scenarios: A review on uncertainties, from earthquake source to vulnerability modelling. Soil Dynamics and Earthquake Engineering 28, (10–11): 890–913. Okada, T., Sugano, T., Ishikawa, T., Ohgi, T., Takai, S. & Hamabe,C. 2005. Structural design methods of buildings for tsunami resistance (SMBTR). The Building Center of Japan, Japan. Papadopoulos, GA., Imamura, F. 2001. A proposal for a new tsunami intensity scale. ITS Proceedings, 569. Park, J., P. Towashiraporn, J. I. Craig & B. J. Goodno. 2009. Seismic fragility analysis of low-rise unreinforced masonry structures. Eng.Struct. 31, (1): 125–37. Peiris, N. 2006. Vulnerability functions for tsunami loss estimation. First European Conference on Earthquake Engineering and Seismology. Penelis, Gr G., A. J. Kappos & K. C. Stylianidis. 2003. Assessment of the seismic vulnerability of unreinforced masonry buildings. Paper presented at 8th International Conference on Structural Studies, Repairs and Maintenance of Heritage Architecture, Chalkidiki, Greece, WIT Press, Southampton, UK. Penelis, Gr G. 2006. An efficient approach for pushover analysis of unreinforced masonry (URM) structures. Journal of Earthquake Engineering 10, (3): 359–79. Reese, S., Cousins, WJ., Powers, WL., Palmer, NG., Tejakusuma, IG. and Nugrahadi, S. 2007. Tsunami vulnerability of buildings and people in South Java – field observations after the July 2006 Java tsunami. Nat. Hazards Earth Syst. Sci., 7, 573–589.

288

Robertson, A. P., & Glass, A. G., 1988. The Silsoe structures building – its design, instrumentation and research facilities. AFRC Inst. Engng. Res., Silsoe, Div. Note DN 1482. Rossetto, T. Allsop, W., Charvet, I. & Robinson, D. (2010). The Generation of Realistic Tsunami using a New Wave Generator. Coastal Engineering. Submitted.Rossetto, T., Elnashai, A. 2003. Derivation of vulnerability functions for European-type RC structures based on observational data. Engineering Structures 25, 1241–1263. Rossetto, T., Elnashai, A. 2005. A new analytical procedure for the derivation of displacement-based vulnerability curves for populations of RC structures. Engineering Structures 27, 397–409. Rossetto, T., Peiris, N., Pomonis, A., Wilkinson, S.M., Del Re, D., Koo, R., Gallocher, S. (2007). The Indian Ocean Tsunami of December 26, 2004: Observations in Sri Lanka and Thailand. Natural Hazards 42, 105–124. Rota, M., A. Pecker, D. Bolognini & R. Pinho. 2005. A methodology for seismic vulnerability of masonry arch bridge walls. Journal of Earthquake Engineering 9, (SPEC. ISS. 2): 331–53. Rota, M., A. Penna & C. L. Strobbia. 2008. Processing italian damage data to derive typological fragility curves. Soil Dynamics and Earthquake Engineering 28, (10–11) (0): 933–47. Ruangrassamee, A., Yanagisawa, H., Foytong, P., Lukkunaprasit, P., Koshimura, S. & Imamura, F. 2006. Investigation of tsunami-induced damage and fragility of buildings in Thaliand after the December 2004 Indian Ocean Tsunami. Earthquake Spectra, 22, S377. Ruiz-García, J. & M. Negrete. 2009. Drift-based fragility assessment of confined masonry walls in seismic zones. Eng.Struct. 31, (1): 170–81. SEOAC. 1995. Performance based seismic engineering of buildings. Vision 2000 Committee, Structural Engineers Association of California, Sacramento, California, USA. Shoji, G. Moriyama, T. 2007. Evaluation of the Structural Fragility of a Bridge Structure Subjected to a Tsunami Wave Load. Journal of Natural Disaster Science, 29, 2, 73–81. Singhal, A. & Kiremidjian, A.S. 1996. A method for probabilistic evaluation of seismic structural damage. Journal of Structural Engineering, ASCE, 122 (12), 1459–1467. Spence, R., E. So, S. Jenny, H. Castella, M. Ewald & E. Booth. 2008. The global earthquake vulnerability estimation system (GEVES): An approach for earthquake risk assessment for insurance applications. Bulletin of Earthquake Engineering 6, (3): 463–83. Stathopoulos, T., Baniotopoulos, C.C. & Zisis, I., 2008. Structural Failure and prevention during catastrophic wind events. Proceedings of Urban Habitat Constructions under Catastrophic Events, COST Action C26, Malta. Strasser, F. O., J. J. Bommer, K. Se¸ ¸ setyan, M. Erdik, Z. Çagnan, J. Irizarry, X. Goula, et al. 2008. A comparative study of european earthquake loss estimation tools for a scenario in Istanbul. J. Earthqu. Eng. 12, (SUPPL. 2): 246–56. Valluzzi, M. R. 2007. On the vulnerability of historical masonry structures: Analysis and mitigation. Materials and Structures 40, (7): 723–43. Vamvatsikos D. & Cornell C.A. 2002. Incremental Dynamic Analysis. Earthquake Engineering and Structural Dynamics, 31(3): 491–514. Vamvatsikos D. & Cornell C.A. 2005. Direct estimation of the seismic demand and capacity of MDOF systems through Incremental Dynamic Analysis of an SDOF Approximation. ASCE Journal of Structural Engineering, 131(4): 589–599. Vamvatsikos D. & Fragiadakis M., 2010. Incremental Dynamic Analysis for seismic performance uncertainty estimation. Earthquake Engineering and Structural Dynamics, 39(2): 141–163. Vamvatsikos D. & Pantazopoulou S.J., 2010, “Simplified vulnerability assessment of historical city cores — the example of the city of Xanthi”, Proceedings of the 9th US National Conference on Earthquake Engineering, Toronto, Canada. Van de Lindt, J.W., Graettinger, A., Gupta, R., Skaggs, T., Pryor, S. & Fridley, K.J. 2007. Performance of wood-frama structures during Hurricane Katrina. Journal of Performance of Constructed Facilities, Vol. 21, No. 2, 108–116. Zisis, I. 2007. Structural monitoring and wind tunnel studies of a low wooden building. MASc Thesis, Concordia University, Montreal, QC, Canada.

289

Chapter 4: Protecting, strengthening and repairing

Urban Habitat Constructions under Catastrophic Events (Final Report) – Mazzolani (Ed). © 2010 Taylor & Francis Group, London, ISBN 978-0-415-60686-8

4.1 Fire damaged structures Y.C. Wang University of Manchester, UK

F. Wald & J. Vácha Czech Technical University in Prague, Czech Republic

M. Hajpál ÉMI, Budapest, Hungary

4.1.1 INTRODUCTION 4.1.1.1 General This chapter is concerned with technical aspects of appraisal of fire damaged steel, concrete and stone structures. The assessment of fire damages to structures follows a similar general process as appraisal of existing structures. It is possible to restore a fire damaged structure to its original load carrying capacity. In making a decision about repairing a fire damaged building, considerations should be given to aesthetic appearance, the reliability of repairs, the views of insurance company and the client, in addition to technical feasibility. For member states of the European Union, safety requirements in case of fire are based on the Construction Products Directive (Council Directive 89/106/EEC: 21.12.1988). The Directive is applied to construction products as the essential requirement in respect of construction works. In Annex I of the Directive, the essential requirements for mechanical resistance and stability, and for fire safety, are summarised. The construction works must be designed and built in such a way that, in the event of an outbreak of fire: • • • • •

The load-bearing capacity of the construction can be assumed for a specific period of time; The generation and spread of fire and smoke within the works are limited; The spread of the fire to neighbouring construction works is limited; Occupants can leave the works or be rescued by other means; The safety of rescue teams is taken into consideration.

The load-bearing capacity of the construction may be modelled on the principles summarised in the parts of the structural Eurocodes which deal with fire. Fire resistance is commonly used to characterize the performance of elements of structure in fire. It may be defined as the time for which elements of a structure satisfactorily perform their required functions under specified fire conditions. These functions may include the ability to avoid collapse, to limit the spread of fire and to support other elements. All construction materials progressively lose their ability to support a load when they are heated. If components of any structure are heated sufficiently they may collapse. The consequences of such a collapse vary, depending on how critical the component is in controlling the overall behaviour of the structure. In order to limit the threat posed by fire to people in a building, and to reduce the damage that a fire may inflict, large buildings may be divided into smaller fire compartments using fire-resisting walls and floors. Parts of a fire compartment may be further divided to protect the building from particular hazards within them. The performance of fire separating elements may rely on the ability of their supporting structure to continue to provide support under fire conditions. The criticality of an element is the degree to which its collapse would affect the performance of the structure as a whole. All of the main components of a structure are generally expected to exhibit fire resistance proportionate to the nature of the perceived risk. The nature of the risk is usually assessed on the basis of the size and 293

proposed use of the building of which the structural element is a part; this is an important part of a fire safety risk analysis. An amplified definition of the fire resistance of a structure or an element is its ability to retain, for a stated period of time, its load-bearing capacity, integrity and insulation, either separately or in combination. As a consequence of European harmonization, fire resistance is increasingly being expressed in terms of R (resistance to collapse, or the ability to maintain load-bearing capacity), E (resistance to fire penetration, or the ability to maintain the fire integrity of the element against the penetration of flames and hot gases) and I (resistance to the transfer of excessive heat, or the ability to provide insulation to limit excessive temperature rise. Design for fire safety has traditionally followed prescriptive rules, but may now apply fire engineering or performance-based approaches, examples of which are given in documents EN 1990 (2002) and EN 1991-1-2 (2002). A fire engineering approach takes account of fire safety in its entirety, and usually provides a more flexible and economical solution than the prescriptive approaches. Within the framework of a fire engineering approach, designing a structure involves four stages: 1. Modelling the fire scenario to determine the heat released from the fire and the resulting atmospheric temperatures within the building. 2. Modelling the heat transfer between the atmosphere and the structure. This involves conduction, convection and radiation, which all contribute to the rise in temperature of the structural materials during the fire. 3. Evaluating the mechanical loading under fire conditions, which differs from the maximum mechanical loading for ambient-temperature design, due to reduced partial safety factors for mechanical loading in fire and changes in mechanical properties of the loadbearing materials. 4. Determination of the response of the structure at elevated temperature. The design recommendations in codes contain simple checks which provide an economic and accessible procedure for the majority of buildings. For complex problems, considerable progress has been made in recent years in understanding how structures behave when heated in fires, and in developing mathematical techniques to model this behaviour, generally using the finite element method which may predict thermal and structural performance. In fire, the behaviour of a structure is more complex than at ambient temperature, because changes in the material properties and thermal movements cause the structural behaviour to become non-linear and inelastic. 4.1.1.2 Assessment Assessment of a fire damaged structure differs from fire resistant design of the structure. In fire resistant design of a structure, the design fire is assumed; the material properties are those at high temperatures and the structure is assessed for reduced structural loads under the fire limit state. In assessment of a fire damaged structure, the engineer has to take into consideration the actual fire that has occurred in the structure; the material properties are those at ambient temperature but after being exposed to high temperatures; the repaired structure should be able to resist loads corresponding to the ultimate limit state, including the additional weight of any repair materials. In fire resistant design of a structure, the engineer obtains data by making suitable assumptions. In assessment of a fire damaged structure, the engineer obtains data by gathering evidences related to the specific damaged structure and the actual fire. It is important that the appraisal process starts as soon as the building can be safely entered and before the removal of debris to preserve vital evidence. Reference Kirby et al. (1986) provides guidance on reinstatement of fire damaged steel and iron framed structures. It gives detailed residual mechanical properties of different types of structural and reinforcing steels, iron and bolts after exposure to different high temperatures, metallurgical evaluation of fire damaged structural steelwork and a number of case studies. Figure 1 is a flow chart for reinstatement of fire damaged steel structures. Reference Concrete Society (1990) provides detailed guidance on assessment of fire damaged concrete structures and design for repair. Detailed information of the effects of high temperatures on structural materials is provided. Different popular methods of assessing fire damaged concrete are described. Detailed guidance is given on how to design and specify repair methods to restore the load carrying capacity of the fire damaged building. A number of detailed examples are included to 294

Figure 1. Appraisal procedure for fire damaged steel structures, see (Kirby et al., 1986).

demonstrate application of the procedures in this document. Figure 2 is a flow chart for assessment and repair of concrete structures. An important aspect of assessing fire damaged structure is to obtain mechanical properties of the structural materials using suitable non-destructive testing methods. Reference CIRIA (1986) provides more detailed explanation of different methods of assessing concrete, some of which are 295

Figure 2. Appraisal procedure for fire damaged concrete structures, see (Concrete Society 1990).

referred to in Reference Concrete Society (1990). This reference is for general use of assessing concrete, but many of the test methods are applicable to fire damaged concrete. Reference ISE (1996) is a general document for appraisal of existing structures, the general procedure of which may be followed in assessment of fire damaged structures. It also provides information on temperature effects on a selection of non-structural materials, which may be used to establish the history of the fire. Reference SCIF (1991) provides a detailed case study of assessment of fire damage to the Broadgate building (a steel framed composite structure) in London, which was extensively damaged 296

by a severe fire during its construction before fire protection to the steelwork was installed. The fire damaged structure was successfully reinstated by replacing the fire damaged floors and columns. The cost of replacing the fire damaged structure was a relatively small fraction (<10%) of the total repair bill, most of which was spent on cleaning the building. Stones are used in many recent and heritage buildings. A study of sandstones at elevated temperatures shows that heating affects the internal structure and mineral composition of natural stones, through influencing the petrophysical parameters (porosity, strength, water adsorption, colour) of the stones. These changes are not always adverse. Hajpál & Török (2004) and Török et al. (2005) describe how the mineralogical composition and texture of natural stones influence their resistance to fire and thermal characteristics. The heat resistance of different quartz sandstones depends on the type of the cementing mineral, the amount of cement (grain/cement ratio), the grain size (fine, medium, coarse) and the grain to grain or matrix to grain contacts. Compact stones show more dramatic change in porosity at elevated temperatures than the less cemented ones. A porous and cement rich stone is more adaptable, being able to accommodate thermal expansion induced additional stresses. Silica cemented, ferruginous or clayey stones are less sensitive than the carbonatic ones, which disintegrate at higher temperatures. 4.1.2 DESIGN PROCEDURE The general procedure of appraisal of a fire damaged structure comprises of the following steps: initial site visit, desk study, detailed collection of evidence, damage assessment and specification of repairs. The purpose of the site visit is to gain an early indication of the scale of damage to the structure and to advise on safety of the building and to recommend measures to protect the general public and other essential personnel. The purpose of the desk study is to collect relevant information (e.g. original design of the building, construction materials, usage before fire, cause of fire, duration of fire, fire spread, contents left unburnt) by examination of physical evidence, interview of the fire brigade and witness. Using the preliminary data gathered, the engineer should establish a strategy for more detailed assessment and data gathering. Fire damages to a structure can be broadly grouped into four categories: no damage/superficial damage, total damage, major damage and reparable damage. No/superficial damage requires no structural repair; total damage leads to scraping of the total structure; major damage requires replacement of the damaged structural members. For these categories, decisions can be made quickly without the need to undertake detailed assessment. Reparable damages are those that may be repaired but there is a high degree of uncertainty about the residual load carrying capacity of the structure. The main objective of damage assessment is to decide with as much confidence as possible the residual mechanical properties of the fire damaged materials so that the fire damaged structure can be restored to its required load carrying capacity. 4.1.3 STRUCTURAL ASPECTS The residual mechanical properties of fire damaged materials may be obtained using the following methods: (1) by direct measurement using Non-Destructive Testing (NDT) and destructive testing; destructive testing should be kept to minimum and should only be used when there is low confidence in NDT results; (2) by direct assessment of maximum material temperatures and links to material residual mechanical properties – temperature relationships; (3) by establishing the fire history, from which the material temperature history may be established using heat transfer; afterwards, using the residual mechanical properties – temperature relationships. Due to uncertainty in results obtained from these different methods, it is important to correlate the different results to improve confidence in them. It is also important to make conservative (safe) assumptions when evaluating residual load carrying capacities of fire damaged structures; for example, assuming simple supports and ignoring any beneficial effects of the restraints. The documents (Kirby et al., 1988, Concrete Society 1990, ISE 1996) described in the Guidelines section of this technical sheet may be consulted to obtain residual mechanical properties – temperature relationships for steel and concrete. Additional data may be obtained from Outinen 297

Figure 3. Tensile test results for structural steel S350GD + Z, the test pieces taken before and after high-temperature compression tests, where the material reached temperatures up to 950◦ C, see Outinen, Mäkeläinen, 2004.

and Mäkeläinen (2004), see e.g. Figure 3; Pang (2006), and Chan (2009) for steel, from Dias (1992) for concrete and from Yan and Wong (2007) for high strength concrete; and from Hajpál and Török (2004), Hajpál (2008) and Török and Hajpál (2005) for sandstones. Materials expand at high temperatures, which may cause brittle materials remote from the fire site to suffer damage. It is important to assess the entire structure for fire damage. For example, expansion of floors directly involved in a fire may damage the walls in remote places from the fire. NDT methods for fire damaged steel include Hardness test and metallurgical microscope. Hardness test is simple and easy, but the hardness test results should not be used to guarantee the material to an appropriate specification, for which coupon tests are required. Microscopic test requires specialist personnel and equipment. It is used only when it is essential, e.g. to provide information on the micro-structure of metal so as to establish an accurate picture of the heating environment. Methods of assessing fire damage to concrete include colour observation (e.g. pink indicating about 300◦ C), visual classification, NDT testing (Schmidt hammer, ultrasonic pulse velocity, thermoluminescence) and destructive testing (cores). It is important to choose the appropriate testing method before detailed assessment starts. The Schmidt hammer test is a simple and cheap approach, but it only gives an indication of the concrete properties near the surface of the structural element and its results may contain large scatters; the ultrasonic pulse velocity method is also easy to use and may be used to detect internal cracks of concrete, but it requires access from both sides of the structure; the thermoluminescence method involves taking a small amount of concrete through a considerable depth; through analysis of the thermoluminescence lost due to heating, the temperature profile through the concrete depth may be established, from which the mechanical properties of concrete may be obtained. Being a material that is made at high temperatures, mild steel recovers much of its initial strength and stiffness after fire exposure. Therefore, a fire damaged steel structure can normally be reinstated. Unless severely distorted to affect appearance, steel structural members can normally be retained. High strength bolts are made by quenching. Exposure to high temperatures above 500◦ C has the similar effect as tempering, which would reduce the residual strength of bolts. Generally, bolts after exposure to high temperatures should be replaced. If the reinstated steel structure requires fire protection, it is important that smoke deposits on the steel surface are removed before application of fire protection materials. Various methods may be employed to repair fire damaged concrete structures, including reconstruction (major repair after extensive damage or sprayed concrete is difficult), sprayed concrete, resin repairs (for repairs to lightly spalled areas), overcladding (non-structural materials such as plasterboard, to restore appearance/restore fire resistance/durability), provision of alternative supports. 298

Figure 4. The coal bridge after the fire.

Figure 5.

Deformed upper stiffening truss.

Figure 6.

Deformed purlins.

4.1.4 CASE STUDY OF A STEEL COAL FEEDING BRIDGE 4.1.4.1 Fire damaged structure In September 2005, fire occurred on the coal feeding bridge in the Opatovice power station, Czech Republic, when unloading of a wagon hit the transport corridor, see Vácha (2006) and Figures 4 to 6. About 100 m of the concrete corridors and 120 m of the steel transport bridges (four truss bridges of span 30 m each) were affected, Coal transport infrastructure was completely burned and the bearing structure was noticeably damaged. The power station depends on coal supply so it was important that the supporting structure was reconstructed to be functional as soon as possible. The steel structure was noticeably affected by the high temperature, which was estimated to be about 1000◦ C, based on its colour during the fire. The four steel coal bridges were similar. There were 30 m span trusses with upper and lower stiffening trusses. One bridge was inclined. The lateral stability of the bridge trusses was achieved by massive rigid frames spaced at 3 m. All the connections were riveted. The fire damaged all sheathing of the walls and roof. The upper part of the bridge suffered local deformations of the upper crossbeam, the truss stiffeners and the end frames, with the maximum deformations around 100 mm. The upper parts of the main trusses were deformed only at a few locations. The lower parts of the structure would not have suffered any structural damage during the fire because the surface painting still remained after the fire. As shown in Figure 4, he inclined bridge was the most damaged. The total deflection of the unloaded structure reached from L/600 to L/500. 4.1.4.2 Reconstruction Instead of replacing all the fire damaged bridges, reconstruction of only parts of the structure was undertaken to minimise the plant down time. The reconstruction started through detailed diagnoses in the following steps: • Mechanical property tests were performed on selected specimens by standard coupon tests. The results of extended measurements of steel heated to 950◦ C are summarised in Table 1. The steel 299

Table 1. Measured yield stress and strength of steel with and without temperature effect, MPa. Yield stress

Strength

Thickness, mm Steel grade

10 S235

20 S275

30 S275

6 S355

10 S235

20 S275

30 S275

6 S355

Affected Unaffected Reduction, %

226 269 16

297 352 16

307 333 8

309 374 17

439 462 5

497 514 3

471 481 2

465 535 11

Figure 7.

• • • •

Replacing of the end frame.

Figure 8. View into the reconstructed bridge.

grade used was S235. It had good weldability before and after the fire and the fire incurred only negligible damage to the yield stress of the steel. In only one place outside the main bridge was steel found to be brittle due to fire attack. The change in microstructure of steel due to heating and cooling was visible on all 54 observed specimens. However, these changes may be accepted because of the slight changes in mechanical properties of the steel. Visual checks of the geometry of the structure and its elements were followed by detailed measurement of all major positions of the structure, including joints of the trusses and the rail of the conveyor. Straightness of all compressed elements of the main trusses was checked by geometric measurements. Chemical analyses were performed to determine contamination of the surfaces so as to prepare cleaning and the following-on corrosion protection.

The major aim for the design and implementation strategy of the reconstruction work was to minimise the reconstruction time. The current stage of the structure was analysed based on the material properties obtained above. The following decisions were taken: • All elements with excessive deformations were replaced; the maximum allowed out of straightness was 10 mm; • When performing structural analysis, the mechanical properties of all steel elements attacked by fire were reduced by 10%; • For the riveted connections affected by fire, but not visibly deformed, the resistance was reduced by 15%; • Second order analyses were adopted and they incorporated the effects of the allowed maximum deformation of 10 mm; • The upper truss stiffeners and the upper crossbeams were changed, see Figure 7; • The upper parts of the props of the end stiffeners were changed as well; • One deformed diagonal member in the lower stiffening truss was changed at each horizontal bridge, see Figure 8; 300

Figure 9. The uncovered bridge during its reconstruction.

Figure 10. The reconstructed bridge during proof of behaviour by loading test.

• For the inclined bridge, all the lower cross beams and lower stiffening trusses were replaced; • Based on detailed calculation, a few riveted connections were strengthened by welding on site; • After the reconstruction, the structure was loaded by an operational test, which proved that the structure recovered its full static behaviour, see Figures 9 and 10.

4.1.5 FURTHER DEVELOPMENT Assessment of a fire damaged structure differs from fire resistant design of the structure. In assessment of a fire damaged structure, consideration is taken into the actual fire that has occurred in the structure; the material properties are those at ambient temperature after being exposed to high temperatures; the repaired structure should be able to resist loads corresponding to the ultimate limit state, including the additional weight of any repair materials. Since a real fire will generate highly non-uniform temperature distribution, the residual mechanical properties of the fire exposed materials may be significantly different in different locations of the structure. Furthermore, it will not be possible to obtain detailed information of mechanical properties of materials through definitive destructive testing. Therefore, in assessing a fire damaged structure, to achieve confidence in material properties, the engineer should correlate data from a variety of sources, including mechanical properties from non-destructive and destructive (to be used with care) tests, temperature history of the fire and temperature history of the materials. The engineer should also take into consideration structures unexposed to the fire but may be severely affected by the fire through thermal deformation. Frequently, fire damaged structures can be successfully repaired to fulfil their original functions. A number of significant references are available to guide the engineer in this process. 301

ACKNOWLEDGEMENTS The fire engineering work of this study and outcome has been achieved with the financial support of the Ministry of Education, Youth and Sports, project no. 1M0579. REFERENCES Chan D., 2009. Fire damage assessment of structural steel in a school, The Structural Engineer, 87 (19) 6 October, pp. 18–20. Concrete Society, 1990. Assessment and Repair of Fire-Damaged Concrete Structures, Technical Report No. 15, The Concrete Society, UK. Construction Industry Research and Information Association (CIRIA), 1986, Testing Concrete in Structures, A Guide to Equipment for Testing Concrete in Structures, Technical Report 143, CIRIA, UK. Dias W.P.S., 1992. Some properties of hardened cement paste and reinforcing bars upon cooling from elevated temperatures, Fire and Materials, 16 (1), pp. 29–35. EN 1990-1-2, 2002. Eurocode 0: Basis of structural design, CEN, Brussels. EN 1991-1-2, 2002. Eurocode 1: Actions on structures. Part 1–2: Actions on structures exposed to fire, CEN, Brussels. Hajpál M. 2008. Heat effect by natural stones used by historical monuments, 11th International Congress on Deterioration and Conservation of Stone (STONE 2008). Hajpál M., Török Á. 2004. Physical and mineralogical changes in sandstones due to fire and heat. Environmental Geology, 46, 3, 306–312. Kirby, B.R., Lapwood, D.G. & Thomson, G., 1986. The Reinstatement of Fire Damaged Steel and Iron Framed Structures, British Steel Corporation (now Corus), London, p. 46. Outinen, J., Mäkeläinen, P., 2004. Mechanical properties of structural steel at elevated temperatures and after cooling Fire and Materials, 28 (2–4), pp. 237–251. Pang P.T.C., 2006. Fire engineering design and post fire assessment, The Structural Engineer 84, (16) 6 October, pp. 23–29. Steel Construction Industry Forum (SCIF), 1991. Structural Fire Engineering: Investigation of Broadgate Phase 8 Fire, Steel Construction Institute, UK. The Construction Products Directive, 1989. Council Directive 89/106/EEC, URL: ec.europa.eu. The Institution of Structural Engineers (ISE), 1996. Appraisal of Existing Structures, Institution of Structural Engineers, London, UK. Török Á., Hajpál M. 2005. Effect of Temperature Changes on the Mineralogy and Physical properties of Sandstones. A Laboratory Study. International Journal for Restoration of Buildings and Monuments, 11, 4, Freiburg, 211–217. Vácha, J., 2006. Reconstruction of coal feeding bridge after fire. in Czech Rekonstrukce zauhlovacích most˚u po požáru,Konstrukce, 01, pp. 22–24. ISSN 1803–8433. Wang Y.C., Wald F., Török A., Hajpál M., 2008. Fire damaged structures, in Technical sheets – Urban habitat constructions under catastrophic events, Print Pražská technika, Czech Technical University in Prague, ISBN 978-80-01-04268-7. Yan, X., Li, H., Wong, Y. L,. 2007. Assessment and repair of fire-damaged high-strength concrete: Strength and durability Journal of Materials in Civil Engineering 19 (6), pp. 462–469.

302

Urban Habitat Constructions under Catastrophic Events (Final Report) – Mazzolani (Ed). © 2010 Taylor & Francis Group, London, ISBN 978-0-415-60686-8

4.2 Innovative seismic protection technologies and case studies M. Kaliske Institute for Structural Analysis, Technische Universität Dresden, Germany

A. Mandara Department of Civil Engineering, Second University of Naples, Italy

4.2.1 INTRODUCTION Current practice in seismic protection of new and existing buildings is today largely oriented to the combined use of advanced materials and technologies, possibly endowed with reversibility features, in order to achieve an optimised performance from all points of view. Innovative materials for the purposes of seismic protection are essentially materials like: Special Steels, Aluminium and Titanium Alloys, Fibre Reinforced Polymers (FRP), Geotextiles, Elastomers and, last but not least, “smart” materials like Shape Memory Alloys (SMA), Piezoelectric Materials, Magnetorheological Fluids, etc. Most of these materials are used to create special devices provided with energy dissipative features, as well as to obtain advanced strengthening systems especially designed for improving both resistance and ductility of structural elements. Likewise, innovative systems can be based on either an increase of dissipated energy (Additional structural damping) or on a reduction of the seismic input energy (Base isolation and Tuned Mass Damper). These techniques are applied by means of either special devices (e.g. Yielding Metal Devices, Friction Devices, Fluid Viscous Dampers, Rubber Bearings, etc.) or advanced bracing systems (Mandara, 2007). As to seismic rehabilitation of existing buildings, the use of modern and advanced strengthening solutions represents an useful tool as, in most cases, existing buildings are conceived in such a way to be scarcely prone to modifications of the structural layout, unless special provisions are adopted. Also, such buildings do not commonly possess structural properties adequate to resist significant seismic actions nor to comply with current structural codification. Also, existing structures may be frequently affected by significant damage and/or degradation of both materials and members, often resulting in a global performance and safety level even lower than at the construction stage. In most of such cases, the new structural requirements cannot be met by means of a simple conservative restoration with “traditional” materials, but ask for more advanced solutions able to satisfy more demanding needs, in particular when a satisfying behaviour under abnormal loading conditions is required. At the same time, reversibility may be essential for the sake of economics (recycling materials for different purposes), for preventing the construction from inappropriate restoration operations or for implementing more effective solutions at a future stage. Within the activity of COST C26 WG2 “Earthquake resistance”, a number of topics have been faced related to innovative seismic protection technologies, leading to point out the main trends of research state-of-the-art and current practice. Such topics are: 1. 2. 3. 4.

Innovative materials; Strengthening systems; Protection strategies; Case studies.

The above topics come out from the intense activity carried out by WG2, based on a 4-year-long fruitful cooperation among WG Members. For each of the above subjects, a short description is given hereafter, aimed at emphasizing the main result of WG2 in this field, as well as to outline some possible future developments. 303

4.2.2 INNOVATIVE MATERIALS 4.2.2.1 Metal materials The use of innovative metal materials, namely stainless steels, copper, titanium and aluminium alloys, is becoming more and more frequent in seismic engineering, including rehabilitation practice. Such materials are mainly intended to complement the well known features of constructional mild steel (e.g. high strength and ductility, lightness, ease of transportation and erection, easy market availability, reversibility, etc.), with some special properties, which can be exploited for creating special energy dissipation devices (Mazzolani & Mandara, 2002). Some of these metals, e.g. titanium alloys, have a very low linear thermal expansion coefficient (6 ÷ 8 × 10–6C◦ −1), which is very similar to that of volcanic or metamorphic rocks, such as granite and marble. This allows titanium elements to be used in redundant systems with no risk to impair the effectiveness of intervention due to thermal changes. This feature has been highly useful in the restoration of Parthenon in Athens and Colonna Antonina in Rome, where titanium clamps have been inserted and hidden into existing stone blocks (Giuffrè & Martines, 1989). They proved to be far more effective than conventional steel elements used before, which had involved many cracks due to corrosion and excess of thermal dilatation. Despite their relatively high cost, innovative metal materials meet increasing application in the field of seismic protection also due to their peculiar features, which generally offer several benefits when the whole lifetime of the project is considered. For example, such metal elements can be easily melted down and re-used for different purposes. This contributes to increase their long term sustainability compared with traditional non-reversible technologies. In the very last years, special devices based on the use of Shape Memory Alloys have been used in combination with steel tying elements for the seismic protection of existing constructions (Indirli, 2000). Shape Memory Alloys (SMA), mostly Ni-Ti or Cu-Al-Zn alloys, may be regarded as “smart” materials, as both their yield stress and modulus of elasticity strongly increase as long as temperature increases within the so-called transformation temperature range, due to a solid martensite-austenite phase change. Such range is limited by Mf and Af , that is by the temperatures where only full martensitic or full austenitic structures can exist, respectively. The above transformation can be induced by either mechanical stress or temperature change, resulting in the capability to recover, spontaneously or by heating, large initial strains due to load (superelastic behaviour, Fig. 1) with a corresponding amount of dissipated energy due to different loading and unloading paths. This allows the construction of seismic protection devices (Croci et al., 2000, Pegon et al., 2000, Dolce & Cardone, 2001). 4.2.2.2 Fibre-reinforced materials Since the middle of the nineties an increasing number of structures have been strengthened by means of externally bonded FRP (Fibre Reinforced Polymers) reinforcement. The main advantages of these materials are their very low weight, a minimum of structural thickness, no corrosion, easy application and simple use on the construction site (Mandara et al., 2002). The composite materials include a matrix and fibres. The matrix has the main role of assuring the united behaviour of

Figure 1. Behavioural principles of SMA (left) and the SMA devices installed in the S. Francesco Basilica, Assisi (Italy) (right).

304

the fibres. In addition, it protects the fibres against environmental and mechanical damages and buckling, grants shear, inter-laminar and in plane resistance. The matrix incorporate epoxy resin so as these materials have the ability to adhere to any kind of surface, excellent ductile capacity, strength, resistance to chemical agents and durability. Fibre Reinforced Polymers (FRP) used as strengthening materials today are typically made of continuous carbon (CFRP), aramid (AFRP) or glass fibres (GFRP). The fibres are produced in the shape of small wires (5 to 20 µm), have high tensile strength and elasticity modulus, low density and show a fragile behaviour (Table 1) (CEB-FIP, 2001). Also, they exhibit good fatigue behaviour, good performance under cyclic loads. In the end, because of high resistance to chemical products, FRP can be seen as durable materials against pollutants which arise under normal service conditions in buildings. Most used FRP in structural applications are the CFRP, but also GFRP have recently gained a great practical importance for applications on substrates with low tensile strength like masonry or natural stones. Also, hybrid fabrics of different types of fibres are offered on the market. The layers of fibres can vary in shape and orientation. They can be disposed in one or two main directions as well as in discontinuous appearances. The properties of these materials are likely to be anisotropic and each layer can fulfil a different function. FRP systems are used in the form of prefabricated elements (straight strips, shaped shells, jackets or angels, rods, wet lay-up systems) or fibre sheet, fabric and fibre tows, and some special systems like prestressed strips, etc. In general, FRP strips can be used for upgrading the elements having a plane surface, while fabrics can be also used where the surface is not regular. FRP strips, fabrics and rods can be used in both concrete and masonry structures. The possibility of selecting the fibre typology, the architecture of sheets and the number of FRP layers gives the designer flexibility in the choice of the reinforcement. The suitability of each system depends on the type of structure to be strengthened. Practical execution and application conditions, like cleanness and temperature, are very important in achieving a good bond. CFRP can be profitably used to strengthen structural members either in bending or shear, but also for column confinement against axial load (Fig. 2). In all these applications, unidirectional strips or fabrics are mostly used. Due to the flexibility of CFRP strips, it is possible to apply also elements of large dimensions without any special joint configurations. In all cases, special attention has to be paid to the preparation of the surface. From the practical point of view, one of the advantages of wet lay-up systems is the possibility to follow also irregular geometrical shapes or curves during application. This feature is useful in old masonry or for arches and vaults. In seismic upgrading of RC framed buildings, the increase in flexural strength of columns is very important to move plastic hinges from columns to beams, because in case of existing gravityload designed buildings, the column over beam ratio of flexural strengths is usually small. This aspect has been investigated within the ILVA-IDEM Project (Mazzolani, 2006), in which several strengthening options for RC building, including FRP wrapping, have been tested on full scale buildings (Fig. 3). The FRP strengthening system was designed in such a way to modify the collapse

Table 1. Tensile characteristics of fibres (CEB-FIP 2001). Fibre Type

Elasticity Modulus (GPa)

Ultimate Strength (MPa)

Ultimate Deformation (%)

Carbon High Strength Very High Strength High Elastic Modulus Very High Elastic Modulus

215–235 215–235 350–500 500–700

3500–4800 3500–6000 2500–3100 2100–2400

1.4–2.0 1.5–2.3 0.5–0.9 0.2–0.4

70 85–90

1900–3000 3500–4800

3.0–4.5 4.5–5.5

70–80 115–130

3500–4100 3500–4000

4.3–5.0 2.5–3.5

Glass Glass E Glass S Aramid Current High Performance

305

Figure 2. Typical applications of CFRP strips and sheets for strengthening RC and masonry elements.

Figure 3. Damage of the pushover test of the initial structure (a); The structure after the upgrading (b); detail of the repairing system (c).

mechanism from a soft story to a global ductile mechanism. Before applying the FRP strengthening, the original RC structure was previously tested. The structure was forced with an increasing topstory lateral displacement, up to the development of a clear plastic collapse mechanism (Fig. 3a). The structure exhibited a top-story sway collapse mechanism, with plastic hinges forming at both ends of each column at the second story (Della Corte et al., 2006a.) After the pushover test of the original structure, the building was straightened up and repaired. Then, the CFRP has been applied as external reinforcement, consisting in pre-formed high-strength carbon strips applied along the columns and CFRP transverse sheets, thus giving additional confinement to the concrete section (Figs. 3b,c). The FRP strengthened structure was subjected to a cyclic loading history, until the displacement capacity of the load-actuators was completely exhausted. The damage pattern consisted of flexural plastic hinges forming at the base of the first-story columns and in the floorbeams constituting the two slabs. The measured lateral strength and stiffness were about 2 times the ones of the original RC structure, with a corresponding increase of ductility. The behaviour of reinforced concrete beams strengthened by thin fibre reinforced concrete (FRC) jackets was also investigated in (Georgiadi-Stefanidi et al., 2008). The FRC jackets were made of high strength concrete with the addition of hooked steel fibers (Fig. 4). The reinforced concrete 306

Figure 4. Test on concrete beams reinforced with FRP jackets and comparison between FEM numerical simulation and experimental results.

beams that were strengthened with the jacket were made of conventional C20/25 concrete. Two cases were considered: a) beams with longitudinal reinforcement and closely spaced stirrups, and b) beams with longitudinal reinforcement and no transverse reinforcement. For the strengthening method, two cases were considered: a) thin FRC jackets with longitudinal and transverse reinforcement, and b) thin FRC jackets reinforced with a light structural steel mesh. The thickness of the FRC jacket was 3 cm. Combining the above cases, four types of beams were produced, which were tested under cyclic 4-point bending. The tested beams were also simulated numerically, taking into account the various nonlinearities of the physical model. The significant contribution of the FRC jackets in strengthening and improving the overall behaviour of the concrete beams is shown by the experimental results (Papatheoharis, 2008) and the predictions of the numerical simulation. Tests were made on: 1) concrete beam with longitudinal reinforcement and closely spaced stirrups, strengthened by FRC jacket with longitudinal and transverse reinforcement; 2) concrete beam with longitudinal reinforcement (no transverse reinforcement), strengthened by FRC jacket with longitudinal and transverse reinforcement; 3) concrete beam with longitudinal and transverse reinforcement, strengthened by FRC jacket reinforced with a light, steel structural mesh; 4) concrete beam with longitudinal reinforcement (no transverse reinforcement), strengthened by FRC jacket reinforced with a light, steel structural mesh. The evaluation of the obtained results verifies the beneficial behaviour of thin FRC jackets reinforced with conventional reinforcement with respect to those reinforced with light structural mesh. The hysteresis loops obtained by members reinforced with FRC jackets with light structural mesh are less stable and pinching occurs after the first few cycles. Also, a very good agreement between experimental and numerical simulation was found (Fig. 4). 4.2.2.3 Textile materials Textile reinforced concrete (TRC) has become a promising alternative for RC structures requiring retrofitting and strengthening because of ageing, damage, and deterioration caused by environmental stresses and earthquake excitations (Sickert et al., 2008). Textile reinforced concrete is a new composite material consisting of textile fabrics and a fine-grained concrete matrix. The textile fabrics are made of filament yarns (rovings) which are connected with the aid of stitching yarn. The filament yarn is a bundle comprised of a large number of single filaments. Filaments may be made of different materials, e.g. alkali-resistant glass (ARG) or carbon fibers. The development of TRC, which first began in the late nineties, is summarized in (Brameshuber, 2006). Their high application potential results from a marginal change in geometry, a high load bearing capacity of textile reinforcement made of alkali-resistant glass and carbon fibres, together with a marginal increase of the dead load and an easy application process. Further advantages are a high durability and a good fire resistance. TRC strengthening layers may be applied independently of the geometry of the existing structure. The use of textiles with rovings aligned with the principal stress direction of the composite is more effective than using the same amount of fibre material in the form of randomly distributed short fibres, as is the case in fibre reinforced concrete (FRC). Two application examples are displayed in Figure 5. Experimental and numerical investigations demonstrate that structural responses of TRC structures are highly dependent on the spatial and temporal variation of the uncertain material and geometric data. For this reason current studies are concentrated on 307

Figure 5. Application of textile reinforced concrete on existing structures.

Figure 6.

Steel wire mesh geometry and texture (a) and chemical anchor (b).

the generalized uncertainty modelling of material, geometric, and load parameters with the aid of fuzzy and fuzzy random functions (Sickert et al., 2007). 4.2.2.4 Metal-based solutions The use of steel wire meshes (SWM) for the strengthening of masonry panels was investigated in (Dogariu et al., 2007, Dogariu et al., 2008), in which the wire mesh is glued using epoxy resin. The system can be applied on one side or both sides of the panel. Such a type of solution can be successfully applied in case of masonry walls, but it is not appropriate in case of masonry vaults and arches. The application technology is rather simple. The mesh (Fig. 6) is produced either as galvanised steel or stainless steel bidirectional fabric. Spacing of the mesh is between 0.05 and 16 mm, while wire diameter is between 0.03 and 3.0 mm. Tensile strength reaches 650–700 N/mm2 , while elongation is about 45–55% in the case of stainless steel wires. For galvanised steel wire, tensile strength is usually in the range of 400–515 N/mm2 . The preparation of resin is similar to the one used for Fibre Reinforced Polymers (FRP). By heating the resin layer, the wire mesh can be removed, hence ensuring intervention reversibility. In order to validate the solution, an experimental program was carried out on 500 × 500 mm and then on 1500 × 1500 mm specimens, under both monotonic and cyclic loading. In the first step six types of wire mesh, zinc coated (ZC) and stainless steel (SS), bonded on one side, were tested. Compared to FRP technique, a thicker fluid resin was selected. Based on the experimental results, the following wire meshes were chosen for large scale specimens: zinc coated (ZC) 0.4 × 1.0 mm (D × W ), stainless steel (SS) 0.4 × 0.5 mm and 0.4 × 1.0 mm to be applied on both sides. The tests were carried out in two different frames, one for monotonic loading and one for cyclic loading. The tests set-up is presented in Figure 7. Loading was applied using displacement control, with lateral drift of the panel being used as control parameter. Diagonal failure mode was observed for all specimens, both under monotonic and cyclic loading. However, due to large in-plane stiffness of masonry walls, the strengthening solution does not avoid completely damage to masonry. A limited 308

Figure 7. Testing frames for monotonic/cyclic loading and envelop curves for SMW and ASP specimens.

amount of damage to masonry has to be allowed in order to take benefit from ductility of the metal. Compared with similar jacketing interventions making use of aluminium shear panels (ASP), the cyclic envelop curves (Fig. 7) show that SWM enables to obtain a good increase of elastic strength, but without increasing the stiffness of the wall, whereas ASP achieves an important increase in terms ultimate displacement that assures a very stable post-cracking behaviour and a large ductility. 4.2.3 STRENGTHENING SYSTEMS Within the WG2 activity, advanced bracing systems such as Buckling Restrained Braces (BRBs) and Eccentric Braces (EB) have been mostly investigated, as recognised the most effective against severe earthquakes. Both of them are reliable and effective devices, which offer an effective solution to the problem of the limited ductility of classic concentric bracing, thanks to either avoidance of global compression buckling, in case of BRBs, or to the plastic behaviour of shear links in EBs. In addition, they can be easily implemented in both new and existing buildings. 4.2.3.1 Buckling restrained braces BRBs were firstly introduced in the ’80s in Japan and later in the USA. They are characterized by the ability of bracing elements to yield inelastically in compression as well as in tension. The BRB technology is currently ongoing a strong development, with a growing number of buildings using buckling restrained braces as primary lateral force-resisting system (Tsai et al., 2004, Wada & Nakashima 2004). In the most classical form, the restraining tube is filled with concrete and an unbonding layer is placed at the contact surface between the core plates and the filling concrete, which is why this version is called ‘unbonded brace’. The unbonding material both ensures the brace to freely slide inside the buckling restraining unit and lets transverse expansion of the brace to take place when the brace yields in compression. ‘Only-steel’ solutions have been also proposed, with two or more steel tubes in direct contact with the yielding steel plates. Contrary to the “unbonded”, this type of BRBs can be designed to be detachable. This aspect implies that is possible to design these systems to be inspected, so that it is possible to control their condition after each seismic event and to allow an ordinary maintenance during the life-time. Other advantages may be low-cost compared with special damping devices and easiness to be assembled into a structure. Steel BRB systems are also a suitable technique to retrofit RC structures, as they can effectively improve lateral strength, stiffness and ductility. This hysteretic behaviour is stable in a wide range of deformation, which has to be defined at the design stage. If the displacement capacity is exceeded because of an unforeseen large demand due to a catastrophic earthquake, then a number of secondary collapse mechanisms can be activated. Accordingly, different performances can be recognized with the possibility to develop ductile behaviour or brittle failure of the device. Different types of BRBs have been proposed in the technical literature, all of them based on the basic concept to use tubes for restraining lateral displacements while allowing axial deformations of the core. Generally speaking, a common BRB is composed by three zones (Fig. 8): the yielding zone, that has a reduced cross 309

Figure 8.

Schematic view of a typical BRB element (Sabelli & Lopez, 2004).

Figure 9. Geometry and details of BRB types investigated in (D’Aniello et al., 2008).

section area within the zone of lateral restrain provided by the sleeve (zone C); the transition zones, which have a larger area than the one of the yielding zone, and similarly restrained (zone B); the connection zones (zone A). The BRB examined in (D’Aniello et al., 2008) is a special detachable “only-steel” device, made of a rectangular steel plate encased in a bolted restraining steel sleeve. In particular, this BRB system has been designed to be hidden inside the typical external walls of existing reinforced concrete structures. A wide experimental program on full scale RC buildings has been carried out within the ILVA-IDEM project (Mazzolani, 2006, Della Corte & Mazzolani, 2006), where several upgrading systems have been compared with each other. Two types of BRBs have been also studied (Della Corte et al., 2005, D’Aniello et al., 2006) (Fig. 9). Based on experimental tests (Della Corte et al., 2005, D’Aniello et al., 2007) the following secondary failure mechanisms have been identified: 1) local buckling of the gusset-plate connections; 2) local buckling of the unrestrained end part of the core; 3) local buckling and related plastic bending of steel plates constituting the restraining sleeve; 4) overall brace buckling. The experimental evidence highlighted an overall satisfactory response of the structure equipped with the proposed BRBs, showing always a large increase of strength and stiffness, even if some local failure mechanisms limited the expected performance in terms of ductility. Figure 10 shows the overall response of the structure equipped with BRB type 1 and type 2, respectively, compared with the response of the original RC building. A great improvement in the terms of both strength and ductility is achieved thanks to BRBs. 310

Figure 10.

BRB braced vs original RC pushover response curves.

Figure 11.

Frame geometry and corresponding pushover curves.

The performance of BRB system has been also studied by (Bordea et al., 2007), considering the seismic upgrading of a RC frame building located in Romania and designed against vertical loads, only. The frame geometry and the obtained cross sections are presented in Figure 11. The following strengthening solutions were considered: 1) steel BRBs only; 2) confinement of the first and second story columns using fiber reinforced polymers (FRP); 3) the combination of the previous two solutions. The BRBs were introduced only in the middle span, as an inverted V braces, pinned at the ends. The design of the BRBs was accomplished according to Eurocode 3, following the procedure described in AISC 2005. Design seismic forces were obtained using spectral analysis with a reduction factor q equal to 6. The core of the buckling restrained brace was considered to be of rectangular shape. Pushover analysis was applied in order to evaluate the differences between the original frame and the retrofit one. Performance of the structure was evaluated in terms of inelastic deformation capacities corresponding to Collapse Prevention (CP) limit state. Development of plastic mechanism was also observed. Strengthening by buckling-restrained braces increased considerably the strength and stiffness of the frame (Fig. 11), decreasing by almost 50% the top displacement demand at the ultimate limit state. The first plastic hinges formed in column, followed by the ones in braces and beams. This strengthening solution reduced the overall damage in the structure, as less plastic hinges formed in reinforced concrete elements at the target displacement. However, seismic performance is still unsatisfactory, as inelastic deformations corresponding to collapse prevention limit state are recorded in columns, braces and beams before reaching the target displacement. As an alternative to strengthening by buckling restrained braces, the possibility to improve seismic performance by confining the columns with FRP in horizontal direction only was investigated. The effect of application of FRP was an increase of axial force capacity of the columns and ductility, but just a slight increase of bending moment capacity. In the end, the use of both FRP and BRB systems was considered. The main effect of the BRB system is an improvement of global force-deformation characteristics (increased strength and stiffness), which results in decreased top 311

Figure 12. The bolted link concept.

displacement demands at the ultimate limit state. On the other hand, FRP technique enhances the local behaviour of columns by increasing their ductility, this being the reason of attaining ultimate deformation after the demand displacement. The analysis has been also carried out assuming q = 3, leading to double the cross-section area of BRBs. As can be observed from Figure 11, the global strength of the system is increased in comparison to the system designed with q = 6. However, the stiffness increases only slightly. Though overall structural response is improved, inelastic deformations in braces and beams are still smaller than displacement demand. In conclusion, the analysis showed that seismic rehabilitation of nonseismic RC frames cannot be accomplished by means of very ductile dissipative bracing system without a proper strengthening of RC members, so that both beams and columns can work mainly in elastic domain, while ductile steel BRB will be responsible for dissipative behaviour. 4.2.3.2 Eccentric braces Likewise BRBs, Eccentric Braces (EB) are believed to be an effective structural system for seismic applications. They are characterised by both excellent stiffness and good ductility. In eccentrically braced frames the dissipative zones are located in link elements. Eccentrically braced frames with removable links connected to the beams using flush-end plate bolted connections have been investigated by (Stratan & Dubina, 2008) (Fig. 12). The use of removable connections between the links and beams/columns enables the replacement of earthquake damaged links, thus reducing the repair costs. An experimental investigation on isolated removable links was recently extended with nearly full-scale tests on eccentrically braced frame with removable links (Stratan & Dubina, 2004). The objectives were to check overall structural performance of the system (avoid inelastic deformations outside links ad replace damaged links) and investigate response of link to beam connection. Eight different tests were performed, by replacing link specimens. Both monotonic and cyclic tests were performed. Under monotonic loading, both positive and negative force was applied. Failure modes and force-deformation plots of two specimens are shown in Figure 13. Test showed that end-plate thickness had a minor influence on the response of links under positive loading (compression in links), but affected seriously the behaviour of links under negative loading (tension in links). They also showed that a too large thickness of the end-plate can reduce the ductility of the removable link, by promoting a brittle failure mode in bolts. In general, experimental tests proved the feasibility of the removable link concept. All eight tests were performed on the same frame, without relevant degradation of non-dissipative members and connections. Replacement of the links were performed easily, even after significant plastic deformations. Further investigation is currently in progress, in order to establish the influence of concrete slabs on both inelastic performance of removable links and possibility to replace damaged links. In case of RC frames, the concrete beams are incapable to perform as a ductile link for the steel bracing system that is inserted in the frame bays. Therefore, it is impossible to adopt for RC frames the common inverted k-brace configuration (typically used in steel frames). Hence, the need to adopt a Y-inverted bracing configuration, with a vertical steel link, can be easily recognized. Besides, bolted connections at the link ends are useful, which could have the advantage to permit replacement of the dissipative members (links) after a damaging earthquake. A possible geometry of this bracing system is illustrated in Figure 14, referring to the ILVA-IDEM Project (Mazzolani, 312

Figure 13.

Failure modes and a typical force-deformation plot of removable link specimens.

Figure 14.

Geometry of EBs tested within the ILVA-IDEM Project.

2006). Three experimental pushover tests have been carried out on this structural unit. In case of EBs it was observed an increase of the lateral capacity from 5.65 to 8.34 times respect to the capacity of the original unbraced structure, while in the same case strengthened with BRBs this was from 4.08 to 4.95 times.

4.2.3.3 Dissipative shear panels Among the advanced bracing systems can be also considered the metal plate shear walls (MPSWs). In the past few decades, MPSWs used as either stiffened or unstiffened thin panels have been introduced as primary lateral load resisting systems in several buildings thanks to their fundamental prerequisites, such as high stiffness and large deformation capacity (Mazzolani et al., 2007). Aiming at emphasizing the potential contribution of MPSWs on the seismic performances of new and existing structures, an intense research activity developed in cooperation between the University of Naples “Federico II” and the University of Chieti-Pescara has been carried out. In this context, a great amount of experimental and numerical results have been achieved, highlighting the effectiveness of both steel and aluminium shear panels susceptible to be adopted as passive protection devices of steel and RC moment resisting frames. MPSWs consist of a vertical series of rectangular bays which are formed by columns intersecting beams at the floor levels and are filled by metal panels (infill plates). Beam-to-column connections can be either simple or momentresisting and the panels can be either stiffened or unstiffened. While stiffened infill plates have longitudinal and transversal stiffeners connected to the surface by means of either welded or bolted connections in order to prevent buckling phenomena in the elastic field (compact shear panels), unstiffened shear panels present instability for low values of the applied lateral load and, therefore, rely on their post-buckling behaviour determined through the activation of a resisting diagonal tension field mechanism (slender shear panels). Both of them respond to seismic loads with a high stiffness, stable hysteresis behaviour and a significant energy dissipation capacity, which is more relevant when compact panels are concerned. The high performances exhibited by MPSWs 313

Figure 15.

Geometrical configuration of tested aluminium MPSW specimens (Mazzolani et al., 2007).

Figure 16.

Numerical–experimental comparison for pure aluminium panels.

have been investigated in (Mazzolani et al., 2007) through three different phases relevant to both compact and slender shear panels: 1) experimental-numerical analysis of four types of stiffened pure aluminium panels (Fig. 15); 2) experimental tests on stiffened bracing-type pure aluminium panels; 3) theoretical-numerical-experimental study on steel and aluminium panels for seismic protection of existing RC buildings. In order to interpret the monotonic and hysteretic behaviour of the tested systems, sophisticated FEM models has been set up (Formisano et al., 2006) (Figure 16). It is clearly evident as the proposed model is able to interpret correctly all the main behavioural phenomena of tested specimens. Therefore, the FEM model has been applied as a sort of virtual laboratory to define the main geometrical and mechanical parameters influencing the global response of the simulated systems. Infill bracing type pure aluminium shear panels (BTPASPs) have been also proposed in (Mazzolani et al., 2007) as an alternative and attractive solution for the passive protection of new and existing structures (Fig. 17) under low intensity seismic events. In fact, on the basis of the aforementioned obtained results, full bay aluminium shear panels showed a good performance, in terms of both energy dissipation and damping capability, for medium-large lateral displacements, while some slipping phenomena were observed for small interstory drift levels. The obtained results clearly emphasize that both panel configurations provide a good hysteretic performance, with quite stable hysteretic cycles also for high deformation levels (Figure 17). It is worth noting that higher values of the equivalent viscous damping factor (about 50%) was achieved for large shear strains. The possibility to use metal shear panels for seismic retrofitting of existing RC buildings has been evaluated within the ILVA-IDEM research Project (Mazzolani et al., 2007). The retrofitting design has been developed in the framework of the performance based design methodology according to the procedures of the ATC-40 guidelines (1996), leading to the configuration shown in Figure 18. Also, a refined finite element model of the selected panels has been implemented by ABAQUS non 314

Figure 17.

Steel frames equipped with BTPASP and corresponding hysteretic cycles.

Figure 18.

Reinforcing interventions on the original RC structure (a) and FEM model of the structure (b).

linear software in order to predict correctly their performance under lateral loads. Furthermore, in order to confirm the validity of the proposed design solution and for evaluating the possible relative interaction problems between the RC structure and the added devices, a global analysis of the retrofitted structure, in which the shear plates have been modelled according to the “strip model” theory, has been performed by means of the SAP 2000 calculation program (De Matteis et al., 2006b). A significant improvement of strength (10 and 11.5 times with steel and aluminium panels, respectively), initial stiffness (2.5 and 2 times with steel and aluminium panels, respectively) and ductile capacity (inter-storey drift greater than 3.5% and 6.5% when steel and aluminium panels have been used, respectively) has been recorded. Based on the above highlights, the performed experimental results shown that steel shear panels can be considered as a profitable system to for improving the lateral load resistance of existing RC structures, whereas the pure aluminium shear panels can be also employed to improve the ductility features and the dissipative capacity of the primary structure.

4.2.4 PROTECTION STRATEGIES 4.2.4.1 Passive structural control As an alternative to the conventional approach to seismic design, based on the ductility resources of structural members and connections, the behaviour of a structure can be controlled by means of suitable auxiliary systems able to modify the natural structural properties (stiffness and damping) under dynamic actions, so as to improve the response of the construction and increase its safety against both serviceability and collapse limit states. Such systems are conceived in such a way to reduce or to dissipate a share of the seismic input energy and, to do this, they make use of the response control concept, aiming at controlling and limiting the dynamic effects on the structural elements by means of special devices. In this view, structural control can be either passive or active depending on whether the action of special devices is independent of or is influenced by 315

Figure 19.

Energy dissipation (a), isolated (b) and TMD systems (c).

Figure 20.

Spectral acceleration as a function of the structural damping (a) and effect of period shifting (b).

Figure 21.

Basic options for structural supplemental damping and base isolation.

the structural response itself. In energy dissipation systems (Fig. 19a) special devices are used in order to reduce the amount of energy dissipated in the structure, with a corresponding increase of energy dissipated by devices. This reduces both structural response and damage in structural elements. Basically, the effect of special devices is to increase the overall damping properties of the system, thus reducing its maximum spectral acceleration (Fig. 20a) and, hence, preventing the main structural members from early collapse. According to the classification of Figure 21, supplemental damping can be obtained with rateindependent or rate-dependent devices. The first type is usually made with Yielding Metal Devices 316

Figure 22.

Devices for energy dissipation: Metal yielding (a,b), Friction (c), Viscous (d), Visco-elastic (e).

Figure 23. Devices for seismic isolation: High Damping Rubber Bearins (a), Elasto-plastic Bearings (b), Friction Pendulum System (c), Wire-rope bearings (d).

(YMD), based on the cycling plasticity of a metal, or Friction Devices (FD), based on the friction between surfaces in contact. Their energy dissipation capability depends on the applied displacement magnitude only. Such devices, also named elastoplastic dampers or plastic threshold devices, absorb seismic energy exploiting the inelastic strain properties of strongly dissipative metals, like steel, lead and some special alloys. Metal elements can have different geometrical shapes (spindle, crescent moon, butterfly, track, triangular plate and X), in order to achieve a more uniform plasticization in the element (Fig. 22). Alternatively, it is possible to use rate-dependent devices, whose effect depends on the velocity of applied action. They are usually based on viscous or visco-elastic materials. With a proper combination of both device types, it is possible to have an effective control of seismic vibrations under both low-intensity and severe earthquakes. As an alternative to energy dissipation systems, it is possible to use isolated systems reducing the earthquake input energy (Fig. 19b). In such systems, special devices are used in order to reduce the amount of the energy transmitted to the main structure (Fig. 23). Basically, the effect of seismic isolation is to shift the fundamental vibration period of the building upward, so as to reduce the value of the maximum spectral acceleration (Fig. 20b). A given amount of input energy can be also dissipated by the isolating devices themselves, when they posses special dissipative features. This prevents the structure from an excess of displacements, with a simultaneous control of damage in 317

Figure 24.

Basic principle of seismic isolation on spheres (Michalopoulos et al., 2008).

structural elements. Ultimately, the target of the isolation system is to act like a sort of low-pass filter of the seismic action with respect to the building, so as that most of earthquake energy is prevented from entering the structure. By means of isolation an increase of the global system deformability is achieved, which causes the magnitude of the ground motion transmitted to the structure, and hence the structural damage, to be drastically reduced. A concomitant increase of energy dissipation of the structure is possible by means of suitable additional dampers, which are also used to reduce the displacements at the serviceability limit state (Fig. 21). Likewise energy dissipation systems, such devices aim at increasing the overall damping properties of the system. A further option is represented by Tuned Mass Dampers (TMD) (Fig. 19b), consisting of a mass, a spring and a damper attached to the structure in order to reduce its dynamic response. The system is characterized by mass, stiffness and damping of the added element. The frequency of the damper is tuned to a particular structural frequency so that when excited, the damper will resonate out of phase with the structural motion. Energy is dissipated though the motion of the additional mass. Within COST C26, an innovative isolation system has been proposed, based on the kinematics of a group of metallic spheres or cylinders placed at the bottom of building foundation between two horizontal steel plates (Michalopoulos et al., 2008). By this system, the horizontal earthquake induced vibration is absorbed by several steel spheres or cylinders and does not propagate to the structure (Fig. 24). By means of the described system effective isolation is provided since the whole horizontal movement of the base does not propagate into the upper structure. As soon as earthquake ends, the system of spheres also returns to its initial position, thanks to the fully similarity of the displacement/movement of the ground, with a similarity ratio 1:2. The movements of the spheres faithfully follow the movement of the ground moving back and forward in the same direction as the ground and they stop immediately when the earthquake stops, just as a seismograph does. The system can be implemented also in bridges, in which case unidirectional cylinders are more suitable. 4.2.4.2 Active structural control A natural evolution of the passive techniques involves the use of “smart” systems, which can be used to create controllable devices to be implemented into actively controlled systems. The goal of this strategy is the creation of a mechanical system having enhanced structural performance, but without adding too much mass or consuming too much power. As shown in the pictorial view of Figure 25, the aim of structural control is to give a “sense of balance” to the structure, so as to give it the capability to self-regulate instantaneously its properties as a function of the structural response. Motion data are taken from a sensor network, analysed in real time by a computer and elaborated in such a way to activate external devices (active control) or just to modify the mechanical properties of these (semi-active-control). The materials used in smart structures often have interesting and unusual properties. Electrostrictive materials, magnetostrictive materials, shape memory alloys, magneto/electrorheological fluids, polymer gels and piezoelectric materials, for example, all can be used to design and develop structures that can be called smart. Active control of structures has been recognized as one of the most challenging and significant areas of research in structural engineering in recent years. A structure with active controllers can modify its behaviour during dynamic loading thanks to the contribution of external energy supply. 318

Figure 25. Representation of SMART system effect and possible implementation of active structural control systems for seismic control of buildings.

Figure 26. Variable Orifice Damper, Variable Friction Damper (Kobori et al., 1993) and ElectroRheological Damper with by-pass (Makris et al., 1996a,b).

Such a structure is also called adaptive (or smart) structure. In an actively controlled structure there is a predetermined number of members actively controlled by means of actuators. Sensors are placed in key points in order to measure displacements and velocities in the directions of predetermined degrees of freedom. Actuators apply in real time the forces required for the appropriate correction of the uncontrolled response, which are determined on the basis of a suitable control law, implemented into the so called control algorithm. Actuators are devices that can apply forces or strains to the structure. In comparison with passive systems, there are some significant advantages associated with smart systems: enhanced effectiveness in motion control, relative insensitivity to site conditions and ground motion, applicability to multi-hazard mitigation situations, selectivity of control objectives. In particular, semi-active systems, namely systems in which the device properties only are changed according to the structural motion, seem to offer the most appealing features for seismic protection of civil structures. Contrary to active systems, in fact, they require a very small amount of energy, which can be easily provided by a battery, thus avoiding the problem of power supply black-out, very common during an earthquake. Smart devices used for semi-active structural control are different from actuators for active control, because can only produce dissipative forces. Semi-active devices include variable orifice dampers, variable friction dampers, controllable tuned liquid dampers, controllable fluid dampers, etc. These devices can be viewed as controllable passive devices, in that the characteristics of the passive device can be changed in real time. In this manner, semi-active devices can produce the desired dissipative control forces. They are characterized by low power requirement, passive working and small dimensions in spite of the great reaction force produced. Variable Orifice Dampers. Variable orifice damper use a controllable, electromechanical variable-orifice valve to vary the flow of hydraulic fluid through a conventional hydraulic fluid damper (Fig. 26a). Variable orifice dampers have been applied to full-scale buildings (Kobori et al., 1993; Kurata et al., 1999, 2000) and bridges (Sack & Patten, 1994, Patten et al., 1999). Variable Friction Dampers. Variable friction dampers generate control forces through friction surfaces so as to control the slippage of the device (Fig. 26b). To date, only analytical studies have 319

Figure 27. Magnetorheological fluid principle (Dyke et al., 1996a), (a) scheme of a MagnetoRheological Damper produced by Lord Corporation (Yang et al., 2001) (b) and the MR devices designed and engineered at the Second University of Naples (Mandara et al., 2008) (c).

been conducted on these devices in the view of their possible application to civil structural control. These devices have, however, been proposed to reduce interstorey drifts of seismically excited frame buildings (Inaudi, 1997). Controllable Tuned Liquid Dampers. Controllable tuned liquid dampers use the motion of a column of fluid, varied with a controllable orifice, to reduce structural responses. These dampers are similar in concept to tuned mass dampers (TMDs), as they absorb the energy of the structure by means of auto-induced vibration. However, whereas TMDs are typically designed for one loading condition, the controllable tuned liquid damper can remain effective for a large variety of loading conditions Controllable Fluid Dampers. Controllable fluid dampers are quite similar to the variable orifice dampers, except for the fact that they use controllable fluids, such as electrorheological (ER) and magnetorheological (MR) fluids. Because of this, they do not require a mechanical servo-valve. Gavin et al. (1996a, 1996b) designed and tested an ER damper that consisted of a rectangular container and a moving plunger comprised of nine rigidly-connected flat plates. Makris et al. (1996a,b) developed an ER damper consisting of an outer cylinder and a double ended piston rod that pushes the ER fluid through an annular duct (Fig. 26c). A number of experimental studies have been conducted to evaluate the performance of MR dampers for vibration reduction under wind and earthquakes. Magnetorheological fluid devices are semi-active control dampers with a magnetorheological fluid inside. The fluid denominated as Magnetorheological consists on a suspensions of micron-sized magnetizable particles in an appropriate carrier liquid like synthetic oil, water or silicone oil. When exposed to a magnetic field, the particles develop a dipole moment aligned with the external field which causes particles to form linear chains parallel to the field, as schematically represented in Figure 27. The main feature of the fluid is its ability to reversibly change from free flowing linear viscous liquids to semisolids having controllable yield strength in a matter of milliseconds. Dyke et al. (1996a,b, 1998), Spencer et al. (1997) proposed MR dampers to reduce the seismic vibration of model building structures. Spencer et al. (2000) incorporated an MR damper with a base isolation system such that the isolation system would be effective under both strong and moderate earthquakes. Johnson et al. (2007) employed the MR damper to reduce wind-induced stay cable vibration. In all cases, experimental results indicate that the MR damper is quite effective for a wide class of applications. For this reason, MR Fluid Dampers are the most promising devices for semi-active control. A MR device has been designed and manufactured at the Second University of Naples (Mandara et al., 2008), within the framework of the European Project PROHITECH (Mazzolani, 2007, 2009). A special type of smart device, based on magnetically controllable elastomer (MCE isolator) capable of adjusting its stiffness depending on shear strain magnitude, was developed in the frame of the European Project VAST-IMAGE by Maurer & Söhne GmbH & Co. KG (Fig. 28). The device was proposed in order to have an optimal stiffness value at both serviceability and ultimate limit 320

Figure 28.

Prototype and scheme of the new MCE device (J. Distl – Maurer Söhne GmbH & Co. KG).

state. The efficiency and the field of application of the new device were tested in the case of the existing three storey reinforced concrete hospital as well as on the idealised one- and multi-storey concrete buildings. The analytical models and the analytical results were evaluated on the basis of the experimentally tested prototype of the new device (Isakovic & Fischinger, 2008).

4.2.5 CASE STUDIES A great importance has been given in COST C26-WG2 to the analysis of real cases, for which specific seismic protection systems have been purposely examined. The study of such cases, which have been investigated in both experimental and numerical way, yielded significant information on the effectiveness of protection strategies. In particular, experimental verification of seismic protection systems and methodologies is an important task in case of implementation of new techniques and innovative materials and devices. The experimental verification can be performed on site and/or in laboratory conditions using either full scale testing or testing of models in reduced scale on shaking table. Non-destructive in-situ tests can be performed in linear range by ambient and forced vibrations, by using appropriate equipment for definition of the dynamic characteristics of the structures, important for verification of the analytical model. Shaking table testing on reduced scale models is one of the most popular experimental methods which offers the possibility to investigate the structure both in linear and non-linear range until collapse, applying representative earthquake excitations. Testing of the models can be performed on both strengthened and original model, by verifying the adopted solution under the same earthquake excitations as for the original model. Numerical analysis is an useful method for prediction of the structural behaviour in earthquake condition, even though it should be calibrated with experimental measurements such as ambient and forced vibration of the full scale structures for linear elastic numerical models, and with the shaking table test for non-linear numerical models. The calibrated numerical models can then be very useful to predict the structural behaviour under strong earthquakes, namely development of cracks and global mechanism of failure, stress distribution, deformations, load capacity and ductility, for both the original model and the strengthened one (Krstevska & Taskov, 2008). Since its establishment in 1965, the Institute of Earthquake Engineering and Engineering Seismology - IZIIS - Skopje performed a large number of test on historical and monumental buildings all around the world. All experimental results have been used for the investigation of seismic resistance of structures strengthened according to different methodologies. A very significant example of the activity carried out at IZIIS - Skopje is represented by the experimental and analytical investigation of Mustafa Pasha Mosque in Skopje, Macedonia (Krstevska et al., 2007) (Fig. 28). Two experimental methods have been used: ambient vibration measurements on the original structure and shaking table tests on a 1:6 reduced scale model of the mosque. The obtained dynamic characteristics of the mosque from ambient vibration tests have been used to calibrate numerical models to predict the damage level of the original mosque as well as of the original reduced scale model. The main objective of the tests was to investigate experimentally the effectiveness of the proposed 321

Figure 29. The Mustafa Pasha Mosque model tested at IZIIS – Skopje, Macedonia.

Figure 30. The Fossanova church model tested at IZIIS - Skopje, Macedonia.

reversible technology for strengthening this type of historical monuments. The performed testing was a part of the activities carried out within the FP6 Research Program PROHITECH – “Earthquake Protection of Historical Buildings by Reversible Mixed Technologies” (Mazzolani, 2007, 2009). Following the main objective, the testing was performed in three phases: 1) Testing of the original model for low intensity level, to provoke small damage; 2) Testing of repaired model and strengthened minaret with FRP, until total collapse of the minaret; 3) Testing of strengthened model with FRP and carbon fiber bars until collapse. A similar study has been also made on the gothic cathedral of Fossanova (Italy) (De Matteis et al., 2008) (Fig. 29). A wide campaign of full scale cyclic tests has been performed on two RC buildings in Naples (Fig. 30) in the framework of ILVA-IDEM Project (Mazzolani, 2006). These tests represented an unique occasion of knowledge, since the studied buildings are “real” constructions, representative of a large part of the European building heritage designed without account for seismic actions (Mazzolani et al., 2007). At first, the full scale buildings have been strongly damaged by applying a large deformation corresponding to a seismic input of great intensity and then they have been repaired by means of several strengthening systems (FRP wrappings, different types of braces, shear panels, etc.). Successively they have been damaged again, so to compare the effectiveness of the adopted systems. Results showed that innovative materials and solution are capable to improve the seismic performance of existing non-seismic constructions to a very great extent. This wide experimental programme has been supplemented by a large numerical activity, mainly aimed at the evaluation of the performance of special systems for which the execution of a direct test is particularly difficult. This is the case of the implementation of smart MR devices into existing buildings, including historical masonry buildings. Such a study, dealing with a proposed solution 322

Figure 31.

Full scale test on RC buildings in Naples (Italy) within the ILVA-IDEM Project.

Figure 32. Implementation of active mass damping with MR devices on a masonry building and dynamic model response time history under Calitri seismic input scaled to PGA 0.2 g.

of active mass damping, has been performed by (Mandara et al., 2008), showing the great potential effectiveness of such strategy (Fig. 31).

4.2.6 CONCLUSIVE REMARKS This chapter summarises in a very brief way the large activity carried out within COST C26 WG2 “Earthquake resistance”. A very great effort, in fact, has been done within this group on the topic of innovative materials and technologies for seismic protection of buildings against severe earthquake. An intense four-year-long discussion among WG2 members has led to point out the main aspects of the problem, at the light of current state-of-the-art and codification. A comprehensive outline of the available strategies for preserving both new and existing buildings with regards to earthquake hazard has been reached, showing that a large variety of means does exist and is readily available to practising engineers. The effectiveness of this provisions has been validated through a large amount of both numerical and experimental research, as well as through direct application to real cases. Also, the possibility to implement such solutions into historical and monumental constructions has been carefully evaluated. The whole of this activity has confirmed that the target of a higher protection level against earthquake is fully achievable with the technological solutions available today. Nevertheless, a need for further research is felt, most of all in those fields, e.g. the new materials and the smart systems, in which the investigation is still at an early stage. In the same way, the current codification should be updated to cover issues related to the protection against abnormal seismic events. The activity carried out within WG2 and the fruitful debate which come out demonstrated the great interest of the scientific community in this field and, at the same time, 323

the availability of research centres all over Europe highly qualified and motivated to promote and undertake further investigation in this field. REFERENCES Bordea, S., Stratan, A., Dogariu, A., Dubina, D. 2007. Seismic upgrade of non-seismic r.c. frames using steel dissipative braces. Proceedings of COST C26 Workshop, “Urban Habitat Constructions under Catastrophic Events”, Prague 30–31 March, 2007. Brameshuber, W. (ed.) 2006. Textile Reinforced Concrete. State-of-the-Art Report of RILEM Technical Committee 201-TRC. Bagneux: RILEM Publ. S.A.R.L. Croci, G., Bonci, A., Viskovic, A. 2000. Use of shape memory alloy devices in the Basilica of St. Francis in Assisi. Proceedings of Final Workshop of ISTECH Project - Shape Memory Alloy Devices for Seismic Protection of Cultural Heritage Structures, Ispra, Italy. D’Aniello, M., Della Corte, G., Mazzolani, F.M. 2007. A special type of buckling-restrained brace for seismic retrofitting of RC buildings: design and testing. XXI C.T.A. Conference, Catania (Italy). D’Aniello M., Della Corte G., Mazzolani F. M., 2008, Response of Buckling Restrained Braces to Catastrophic Seismic Events. Proceedings of COST C26 Symposium “Urban Habitat Constructions under Catastrophic Events”, Malta 23–25 October, 2008. Della Corte, G., Barecchia E., Mazzolani, F.M. 2006. SeismicUpgrading of RC buildings by FRP: full scale tests of a real structure. Journal of Materials in Civil Engineering, ASCE, vol.18, No.5, 659–669. Della Corte, G., D’Aniello, M., Mazzolani, F.M., 2005. Seismic Upgrading of RC buildings using Buckling Restrained Braces: full-scale experimental tests. XX C.T.A. Conference – First Int. Workshop on Advances in Steel Constructions, Ischia (Italy), 26–28 September 2005. Della Corte, G., Mazzolani, F.M., 2006. Full-scale lateral-loading tests of a real masonry-infilled RC building. Proceedings of the Second Fib Congress, Naples (Italy) 5–8 June 2006. De Matteis, G., Formisano, A., Panico, S., Calderoni, B., Mazzolani, F.M. 2006. Metal shear panels. Seismic upgrading of RC buildings by advanced techniques – The ILVA-IDEM Research Project. Mazzolani, F. M. Coordinator & Editor, Polimetrica International Scientific Publisher, Monza, Italy, pp. 361–449. De Matteis G., Mazzolani F.M., Krstevska L., Tashkov L. 2008. Seismic analysis and strengthening intervention of the Fossanova gothic church: numerical and experimental activity. Proceedings of COST C26 Symposium “Urban Habitat Constructions under Catastrophic Events”, Malta 23–25 October, 2008. Dolce, M. & Cardone, D. 2001. Mechanical Behaviour of Shape Memory Alloys for Seismic Applications. International Journal of Mechanical Sciences, Vol.43. pp. 2631–2656. Dogariu A., Dubina D., Campitiello F., De Matteis G. 2008. Strengthening of masonry walls by innovative metal based techniques. Proceedings of COST C26 Symposium “Urban Habitat Constructions under Catastrophic Events”, Malta 23–25 October, 2008. Dogariu, A., Stratan, A., Dubina, D., Nagy-Gyorgy, T., Daescu, C., Stoian, V. 2007. Proceedings of COST C26 Workshop, “Urban Habitat Constructions under Catastrophic Events”, Prague 30–31 March, 2007. Dyke, S.J., Spencer Jr., B.F., Jr.,Sain, M.K., Carlson, J.D., 1996a. Modeling and control of magnetorheological dampers for seismic response reduction. Smart Materials and Structures, 5, 565–575. Dyke, S.J., Spencer Jr., B.F., Quast, P., Kaspari Jr., D.C., Sain, M.K., 1996b. Implementation of an Active Mass Driver Using Acceleration Feedback Control, Microcomputers in Civil Engineering: Special Issue on Active and Hybrid Structural Control, 11, 304–323. Dyke, S.J., Spencer Jr., B.F., Sain, M.K. and Carlson, J.D., 1998. An experimental study of MR dampers for seismic protection, Smart Mat. and Struct., 7, pp. 693–703. Formisano, A., Mazzolani, F.M., Brando, G., De Matteis, G. 2006. Numerical evaluation of the hysteretic performance of pure aluminium shear panels”. Proc. of the 5th International Conference STESSA ’06, Balkema, 211–217, Yokohama, Japan. Gavin, H.P., Hanson, R.D., Filisko, F.E., 1996a. Electrorheological dampers, part 1: analysis and design. Journal of Applied Mechanics, ASME; 63(9):669–75. Gavin, H.P., Hanson, R.D., Filisko, F.E., 1996b. Electrorheological dampers, part 2: testing and modeling. Journal of Applied Mechanics, ASME;63(9):676–82. Georgiadi-Stefanidi, K., Mistakidis, E., Perdikaris, P.C. 2008. Thin fibre-reinforced concrete jackets for improving the seismic response of reinforced concrete members: experimental and numerical results. Proc. of COST C26 Symp. “Urban Habitat Constructions under Catastrophic Events”, Malta 23–25 Oct. 2008. Giuffrè A.& Martines G. 1989. Impiego del Titanio nel Consolidamento del Capitello della Colonna Antonina. Proceedings of the Convegno e Mostra A.N.I.A.SPE.R. Rome.

324

Inaudi, J.A., 1997. Modulated Homogeneous Friction: A Semi-Active Damping Strategy. Earthquake Engrg., and Struct. Dyn., 26, 361–376. Indirli. M. 2000. The Demo-Intervention of the ISTECH Project: the Bell Tower of S. Giorgio in Trignano (Italy). Proceedings of the Final Workshop of ISTECH Project, Ispra (Italy). Isakovi´c, T. & Fischinger, M. 2008. Seismic protection of buildings using innovative isolators based on magnetically controlled elastomer. Proc. of COST C26 Symposium “Urban Habitat Constructions under Catastrophic Events”, Malta 23–25 Oct. 2008. Johnson, E.A., Baker, G.A., Spencer, B.F., Jr., Fujino, Y. 2007. Semiactive Damping of Stay Cables. Journal of Engrg. Mech., ASCE„ Vol. 133, Issue 1, pp. 1–11. Kobori, T., Takahashi, M., Nasu, T., Niwa, N., Ogasawara, K., 1993. Seismic Response Controlled Structure with Active Variable Stiffness System, Earthquake Engrg., and Struct. Dyn., 22, 925–941. Krstevska L., Taskov Lj., Gramatikov K., Landolfo R., Mammana O., Portioli F. Mazzolani F.M. 2007. Experimental and numerical investigations on the Mustafa Pasha Mosque large scale model. Proc. of COST C26 Workshop, “Urban Habitat Constructions under Catastrophic Events”, Prague 30–31 March, 2007. Krstevska, L., Taskov, L. 2008. Verification of effectiveness of seismic protection and retrofit techniques by experimental testing. Proceedings of COST C26 Symposium “Urban Habitat Constructions under Catastrophic Events”, Malta 23–25 October, 2008. Kurata, N., Kobori, T., Takahashi, M., Ishibashi, T., Niwa, N., Tagami, J., Midorikawa, H., 2000. Forced Vibration Test of a Building with Semi-Active Damper System. Earthquake Engrg., and Struct. Dyn., 29, 629–645. Kurata, N., Kobori, T., Takahashi, M., Niwa, N., Midorikawa H., 1999. Actual Seismic Response Controlled Building with Semi-Active Damper System, Earthquake Engrg., and Struct. Dyn., 28, 1427–1447. Makris, N., Burton, S.A., Hill, D., Jordan, M., 1996a. An electrorheological damper with annular duct. Proceedings of Structures Congress XIV, Chicago, IL: 1197–204. Makris, N., Burton, S.A., Hill, D., Jordan, M., 1996b. Analysis and design of ER damper for seismic protection of structures. Journal of Engineering Mechanics, ASCE; 122(10):1003–11. Mandara, A. 2007. Innovative materials and technologies for existing and new buildings in seismic areas. Proceedings of COST C26 Workshop “Urban Habitat Constructions under Catastrophic Events”, Prague, 30–31 March, 2007. Mandara A., Muzeau J.P., Perdikaris P., Piazza M., Schaur C. (2002). Repairing and Strengthening for New Requirements: Use of Mixed Technologies. 1st COST-C12 Seminar, Lisbon (Portugal). Mandara A., Ramundo F., Spina G. 2008. Smart technologies in the seismic protection of existing buildings: Part 1: General concepts: Part 2: Applications. Proceedings of COST C26 Symposium “Urban Habitat Constructions under Catastrophic Events”, Malta 23–25 October, 2008. Mazzolani, F.M. (Co-ord. and Ed.), 2006. Seismic upgrading of RC buildings by advanced techniques. The ILVA-IDEM research project”. Polimetrica International Scientific Publisher, Monza, Italy. Mazzolani, F.M. 2007. Earthquake protection of historical buildings. Proceedings of COST C26 Workshop, “Urban Habitat Constructions under Catastrophic Events”, Prague 30–31 March, 2007. Mazzolani, F.M. 2009. Seismic protection of historical buildings: The PROHITECH project. Proc. of the 1st Int. Conference on Protection of Historical Buildings. Rome, 21–24 June, 2009. Mazzolani, F.M., Della Corte, G., Barecchia, E., D’Aniello, M. 2007. Experimental tests on seismic upgrading techniques for RC buildings. Proceedings of COST C26 Workshop, “Urban Habitat Constructions under Catastrophic Events”, Prague 30–31 March, 2007. Mazzolani, F.M., De Matteis, G., Panico, S., Formisano, A., Brando, G. 2007. Shear panels for seismic upgrading of new and existing structures. Proceedings of COST C26 Workshop, “Urban Habitat Constructions under Catastrophic Events”, Prague 30–31 March, 2007. Mazzolani, F.M. & Mandara, A. 2002. Modern Trends in the use of special metals for the improvement of historical and monumental structures. Eng. Structures Vol.24. pp. 843–856. Michalopoulos A., Nikolaidis Th., Baniotopoulos C.C. 2008. Seismic protection of new and existing buildings using an innovative isolation system. Proceedings of COST C26 Symposium “Urban Habitat Constructions under Catastrophic Events”, Malta 23–25 October, 2008. Papatheoharis, T., Perdikaris, P.C., Tzaros K. 2008. Experimental Investigation of the Response of FiberReinforced Concrete Beams under Static and Cyclic Bending, Proc. of 3rd Greek National Conference of Seismic Engineering and Technology, Athens. Patten, W.N, Sun, J., Li, G., Kuehn, J., Song, G. 1999. Field Test of an Intelligent Stiffener for Bridges at the I-35 Walnut Creek Bridge. Earthquake Engng. Struct. Dyn. 28, 109–126. Pegon, P., Armelle, A., Pinto, A., Renda, V. 2000. The ELSA Laboratory and the Protection of Cultural Heritage. Proceedings of the 5th Int. Congress on Restoration of Architectural Heritage, Florence.

325

Sabelli, R., Lopez, W. 2004. Design of Buckling-Restrained Braced Frames, Modern Steel Construction, 44(3): 67–73, 2004. Sack, R.L. & Patten, W. (1994). Semiactive Hydraulic Structural Control, Proc. Int. Workshop on Struct. Control, USC Publication Number CE-9311, pp. 417–431, 1994. Sickert J-U., Kaliske M., Graf W. 2008. Strengthening of earthquake damaged structures by means of textile reinforced concrete. Proceedings of COST C26 Symposium “Urban Habitat Constructions under Catastrophic Events”, Malta 23–25 October, 2008. Sickert, J.-U., Möller, B. Graf, W. 2007. Fuzzy Stochastic Finite Element Analysis. In: Kanda, J., Takada, T. & Furuta, H. (eds.) Application of Statistics and Probabilistics in Civil Engineering, Proc. of the ICASP 10, Tokyo. London: Taylor & Francis. Spencer BF, Dyke SJ, Sain MK, Carlson JD., 1997. Phenomenological model for magnetorheological dampers. Journal of Engineering Mechanics; 123(3):230–8. Spencer Jr., B.F., Johnson, E.A., Ramallo, J.C., 2000. Smart Isolation for Seismic Control. JSME International Journal: Special Issue on Frontiers of Motion and Vibration Control, Series C, 43(3), 704–711. Stratan, A. & Dubina, D., 2004. Bolted links for eccentrically braced steel frames. Proc. of Connections in Steel Structures - Innovative Steel Connections, June 3-5, 2004, Amsterdam, Netherlands. Stratan, A. & Dubina, D. 2008. Removable bolted links for eccentrically braced frames. Proc. of COST C26 Symposium “Urban Habitat Constructions under Catastrophic Events”, Malta 23–25 October, 2008. Tsai, K.C., Lai, J.W., Hwang, Y.C., Lin, S.L., Wong, Y.T., 2004. Research and application of double-core buckling restrained braces in Taiwan. Proc. of the 13th World Conf. on Earthquake Eng., Vancouver. Wada, A. & Nakashima, M., 2004. From infancy to maturity of buckling restrained braces research. Proceedings of the 13th World Conference on Earthquake Engineering, Vancouver (Canada). Yang, G., Spencer Jr., B.F., Carlson, J.D., Sain, M.K., 2001. Dynamic modeling and performance considerations on full-scale MR fluid dampers. Proc. 8th International Conference on Structural Safety and Reliability, Newport Beach, CA.

326

Urban Habitat Constructions under Catastrophic Events (Final Report) – Mazzolani (Ed). © 2010 Taylor & Francis Group, London, ISBN 978-0-415-60686-8

4.3 Blast loading assessment and mitigation in the context of the protection of constructions in an urban environment Peter D. Smith Department of Engineering Systems and Management, Cranfield University, Defence Academy of the United Kingdom, Shrivenham, UK

4.3.1 INTRODUCTION As Remennikov [2002] states: “Civil engineers today need guidance on how to design structural systems to withstand various acts of terrorism.” This paper seeks to set out some basic guidance for the civil engineer who needs to know about: (i) the characteristics of a blast load produced by, for example, the detonation of a quantity of high explosive material contained in a terrorist-deployed Vehicle Borne Improvised Explosive Device (VBIED), (ii) techniques to mitigate the effects of such a load that will allow him to design buildings that, though robust, are far from being fortress-like and (iii) features that buildings should incorporate to reduced the damaging effects of blast loading. Firstly, attention is given to the assessment of blast loads on buildings in an urban environment starting with the simplest techniques that are generally based on the calculation of scaled distance which, though useful, have their limitations in an urban context. Secondly, therefore, empirical approaches to blast load estimation that attempt to account for shielding and channeling effects in an urban environment are briefly considered before the use of numerical simulation for blast load assessment is discussed. Thirdly, the paper discusses the desirability of the creation of both ‘real’ and ‘virtual’ stand-off. The first might be achieved by the appropriate deployment of fixed, strong and immovable barriers and obstacles such as bollards etc. designed to keep the threat at some distance from the building while ‘virtual’ stand-off might be developed by the construction of blast mitigating devices such as blast walls positioned between the threat location and the building. Finally, the requirements and techniques for the development of a robust building are introduced, including the provision of a building façade that prevents the ingress of a blast load into the building. Particular reference is made to the incorporation of glazing systems that, though they may suffer damage, do not rupture.

4.3.2 BLAST LOAD ASSESSMENT BASED ON SCALED DISTANCE IN SIMPLE GEOMETRIES When a single isolated building is loaded by the blast wave produced by the detonation of a quantity of high explosive, calculation of the pressure-time history experienced by the building is generally relatively straightforward, particularly for the side of the building directly facing the blast. Loading information can be obtained from graphs of blast resultants that are contained in manuals such as UFC 3-340-01 [2002] or UFC3-340-02 [2008] and texts such as Baker et al. [1983], Smith and Hetherington [1994] or Cormie et al. [2009]. Limited distribution software such as ConWep [2001] and the more readily accessible Blast Effects Computer [2000] have user-friendly interfaces 327

Figure 1. Peak incident and reflected overpressure vs range for 200 kg TNT equivalent charge generated using ConWep.

and automate the process of blast load assessment for simple geometries. Both programs use the data provided by Kingery and Bulmash [1984] for the calculation of important quantities such as peak overpressure, specific impulse and positive phase duration. If ConWep is used, the mass of explosive, W , and the standoff distance to the building of interest, R, is input to the program and the quantity scaled distance, Z (= R/W 1/3 ), is evaluated allowing access to the blast resultants database. An example of the ConWep program’s output showing the variation of incident and reflected pressure with range from a hemispherical charge comprised of 200 kg TNT is shown in Figure 1.

4.3.3 BLAST LOAD ASSESSMENT IN MORE COMPLEX GEOMETRIES If a device such as a VBIED is detonated in a more complex urban geometry where there are many buildings near the point of detonation, assessment of the loading experienced by a particular building becomes more difficult. Such assessment becomes even more complicated should the façades of buildings partly or completely fail, allowing the blast to enter. Smith & Rose [2006] illustrates the problems of accurate blast load estimation in complex geometries. In this study a 1/50th scale model of a straight city street was constructed with buildings, represented by small reinforced concrete beams, along each side. Two separate buildings, made of steel plate, were located at each end of the street and a small charge (replicating a sizeable VBIED) was detonated in the middle of the street at different distances along it, as illustrated in Figure 2. Reflected pressure-time histories were measured on the façade of Building A and compared with those when the buildings along the street were absent. Figure 3 shows how the peak reflected pressure, Pr, varied with scaled distance, Z. The presence of the street buildings increased reflected pressure (and also the associated impulse) by a factor of up to about four, thus clearly demonstrating the blast ‘channelling’ effects produced by the buildings along a street. In reality, this enhancement would be somewhat less than measured here, because of the likelihood of glazing failures in the vicinity of the explosion, thus reducing the intensity of the loading on Building A. However, even allowing for some façade failure, the important point is that this enhancement could not have been accurately predicted using simple tools based on the ‘free field’ calculations described in UFC3-340-02 [2008] and Kingery & Bulmash [1984]. 328

Figure 2. VBIED in city street.

Figure 3. Peak reflected overpressure on façade of Building A with and without buildings along the street for various charge locations in the street.

To acquire a better understanding of the effects of urban geometry, series of both experimental and numerical investigations were conducted by the author and others. In summary, these involved: (i) the detonation of a VBIED at a crossroads, T-junction, 90◦ bend, straight street, and cul-desac (or ‘dead end’) and the measurement of blast resultants on the façades of buildings along adjacent streets. The study demonstrated that the more confined the space in which a detonation occurs, the greater the loading experienced by adjacent buildings, with the cul-de-sac geometry producing the highest loading and the cross-roads geometry the lowest. (ii) the measurement of blast resultants on the street level façades of buildings in straight streets of various width, w, bounded by buildings of different height, h. The study demonstrated that if scaled width (w/W 1/3 ) exceeded 4.8 m/kg1/3 the street is sufficiently wide such that the loading experienced by buildings on one side is not affected by reflections from buildings on the other side of the street. If scaled height (h/W 1/3 ) exceeds 3.2 m/kg1/3 buildings are effectively infinitely tall and there is little significant enhancement at street level of the positive phase impulse. Finally, for Z > 2.0 kg/m1/3 , negative phase impulse exceeds the positive phase impulse for all street widths and all scaled building heights. This observation may go some way to explaining the anecdotal evidence that much of the glazing in city streets is drawn into the street by the passage of a blast wave following an explosion. (iii) studying the effect of detonating a charge at some distance from a road feature such as a T-junction. It was determined that the distance of the charge from the junction influences the 329

Figure 4. Impulse reduction at a fixed location along a street as a function of the porosity of a façade failing adjacent to an explosion.

extent to which the blast diffracts at the junction (and enters the other streets leading off the junction). The larger the distance of the charge from the junction, then the greater the degree of diffraction that occurs at the junction, as opposed to reflection and transmission back down the street in which the charge is located. (iv) considering the effect of building façade failure in which it was found that, as the percentage of frangible material in a façade (termed ‘porosity’) adjacent to an explosion increased, the intensity of the blast further along the street diminished as energy from the blast entered the building via the broken façade. There was an approximately linear decay in impulse down the street with increasing porosity as shown in Figure 4. (v) assessing the effect of the blast loading entering a building via a failed façade. The study indicated that the resultant internal loading could be of sufficient magnitude to cause damage both to the building fabric and the building’s occupants. (vi) quantifying the effects of (a) ‘shielding’ (i.e. the reduction in blast loading due to the presence of another building being located between the explosion and the building of interest) and (b) ‘channelling’ (i.e. the increase in blast loading due to multiple reflections on buildings adjacent to the building of interest). In these studies, arrays of buildings included (i) simple terraced housing (i.e. straight rows of interconnected houses in a series of parallel streets), (ii) symmetrical arrays of buildings with a charge detonated outside the array and (iii) regular and random arrays of buildings with a charge detonated within the array. The more complex the array, the more difficult it became to define the factors influencing the resulting blast loads produced, though it did seem that the greatest shielding effect was produced by buildings closest to the point of detonation. However, when the results of all these studies were combined it was apparent that, no matter what the geometry of a building array, the net result outside the array was an approximately 10% reduction in overpressure and impulse. This is because the resultant load is a combination of shielding and channelling effects occurring in parallel as illustrated schematically in Figure 5. Further details of the studies discussed above can be found in Smith & Rose [2006] from which it can be concluded that, for simple geometries, simple blast calculation tools could be acceptable. In the case of blast propagation along relatively simple-geometry city streets, rules can be formulated to predict blast resultants on building façades but, for more complex city street layouts, such rules become difficult to develop. It was also clear that, when buildings bounding streets respond and façades fail, any such rules must be altered. Also, when blast propagates through arrays of simple buildings, both shielding and channelling effects occur. Because of the complexity of ‘real’ cityscapes, reasonably accurate prediction of blast resultants may require numerical simulation. Thus, Smith et al. [2005] studied blast propagation in the generic cityscape show in Figure 6 both experimentally and numerically using the code Air3d. For a particular VBIED location (as shown in Figure 7) measurements of pressure-time histories were made on the portion of façade of Building 7 adjacent to the right hand wall of Building 4. 330

Figure 5.

Schematic representation of ‘shielding’ and ‘channelling’ effects.

Figure 6. Generic city scale (nb roofs of two triangular buildings not shown) showing various VBIED detonation locations.

The Air3d numerical study that was conducted further emphasised the complex nature of the blast loading that develops in real urban landscapes. Figure 9 shows Air3d-generated pressure (in kPa) and Figure 10 impulse (in kPa-msec) contours on the façade of Building 6 where it is seen that the loading is, as expected, non-uniform. However, it should also be noted that regions of high and low loading do not necessarily occur where they might intuitively be expected: ‘hot spots’ occur 331

Figure 7. Location of VBIED in generic cityscape shows three superimposed pressure-time histories from three separate experiments, demonstrating both the complexity of the records and their repeatability – the complex nature of these records is a ‘real’ effect.

Figure 8.

Pressure-time histories recorded on façade of Building 7.

Figure 9. Air3d calculation of overpressure on façade of Building 6.

Figure 10. Air3d calculation of impulse on façade of Building 6.

332

where a building surface might be expected to be shielded and relatively low loads are evident where a direct line from charge to ‘target’ might be expected to produce a higher load.

4.3.4 ‘REAL’ AND ‘VIRTUAL’ STAND-OFF 4.3.4.1 The effect of stand-off When an explosive device is detonated on the ground (i.e. as a ‘surface burst’), in the absence of any surrounding obstacles, the blast wave so created propagates away from the point of detonation with a hemispherical wave front. If the blast wave-forming energy released by the detonation (the majority of the explosive’s energy content) is E and the wave front has reached a radius r, the energy per unit volume within the wave front is E/[2 /3 πr 3 ] measured in Joules per cubic metre (J/m3 ). This quantity can be expressed as ‘N-m/m3 ’ which can be simplified to ‘N/m2 ’ which, of course, is the standard SI unit of pressure, the Pascal. Thus, it is evident that as blast wave front radius increases, the peak overpressure at the wave front, one of the most significant quantities defining the blast wave, decreases with distance from the point of detonation at a rate that is related to the inverse of the cube of the distance from the explosion. Put simply, this means that peak overpressure (and the associated impulse) decreases rapidly with range. This reduction can clearly be seen by inspection of Figure 1 for a 200 kg TNT hemispherical surface burst: as scaled distance Z changes from 1 m/kg1/3 to 10 m/kg1/3 (ie from a range of 5.84 m to 58.4 m from the point of detonation) the peak overpressure at the wave front falls from 20 MPa to 0.4 MPa. Thus, a change by a factor of 10 in range produces a reduction in pressure by a factor of 50. This simple demonstration suggests that the easiest way to reduce the damaging effects of a blast load on a building is to keep the point of detonation as far away from the building as possible: the creation of ‘stand-off ’ will lead to a significant reduction in blast overpressure and impulse loading on the building. This desire is sometimes difficult to fulfil in an urban environment where real estate is at a premium. The following sections discuss means whereby ‘real’ stand-off can be created by systems designed to keep an explosive threat at a distance from a target building and ‘virtual’ stand-off which might be achieved by the deployment of structures such as blast walls. 4.3.4.2 Creation of ‘real’ stand-off It is often the case that the threat to a building is in the form of a VBIED whose close approach to a target building must be prevented. To counter such a threat, ‘real’ stand-off can be created by the use of a Vehicle Security Barrier (VSB) which provides a substantial obstacle to the progress of the VBIED. VSBs are available in a number of structural forms and are either ‘passive’ or ‘active’ systems, the latter being either externally or manually powered. Passive systems include static bollards, planters and strengthened street furniture (sometimes described as ‘architectural solutions’), bunds (i.e. mounds) and ditches and perimeter systems based on the use of wire rope. Even trees of sufficient girth (and hence likely to be robust), provided the spacing between them does not allow passage of a potential VBIED, might be suitable as a passive VSB. Figures 11(a), (b) and (c) show typical examples of static bollards. When installing discrete VSBs, they should be spaced such that the maximum clear distance between them is no greater than 1200 mm to limit the opportunity for a hostile vehicle to encroach through the barrier line, whilst providing sufficient access for pushchairs and wheelchairs etc. Those shown in Figure 11(b) taper from base to top: it is essential that the 1200 mm dimension be measured at a height of 600 mm above the finished ground level. Figure 11(d) shows a typical heavyweight planter. Active systems include retracting and rising bollards and road blockers, rising and dropping arm barriers and sliding and hinge gates, examples of which are shown in Figure 12. The choice of which system (either active or passive) depends on a number of ‘scenario specific’ factors. As stated in Cormie et al. [2009]: “When considering the most effective barrier configuration for a site, the threats to be mitigated ………………….must first be clearly identified. Once identified, the potential vulnerabilities of each configuration against the defined threats may be assessed.” Further information on vehicle-borne threats and the principles of hostile vehicle mitigation will be found in Chapter 11 of Cormie et al. [2009]. 333

Figure 11.

Passive measures for maintaining ‘real’ standoff (from Cormie et al. [2009]).

Figure 12. (a) Temporary modular hinge gate and linked surface-mounted barriers (b) Retracting bollards [note static bollards at kerb edge] (c) Rising arm barrier (from Cormie et al. [2009]).

334

Figure 13.

New US embassy in London [© KieranTimberlake/studio amd] (from USEmbassy.org.uk [2010]).

4.3.4.3 Creation of ‘virtual’ stand-off The term ‘virtual’ stand-off is here taken to mean that the actual distance from the point of detonation of an explosive to a building has been effectively increased. This could be achieved by the interposing of some blast mitigating system between the explosion and the building. The effect of the blast on the building is reduced by the mitigating system since its intensity is now that which would have been developed if the detonation had occurred further from the building: the actual stand-off distance has been enhanced to a ‘virtual’ stand-off. Perhaps the most commonly used mitigating system is the ‘blast wall’. In the context of this paper, a blast wall is a physical barrier separating a building from an explosive threat that produces a blast load capable of damaging the building; the wall reduces the intensity of the blast loading on the building being protected. It is most likely that such a barrier would be highly robust, being both massive and of strong construction, undergoing little by way of serious damage from the blast produced by the explosive device. In simple terms, the wall reflects blast energy back towards the explosive source and this energy is thus unavailable to damage the building being protected. However, it is worth noting that, when a blast wall does undergo permanent, damaging deformation, energy from the blast is, in addition, absorbed by the wall and is thus not available to damage the building. Therefore, it is worth considering that blast walls could also be relatively lightweight and weak and still offer some degree of protection because a high level of deformation of such a wall could absorb a significant amount of the blast wave energy from the explosive device. Thus, depending on the design of wall deployed, a combination of energy reflection and energy absorption will mitigate the loading developed behind the wall. It is important to realise, however, that partial (or complete) wall failure should not produce fragments that could themselves cause damage to the building and its occupants. The use of massively strong permanent blast walls (probably made from in situ reinforced concrete) may not be a practical solution for installation in the central business district of a large city. Often the necessary space is unavailable and, even it was, the construction of such a wall could project an unacceptable image of the building’s occupiers as being intimidated by a possible threat. However, the same argument does not apply to buildings occupying more generously-sized real estate where a combination of physical stand-off coupled with a blast wall (likely to be of a design complementing that of the building) to provide extra ‘virtual’ stand-off offers a sound protective strategy. Such an approach seems to be that adopted by KierenTimberlake of Philadelphia, Pennsylvania, USA, the winner of the contest to design the new United States of America’s Embassy in the United Kingdom on the River Thames in London (USEmbassy.org.uk [2010]) as shown in Figure 13. Here, a large site coupled with a perimeter wall are components of the protective system for the building. 335

Figure 14. Peak overpressure behind (i) a mound and (ii) blast walls of height H of different configurations vs scaled range (R/W 1/3 ) (from UFC3-340-01 [2002]).

There are situations where the threat to a building (which could be in a central business district, on the periphery of an urban environment or even in a relatively isolated location) is transient as would be the case if the building was, say, the location of a gathering of important political leaders. In this situation, in addition to the deployment af a temporary VSB as illustrated in Figure 12(a) above, consideration might be given to the use of temporary blast walls that might use water as the mitigating material. Smith [2010] presents a review of the use of blast walls for structural protection against high explosive threats. The paper discusses early work on blast walls conducted in the United States aimed at the protection of US overseas missions. These investigations, augmented by other studies (including those conducted by the author and others at the Defence Academy of the United Kingdom) has been collated and is presented in the form of a series of graphs of either peak overpressure or impulse vs scaled range in UFC3-340-01 [2002]. Figure 14 shows one of these graphs where it will be seen that the introduction of a blast wall (of height H ) between the explosion (of W kg TNT equivalent) and the building produces a reduction in blast resultant compared without a wall (‘free-field’ on the graph). Even a mound of material produces some degree of blast mitigation. This work relates to walls that, though they may suffer damage, remain essentially intact. Rose et al. [1998] reports on studies of walls that completely fail under the action of a blast load. Materials used for construction of the walls included sand, plastic sheeting and water. The work suggested that, provided a blast wall can survive sufficiently long to interact with the blast wave, some reduction in the intensity of the blast will occur. Of the materials investigated, water was one of the most promising because, not only did it produce a significant reduction in blast resultants behind the wall, but also fragments from the explosive device were captured by the water. Figure 15 shows some examples of blast wall designs. 4.3.5 ROBUST BUILDINGS 4.3.5.1 Requirements for robustness When a building is designed with a capability to withstand blast, the provision of adequate blastresistance for individual structural elements is, of course, very important. However, equally 336

Figure 15. Blast wall designs: permanent (a) and temporary (b) reinforced concrete walls; temporary sand-filled military (c) and civilian walls (d); water-filled wall.

significant is the need to consider the entire building’s response to the damage caused to individual structural elements: the ‘global’ effects of what might seem relatively ‘local’ damage must be understood. For example, if a building does not have adequate framing, the loss of a loadcarrying member due to blast means that alternative load paths are not available to redistribute the load previously carried by the failed element and a wider collapse could result. The provision of 337

continuity between members and tying can limit the extent of collapse resulting from the loss of a structural member, but it is important to be able to assess whether the extent of any collapse is likely to be “disproportionate”. In 1968 an internal gas explosion resulted in the progressive collapse of one corner of a multistorey block of flats at Ronan Point in London [1968] which led to the relevant section of the UK Building Regulations (DCLG [2004]) being amended. A section on disproportionate collapse due to accidental loading was included which states that: “the building shall be constructed so that in the event of an accident the building will not suffer collapse to an extent disproportionate to the cause.” This amendment did not imply a requirement that buildings should to be designed to resist blast loads. Rather, it is intended to ensure buildings are provided with a degree of robustness that allows the redistribution of loads and that structural frames are better able to withstand the loss of one or more elements, even if the resulting damage exceeds the design basis of the structure. For tall buildings, DCLG [2004] adopts a measure of disproportionate damage that says that the building should remain stable and the risk of collapse of the area of floor at that storey should be the smaller of 15% or 70 m2 of the floor area following the notional removal of each supporting column and each beam supporting one or more columns, or any nominal length of load-bearing wall. Further, there is a requirement that any collapse should not extend further than immediately adjacent storeys. Figure 16 shows examples of the collapse of buildings subjected to dynamic loading (i) designed prior to the amendment to UK Building Regulations [a, b] (ii) designed after the amendment [c] and (iii) designed to the US document ACI 318-71 as it was in 1977 [d]. 4.3.5.2 Protection and disproportionate collapse In the context of blast damage caused by acts of terrorism, such stringent provisions are not typically applicable. However, designers of buildings that could be the targets of such attacks should ensure that ‘disproportionate’ building response and possibly collapse is prevented. It is encouraging to note that most structural steel and in situ reinforced concrete frames have the potential to perform well under blast loads including those from large VBIEDs. The following is extracted from Chapter 10 of Cormie et al. [2009]: “Local damage may well be severe – it is unlikely that structural slabs closest to the explosion will survive the blast, and this damage may extend over a number of storeys. One or more columns may be severely deformed or completely severed in the explosion. However, for all but the most missioncritical facilities, the designer’s aim should be to strike a balance between economic design and explosion protection, and thus to avoid a level of damage which is deemed to be disproportionate, rather than to eliminate damage completely. If the primary structure of a building is well-designed and robustly detailed, it will inevitably exhibit a degree of structural distress especially if columns and transfer beams are damaged, but will otherwise respond well to the explosion. When sufficient attention is paid to good design and detailing of the structure, the frame should act to arrest a developing collapse through redistribution of the load. In doing so, the extent of structural collapse resulting from an explosion is limited as will be the risks to occupants associated with the structure, which are generally substantially lower than the risks due to glazing failure.” 4.3.5.3 Design methods for structural robustness Three basic approaches to design for structural robustness can be identified which are generally common to the different international codes and standards. The following extract from Chapter 10 of Cormie et al. [2009] summarizes them thus: “(i) Tie-force based design methods: prescriptive (rule-based) approaches by which the structure is usually deemed to satisfy robustness requirements through minimum levels of ductility, continuity and tying. (ii) Alternate loadpath methods: quantitative approaches whereby the structure is shown to possess adequate robustness against collapse to satisfy the code requirements. (iii) Key element design: typically used as the method of last resort, a quantitative design approach for designing elements, the removal of which would lead to a collapse defined as disproportionate, for an accidental loadcase. It varies whether a prescriptive load is defined for use in 338

Figure 16. (a) Progressive collapse at Ronan Point, London: 18th floor gas explosion 1968; (b) Chamber of Shipping, London built prior to post-Ronan Point Building Regs amendments: VBIED attack 1992; (c) Kansallis House, Bishopsgate, London, design incorporated the post-Ronan point tying requirements: VBIED attack 1993 (d) Murrah Building, Oklahoma City, USA designed to the American Concrete Institute code ACI 318-7. A transfer beam destruction promoted a progressive collapse: VBIED attack 1995 [from Cormie et al. [2009]]

this circumstance as is the case in the UK Building Regulations, or whether the accidental loadcase is derived from the actual loads due to a specific threat as is the case in some more recent guidance (notably the UFC criteria, UFC4-023-03 [2008]). If prescriptive, the magnitude of the accidental loadcase also varies but is generally based on the 34 kPa adopted in the UK codes.” The reader is directed to the detailed discussion of these approaches provided in Chapter 10 of Cormie et al. [2009]. 339

Figure 17. (a) shards from failed annealed glass (b) dice-like fragments from failed tempered glass (from Cormie et al. [2009])

4.3.5.4 Robust façades The approaches to achieving buildings of robust construction have been outlined above from which it is apparent that modern buildings, designed to meet the requirements of current design codes, are unlikely to suffer collapse (or even unrepairable damage) from the effects of an explosion, particularly if coupled with the techniques for stand-off provision outlined above. However, the effects of building façade failure discussed above indicate that preventing blast from entering a building is of paramount importance in both protecting building occupants and maintaining the overall functionality of the building by preserving the integrity of ventilation and computing systems etc. All buildings have façades which have glazed windows to a greater or lesser extent; those of modern design and construction often have façades whose areas are predominantly glazed. The design of an appropriately robust glazing system is of paramount importance in keeping the blast out of the building. The choice of glazing material and its framing system is key to the efficacy of a building’s glazing. The two commonest types of glass have severe limitations as blast resistive materials. Annealed (‘float’ or ‘plain’) glass fails by breaking into sharp shards which are likely to travel at high velocity into the building. Toughened glass (obtained by re-heating annealed glass to a plastic state, and then cooling it in a controlled manner) has a permanent compressive stresses at its surface which results in it being from four to six times stronger than annealed glass. However, on failure under blast loading, high velocity ‘dice-like’ fragments could be projected into the building. These types of glazing could have their blast-resistance enhanced by the retrospective application of Anti-Shatter Film (ASF) with or without Bomb Blast Net curtains (BBNC), but these should be regarded as only palliative measures that would not be incorporated into a new building. The sharp shards produced by failing annealed glass and the dice-like fragments from tempered glass are shown in Figure 17(a) and (b) respectively. Best blast-resistance is obtained from laminated glass which is a composite comprising alternating layers of glass and a tough, flexible interlayer (usually polyvinyl butyral [pvb]) which bonds the glass sheets together. Under a blast load, the glass cracks but fragments remain bonded to the plastic interlayer and the production of high velocity hazardous shards is eliminated. Though such a pane could be installed in a normal frame, better performance can be obtained by fixing it in a frame that will retain the pane, with the frame itself being strongly fixed to the building frame. An example of the failure of a laminated glazing system is shown in Figure 18 where it is seen that, though the pane has undergone considerable distortion, it has not been ruptured and is wholly retained in its frame meaning that blast has not entered the building. For details on the design and performance of blast-resistant glazing systems, the reader is referred to Cormie et al. [2009] where, in Chapter 9 ‘Design of Glazing’, the more than thirty years of research into the effects of blast on glass carried out in the UK and the US is summarised. 340

Figure 18. Severely cracked laminated panes with pvb interlayer stretched but not torn, completely retained in frames. Blast has been excluded from the building’s interior (from Cormie et al. [2009])

The chapter provides information on the different types of glazing and their behaviour under blast loading and the different levels of blast enhancement that can be achieved including the requirements of suitable blast-resistant framing systems. The design of laminated glazing systems for blast resistance is addressed with two numerical examples and there is information about edge reaction forces and the way glazing hazard is classified. It is worth noting that the work summarised in Chapter 9 of Cormie et al. [2009] contributed to the development of the document by the International Organisation for Standardisation [2007] relating to blast resistant glazing and that, using results of the extensive UK testing program reported in the Glazing Hazard Guide [1997] together with US data, the program HAZL [2004] was developed. This piece of software performs a single degree of freedom analysis to calculate the response to blast loading of different glazing systems (including monolithic glass or plastic windows, laminated windows, insulated glass units and windows retrofitted with ASF) together with a prediction of fragment trajectory. 4.3.6 CONCLUSIONS This paper has sought to provide a summary of current knowledge and understanding of the factors that are important in the development of blast loading assessment and mitigation in the context of the protection of constructions in an urban environment. It is clear that an understanding of the threat and the loading that it can generate is of primary importance and the paper reviews the methods available for such load prediction. Two approaches to mitigating the effect of blast loading can be identified. The first is based on providing as great a stand-off distance as possible between the explosion and the building: techniques for developing both ‘real and ‘virtual’ stand-off are discussed. The second approach is to strengthen the building’s fabric by ensuring (i) that is of robust construction and will not suffer disproportionate collapse if subject to a blast load and (ii) it has a façade that will not readily be breached and will keep blast from entering the building. Application of these approaches will reduce the level of building damage and increase the safety of the building’s occupants. REFERENCES Baker W.E., Cox P.A., Westine P.S., Kulesz J.J., Strehlow R.A., 1983, Explosion hazards and evaluation, New York: Elsevier. BECV4, 2000, www.globalsecurity.org/military/library/report/2000/BECV4.xls (Accessed 24/02/10). ConWep: Conventional weapons effects program. v2.1.0.8, 2002, pdc.usace.army.mil/software/conwep (Accessed 16/4/10). Cormie D., Mays G.C., Smith P.D. (eds), 2009, Blast effects on buildings [2nd Edn], London: Thomas Telford. Department of Communities and Local Government, Building Regulations 2000. Approved Document A – Structure. Part A3 – Disproportionate Collapse (2004 edition including 2004 amendments), 2004, London: DCLG.

341

Glazing Hazard Guide (Restricted), 1997, Security Facilities Executive Special Services Group – Explosion Protection, Cabinet Office. HAZL – Window fragment hazard level analysis software, 2004, pdc.usace.army.mil/software/hazl/ [Accessed 22/03/10] International Organization for Standardization. Glass in building – Explosion-resistant security glazing –Test and classification for arena air-blast loading. ISO 16933, 2007, Geneva: International Organization for Standardization. Kingery C.N., Bulmash G., 1984, Airblast Parameters from TNT Spherical Air Burst and Hemispherical Surface Burst, Report ARBL-TR-02555, U.S. Army BRL, Aberdeen Proving Ground, MD. Ministry of Housing & Local Government., 1968, Report of the inquiry into the collapse of flats at Ronan Point, Canning Town. London: HMSO. Remennikov A.M., 2002, Blast resistant consulting: a new challenge for structural engineers. Australian Journal of Structural Engineering Vol 4(2), pp. 121–134. Rose T.A., Smith P.D., Mays G.C., 1998, Protection of structures against airblast using barriers of limited robustness. Institution of Civil Engineers Structures and Buildings Journal V128. May, pp. 167–176. Smith P.D., Hetherington J.G., 1994, Blast and ballistic loading of structures Oxford: Butterworth-Heinemann. Smith P.D., Whalen G.P., Feng L.J., Rose T.A., 2001Blast loading on buildings from explosions in city streets Proceedings of the Institution of Civil Engineers Structures and Buildings Journal V146 (1), February, pp. 47–55. Smith P.D., Rose T.A., 2006, Blast wave propagation in city streets – an overview Progress in Structural Engineering and Materials, V8(1), Jan/Mar, pp. 16–28. Smith P.D., Rose T.A., Brittle M.A., 2005, Analysis of a generic cityscape using an adaptive mesh CFD code Proceedings of the 12th International Symposium on Interaction of the Effects of Munitions with Structures, New Orleans, USA, 13th–16th September 2005. Smith P.D., 2010, Blast walls for structural protection against high explosive threats: a review. International Journal of Protective Structures V1, No 1, March, pp. 67–84. UFC3-340-01 Design and analysis of hardened structures to conventional weapons effects, 2002, US Army Corps of Engineers, Naval Facilities Engineering Command, Air Force Civil Engineer Support Agency. Defense Special Weapons Agency, Washington DC, June 2002. UFC 3-340-02 Design of structures to resist the effects of accidental explosions. 2008, US Army Corps of Engineers, Naval Facilities Engineering Command, Air Force Civil Engineer Support Agency. Defense Special Weapons Agency, Washington DC, Dec 2008. UFC4-023-03.Design of buildings to resist progressive collapse,. 2008, U.S. Department of Defense, Unified Facilities Criteria Final draft. Department of Defense, Washington, DC, 23 June 2008. US Embassy, United Kingdom,2010, www.usembassy.org.uk/new_embassy/new_embassy5.html (Accessed 10/03/10).

4.3 APPENDIX COST ACTION C26 RESEARCH INTO THE PROTECTION OF STRUCTURES AGAINST IMPACT AND EXPLOSION COST Action C26 contributors have been engaged on various studies relating to the protection of structures against impact and explosion. These include the work of Casadei & Agneloni [1] concerned with structural strengthening using fibre reinforced polymer and polyurea layers and Sendova et al. [2] who investigated the potential blast resistance of a seismically strengthened building. Quek et al. [3] studied the performance of cementitious panels for use in construction when subjected to ballistic impact while Cadoni et al. [4] considered the behaviour of cementitious materials under high dynamic loading. [1] Casadei P., Agneloni E., Elastic systems for dynamic retrofitting (ESDR) of structures. [2] Sendova V., Jekic G., Tashkov L., Effectiveness of seismic engineering strengthening of monuments for their blast resistance. [3] Quek S.T., LinV.W.J.,Lee S.C.,Maalej M., Numerical study of functionally-graded cementitious panels subjcetd to small projectile impact. [4] Cadoni E., Caverzan A., di Prisco M., Behaviour of High Performance Fibre Reinforced Cementitious Composites under high dynamic loading and fire for safe tunnels.

342

Urban Habitat Constructions under Catastrophic Events (Final Report) – Mazzolani (Ed). © 2010 Taylor & Francis Group, London, ISBN 978-0-415-60686-8

4.4 Mitigation options for natural hazards, with a special focus on volcanic eruptions M. Indirli ENEA, Bologna, Italy

E. Nigro Department of Structural Engineering, University of Naples “Federico II”, Naples, Italy

L. Kouris Department of Civil Engineering, Aristotle University of Thessaloniki, Greece

F. Romanelli Department of Geosciences, University of Trieste, Italy

G. Zuccaro PLINIVS Centre and University of Naples “Federico II”, Naples, Italy

4.4.1 INTRODUCTION The main risk assessment procedures (hazard identification, hazard profile, combination of hazards scenarios, inventory assets, estimate losses, mitigation options) have been already discussed, from a general point of view, in the Section II.4. 4.4.2 THE EFFECT OF WATER 4.4.2.1 Description Floods concern more than 65 million people per year in the world; they cause the major economical damages and are responsible for about 60% of the fatalities due to natural catastrophes in the world. River floods River floods are one of the main hazards encountered by people living in the whole European Union. Since these floods can take numerous forms (such as flash floods, estuarine floods, mud floods, etc.), almost all kinds of landscapes can be impacted. In the past decades, an increase in terms of frequency and of importance has been noticed, related to numerous causes: modification of land occupation, erratic river banks management, and the effects of climate changes on storm events frequency. Coastal floods Highly energetic wave regimes, some negative effects from coastal interventions, littoral occupation, exterior interventions in harbours, the weakening of river sediment supply, the generalized sea level rise and other effects of climate changes can be pointed out as the main causes of the increasing number of constructions exposed to waves. Overtopping and flooding are being more frequent events on the coastal zones, jeopardizing buildings and infrastructures. Erosion Most of the European coastal states are affected by coastal erosion, some of them in spite of the coastal protection works done. In addition, another 4 700 km have become artificially stabilized. 343

The main causes for coastal erosion are the generalized sea level rise, caused by climate change, some negative effects from coastal interventions, littoral occupation, exterior interventions in harbours, and river sediment supply reduction. Tsunamis Tsunamis are series of waves created by the fast displacement of huge volumes of water (an ocean for instance) strongly and rapidly affected by a natural phenomenon at a huge scale. Generally, they can be initiated by earthquakes, submarine volcanic eruptions or landslides (seabed slides) for example, but not by strong winds whose impulsion is not short and strong enough. The effects of a tsunami can be classified as insignificant to catastrophic regarding coastal population and constructions. The energy of the phenomenon is sufficient to project any kinds of objects found on its path (ships for instance but also any kind of debris) and sometimes far from the coast. Their combination is powerful enough to shear weak brittle houses at their base or to submerge and to create bending actions on rather high constructions depending of the wave height. Details can be find in: Coelho et al., 2004, 2008 and 2010; Rossetto et al., 2010. 4.4.2.2 Mitigation actions against water effects The European Parliament presented a Directive (2007/60/CE) in 23 October 2007, related with the evaluation and management of flood risks. The document intends to reduce the risk of damaging consequences associated to floods in human being health and lives, environment, cultural patrimony, economic activities and infra-structures (Directive, 2007). In the USA, the Federal Emergency Management Agency recently (June 2008) published guidelines for the design of structures for vertical evacuation from tsunami (FEMA, 2008). In the UK, a new tsunami generator is being built that will be capable of generating a complete tsunami within a physical model (Allsop et al., 2008). One of the usually preferred solutions to solve erosion problems is beach artificial nourishment. This can represents a very expensive solution, when the sedimentary deficit is very high, and there is not any sand deposit availability to such high values. However, it is essential to try to mitigate coastal erosion processes in specific locations. At the moment, the so-called hard coastal defences are indispensable to protect some of the existing settlements, but should be foreseen an adequate plan of monitoring of the existent coastal defence structures. Regarding the political point of view, it is crucial to regulate urban seafront extension. In some cases, the policy options of managed realignment – identify a new line of defence and re-settle the populations in the hinterland – have to be considered. The solution for coastal erosion problems must pass through a compromise between the passive acceptance of erosion, some beach nourishment and coastal intervention for urban front protection. With regards to tsunami and coastal floods, increasing numbers of coastal properties or assets are exposed to wave attack due to changes to land use and increases of sea levels (and/or wave action). Improved management of anthropogenic impacts may help reach goals such as the reduction of human lives losses, reduction of damage in structures and coastal buildings, preservation of natural environments, increase in evacuation capacities, location of new structures and buildings out of danger areas, and relocation of existing structures and buildings. With the advent of new techniques such as Geographical Information System (GIS) and Remote Sensing (RS) for data processing and classifying coastlines, it is needed to consider the incorporation of a huge amount of data into a global open-ended and interrelated coastal classification system, in order to evaluate inundation maps and their impact towards the natural and built environment. Another aspect is related to the role of existing ecosystems (coral reef, seagrass bed, mangrove forests, beaches and dunes, beach forest and other dense forest) and their influence on the intensity of flood damage, as “buffers” for normal waves, storm waves and tsunami. However, the benefits of these “buffers” is strongly debated. For instance, based on data from the 2004 Indian Ocean disaster, Kathiresan and Rajendran, (2005), Liu et al. (2005), Marris (2005), Morton et al. (2006), and Usha et al. (2009) observed that intact sand dunes, rock platforms, mangrove forests, coral reefs and barrier islands all offered enhanced protection against wave impact, flooding and scour. Feagin et al. (2010), throw doubt on some of these findings and call for controlled physical experiments to be carried out to investigate the effects of vegetation on coastal defense from storm surges and tsunami (Rossetto et al. 2010). 344

In any case, the scientific community agree on highlights gaps in knowledge and the necessity of further research. The main mitigation strategies against flooding are the following: – mapping of the flood prone area; – land use control, meaning that no major development should be allowed in the areas subjected to flooding; – construction of strong engineered structures to withstand flood forces; – flood control in order to reduce flood damages, including flood reduction, diversion and proofing. Approaches to limit disruption and damage from flooding have changed significantly in recent years. However, this approach needs integrating: enhanced defences and warning systems with improved understanding of the river system and better governance, emergency planning and disaster management actions. Because the risk is increasing in most of the urban areas due to their fast growing extension involving larger zones, it is necessary to provide efficient models of calculation in order to define different likely scenarios to help in choosing the best strategies. In each country, many institutions are working in that goal to create realistic rescue plans. Flood risk is generally highly localized and as a result, difficult to quantify. Existing maps should be updated, taking into account the effects of climate changes and recent urbanization. Land use plans with identified flood hazard areas, inundation areas, evacuation plans and rehabilitation areas are required in all the regions in the world where likely floods exist. In the industrialized countries, this work is more or less operational and dynamic models allow the development of flood hazards to be evaluated and the safety measures established. Nevertheless, as an example, the consequences of the Hurricane Katrina on New Orleans in August 2005 have shown that, even if the US which can be considered as a modern country, the predictive system and the protected measures were not sufficient, that the levees were not designed to be safe enough and that the rescue organisation was not well prepared in front of such a disaster. In low level countries, the sporadic urbanisation, the unsustainable constructions and the inadequate drainage systems are increasingly causing flood disasters. Integrated and holistic plans for larger water basin areas are strongly needed. As the climate evolution seems to be difficult to be predicted, it becomes imperative for the countries to increase their efforts in monitoring and preparing emergency models. Flood modelling software can help to simulate various scenarios and take proactive steps. It is possible to consider which zones or streets risk getting flooded first, the sewer conducts that will overflow first and plan the investments needed to prevent them. Wastewater software can be used to manage sewerage effectively. To avoid streaming and to define realistic limits of permeability in urban areas by creating water storage possibilities, as green spaces or buffer zones for instance. The elevation of the sea level and the effects of glacier melting have to be taken into account in the modelling. In the significant case is given by the hurricane Katrina, an updated code has been provided (FEMA, 2006a-b); the main recommendations (both for flood and wind) include the adoption of updated building codes (IBC, 2006; IRC, 2006; NFPA 5000, 2006), incorporating flood load (ASCE, 2006a) and flood-resistant construction standards (ASCE, 2006b), with particular regard to foundations.

4.4.3 THE EFFECT OF SNOW 4.4.3.1 Description 4.4.3.1.1 Exceptional snow loads Extreme snow is a natural action resulting from heavy snow falls in areas where snow is usual environmental load or an action resulting from any snow loads in regions not normally exposed to snow falls. Repeated snow events that do not have time to melt and the rain that saturates the snow, which greatly increase its weight, can accumulate and significantly surpass the roof design’s live load and can causes a roof structure to fail. Snow covers on roofs are susceptible to drift action, which leads to removal of snow from some areas and an accumulation in others, and can bring to the extreme design states of snow loads. The collapse of roofs due to heavy snow accumulation may be considered as a catastrophic event and the risk-based approach may be utilized for safeguarding constructions against extreme snow actions. 345

For altitudes smaller than 1500 m, exceptional snow loads are specified in EN1991-1-3:2003, and the code is based on the assumption of a return period equal to 50 years. Exceptional snow loads are considered as accidental loads (EN, 1991a-b). 4.4.3.1.2 Avalanches Avalanches are one of the infrequent actions not taken into account in the Eurocodes. It is possible to identify two main types of avalanches (Givry and Perfettini, 2004) depending on the state of the snow: dry snow (or powder avalanche) and wet snow avalanche. The difference between these two kinds of phenomena defines their mode of failure, their way of displacement down the slope and their relative power. Being slower, wet snow avalanches appear to be less dangerous for humans than dry avalanches, but regarding construction it is the opposite. Nevertheless, if the robustness of the construction can be strongly affected by wet snow avalanches, the openings are affected by dry snow avalanches due to its related high pressure. 4.4.3.2 Mitigation actions against snow effects The design of building structures is largely based on statistics of extreme or catastrophic events which are usually utilized by applying the theory of extremes. Extreme snow loading accounts for several roof collapses each year. Lightweight roof structures, especially long span flat roofs of shopping centers, sport and concert halls, stadiums, railway and bus stations, metal dome roofs of tanks and so on, are the most frequent types of constructions collapsed during recent years (Zuranski, 2007, Pavlov and Vostrov, 2005). Snow loads on roofs depend on different climatic variables (the amount and type of snowfall, the specific gravity and other snow properties, wind, air temperature, amount of sunshine, etc.), on roof variables (shape, thermal properties, etc.), on site exposure and surrounding environment variables. Calculation of extreme snow loads is largely based on statistics of extreme events which are conventionally utilized by the theory of extremes. Because this method may give very misleading results, purely empirical fit to the observed extremes is recommended (Wolinski, 2007 a-b) Building structures may be considered as structural systems, i.e. bounded groups of interrelated interdependent or interacting elements forming an entity that achieves a defined objectives. Therefore, the approach based on the generic system characteristics such as exposure, robustness and vulnerability may be utilized for design and assessment of concrete structures (Steward & Melchers, 1997; Faber et al., 2006). An exposure is related to any event with the potential to cause damage to the structure (loads, corrosion, errors or other disturbances). Robustness is considered to be a measure of the degree to which the specified or unpredictable perturbations influence performance of a structure and is characterized by means of the risk associated with all indirect consequences of its failure or collapse. Vulnerability of a structure is defined as the measure of extend to which changes would harm a structure and characterizes the direct risk associated with its damage or failure. New methods, mainly the purely empirical method of determining extreme snow loads for structural design should be tried in order to compare with the extreme value theory via Gumbel model which is widely used in building codes to estimate snow loads. The approach to design and assessment of the lightweight roof structures subjected to infrequent loading conditions (heavy snow storms, unusual patterns of snow cover, combined snow, wind and ice actions, etc.) based on assessment of risk characteristics should be introduced as a helpful supplement to design and assessment procedures. Other details are given in: Ellingwood and O’Rurke 1985; Gooch, 2002; Mihashi et al., 1989. Avalanche mitigation can be done trough active and passive defences: a) Active defences: they aim to reduce the avalanche occurrences; regarding this aspect, it is for instance possible: to activate small explosions in order to create little avalanches to clean the slopes; to prevent skiing on slopes when a risk exists that an avalanche could be initiated by a skier. b) Passive defences: some actions can be provided to reduce the avalanche impacts; for instance, it can be efficient to build avalanches barriers and walls, to optimise the design and the orientation of constructions or to organise mountain rescue. A major possible development is relevant to the prediction of the occurrence of the avalanche phenomenon (Burlet, 1999) at the Ski resort scale. It is based on the idea that mechanical models 346

can help avalanches predicting as mechanical models do for land slopes. The development of this approach needs the knowledge of the snow layers (type, thickness and mechanical properties) at each time where the risk assessment is needed. Another possible development concerns the avalanche dynamics and more precisely its impact on works (Ma, 2008) at the slope scale. It aims to estimate the avalanche action on an avalanche wall used to protect a road. It is based on the idea that an avalanche can be modelled as a granular flowing. Details can be find in: Muzeau et al., 2007; Platzer, 2006; Talon et al., 2010a-b.

4.4.4 THE EFFECT OF WIND 4.4.4.1 Description 4.4.4.1.1 Extreme winds Cyclones, hurricanes, tornados or typhoons are extreme winds whose dynamic action leads generally to severe damages on constructions. To be initiated, tropical cyclones need certain thermodynamic conditions to be respected above a large mass of warm water. Therefore, they form above seas or oceans. They are named hurricanes in the Atlantic Ocean and typhoons in the Pacific Ocean. Tornados are initiated above the earth during a severe storm when special thermodynamic conditions are found between huge cloudy masses and winds. 4.4.4.1.2 Cyclones Three types of cyclonic perturbations are commonly defined: tropical depressions, tropical storms and tropical cyclones (Chaboud, 2003). A tropical cyclone is constituted by an eye at its centre, which is a relatively warm and calm zone, surrounded by an area about 16–80 km wide in which the strongest thunderstorms and winds circulate around it. Up to now, the extreme wind speed due to a tropical cyclone is estimated to be equal to 305 km/h. To be initiated and sustained, tropical cyclones need large unstable volumes of warm water (more than 26◦ C over 60 m in depth) so, their strength decreases over land because of the lack of water. 4.4.4.1.3 Tornados Much smaller than a tropical cyclone regarding its influence diameter, a tornado is a violently rotating column of air starting from the lower part of a cumulonimbus cloud and in contact with the earth. The damages on constructions are generally localised but very important due to the high speed of the rotation. In most cases, a cloud of debris collected on the way, moves around the funnel at its lower part and it contributes to increase the damages. A powerful tornado may extract light constructions from their foundations. Most of the tornadoes create a very localised strong wind whose speed may reach 175 km/h. Their lower part is generally about 100 m and they travel only on a small distance (about 10 km) before they dissipate. Nevertheless, much more powerful tornadoes have been observed: with a wind speed close to 500 km/h, with a base diameter close to 1.5 km and whose way on the ground may be longer than 100 kilometres. 4.4.4.2 Mitigation actions against wind effects The tremendous losses year by year indicate that structural safety and serviceability criteria are not fulfilled. Experience from past disasters must be considered along with advanced wind monitoring related research for safer and more complete design and construction codes. Structural engineers face a new challenge in the new millennium, due to the effect of rapid climatic changes, which influences directly the engineering society which needs to adjust to the unexpected and severe natural catastrophic events. Building codes are supposed to be strict enough to protect buildings from high wind loads. In some cases, excessive damage is noticed during hurricane events in spite of the fact that the recorded wind speeds were lower than the maximum recommended by the code design wind speed (Kareem, 1984). Local and international building codes have been revised several times over the last few decades, but the losses and damages due to wind are still considerably high (ABI, 2003; Munchener Ruck; 2006 and 2007). These facts indicate the necessity of advanced research in the area of extreme wind action and low-rise buildings interaction. The nature of extreme 347

wind phenomena is very complex and researchers must focus on the understanding of such events. Notwithstanding the relevance of these questions, the structural response of low-rise structures subjected to extreme wind loads must be evaluated through model and full-scale studies, in order to be able to conclude about safe and economical design parameters. Other details are given in: Doudak et al., 2005a-c; Holmes, 2001; Islam and Peterson, 2003; Minor, 2005; Simiu and Scanlan, 1996; Witham, 2005; Zisis, 2006; Zizis and Stathopoulos, 2008; Solari, 2007.

4.4.5 THE EFFECT OF LANDSLIDES, ROCKFALLS AND FLOWSLIDES 4.4.5.1 Description Landslide describes a wide variety of processes that result in the downward and outward movement of slope-forming materials including rock, soil, artificial fill, or a combination of these. The materials may move by falling, toppling, sliding, spreading, or flowing. The various types of landslides can be differentiated by the kinds of material involved and the mode of movement. Other classification systems incorporate additional variables, such as the rate of movement and the water, air, or ice content of the landslide material. Although landslides are primarily associated with mountainous regions, they can also occur in areas of generally low relief. In low-relief areas, landslides occur as cut-andfill failures (roadway and building excavations), river bluff failures, lateral spreading landslides, collapse of mine-waste piles (especially coal), and a wide variety of slope failures associated with quarries and open-pit mines. The conventional stability analysis of slopes where sliding is possible along some definable surface is usually preformed by calculating the factor of safety, i.e. by comparing the shearing resistance available along the failure surface. With the shearing stresses imposed on the failure surface. Most analytical methods are based on limit equilibrium, with typical failure forms such as infinite slope or finite slope with planar or curved failure surface considered. However, the recently introduced performance-based approach, emphasis is placed not on whether the slope is stable or unstable, but on the magnitude of deformation after failure. Several techniques are currently available to asses the post-failure velocity and travel distance of the moving mass. The basic model assumes that during the shaking a slope will suffer displacement only when the ground acceleration exceeds a threshold value, the critical acceleration, which can be derived from the static factor of safety of the slope in question. The sliding mass will continue to move until the shaking drops below the critical acceleration. If the cumulative displacement caused by shaking, known as Newmark displacement is sufficient to cause a reduction in the shear strength of the soil or rock mass then a re-calculation of the slope stability is carried out using residual shear strength parameters to establish whether failure occurs. Thus the analysis is bi-linear, allowing for a change in the strength parameters of the slope forming materials based on the deformation of the slope. 4.4.5.2 Mitigation actions against landslides and flowslides 4.4.5.2.1 Landslides as a secondary event of earthquakes and eruptions Landslide and flowslides can be surely considered as one of the most dangerous slope movements, for their capability to produce casualties and remarkable economic damage. Such phenomena are widespread in many countries and they involve different kind of soils, generally in a loose state, which in the post failure stage collapse and rapidly reach the toe of the slope; the initial mobilised mass often increases during its path downslope either by inducing additional slope failure and/or by eroding the stable in place soils. Significant examples of this type of slope movements occurred in several areas of the world, as those periodically occurring in the Campania Region triggered by critical rainfall events. They involve unsaturated pyroclastic soils – originated by the explosive phases of the Somma-Vesuvius volcano – which mantle the limestone and tuffaceous slopes over an area of about 3000 km2 . However, landslides may be also secondary events of earthquakes or eruptions, as several historic data show. For example, one of the most significant effects of the 1994 Northridge, California, earthquake and of the 2001 El Salvador earthquakes was the triggering of thousands of landslides over a broad area. Some of these landslides damaged and destroyed homes and other structures, 348

blocked roads, disrupted pipelines and caused other serious damages. A further type of landslides may be produced by “rapid avalanches” of intimately mixed snow and hot pyroclastic debris during a volcano eruptions, as for the eruptions at Mount St. Helens, Nevado del Ruiz, and Redoubt Volcano between 1982 and 1989. The types of landslides previously described may be events induced by the eventual eruption of the Vesuvius and by the connected seismic motions. In Chapter 3.4 of this Report the possible effects of such kinds of landslides on the urban areas are investigated and mechanical models deduced utilising also hydrodynamic concepts are introduced; the models are capable to interpret the effects of the landslide impact on the constructions and the collapse mechanisms of various types of structures. Based on the quoted mechanical models and hydrodynamic concepts, the main guidelines of two technical codes devoted respectively to the rebuilding and reparation of constructions in areas with high landslide risk may be applied in order to obtain a risk mitigation for the effects of landslides. 4.4.5.2.2 A technical code on the structural design in urban areas with high debris-flow-risk The serious damages produced in the urban areas by the hydrogeological disaster occurred in Campania put in evidence into the civil community the fundamental problem of a suitable territory use, with particular care to the possibility of reparation or rebuilding of the constructions damaged or destroyed by landslides. Thanks to the significant work of the Operating Unit stated by the Civil Protection Agency at the University of Salerno, the immediate answer was the definition of a “red line”, which perimeter individuated the areas with high debris-flow-risk on the base of their geomorphologic characteristics. Based on the interpretation of the surveyed building damages and mechanical characteristics of the dynamic impact of the debris flows on the constructions, the main guidelines of a technical code devoted to the rebuilding of constructions in areas with high landslide risk is reported in the following. Among other things, the criteria to evaluate the design actions due to debris flow impact and the types of buildings and structural systems more efficient to sustain the same actions are suggested. The mechanical models described in Chapter 3.4 have represented a worthwhile contribution to the definition of the technical code, approved by the Campania Regional Government (Italy), in order to allow suitable operations of rebuilding of in urban areas with high landslide risk. Being socially unacceptable to delocalise the population hit by the above described disaster, Campania Region Authorities have provided the areas having a significant debris-flow-risk with a specific code, partially based on the results summarized in Chapter 3.4, to allow to rebuild constructions capable to resist to debris flows phenomena. The code in argument refers to the areas of the cities, where there is a residual risk of debris flows, also taking into account the effects of the foreseen protection works able to reduce or eliminate the debris-flows-risk. These protection works are mainly: • naturalistic engineering works, to prevent irregular concentration of water flows or local soil collapse, which could constitute the primer of more general landslides; • hydraulic works, to regularize the flows and to prevent the detachment of unstable pyroclastic layers along the valleys; • dams, to contain the debris flows phenomena and to protect the built-up area. 4.4.5.2.3 Main aspects of the Technical Code The Technical Code on rebuilding (Campania Region Government, 2001) is based on the fundamental remark that it results practically impossible that ordinary built constructions resist to the hydrodynamic action due to the debris flow impact. Therefore, when it is impossible to protect the construction with specific works or to deflect the flow, the defence strategy consists in reducing the impact surfaces adopting a construction typology with isolated columns at the ground floor. In this case it is necessary to quantify the impact actions, mainly deriving from the impact velocity. With this purpose, the territory of interest has been subdivided in areas with different expected velocity: A) High velocity expected zone: B) Medium velocity expected zone: C) Low velocity expected zone:

10 m/sec 7 m/sec 5 m/sec 349

The consequent horizontal actions are generally very strong compared with ordinary seismic or wind actions; in fact, orientation values of the debris flows action are the following: A) High risk zones: hydrodynamic pressure: B) Medium risk zones: hydrodynamic pressure: C) Low risk zones: hydrodynamic pressure:

150.0 kN/m2 73.5 kN/m2 37.5 kN/m2

The main contents of the Technical Code can be summarised in the following points. 4.4.5.2.4 Evaluation of the debris flow impact actions The actions produced by the impact of the debris flows on obstacles depend on velocity, density, height and direction of the stream, on shape and dimensions of the impacted obstacle, on the presence of masses transported by the flow. The resultant thrust is expressed as the sum of the dynamic and static thrusts:

being, with reference to walls of rectangular shape (see Figures 1a,b): • • • • • • • • •

ρc v θ g ho = hc + d, b hc

• d

debris flow density (ρc = 1500 kg/m3 , unless more accurate evaluations) debris flow velocity (in ms−1 ): angle of the flow direction with respect to the axis normal to the impacted surface gravity acceleration (9.81 ms−2 ) height of the impacted surface (in m) width of the impacted surface (in m) height of the debris flow (hc = 3.00 m in all the zones, unless more accurate evaluations) depth of the laying plane of the surface with respect to the external land plane (in m)

In the case of fixed obstacle completely immersed into the flow, as well as r.c. columns at the ground floor, the resultant dynamic thrust on the obstacle holds:

Cf being the shape coefficient of the obstacle, variable in the range (0.5 ÷ 1.1). 4.4.5.2.5 Safety checks The above defined actions due to debris flow are considered “design values”, so that they have to be utilised in the ultimate limit state checks adding to the other permanent and variable design actions, unless wind and seismic ones. If the check of the structure is performed according to the “allowable stress method”, the debris flow actions have to be reduced by dividing for the load partial safety factor γc,f = 1.5. 4.4.5.2.6 Constructive provisions In order to reduce the actions and the effects of the debris flows on structures, the following provisions hold: • the structure has to be designed with reference to the actions previously defined; • the impact surfaces of the construction have to be reduced adopting appropriate shapes and planimetric distributions; • the impact surfaces at the ground floor have to be reduced, distinguishing the cases of tangential action and orthogonal action of the debris flows: – in the first case, buildings with external walls at the ground floor are allowed, making them able to sustain the hydrostatic pressure; 350

Figure 1a.

Dynamic thrust on a walls of rectangular shape.

Figure 1b.

Static thrust on a walls of rectangular shape.

– in the second case, buildings with external walls at the ground floor must be avoided, realising framed structures with isolated columns at ground floor able to sustain both hydrostatic and hydrodynamic actions. 4.4.5.2.7 Other topics Further topics treated in the code provisions concern: • the design of direct and indirect foundations of buildings, taking into account the horizontal actions due to debris flow impact; 351

• the design of the structure, reducing the impact surface of the single members by means of appropriate cross-section shape and planimetric position; • the design of structures, avoiding distance between columns less than 5.0 meters in order to prevent the formation of accidental obstructions to the free flow. 4.4.5.3 Technical Code on repairing of existing buildings The Technical Code on repairing of existing buildings (Campania Region Government, 2002) assumes that these buildings are not able to sustain the orthogonal dynamic impact of debris flows. Therefore, the repair of damaged buildings located in areas with high debris flow risk is only allowed in the case of tangential impact of the flow against the building, thanks to the local orography or protection systems. In such a case the flow action can be assumed as equivalent to hydrostatic pressure due to fluid with high mass density (ρc = 1500 kg/m3 ). Finally, it has to be remarked that areas with high debris flow risk are inhabited from centuries or millennia, so that the dislocation of the people is not socially acceptable and it appears simultaneously necessary to reduce the risk level. This remarks led to the introduction of specific technical codes devoted to rebuilding and repairing of the constructions located in some areas of the Campania Region (Italy) interested by debris flow phenomena in the last years. These codes do not substitute the planned works devoted to the reduction of the area risk, but have the aim to reduce the vulnerability of the constructions, being aware that, notwithstanding the defence works, a residual risk remains, which should be eliminated only with an un-acceptable impact on the environment. 4.4.5.4 Final remarks The general methodology, the criteria and the failure models described in the Chapter 3.4 and in the present one with reference to the case of debris flow impact on buildings may be extended to other cases of interest for the Risk Assessment for Catastrophic Scenarios in Urban Areas. For example, the phenomenon of pyroclastic flows is quite similar to the described one and requires the knowledge of the same structural data concerning buildings in order to assess their vulnerability. Also the Technical Codes, summarily illustrated in the previous paragraphs, may represent a reference in order to write guidelines concerning the refurbishment or retroffiting of existing buildings to resist to catastrophic events. Details can be found in Cascini and Ferlisi, 2003; Cascini et al., 2003; Faella and Nigro, 2003a-c.

4.4.6 THE EFFECT OF EARTHQUAKE This topic is widely discussed in other parts of this final report. It is necessary to stress that the adoption of revised set of rules by several Government Authorities is a step already achieved in many earthquake-prone countries, especially after the Northridge (1994) and Kobe (1995) seismic events, but also following the primary school collapse of San Giuliano di Puglia, Molise Region 2002, Italy.

4.4.7 THE EFFECT OF VOLCANIC ACTIONS Several other sections are dedicated to the volcanic question. This part, after a brief description of the effects, focuses in particular the mitigation options. 4.4.7.1 Description The effects of a volcanic eruption on built environment have been investigated in the last 15 years, defining a comprehensive framework of studies, surveys and simulations that include all the different eruptive phenomena and their possible impacts on existing buildings and infrastructure. Nevertheless, in order to define a design methodology for the technological retrofit of structures in volcanic risk-prone areas, a different approach is needed, starting from the basic consideration 352

that the cumulative effects given by a complex eruptive scenario (such as a Sub-Plinian or Plinian eruption) produce extremely variable impacts on constructions, depending by the specific time history of the event and by the building typologies and their level of vulnerability. This peculiar approach has been recently formalized in order to evaluate the impact of a Sub-Plinian eruption in Vesuvius area (Zuccaro et al., 2008), through the development of a numeric model for the definition of impact scenarios. The study of building technologies for the mitigation of volcanic risk is strongly connected to the model. It has been developed as a part of SPeeD project, funded by Italian Civil Protection Department. The parameters assumed for the design of specific technical solutions come from the elaboration of data by SPeeD and Exploris project, and the proposed retrofit technologies are directly referred to conventional building types inside Vesuvius and Campi Flegrei area. The first part of the study analyses the mitigation strategies referred to each single eruptive phenomenon, in order to define different technical options for the mitigation. The second part proposes a comparative methodology for the assessment and the choice between possible solutions based on conventional and innovative technologies, including the use of advanced materials, which must meet specific requirements of safety, reliability, durability and integrability. A specific attention is given to building site issues related to logistics and on-site operations, taking into account the need to make widespread interventions in risk-prone areas and preferring solutions – with equal performance levels – characterized by quickness and easiness of installation or by low operating costs.

4.4.7.2 Mitigation actions against volcanic effects on structures 4.4.7.2.1 EQ – Earthquake The seismic events that characterize an eruptive phenomenon can be generally considered of low to medium intensity. Nevertheless, the cumulative damage caused by the sequence of earthquakes in various stages of the eruption produces a progressive increase in the level of expected damage. According to the sequence of phenomena characterizing the eruptive event, more conditions can occur and raise the damage caused by the earthquake. In particular, the ash fall creates a progressive overload on the roofs, and even when it doesn’t result in a partial collapse of the floor, it brings to an increase of reactive mass of the building, thus modifying the response to seismic action. The building types with high vulnerability, with particular reference to masonry structures, would then suffer more damage than for a single event comparable to the maximum intensity expected in case of a Sub-Plinian eruption. Generally speaking, considering the high seismic vulnerability levels and the construction density in Vesuvius area, cost-effective mitigation measures should be provided. It is possible to choose cheap and reliable technical solutions (such as iron chains in masonry buildings, the insertion of infill panels or resistant elements in soft floors of reinforced concrete buildings), but also to adopt, in case of seismic reinforcement, specific solutions able to respond effectively also to other volcanic phenomena, such as pyroclastic flows or ash fall. In this context, one solution is the construction of pitched roofs by overlapping light structures in CFS (Cold Formed Steel). This allows to chain vertical structures by increasing the resistance to seismic actions (box behaviour) and simultaneously prevent the deposit of ashes and the structural risks related to overloading of the roof, also in consideration of a possible earthquake following the ash fall phase. At the same time, should be avoided the employ of widely used reinforcement systems not satisfying the conditions of volcanic risk, such as FRP (Fiber Reinforced Polymers) in proximal areas, whose effectiveness is seriously reduced by the possible impact of pyroclastic flows. In fact, the high temperatures produced could affect the polymer matrix, whose physical and mechanical properties degrade in range above 60–80◦ C, with the consequent failure of the system caused by the loss of adhesion of the reinforcement to the walls. In this cases alternative technologies should be adopted, compatible with the environmental conditions related to a volcanic event, such as FRCM (Fiber Reinforced Cementitious Matrix) systems, able to withstand high temperatures while preserving the mechanical properties. Global mitigation strategies related to seismic risk in case of a volcanic event may include planning for widespread interventions, defining the areas that require priority actions, such as the building curtains facing the main transport routes and escape routes identified by the Civil Protection Emergency Plan, in order to ensure safe evacuation routes during unrest phase, characterized by increased seismic intensity. 353

Figure 2. Technical solution for the protection of openings.

4.4.7.2.2 PF – Pyroclastic Flows Pyroclastic flows can produce high damages to the built environment in areas near to the vent. Although they would have a limited action range, the effects can be critical because of the combination of mechanical impact and thermal stress on the vertical surfaces of buildings. The main damages come from the impact on openings, particularly vulnerable to pyroclastic flows. In these cases, although not resulting a static failure of the building, a fire risk is associated with the flow passage inside the building following the crash of the openings. In the case of Vesuvius and Campi Flegrei, pyroclastic flows can cause lateral pressure impact within a range of 0.5 and 10 kPa, and thermal stresses ranging between 150 and 450◦ C. In Campi Flegrei the probable location of the vent, near to densely populated areas (including the west area of Naples), the impact of pyroclastic flows would be particularly serious, while in the case of Vesuvius is expected a decay of the initial power due to the distance of the built areas from the vent. Mitigation strategies mainly concern the reinforcement of infill panels in r.c. buildings and measures for the protection of openings (Figure 2). When reinforcing infill panels, the goal is to increase the impact resistance while withstanding the high temperatures produced by the flow. Currently used techniques for the seismic reinforcement of infill panels are generally effective to prevent them from breaking due to pyroclastic flow, however, as noticed above, the employ of currently used technologies that are particularly sensitive to temperature should be avoided. In the absence of specific constraints to envelope system modification, the goal of increasing infill panels impact resistance may be achieved by overlaying existing facades with coatings made of advanced materials offering high thermal and mechanical performances in very low thickness. It is the case of UHPC (Ultra High Performance Concrete) components, which can be cast in very large panels and show high durability and resistance to aggressive environment. These operations allow also to obtain additional performances, such as the increase of shear strength in the plane, where the panel is placed within the structural grid, or the increase of thermal resistance, where combined with a layer of insulation or with a ventilated facade system. The use of low thickness UHPC panels may also be suitable for the construction of temporary and removable systems to protect archaeological areas and sites of historical and artistic interest subject to the risk of pyroclastic flows. Protection of openings is an essential mitigation measure in relation to pyroclastic flows, as it allows minimizing the risk of fire related to penetration of the flow inside the buildings. At the same time the technical solutions provided should be able to withstand the mechanical stresses related to the pressure of the flow itself, but also to the potential presence of debris that can impact as “bullets” on openings surface. Borrowing technologies used in tropical areas for hurricanes protection it is possible to define different solutions, made with removable components or integrated into the 354

Table 1. Vulnerability of common roofing typologies. Vuln. classes A_rf B_rf

C1_rf C2_rf D_rf

Roofing type

LoadkPa

Weak pitched wooden roof Standard wooden flat roof Flat floor with steel beams and brick vaults Sap floors Flat floor with steel beams and hollow bricks R.C flat slab (more than 20 year old) R.C flat slab (less than 20 year old) Last generation R.C. flat slab Last generation R.R. pitched slab Last generation steel pitched roof

Collapse prob. %

2,0 3,0

50 50

5,0

60

7,0

51

12,0

50

shutting systems. In the first case, it is possible to overlay steel or kevlar sheet to existing openings, anchored along the external perimeter. Protection systems integrated into the shutting systems, unlike the removable panels, are not always able to assure an effective response to the impact of the flow, but are suitable for medium ranges of temperature and pressure or for short exposition time. It is also possible to apply special protective films on glass surfaces that can provide protection from fire and explosion. Fire safety shutters, steel or aluminum, associate the heat resistance with adequate mechanical strength. In some cases, a combination of protective films and special shutters should be provided, in order to reach the required levels of temperature resistance and mechanical strength. 4.4.7.2.3 AF – Ash fall Ash fall is one of the eruptive phenomena with greater risk for existing buildings and infrastructure, as the expected impact involves (with different levels of intensity) a very large area, which definition is strictly linked to the direction and intensity of the wind, as well as to the type of eruption. In the case of Vesuvius and Campi Flegrei, the scenarios show an increase of roof loads (Table 1) due to ash fall between 1000–3000 kg/sqm inside the red zone and between to 300–400 kg/sqm for distances up to 30 km from the vent. Different types of damage may also occur in distal areas (more than 100 km from the vent), where the ash deposits are not likely to cause structural problems to buildings, but still could affect transportation networks and HVAC systems (ashes infiltration in filters and ducts). In case of eruption of Campi Flegrei, the direction with higher risk is the whole urban center of Naples, where the population is more than twice the area of the villages around Vesuvius. Ash deposit on roads and transport networks can cause considerable damages especially in proximal areas, causing localized or extended interruptions with direct effects on emergency management. Mitigation strategies, beside the need to develop an operational plan for the removal of ash on roofs and transport networks, mainly concern the repairing and reinforcement of roofing systems in order to increase the load carrying capacity (Figure 3). Pitched roofs with wooden or steel structure, reducing the deposits of ashes, would be at risk only in proximal areas where the surface of the cover present disconnections or missing parts. In this case, given the adequate inherent fire resistance of commonly used coating materials (typically clay tiles or panels of steel sheet) is enough to replace the missing elements in order to prevent the passage of hot ashes under the roof covering. In case of flat roofs it is possible to identify two main types of intervention: the reinforcement of the roof slab in order to increase the resistance according to the expected overload, or the realization of a sloped roof over the existing one. In the first case, it is necessary to define the characteristic flexural strength of different types of existing roofs in areas at risk (concrete and bricks, steel or wooden beams and hollow bricks or brick vaults, “Sap” floors, etc.), thus determining the capacity to withstand to overloads produced by ash. It is then possible to apply conventional technologies, such as integration of reinforced concrete slabs placed on the existing floors and connected to existing beams, or innovative solutions, 355

Figure 3. Technical solution for the mitigation of ash fall impact on roofs through the employ of CFS structures (Alborelli, 2009).

including for instance the use of FRP (Fiber Reinforced Polymers) and FRCM (Fiber Reinforced Cementitious Matrix) systems for reinforcement of beams and joists. The main advantages of such interventions include the possibility of not modifying the existing roofing system. In the second case a very effective solution is to build truss or lattice structure in CFS (Cold Formed Steel) on top of the existing roof, in order to create a sloped surface. The mechanical properties and lightness of CFS structures allow the realization of a strong roofing system without a high overload on the underlying structure. The coating can be made of steel sheet, with the possibility of providing additional layers in order to offer additional benefits to the intervention of structural retrofit, such as the insertion of insulation or micro-ventilation system for energy conservation, or the integration of photovoltaic thin film for the production of electricity. Such actions may be also connected with housing refurbishment programs, allowing for instance the increase of building volume for intervention of volcanic and seismic mitigation. The realization of lightweight structures for protection from ash fall may be an appropriate solution not only for buildings but also for the several areas of historic and artistic interest (such as Pompeii, Herculaneum, Oplonti, Stabiae, etc.), which might be seriously compromised after an eruption of Vesuvius. In these areas, however, the mitigation may be invasive in terms of visual impact, and it is possible to develop provisional removable shingles. An alternative to steel roofing is the realization of UHPC (Ultra High Performance Concrete) shells, characterized by very high mechanical properties, durability, resistance to high temperatures and fire, with very low thickness required (up to 2 cm for spans of 5 m), offering effective and innovative technical solutions in terms of aesthetics and design. 4.4.7.2.4 LH – Lahars The lahars are a relevant risk factor for buildings and structures in volcanic areas. The same phenomenon may have specific characteristics depending on some variables. Damage to buildings caused by lahars can be connected to different factors. Hydrostatic and dynamic strength determine the amount of lateral forces that can bring to failure and collapse of technical elements such as openings and cladding. The density and velocity of the flow determines the magnitude of dynamic forces, while hydrostatic forces depend on the height and composition of the flow. Minor mudslides can cause abrasions of the external finishing of buildings and damage to surfaces and furnishings in case of penetration of the flow in the interior. Local effects may be caused by the transport of medium and large debris, rocks, but also uprooted trees, motorbikes and cars that can act as missiles on buildings exposed. Depending on the magnitude of the phenomenon and orographic conditions of the site, buildings of medium-low height can be buried by lahar. Further damage can be caused to structural parts of both masonry and reinforced concrete buildings, causing even serious cracks and damages, with structural failure involving foundations, due to erosion and soil liquefaction. Structural and non-structural metal elements can also be seriously damaged by the acidity of the flow. 356

The response of structures and buildings technical elements to the action of lateral forces produced by lahars depends mainly on construction type and materials employed, as well as specific characteristics such as size in plan and elevation, number, size and position of openings, spatial distribution and presence of protective elements around the building able to divert the flow, etc. Generally speaking, structures, infill panels and ground floor openings are the technical elements most at risk in case of lahars. The reinforcement of these elements yet does not guarantee the survival of the building in case of direct impact with mudslide and debris, especially in the case of compact urban areas, where a “tunnel effect” can increase speed and height of the flow after the passage inside particularly narrow roads. For this reason the most effective mitigation strategies are related to environmental engineering interventions, to be made in risk prone areas and designed to contain or divert lahars. Measures such as retention basins, alternative artificial canals, high-strength reinforced concrete containing structures, may be appropriate solutions to mitigate risk from lahars, reducing the entity of the phenomenon in residential areas and increasing the probability of survival of the buildings.

4.4.8 DESIGN APPROACH FOR THE MITIGATION OF VOLCANIC RISK 4.4.8.1 General approach Preceding paragraphs illustrate mitigation strategies in relation to each volcanic phenomenology (EQ, PF, AF, LH). However, to ensure the effectiveness of technical options and feasibility of specific interventions, it is necessary to define a design approach that takes into account all the factors involved and the complexity resulting from the combination of effects in relation to the eruptive scenario. This approach allows a proper assessment of the effectiveness of the intervention in relation to the possible impact of eruption on the entire construction, taking into account the several factors that determine the vulnerability of a building. In particular, it has to be assumed that the survival of a building in case of eruption starts from increasing ability to respond to seismic actions, which are the first critical factor to face, also considering the expected time history. In relation to this objective, it has to be noticed that an economically sustainable intervention of seismic improvement (typically, the maximum cost should not exceed 60% of the cost of building reconstruction) determines a two classes increase of the building seismic vulnerability (i.e. A to C or B to D) and raises the chances of survival to earthquakes precursors of the eruption (between V-IX, EMS ’98). However, to fully evaluate the effectiveness of such interventions in relation to the overall expected scenario, the proposed seismic reinforcement should be verified also for its contribution to resistance to roofing overload (AF) and lateral pressure (PF-LH), in order to take any corrective measure or provide additional mitigation solutions. For example, it is possible to assume that an intervention of seismic reinforcement which aims to increase the ductility of the structural system (i.e. by junction bonding of a R.C. building with FRP or FRCM systems), may be less suitable for mitigation from pyroclastic flow compared to an intervention able to increase the stiffness of the structural system, given the non-cyclic lateral action of the flow. At the same time, seismic reinforcement of horizontal structures involving roofing system, should be verified considering the overloads due to ash fall, possibly focusing on solutions that provide the superposition of a pitched roof able to enchain the perimeter walls. A similar consideration can be made about the resistance of buildings to lateral pressures produced by pyroclastic flows and lahars. Once verified the resistance to different ranges of pressure provided by the structural elements in relation to the building type (see Table 2), this should be taken as a benchmark to define the design resistance of non-structural elements (openings and infill panels). An interesting case study is related to the evaluation of mitigation scenarios for ash fall. A cost-benefit analysis was carried out involving 11 of the 59 municipalities in the areas surrounding Vesuvius involved in the AF areas. The aim is to “ensafe” about 50% of the buildings through the realization of a pitched roof over the existing flat one through CFS (Cold Formed Steel) technologies. This solution can significantly reduce the number of victims for roofs collapse, assuming that the people occupying unsafe buildings can find a shelter in buildings subject to mitigation action. 357

Table 2. Resistance to lateral pressure of structural elements for building type. Critical pressure kPa

Technical element Wooden seasonal structures 3–4 floors weak masonry buildings with deformable floors 4+ floors weak or strong masonry building 6+ floors r.c. buildings Weak tuff walls (thickness ≤40 cm, span>4 m) 4–6 floors r.c. buildings Non aseismic weak r.c. buildings 1–3 floors r.c. buildings Medium strength tufo walls (thickness ≥40 cm, span>4 m) Non aseismic strong r.c. buildings 1–2 floors weak masonry buildings with deformable floors 3–4 floors masonry buildings with rigid floors 1–2 floors masonry buildings with rigid floors

3,5 3,5–5 4–5 4–7,5 4,5–6 4,5–8 6,5–9 7–9 11–18 14–19

The intervention is provided only for the Municipalities where the vulnerable roofs areas exceed 50% and the collapsed roof areas exceed 5%. The study has shown that compared with a total investment of around 182 million Euros is possible to reduce of about 35% the number of roof collapsed after the ash fall. Assessment of the intervention strategy in terms of cost-effectiveness takes into account many different parameters. The priority is to define one or more “mitigation scenarios”, according to a preliminary evaluation that takes into account issues such as technical, economic and social consequences related to any given scenario. Depending on the different scenarios it is possible to predict the level of effectiveness of the envisaged strategies, considering also other parameters such time and cost of rehabilitation.

4.4.8.2 Comparison and choice of technical options through indicators As stated above, in order to assess the effectiveness of mitigation actions, it is necessary a comprehensive analysis of technological options, considering the performances expressed by the employed materials and the response to primary requirements of safety, reliability, durability and integrability. Beside these considerations, it is necessary to identify additional selection criteria for different technical options. The study considers six key indicators: quick installation; storability; lightness; cost; preservation of constructive and architectural features; multifunctionality (ability of the technical solution to respond to different phenomena). The technical sheets produced report a synthetic assessment based on these six indicators, expressing a qualitative judgment that highlights the level of response given by any single technical solution for each of the parameters identified. Figure 4 shows one of the over 30 sheets of technical solutions for the mitigation of volcanic risk on buildings developed in SPeeD project, based on data and scenarios defined by the numeric model. The study is focused in particular on conventional or innovative structural reinforcement technologies for masonry and r.c. buildings; on innovative solutions for roofing and infill panels protection (including the use of advanced materials); on permanent and temporary intervention for the protection of the openings. The sheets are classified in four categories: SE – Interventions on elevation structures SV – Interventions on vertical surfaces SO – Interventions on horizontal structures AP – Interventions on openings 358

Figure 4. Technical sheet of mitigation solution developed in SPeeD project.

4.4.9 FINAL REMARKS The mitigation of volcanic risk on buildings and infrastructure can significantly reduce the expected damage after an eruptive event. Even the impacts of high destructive type of eruptions, such as SubPlinian, can be strongly reduced by the application of one or more mitigation measures, responding to the different phenomena involved. It is therefore necessary to start from a comprehensive knowledge of the construction types available in risk-prone areas, providing specific interventions that take into account the cumulative effects given by the expected time history of the event. Hence, an effective design approach aims to put in relation technological features of existing buildings, parameters and data from probable scenarios, opportunities given by mixing together conventional technologies and advanced materials. Furthermore, considering the economical, political and social “weight” of the strategies for the mitigation of volcanic risk in densely populated areas, a valid evaluation method of the effectiveness of the proposed solutions can give scientific support to strategic choices and emergency plans. Therefore, tools for assessment and comparison between different solutions and “mitigation scenarios” are needed, trying to put together the different factors involved, such as economical and social sustainability, cultural and historical value, implication on emergency plans and on post-eruption rehabilitation and reconstruction interventions. 359

The main reference is Zuccaro and Leone, 2010. Other details are given in: Acker, 2004; Bellomo and D’Ambrosio, 2010; Macedonio et al., 2010; UNDRO, 1991; Spence et al., 2004.

4.4.10 MITIGATION ACTIONS AGAINST VOLCANIC EFFECTS: FINAL GENERAL CONSIDERATIONS Mount Vesuvius is an active volcano surrounded by a densely populated area, now quiescent. However, computer simulations predict that there is a high probability of at least a subplinian eruption occurring in this century. Even if the COST Action C26 (COST, 2006) is not addressed to the issue of the evacuation plan managed by the Italian Civil Protection (Protezione Civile, 2010), and the study has been restricted to the modelling of loads acting on structures and the corresponding construction response, some general considerations are needed at the end of this Section. Due to the difficulty to interpret the premonitory signals coming from the volcano immediately before its eruption, it is not really sure that the emergency plan (foreseeing the evacuation of about 600,000 people away from the eighteen municipalities of the Vesuvius “red zone”) could be managed with a sufficient amount of time (measurable in terms of days or weeks), without a heavy risk of wrong alert; on the contrary, it is necessary to take into account that the period available for the emergency operations can be limited only to a few amount of hours. In this case, the organization of a regular evacuation procedure can be rather difficult, especially in a very crowded region. For the above said reasons, a rigorous policy of prevention (based on a multidisciplinary approach) should be considered indispensable, hoping that Vesuvius give us enough time to promote it, in order to minimize the impact of a great eruption on population and environment. The principal objective is to identify safe areas where people can live, in a secure cohabitation with the volcano. Therefore, the main goals of future researches, to be planned or improved, are the following (Dobran, 2006 and 2007): a) the development of accurate volcanic models (physical and mathematical), assessing future eruption scenarios and their consequences on the surrounding territory; b) the assessment of the global vulnerability and potential damage induced by the volcano on the entire system (population, built environment, infrastructure, etc.); c) the production of volcanic risk-reduction guidelines for communities and local/national governments; d) the promotion of a socio-cultural methodology enhancing consciousness and auto-regulation of the territory. Of course, the most important result of these studies should be the identification of alternative people settlements and the reorganization of the entire infrastructural network in the whole region, relieving the current situation to more manageable scenarios. In order to avoid a potential immense tragedy, this tremendous effort needs a multidisciplinary cooperation of the scientific community as a whole in the next decade, together with a strong institutional support in Italy and Europe.

REFERENCES ABI, 2003. The Vulnerability of UK Property to Windstorm damage, Association of British Insurers, London, UK. Acker, P., Behloul, M. 2004. Ductal Technology: A Large Spectrum of Properties, A Wide Range of Applications. Ultra High Performance Concrete; Proc. Intern. Symp., Kassel, 13–15 September 2004: 11–23. Alborelli, E. 2009. Messa in opera di misure di mitigazione per edifici soggetti a precipitazione di materiale piroclastico. Tecnologie sostenibili per la riqualificazione delle coperture, degree tesis, Università degli Studi di Napoli Federico II, Facoltà di Architettura. Allsop, W., Rossetto, T., Richardson, S. and Robinson, D.I., 2008. A Unique Tsunami Generator for Physical Modelling of Violent Flows and their Impact. Proceedings of the 14th World Conference in Earthquake Engineering, Beijing, China. In Publication. ASCE, 2006a. ASCE 7-05. American Society of Civil Engineers, flood load.

360

ASCE, 2006b. ASCE 24-05. American Society of Civil Engineers, flood-resistant construction standards. Bellomo, M. D’Ambrosio V. (ed.) 2010. Fibrorinforzati in architettura. Le tecnologie FRP e FRCM nel recupero delle strutture in c.a., Napoli: CLEAN. Burlet, J.L., Bacconnet, C., Boissier, D., Gourvès, R., 1999. Modelling of the snow cover variability with a Gaussian stochastic field. 8th ICASP, pp 869, 875 – Sydney, 12,15 décembre 1999 – R.E. Melchers et M. G. Stewart Ed., Balkema. COST C26, 2006. COST, European COoperation in the field of Scientific and Technical research, Transport and Urban Development, COST Action C26: “Urban Habitat Constructions Under Catastrophic Events”. 2006–2010. Campania Region Government, 2001. Code Provisions to Rebuild Private Residential Buildings Damaged by Debris Flow Occurred on 5th and 6th May 1998 in the Cities of Sarno, Siano, Bracigliano, Quindici and San Felice a Cancello. Ordinance n. 1991 of 9 November 2001. Napoli (Italy). Campania Region Government, 2002. Code Provisions to Repair Private Residential Buildings Damaged by Debris Flow Occurred on 5th and 6th May 1998 in the Cities of Sarno, Siano, Bracigliano, Quindici and San Felice a Cancello. Ordinance n. 2631 of 24 December 2002. Napoli (Italy). Cascini L., Ferlisi S., 2003. Occurrence and Consequences of Flowslides: a Case Study, in L. Picarelli (Ed.), Proceedings of Int. Conf. on Fast Slope Movements-Prediction and Prevention for Risk Mitigation (FSM2003), Napoli (Italy), 11–13 May 2003. Patron Ed.: Bologna. Cascini L., Sorbino G., Cuomo S., 2003. Modelling of flowslides triggering in pyroclastic solis, in L. Picarelli (Ed.), Proceedings of Int. Conf. on Fast Slope Movements-Prediction and Prevention for Risk Mitigation (FSM2003), Napoli (Italy), 11–13 May 2003. Patron Ed.: Bologna. Chaboud, R. 2003. Cyclones. Encyclopédia Universalis. DVD-ROM, France. Coelho, C, Granjo, M.J. & Silva, C.S. 2008. Map of Costal Zone Vulnerabilities to Wave Actions – Application to Aveiro District (Portugal), (in press), International Symposium on Integrated Coastal Zone Management, 11–14 June 2007, Arendal, Norway. Coelho, C., Pinto, F.T., Gomes, F.V. & Barbosa, J.P. 2004. Coastal Evolution and Coastal Works in the Southern Part of Aveiro Lagoon Inlet, Portugal, Coastal Engineering 2004, Proceedings of the 29th International Conference on Coastal Engineering, Lisbon, Portugal. Directive, 2007. Directive 2007/60/EC of the European Parliament and of the Council of the 23rd October 2007 on the assessment and management of flood risks. Official Journal of the European Union, L 288/27. Dobran F. 2006. VESUVIUS. Education, security and prosperity. Development in volcanology. Flavio Dobran Eds. Elsevier. Dobran F. 2007. Urban Habitat Constructions Around Vesuvius. Environmental Risk and Engineering Challenges. Proc. of COST Action C26 Seminar on Urban Habitat Constructions Under Catastrophic Events, Prague, 30–31 March 2007. Doudak G., Ghallagher A., Kasal B., McClure G., Mohammad M., Smith I., Stathopoulos T., Zisis I., 2005a. Towards wind load paths on wood buildings, EACWE4, The Fourth European & African Conference on Wind Engineering, Prague, Paper #285. Doudak G., McClure G., Smith I., Hu L. and Stathopoulos T., 2005b. Monitoring Structural Response of a Wooden Light-Frame Industrial Shed Building to Environmental Loads, Journal of Structural Engineering, ASCE, Volume 131, No. 5, pp. 794–805. Doudak G., 2005c. Field Determination and Modelling of Load Paths in Wood Light-Frame Structures, PhD Thesis, Department of Civil Engineering and Applied Mechanics, McGill University, Montreal, Canada. Ellingwood, B. & O’Rurke M., 1985. Probabilistic Models of Snow Loads on Structures, Structural Safety 2, pp. 291–299. Faber M., et. al., 2006. On the qualification of robustness of structures. Proceedings of OMAE2006, 25th Offshore Mechanics and Arctic Engineering Conference, Hamburg, Germany, June 4–9, 2006, (OMAEE 2006-92095). Faella C., Nigro E., 2003a. Dynamic Impact of the Debris Flows on the Constructions during the Hydrogeological Disaster in Campania-1998: Description and Analysis of the Damages, in L. Picarelli (Ed.), Proceedings of Int. Conf. on Fast Slope Movements-Prediction and Prevention for Risk Mitigation (FSM2003), Napoli (Italy), 11–13 May 2003. Patron Ed.: Bologna. Faella C., Nigro E., 2003b. Dynamic Impact of the Debris Flows on the Constructions during the Hydrogeological Disaster in Campania-1998: Failure Mechanical Models and Evaluation of the Impact Velocity, in L. Picarelli (Ed.), Proceedings of Int. Conf. on Fast Slope Movements-Prediction and Prevention for Risk Mitigation (FSM2003), Napoli (Italy), 11–13 May 2003. Patron Ed.: Bologna. Faella C., Nigro E., 2003c. A Hypothesis of Technical Code on the Structural Design and Repairing in Urban Areas with High Debris Flow Risk, in L. Picarelli (Ed.), Proceedings of Int. Conf. on Fast Slope MovementsPrediction and Prevention for Risk Mitigation (FSM2003), Napoli (Italy), 11–13 May 2003. Patron Ed.: Bologna.

361

Feagin, RA., Mukherjee, N., Shanker, K., Baird, AH., Cinner, J., Kerr, AM., Koedam, N., Sridhar, A., Arthur, R., Jayatissa, LP., Lo Seen, D., Menon, M., Rodriguez, S., Shamsuddoha, M. and Dahdouh-Guebas, F. (2010). Shelter from the storm? Use and misuse of coastal vegetation bioshields for managing natural disasters. Conservation Letters, 3, 1–11. FEMA 2006a. FEMA 549. Hurricane Katrina in the Gulf Coast, Mitigation Assessment Team Report, Building Performance Observations, Recommendations, and Technical Guidance, FEMA (Federal Emergency Management Agency) Report, 2006. FEMA 2006b. FEMA 550. Recommended Residential Construction for the Gulf Coast: Building on Strong and Safe Foundations, FEMA (Federal Emergency Management Agency) Report, 2006. FEMA, 2008. Guidelines for Design of Structures for Vertical Evacuation from Tsunamis, FEMA P646 Report, Federal Emergency management Agency, Washington D.C. Givry, M., Perfettini, P. 2004. Construire en montagne – La prise en compte du risque d’avalanche. Ministère de l’écologie et du développement durable. Ministère de l’équipement, des transports, du logement, du tourisme et de la mer. Available on http://www.avalanches.fr. Gooch C., 2002. Winter snow loads. Cornell University. Holmes J.D., 2001. Wind Loading of Structures, Spon Press, New York. IBC, 2006. International Building Code, USA. IRC, 2006. International Residential Code, USA. Islam T. and Peterson R. E., Climatology of landfalling tropical cyclones in Bangladesh 1877–2003, Natural Hazards, DOI 10.1007/s11069-008-9252-4. Kareem A., 1984. Wind Speed – Damage Correlation in Hurricane Alicia, Hurricane Alicia: One Year Later, Specialty Conference, ASCE – EMD, Galveston, Texas, pp. 81–93 Kathiresan, K. and Rajendran, N., 2005. Coastal mangrove forests mitigated tsunami. Estuarine and Coastal Shelf Science, 65, 601–606. Liu, P.L.F.; Lynett, P.; Fernando, H.; Jaffe, B.E.; Fritz, H.; Higman, B.; Morton, R.; Goff, J., and Synolakis, C., 2005. Observations by the International Tsunami Survey Team in Sri Lanka. Science, 308(5728), 1595–1595. Ma, Y., 2008. Analyse de l’effet d’une avalanche de neige sur un ouvrage de protection type galerie paravalanche. Expérimentation, modélisation. PhD Thesis, University of Savoie, September 2008. Macedonio, G. Costa, A. Folch, A. 2008. Ash fallout scenarios at Vesuvius: numerical simulations and hazard assessment implications. Journal of Volcanology and Geothermal Research 178 (3): 366–377. Marris, E., 2005. Tsunami damage was enhanced by coral theft. Nature, 436(7054), 1071–1071. Mihashi H., Takahashi T., Izumi M., 1989. Wind effects on snow loads on roofs. Proceedings of 1st. Int. Conference on Snow Engineering, CRREL Special report 86-6, pp. 168–167. Minor J., 2005. Lessons Learned from Failures of the Building Envelope inWindstorms, Journal ofArchitectural Engineering, ASCE, Volume 11, No. 1, pp. 10–13. Morton, R.A., Goff, J.R., and Nichol, S.L., 2006. Impacts of the 2004 Indian Ocean tsunami on the southwest coasts of Sri Lanka. Coastal Sediments Conference (New Orleans, Louisiana), Volume 2, Paper 20, 14p. Munchener Ruck, 2006. Knowledge Series, Hurricanes-More intensive, more frequent, more expensive: Insurance in a time of changing risks. Munich Re Group. Munchener Ruck, 2007. Topics Geo, Natural catastrophes 2007, Analyses, assessments, positions. Munich Re Group. Muzeau, J.P., Bouchaïr, A., Sesov, V., Coelho, C., 2007. Identification & classification of exposure events. Exceptional or infrequent event scenarios. Workshop COST C26: Urban Habitat Constructions under Catastrophic Events, Prague, March 2007. NFPA, 2006. NFPA 5000. Building Construction and Safety Code, published by the National Fire Protection Association. Pavlov A. & Vostrov V., 2005. Extreme snow loads on dome roofs of buildings and structures. IABSE Symposium “Structures and Extreme Events”, Lisbon, Sept. 14–17, 2005, IABSE Reports, Volume 90, pp. 198–199 (full version on CD-ROM). Platzer, K., 2006. Experimental investigation of avalanche forces acting on snow sheds. PhD.Thesis, University of Natural Resources and Applied Life Sciences, Vienna, Austria. Protezione Civile, 2010. Rischio vulcanico. Vesuvio. Piano Nazionale di emergenza. http://www. protezionecivile.it/minisite/index.php?dir_pk=250&cms_pk=1440&n_page=3. Rossetto T., Coelho C., Carlier J.P., Allsop W., Lloyd T.O., 2010a. Tsunami impact evaluation for coastal areas. Proceedings of COST Action C26 Final International Conference on Urban habitat construction under catastrophic events, Naples, 16–18 September 2010. Simiu E. and Scanlan R. H., 1996. Wind Effects on Structures: Fundamentals and Applications to Design, John Wiley & Sons, New York.

362

Solari G., 2007. Lessons from catastrophic events in the evolution of bridge and wind engineering. Proc. of COST Action C26 Seminar on Urban Habitat Constructions Under Catastrophic Events, Prague, 30–31 March 2007. Spence, R.J.S. Zuccaro, G. Petrazzuoli, S. Baxter, P.J. 2004. Resistance of buildings to pyroclastic flows: analytical and experimental studies and their applicaton to Vesuvius. Natural Hazards Review 5 (48): 48–59. Steward M.G. & Melchers R. E., 1997. Probabilistic risk assessment of engineering systems. London, Chapman Hall. Talon A., Muzeau J.P., Carlier J. Ph., 2010a. Avalanche actions on constructions.Proceedings of COST Action C26 Final International Conference on Urban habitat construction under catastrophic events, Naples, 16–18 September 2010. Talon A., Muzeau J.P., Carlier J. Ph., 2010a. Avalanche risk assessment in populated areas.Proceedings of COST Action C26 Final International Conference onUrban habitat construction under catastrophic events, Naples, 16–18 September 2010. UNDRO, 1991. Office of the United Nations Disaster Relief Organization (UNDRO). Mitigating Natural Disasters, Phenomena, Effects and Options, New York: United Nations. Usha, T., Murthy, M., Reddy, N., Murty, T. 2009. Vulnerability Assessment of Car Nicobar to Tsunami Hazard Using Numerical Model. Science of Tsunami Hazards, Vol. 28, No. 1, pp. 15–34. Witham C.S., 2005. Volcanic disasters and incidents: A new database, Journal of Volcanology and Geothermal Research, Volume 148, Issues 3–4, Pages 191–233. Wolinski Sz., 2007a. Framework for assessment of actions on structures due to catastrophic events. In monograph Research problems of building structures. Vol. III, Polish Academy of Sciences, Bialystok, pp. 51–58 (in Polish). Wolinski Sz., 2007b. Identification and classification of response of urban habitat constructions subjected to extraordinary loading conditions. Proceedings of the COST ACTION C26 Workshop, Prague, 30–31 March, 2007, Czech Technical University in Prague, pp. 368–371. Zisis I., 2006. Structural Monitoring and Wind Tunnel Studies of a Low Wooden Building, MASc Thesis, Civil, Building and Environmental Engineering Department, Concordia University, Montreal, Canada. Zisis, I. and Stathopoulos, T., 2008. Wind-induced cladding and structural loads on a low wooden building, Journal of Structural Engineering, ASCE, accepted May 2008. Zuccaro, G. Cacace, F. Spence, R.J.S. Baxter, P.J. 2008. Impact of explosive eruption scenarios at Vesuvius. Journal of Volcanology and Geothermal Research 178 (3): 416–453. Zuccaro G., Leone M.F., 2010. Building technologies for the mitigation of volcanic risk. Proceedings of COST Action C26 Final International Conference on Urban habitat construction under catastrophic events, Naples, 16–18 September 2010. Zuranski J., 2007. Failures of roofs under snow loads in Poland. Proceedings of the XXIII Conference“Structural Failures 2007”, 23–26 May 2007, Szczecin, Poland, pp. 357–364 (in Polish).

363

Urban Habitat Constructions under Catastrophic Events (Final Report) – Mazzolani (Ed). © 2010 Taylor & Francis Group, London, ISBN 978-0-415-60686-8

4.5 Avalanche risk assessment in populated areas Aurélie Talon & Jean-Pierre Muzeau LaMI – Polytech’Clermont-Ferrand (CUST), Blaise Pascal University Clermont-Ferrand, France

4.5.1 INTRODUCTION Each year, avalanches cause very important economical, sociological and environmental impacts; then, it is of main interest to develop efficient risk mitigation actions. This mitigation efficiency depends on the accurate knowledge of the studied system. Different scales of risk analysis may be considered (§ 4.2); the study scale will determinate the studied consequences, the associated responsibilities and the action possibilities. The risk is the combination of hazard occurrence and issue damages. In the avalanche risk context, hazard is the avalanche itself and issues can be persons, structures, infrastructures, communications, environment and economy. A risk analysis compounds five phases: – – – – – –

Definition of the avalanche risk analysis scale (§ 4.3) Leading of system analysis (§ 4.4) Identification of risk scenarios (§ 4.5) Quantification of avalanche hazard (§ 4.6) Quantification of avalanche consequences (§ 4.7) Proposition of mitigation techniques regarding the quantified risk scenarios and the possible actions (§ 4.8)

4.5.2 SCALE OF AVALANCHE RISK ANALYSIS The definition of risk analysis scale is very important as it conditions the risk analysis result and the efficiency of the study process. Indeed, a risk analysis at the mountain scale will generate an important loss of time if it is expected to know the damage risk of a resort building regarding a potential avalanche hazard that may progress in a particular slope; in this case, the risk analysis should be made at the slope scale. To simplify, it is possible to relate the object and the scales of this kind of studies: – at the mountain scale: global environmental impacts, – at the massif scale: local environmental impacts, – at the slope scale: economical and sociological impacts on persons, structures, infrastructures, communications, etc. – at the snowy coat: behavioural knowledge of the avalanche departure, flow and deposit. Considering a risk analysis at the civil engineering point of view, the scales of main interests are the slope scale and the snowy coat scale. 4.5.3 SYSTEM ANALYSIS The goal of the system analysis is to understand and to analytically model the behaviour of a case study in terms of elements that constitute this case study and the functions ensured by these elements. 365

The three sub-phases of a system analysis are: a) to define the system limits and granularity, b) to lead a structural analysis, c) to provide a functional analysis. The definition of system limits aims at focusing the risk analysis on a spatially and temporally limited domain. The granularity definition goal is to take into account the efficient scale of description of the study: not too fine in order to avoid a too long study and not too global in order to avoid missing relevant risk scenarios. For instance, it is possible to study the snowy coat breaking (due to snow overload) during a snowfall episode. The structural analysis aims at describing the elements composing a case study. For example, at the massif scale, a mountain may be decomposed into workable massifs and into unworkable massifs. At the slope scale, a slope may be composed of: – – – – – – – – – – –

an accumulation basin, a gorge, a cone dejection, fauna and flora, skiers, ski lifts, station personal, habitants, communication network, buildings, emergency equipment.

The functional analysis objective is to describe in terms of functions the behaviour of the elements defined during the structural analysis. For instance, considering the previous massif scale, the function of workable massifs needs “to be sure for humans” and “to be profitable” and the function of unworkable massif needs “not to make damage on workable massifs”. Considering a slope scale, the main functions are: – For skiable slopes: • to carry out the snowy coat, • to be secure, • to have a good quality, • to be supervised, • to be groomed each day, • to have enough snow. – For non skiable slopes: • to carry out the snowy coat, • to be secure, – For buildings: • to resist, • to shelter persons. – For persons: • to be informed, • to be secure, • to be responsible. – For infrastructures: • to resist, • to transport flows (persons, fluids, etc.). The system analysis is the essential starting point of efficient risk scenario identification. 366

Figure 1.

Example of a scenario of skier death.

4.5.4 RISK SCENARIOS Risk scenarios are basically the chains of events starting from an avalanche departure that lead to catastrophic damages on issues: persons, structures, infrastructures, communications networks, etc. Beginning with the system analysis several methods are available to build risk scenarios: – Fault Tree Method for avalanche feedback that provides explanation of causes combinations that generate damages (Villemeur, 1988), – Event Tree Method for foreseeing consequences generated by an avalanche event (Zwingelstein, 95), – Failure Mode and Effects Analysis for describing precisely avalanche phenomenon and their chaining (Faucher, 2004), – ... A simplified example of a risk scenario that leads to the death of a skier is presented in the following Figure 1. When all the risk scenarios are identified for a considered case study, the next phases are the quantification of those scenarios: (1) quantification of the occurrence of the departure event, of the avalanche, (2) quantification of the consequences when the avalanche has occurred.

4.5.5 QUANTIFICATION OF AVALANCHE HAZARD Most of the studies carried in the domain of quantification of avalanches aim at collecting knowledge on the snowy coat all over the massif and the slopes in order to efficiently foresee the breaking phenomenon of this snowy coat that provides avalanche departure. Three research axes in the domain of avalanche hazard quantification are presented here: – modelling of spatial snow deposit and flow, – measuring of snowy coat properties, – investigation of occurred avalanche. 4.5.5.1 Modelling of spatial snow deposit and flow This kind of study is leaded at the massif scale. This section presents three software that can be used to characterize the snowy coat at massif scale: the SAFRAN software (Durand et al., 1993), the CROCUS software (Brun et al., 1992) and the MEPRA software (Giraud, 1991). The SAFRAN software provides a situation analysis by: – calculation of meteorological parameters on the massif, – calculation versus altitude, slope and exposition, – description of a situation at a given moment. 367

Figure 2.

Example of penetromeric profile, stratigraphic profile and Pandalp device (Burlet, 2002).

The CROCUS software allows realizing simulation of the snowy coat evolution by: – realizing calculations with physical bases, – integrating modifications undergone by each stratum: • shape and size of grains, • density, • humidity, • ... The MEPRA software provides foresee-aid by: – estimating the stability of the snowy coat, – defining the nature of avalanche risk by altitude stratum and by intensity level. 4.5.5.2 Measuring of snowy coat properties These research tasks are performed at the snowy coat scale. A set of tests and a measuring device have been developed in the context of a partnership between the Blaise Pascal University of Clermont-Ferrand and the Sol Solution Company in order to characterize snowy coat properties: cohesion, unit weight and point resistance. Figure 2 represents survey results (penetrometric profile and stratigraphic profile) and the developed measuring device (Pandalp) that allows the evaluation of point résistance versus depth. Those results provide punctual but very useful information. 4.5.5.3 Investigation of occurred avalanches The avalanche investigation begins at the end of the XIXe century in the French Savoie region, through a permanent investigation on avalanches (EPA). Following an avalanche that killed 39 teenagers at Val d’Isère (February 10, 1970), an inventory map of all avalanche sites has been made on a decision of Minister Council: the location map of avalanche phenomenon (CLPA). The EPA and CLPA are based on the investigation of avalanche sites and the collection of evidences. The EPA is a chronicle of avalanche events on selected sites. The number of sites for which EPA is realized is limited. Often easily observable, those sites have been originally chosen versus the damages generated in forests. Nowadays, the sites are chosen for the human issues and the scientific knowledge in time of avalanches. 4200 EPA sites are available; there are referenced on EPA observation maps. Ground officers of the ONF (Forest National Office) note the characteristics of event in their avalanche notebook for each avalanche event: date, snow cover, departure altitude, stopping altitude, avalanche type. . . All that information is reported on an avalanche advice. This one is sent to the CEMAGREF institute who implements them into a database. More than 70 000 events are available on the www.avalanches.fr website. Figure 3 represents the national map of the EPA disposal, where the green points represent the site under observation or already observed. The location map of avalanche phenomenon (CLPA, see Figure 4) is an informative document drawn at a 1/25000 scale which describes the maximal amplitudes of avalanche phenomenon 368

Figure 3. National map of EPA system (Avalanche, 2008).

Figure 5. Repartition of the average number of fatal accidents and the number of deaths by avalanche type – Period from October 2007 to September 2008 (Avalanche, 2008).

Figure 4.

Example of CPLA (Avalanche, 2008).

Figure 6. Repartition of avalanches accidents by slope downgrade (SLF, 2008).

occurred in the past and observed with precision and certainty. The CLPA essentially indicates the avalanche influence that is to say the maximal extension of known events. The CLPA is upgraded each year. Every ten years, a summary investigation of detailed upgrading is lead. 4.5.5.4 Quantification of avalanche consequences The quantification of avalanche consequences is generally based on statistics of the avalanche events that generated damages. Figure 5 presents an example of statistics performed by the French ANENA organisation and Figure 6 shows another statistics example provided by the Swiss SLF institute. 4.5.6 MITIGATION TECHNIQUES When risk scenarios are identified, quantified (departure occurrence and consequences effects) and classified by order of importance, it becomes possible to reduce the risk probability (prevention action) or the risk gravity (protection action) in order to put a snowy slope into secure conditions. Two categories of construction disposals are available to protect buildings against 369

Figure 7.

Mutual protection by building grouping.

Figure 8.

Principle of non-increasing the risk for the neighbourhood.

Figure 9.

Principle of prevision of an avalanche outlet.

avalanche risk: overall disposals and specific disposals for each construction. The build principles that correspond to overall disposals are: – – – –

building grouping (Figure 7), orientation and shape of buildings, non-increasing of the risk for the neighbourhood (Figure 8), prevision of an avalanche outlet (Figure 9). The building principles that correspond to specific disposals for each construction are:

– – – –

foresee an access and an entrance on the non-exposed facades, design facades without hold-in corner when there are facing the avalanche-prone slopes, no storage of polluting or dangerous product in poorly resistant constructions, foresee an appropriate distribution of the places: the more vulnerable places have to be localized upstream in the avalanche direction. Mainly four types of devices that allow preventive start of avalanches are available:

– The Catex, which is a cable rotating supported by towers that allows to place the explosive above snowy coat (Figure 10), – The Gazex, which is a gas burst (mix of propane and oxygen) (Figure 11), – The “Avalancheur”, which is a pneumatic bowler of explosive arrows (Figure 12), – The “Alvalhex balloon”, which is a device that can generate the departure of an avalanche by explosion, above the snowy coat, of a balloon blown up with a mix of hydrogen and oxygen. This explosion generates a spherical blast wave (Figure 13). 370

Figure 10.

Catex (Givry et al., 2004).

Figure 12.

“Avalancheur” (Givry et al., 2004).

Figure 11.

Cazex (Givry et al., 2004).

Figure 13. Avalhex balloon (Givry et al., 2004).

The protection works are classified into two main categories, depending on their location in the departure zone (active protection) or in the flow or deposit zones (passive protection). In both cases, these actions can be provided permanently (without human intervention) or temporarily (with decision taken). The main works of permanent active protection are: reforestation on seat, wind barrier, snow barrier, buzzard roof, wind transfer, tire racks, wicker racks, and fillets. The main works of permanent active protection are: galleries, stems, deflectors, stopping dike, stakes, and road detector of avalanche. The major part of temporary protections consists in the application of several rules to the population in order to prevent a direct exposure to the avalanche risk: traffic restrictions and regulation. This may be done in order to proceed to an evacuation, or on the contrary to maintain people in a safe location until the end of the critical period. The information is coordinated and spread at two levels: the one of the department (vigilance map of de Météo France) and the one of the ski resort (hoarding, yellow and black flags).

4.5.7 CONCLUSIONS The avalanche risk analysis may be described in several main phases: – definition of the study scale and limits, – determination of the constitutive elements and their function, 371

– identification of the risk scenarios, – quantification of those scenarios in terms of occurrence and consequences, – proposing of mitigation actions. Mitigation actions will be all the more efficient than the risk analysis is well lead. REFERENCES Avalanche. Programmes d’études des avalanches. URL: http://www.avalanches.fr (2008) [On line]. Brun E., David P., Sudul M., Brunot G. A numerical model to simulate snow-cover stratigraphy for operational avalanche forcasting. Journal of Glaciology, IGS, Cambridge, (1992) n◦ 32, pp. 13–22. Burlet J.L. Mécanique de la neige et variabilité – Application à la prévision du risque d’avalanche. Civil Engineering Thesis (2002). Université Blaise Pascal Clermont-Ferrand II. Durand Y., Brun E., Mérindol L., Guyomarc’h G., Lesaffre B., Martin E. A meteorological estimation of relevant parameters of snow. Annals of Glaciology, IGS, Cambridge, (1993) n◦ 18, pp. 65–71. Faucher J. Pratique de l’AMDEC. Paris: DUNOD, (2004), 177 p. Giraud G. MEPRA : modèle expert d’aide à la prévision du risque d’avalanches. Actes du symposium de Chamonix, juin 1991, ANENA (1991) Grenoble, pp. 248–254. Givry M., Perfettini P. Construire en montagne – la prise en compte du risque d’avalanche. Ministère de l’écologie et du développement durable et Ministère de l’équipement, des transports, du logement, du tourisme et de la mer (2004). SLF. URL : http://www.slf.ch (2008) [On line]. Villemeur A. Sûreté de fonctionnement des systèmes industriels. Paris : Eyrolles, (1988) 798 p. Zwingelstein G. Diagnostic des défaillances. Paris : Hermès, (1995) 601 p.

372

Chapter 5: Strategy and guidelines for damage prevention

Urban Habitat Constructions under Catastrophic Events (Final Report) – Mazzolani (Ed). © 2010 Taylor & Francis Group, London, ISBN 978-0-415-60686-8

5.1 Fire design in Europe M. Heinisuo & M. Laasonen Tampere University of Technology, Finland

J. Outinen Rautaruukki Oyj, Finland

5.1.1 INTRODUCTION European countries have started applying Eurocodes to fire design. In the past they used to have separate structural and fire codes, which are now integrated into a single standard system. Earlier, let’s say, fire loads and mechanical loads were defined in different sources, but now both are in given in the EN 1991 series. Performance based fire design was created long ago but is still a new concept in many countries. In Sweden and UK, the countries that have most experience from it, local authorities still have a great influence on which solutions are approved locally especially as concerns fire design. Performance based design is frequently used to design the evacuation and exit routes of buildings and to estimate smoke and toxin propagation. Due to the huge improvements in methods, computing facilities and know-how of engineers, performance based design can now be used to estimate temperatures in fire compartments and load bearing structures. That allows calculating the resistance of structures in fire. The developments in fire design have led to close co-operation between authorities, clients, architects and fire safety, mechanical and structural engineers in actual projects. Co-operation is always important, and performance based design intensifies it in projects. It results in better fire safety of buildings, because the fire scenarios and related risks have to be analysed carefully when using performance based fire design. This study shows the situation with respect to fire Eurocodes and performance based design in 10 European countries presently. A case study that promotes the design of load bearing structures exposed to fire is also included. The fire scenarios and corresponding design fire packages for a large sports centre built in Helsinki, Finland are presented as well as the results of the fire simulations. 5.1.2 EUROPEAN RULES 5.1.2.1 Scope of the study European countries are starting to apply the EN standards for buildings. They include principles and application rules for fire resistance design of buildings which also allow performance based fire design. What is the situation with respect to applying the EN standards for fire resistance design and, especially, performance based design? What kinds of applications of performance based fire design do we have experiences from? These were the main questions to which we sought answers from members of the COST C26 during spring 2008. A short summary of the results follows. The questions and further details can be found in Heinisuo (2009). Answers were received from ten countries (three from UK). Table 1 shows the introduction of EN standards and the countries which answered the questions. 5.1.2.2 Annexes to EN 1991-1-2 The national fire codes of Czech Republic, UK, Finland, Hungary, and Italy cover performance based design. It can be used in Belgium if an exception to the Fire Regulations is approved by 375

Table 1. Introduction of fire ENs in ten European countries.

Belgium Czech Republic Finland France Hungary Italy Poland Portugal Romania UK

1991-1-2

1992-1-2

1993-1-2

1994-1-2

1995-1-2

1996-1-2

1999-1-2

2008 2006 2007

2008 2006 2007

2008 2006 2007

2008 2006 2007

2008 2006 2007

2008 2006 2008

2008 2008 2008

2005 2008 2006 2008 2007 2002

2005 2008 2008 2008 2008 2004

2005 2008 2007 2008 2007 2005

2005 2008 2008 2008 2007 2005

2005 2008 2008 2008 2007 2004

2005 2008 2009 2008 2008 2005

2007 2010 2008 2008 2007

the Minister of the Interior. France allows its partial application with respect to fire resistance and smoke propagation design. UK has used performance based fire design the longest. When the ENs are used for fire design, the essential data for performing calculations is found in the Annexes of Standard EN 1991-1-2 (2002). EN 1991-1-2 has seven annexes (A-G). All the annexes are informative only, meaning that any country can substitute them by their own rules, or restrict their use. The ten countries accept Annexes B (thermal actions for external members – simplified calculation method) and G (configuration factor) as they are. To the others, they have made some modifications. Two countries (Belgium, France) allow usingAnnexA (parametric temperature-time curves) only in the preliminary design stage. In France the use of Annexes C (localised fire) and D (advanced fire models) requires a peer review. Use of Annex E (fire load densities) and Annex C is subject to restrictions. E.g. Finland only allows using Chapter E.4. In France Annex E is banned entirely. Five countries (Belgium, Finland, France, Portugal, and UK) do not allow use of Annex F that deals with equivalent time of fire exposure. In UK almost all the annexes have been substituted by the national PD 6688-1-2:2007 standard. The detailed question dealt with the National Annex for EN 1991-1-2, clause 4.3.1(2) dealing with the representative values of variable actions in fire. Both combination factors, ψ1 and ψ2 , are used in Europe. This means that, for example, the snow load in fire is defined differently in various parts of Europe. In Italy and Romania ψ2 is used for variable actions in fire. France and Portugal (and Spain, Estonia, Slovenia according to the DIFISEK+ project) use ψ1 . In UK ψ2 is used for EQU cases and ψ2 for STR cases. Belgium (and Netherlands and Luxembourg according to the DIFISEK + project) uses ψ2 , except for wind when it uses ψ1 . The Czech Republic uses ψ2 , except for wind and snow where the choice is ψ1 . Finland uses ψ2 for live loads and ψ1 for wind, snow and ice actions. 5.1.2.3 Current projects Performance based fire design has been used in real projects in 8 countries, 10 of which answered the query. Performance based fire design has been used to: – – – – – – – – – –

determine the fire resistance of structures, perform evacuation calculations, perform smoke control, perform risk analyses, optimise the fire protection of structures, study local fires, study external flames, study equivalent times of exposures, demonstrate adequacy of fire fighting provisions, demonstrate functioning of escape routes, 376

– demonstrate the existence of an acceptable standard of safety for complex buildings with large numbers of people and/or large open spaces. There are many examples of different kinds of applications of performance based fire design in Europe. Typically performance based fire design has been used in large projects, but it is now gaining ground also in smaller ones. Typical projects using it include: – – – – – – – – – – – – – – – –

shopping centres, office buildings, airports, hospitals, residential buildings, stadiums, music halls, underground facilities, industrial buildings, historical buildings, high rise buildings, car parks, libraries, churches, monumental buildings, warehouses.

One interesting project involved a cruiser. When performance based fire design is used, approval of the design by the third party may or may not be required. Typically no specific qualification is required of the fire designer, but many countries may require some proof of competence in the future. E.g. Finland grants certificates to skilled fire safety and structural engineers, and authorities frequently require their participation in projects where performance based fire design is used. Requirements for the documentation of fire design are set by 50% of the countries. The main problems related to performance based fire design are: lack of experience and confidence of authorities, definition of design fires and parameters in some cases, and lack of design tools. Heinisuo (2009) lists the software used in fire design of: – fire simulations, – evacuation simulations and – resistance checks of members. Some ideas about the level of education of fire design engineers were gleaned from the answers. Generally the education of fire engineers seems to be of a very low level in the 10 countries involved. In Czech Republic, France and Poland some improvements are planned. Sweden is a country where sector education is first rate.

5.1.3 CASE STUDY 5.1.3.1 Salmisaari Sports Centre Salmisaari Sports Centre is located in the middle of Helsinki, Finland. The building will be ready for use in April 2010. The floor area of the building is 22,200 m2 and its volume 167,700 m3 . The main contractor is YIT Rakennus Oy. The architects and consulting structural and fire engineers are Arkkitehtitoimisto Pekka Lukkaroinen Oy and Finnmap Consulting Oy and L2 Paloturvallisuus Oy, respectively. The load bearing structures were delivered by Ruukki Construction Oy. The length, width and height of the building are about 136, 35 and 36 metres. There are four stories, each 8–10 metres high. Each storey has a space about 30 m wide supported by 30 m span trusses located at every 5 metres. These trusses are innovative structures used in some Finnish projects: the top chord is made of a welded slim floor box beam that supports pre-stressed hollow core concrete slabs, the braces are of tubular steel and the bottom chord is a flat steel bar. The 377

Figure 1.

General view and the examined space.

trusses are about 3 m high. That leaves a lot of space for installations below the floors. The columns supporting the trusses are reinforced concrete filled steel tubes. A general view and the space examined in this study are shown in Figure 1. Performance based fire design was applied in this project only to trusses. Fire actions were determined for the parts of the building topped by trusses. The intended uses of the spaces below the trusses are: – First floor: Two ice hockey rinks (total area 4200 m2 ). – Second floor: Bowling, martial arts and restaurants (2000 m2 ). – Third floor: Adventure place for children (2000 m2 ), beach volley (780 m2 ) and badminton (570 m2 ). – Fourth floor: Dancing (900 m2 ). – Climbing wall, area 170 m2 , and max. height 30 m. Fire actions were determined for the intended uses of the spaces, and for the following special cases: – – – – – – –

Ice resurfacing machine fire, Storage fire with flashover, Coat-rack fire, Plastic slide fire, Stage fire (abnormal use), Stand fire (abnormal use), Climbing equipment fire.

The fire safety plan was prepared by the fire engineers of the project. Fire compartments were partitioned using EI60 structures. The fire compartments consist of stairwells, exit areas, storage spaces, offices, saunas, dressing rooms and special facilities. According to the safety plan, the building should have the following fire safety equipment: – Initial extinguishing equipment, consisting of: one portable extinguisher per 300 m2 or hose reels. – Automatic alarm system covering the whole building. – Smoke extraction, mainly by the fire brigade. – Automatic sprinkler system. According to CEA (1998) requirements, the sprinkler system should be able to detect and put out a fire in its early stage, or to restrict the spread of fire until the fire brigade arrives. Fire actions are determined based on fires which may occur in different spaces (during intended use, special use and abnormal use). The effects of the sprinkler system are taken into account when defining design fires. Traditionally the effects of the actions of the fire brigade and other fire fighting measures are not taken into account in defining design fires. Fire brigade actions are taken into account in the following references: Tillander et al. (2009), Karhula & Hietaniemi (2008), NFPA (1996), Barry (2002) and Hietaniemi (2008). 378

Figure 2.

Defect flow of sprinklers.

A summary of the definitions of design fires used in the performance based fire design of this project is given below. More details are given in a report by Hietaniemi (2009). In Finland it is not possible to use Annex E.1 of EN 1991-1-2 (not applicable) to define the fire activation risk due to the size of the compartment and the type of occupancy, which is why probability analysis was used in this study. Fire load densities were determined based on national fire load classifications of occupancies and by conducting a fire load survey using both analysis and synthesis of experimental data as well as modelling and fire simulation. The fire scenarios and all details of the fire load calculations were approved by the local authorities, the client and the fire safety and structural engineers of the project before fire simulations and structural calculations were done. 5.1.3.2 Fire actions based on intended uses of spaces The following properties are supposed to be valid for the sprinklers: – RTI = 110 m1/2 s1/2 – Activation temperature is 67◦ C – Protection area Ar of one sprinkler is 12 m2 . The defect frequency of sprinklers is 3% according to International Fire Engineering Guidelines (2005). Assuming a floor area of 5000 m2 , and a 12 m2 protection area, about 500 sprinklers are required on that floor. Then the probability is that one of those sprinkler heads is defective. Let us then suppose that this defective sprinkler head is just above the starting point of the fire. The resulting fire scenario would be a so-called local fire in the sprinklered building where: – The other sprinklers restrict the fire to the protection area of one sprinkler. – Fire intensity is defined by the use of the space under consideration, as shown later. Let us then consider the failure of the entire sprinkler system. That can be estimated by the defect flow of Figure 2. The sources of the initial data are the following: – Pump defect, Isaksson et al. (1998), – Duct defect and water source defect, Isaksson et al. (1998), – Installation defect, Korpela (2002). The probability for failure of the entire sprinkler system is about 0.00197 ≈ 0.2% according to the estimate. On that basis a second fire scenario is created involving a so-called global fire in the sprinklered building: – After sprinkler activation the fire intensity is doubled from the value defined based on the use of the space at sprinkler activation time and it remains constant. The doubling provides the extra safety required by authorities in this case. So we end up with two fire scenarios, the first one based on local sprinkler defects and the second one on the failure of the entire sprinkler system. They are graphically demonstrated in Figure 3. In the first case the fire decays either due to a lack of oxygen or combustible material in the space. The fire is local within an area of 12 m2 and should be applied to the most severe locations in the building. The second fire is not dying down and engulfs the whole floor under consideration. 379

Figure 3.

Schematic fire loads (heat release rates, HRR in MW), local (left) and global (right) fires.

Table 2. Probability of fire activations and sprinkler defects during 50 years of special uses of spaces. First floor

Ar (m2 ) Ar /Atot Fire activation Sprinkler defect

Second floor

Third floor

Fourth floor

Ice machine

Storage

Coat-rack

Plastic slide

Stand (abnormal)

Stage (abnormal)

4200 0.221 6.8 E-01 3.3 E-02

45 0.002 1.2 E-02 3.6 E-04

12 0.0006 3.2 E-03 9.7 E-05

2000 0.10 4.2 E-01 1.6 E-02

260 0.013 5.8 E-03 1.7 E-04

900 0.047 2.0 E-02 6.0 E-04

The local and global fires in a sprinklered building defined above were assumed to occur at the most severe locations in the building. The special uses, including abnormal uses, and corresponding fires were also assumed to occur in the building. The probabilities of these fire activations resulting from the size of the compartments and the occupancies are given in Table 2. The probabilities were calculated based on Tillander et al. (2009) for a 50 year period. Probabilities for abnormal uses were calculated assuming their occurrence once a month. The probabilities for local and global fires are given in Hietaniemi (2009). Next we shall consider the fire load intensities q” [MJ/m2 ] for local and global fires. These can be estimated based on local regulations and experimental data. Finnish regulations (Ympäristöministeriö (2002)) state that the value for stores should be more than 1200 MJ/m2 . For shops, exhibitions halls and libraries its proper range is 600–1200 MJ/m2 . For restaurants, smaller than 300 m2 shops, offices, schools, sports halls, theatres, churches, and similar buildings the value is below 600 MJ/m2 . Based on the above, the maximum value for sporting areas is 600 MJ/m2 . Measured data (International Fire Engineering Guidelines (2005)) yielded 421 fire load intensities for production spaces, which are clearly higher than in our cases. The mean of the sample was 530 MJ/m2 and the deviation 540 MJ/m2 . The 3-parameter gamma distribution was used with the following results: 80% fractile = 600 MJ/m2 and 95% fractile = 1100 MJ/m2 , see Figure 4. Based on these estimations, the following fire load intensities were used in this study: – 600 MJ/m2 for the spaces meant for sporting (80% fractile for generic fire intensity distribution). – 1100 MJ/m2 for other spaces excluding stores (95% fractile for generic fire load intensity distribution). Next we shall consider the corresponding fire release rates (HRRPUA, Heat Release Rate Per Unit Area). Table 3 draws on data from Hietaniemi (2007). It presents the fire load intensities and corresponding fire release rates. The origin of each data line is given in Hietaniemi (2007). The close correlation between fire load intensity and heat release rate per unit area is shown by Figure 5. 380

Figure 4.

Fire load intensity distribution for production space.

Table 3. Sample of HRRPUA and q” values. Item

tg [s]

HRRPUA [kW/m2 ]

q” [MJ/m2 ]

Wood pile (4 pieces) Stack of pallets (2 pieces) One plastic chair Stack of plastic chairs Two stacks of plastic chairs Sports bags Fair stand Litter basket (2 pieces) Carton Work point in office (8 pieces) Television Washing machine Washing machine in cabinet Refrigerator Polyester coat Coat-rack (2 pieces) Shoe store Speciality shop Armchair Sofa (2 pieces) Unprotected mattress Protected mattress

209–409 600–900 900 110 110 420 150 140–1450 150 115–225

469–2156 3062–4105 600 7600 4300 1324 1966 1200–1400 1966 820–1799 930 1422 1483 1921 250 188–190 2500 2900 5480 3120–3375 527 34

703–1561 1500–2250 160 1140 1450 1829 1203 400–422 1230 376–914 500 639 1054 1031 40 90–125 1760 2900 980 727–940 126 3

273 563 660 720 150–210 80 71 120 110–110 145 360

Figure 5. HRRPUA versus fire load intensity.

381

Figure 6. HRRPUA for two types of spaces, shading indicates values between 5% and 95% fractiles.

Thus, fire release rates can be estimated based on fire load intensities: – For sporting areas the mean fire release rate is 1000 kW/m2 and its 5% and 95% fractiles are 800 kW/m2 and 1100 kW/m2 (see Figure 6 a). – For other spaces (excluding stores) it is 1900 kW/m2 and its 5% and 95% fractiles are 1600 kW/m2 and 2100 kW/m2 (see Figure 6 b). Next, we fit the local fire (Ar = 12 m2 ) curve of Figure 3 to the data above. In the growing phase we use the t 2 curve including the time tg needed to reach a heat release rate of 1 MW. In the decay phase we use an exponential curve including a creeping factor of 30% of total heat release based on experimental values. Fire intensity is: The height Hf of the fire source can be estimated using the equation 1.

where Q0 = 1 MW, tg = 150 s, τ is the creeping factor and t1 and t2 are the limit times for uniform fire intensity. The result for the sports area is shown in Figure 7 a) and for other areas in Figure 7 b). The maximum HRR for the sports area is little below 15 MW and for other spaces little over 25 MW. The HRR for the corresponding global fire is shown in Figure 8. In studying structures above the fire, the simplified geometrical model for modelling the local fire uses a square (3 × 4 = 12 m2 ) at a specific height from the floor, and the fire burns only on the top surface. Special cases where the fire source was supposed to be 5 m above the floor were considered too. The height Hf of the fire source can be estimated using the equation 2.

where q” is the fire load intensity, Hc is the calorific value of the material (supposed to be within 25–44 MJ/kg), η is the factor that accounts for the solidity of the material (one for a solid 382

Figure 7.

Local design fires and their parameters.

Figure 8.

Global design fire.

material, zero for a loose material) and ρfuel is the density of the material (supposed to be within 900–1200 kg/m3 ). IF we suppose for simplicity uniform distribution of all the quantities within the ranges shown above and a 10–90% range for the factor η. Then, based on 1000 Monte-Carlo simulations we find that for 600 MJ/m2 the value Hf is smaller than about 20 cm (Figure 9 a) and for 1100 MJ/m2 the value Hf is smaller than about 50 cm (Figure 9 b). Traditionally the value Hf = 0.5 m is used for both cases. The fire source area used in the simulations is shown in Figure 10. Next, we shall consider the design fires for special uses. 5.1.3.3 Ice resurfacing machine fire Two kinds of approaches were used to define the design fire for this case: simulation with the FDS 5 program and estimation with a general fire model for vehicles (Hietaniemi 2007). The goal was to define the design fire for the ICECAT (2008) machine shown in Figure 11. The machine contains the following combustible materials: plastics (ABS), glass-reinforced plastic (GRP) and rubber. The properties of ABS were derived from Lyon & Walters (2001) and Scudamore et al. (1991), those of GRP from Mouritz & Mathys (2006) and those of rubber from Iqbal et al. (2004), Chapter 7. 383

Figure 9.

Distributions for the burning item.

Figure 10.

Geometrical model for burning item (local fire).

Figure 11.

Ice resurfacing machine.

Figure 12.

FDS 5 model for predicting the fire load of ice resurfacing machine.

The machine was modelled with FDS 5 using cubes fit to the grid size and amount, distances, total size and mass of the cubes fit to the machine data. The FDS 5 model is shown in Figure 12. The thermal properties used in the simulation were typical for plastics: density 1100 kg/m3 , thermal conductivity 0.2 WK−1 m−1 and specific heat 1500 JK−1 kg−1 . Combustion time is estimated 384

Figure 13.

Distributions of fire load, effective net caloric value and HRRPUA.

Figure 14.

FDS 5 prediction and general vehicle model prediction for the ice resurfacing machine fire.

at 30 s and combustion temperature at 320◦ C. The fire is supposed to reach its maximum intensity in 60 s. The simulation is based on normal distributed fire load [Q, MJ], effective net caloric value [EHC, MJ/kg] and heat release rate per unit area [HRRPUA, kW/m2 ]. Their distributions are shown in Figure 13. In the simulations the following 95% fractiles were used as input. Their standard deviations are shown in parentheses: – Q: 16700 (750) MJ, – EHC: 35 (1) MJ/kg, – HRRPUA: 700 (70) kW/m2 . In some cases other fractiles were used to determine the effect of the input on the result. The result of the simulation is shown in Figure 14. Figure 14 also shows the result based on Hietaniemi (2007) using the 95% fractiles 2225 MJ/m2 for the fire load and 1725 kW/m2 for the heat release rate. The final design fires for the ice resurfacing machine were determined based on these analyses. They are presented in Figure 15. Figure 15 a) presents the local fire and Figure 15 b) the global fire where after the total collapse of the sprinkler system the heat release rate doubles and then remains constant. 5.1.3.4 Storage fire with flashover The large compartment comprises storage spaces which should be divided into individual compartments using EI60 structures. However, the doors of the spaces open into the large compartment which is why the scenario where the door is open during the fire was chosen. The storage space was modelled as a single sprinklered floor area because that represents the most severe situation as flames come out of the storage door. The fire load was modelled using 64 burning units each equalling a cell of the FDS grid. The heat release rate from each surface of 385

Figure 15.

Local and global ice machine design fires. Vertical axis shows time in minutes.

Figure 16.

FDS 5 model for predicting the storage fire load and an example of flaming through the door.

Figure 17.

Design fire for storage with flashover. Time in minutes.

each unit was 500 kW/m2 . The net caloric value was 35 MJ/kg and the total fire load 30,000 MJ. The FDS 5 model and an example of the flaming through the door are presented in Figure 16.

5.1.3.5 Coat-rack fire with local flashover The definition of the design fire for this case started by modifying the FDS 5 model to simulate closely the experiments of Hadjisophocleous & Zalok (2004). The HRRPUA was 160 kW/m2 and the EHC was 30 MJ/kg. The geometrical model, the FDS 5 model and examples of the fire are presented in Figure 18. The fire loads for the basic case and two variations are presented in Figure 19. The first variation is calculated using double the HRRPUA [kW/m2 ] value of the basic case. The second variation is calculated using double the fire load intensity [MJ/m2 ] of the basic case. The fire of Figure 19 c) was used in the final simulations of the building fires. 386

Figure 18.

Geometrical model, FDS 5 model and examples of fires to predict the coat-rack fire.

Figure 19. The basic case (a) and two variations: doubled heat release rate (b), and doubled fire load intensity (c).

Figure 20.

Plastic slide and its simplified geometrical model.

5.1.3.6 Plastic slide fire The most hazardous object in the adventure space for children in case of fire is the plastic slide which is high and contains a lot of combustible materials. The slide and its simplified model are presented in Figure 20. More refined models of parts of the slide are presented in Figure 21. Two possible ignition locations were considered as shown in Figure 22 a). The corresponding fires are presented in Figure 22 b). The corresponding design fires are presented in Figure 23 a) and b). The design fire presented in Figure 24 was used for global fire. The design fire used in this case was much larger than the doubled fire load after sprinkler activation (about 5 minutes in Figure 24.). 387

Figure 21.

Description of slide and related quantity data.

Figure 22. Two ignition locations (a) and corresponding fires (b) for predicting the slide fire.

Figure 23.

Local design fires for plastic slide with two ignition locations.

388

Figure 24.

Global design fire for plastic slide, time in min.

Figure 25.

Stage model.

5.1.3.7 Stage fire The stage is not a permanent structure and is not normally in use. However, it may be needed in the dance, which is why this scenario was also considered. Stage load was defined for the area of one sprinkler (12 m2 ). The geometrical representation of the stage and the quantity data for calculating the fire load are given in Figure 25. The quantity data and the corresponding fire are shown in Table 4. The attributes of the single homogeneously burning stage material for the whole area are: – – – –

HRRPUA = 1000.0 kW/m2 THICKNESS = 0.05 m DENSITY = 1200.0 kg/m3 HEAT_OF_COMBUSTION = 30.0 MJ/kg The fire load of the stage is presented in Figure 26 with the fire load of a global fire (red line).

5.1.3.8 Stand fire The stand is not a permanent structure. Temporary stands are needed for spectators of beach volley and badminton matches. The stand is made of plywood and plastics. Its geometrical model is given in Figure 27. The quantity data and corresponding fire load calculations are shown in Table 5. 389

Table 4. Quantity data of stage fire.

Speaker Amplifiers Cables Platform Back wall Curtain

Density [kg/m3 ]

Heating value [MJ/kg]

HRR [kW/m2 ]

V [m3 ]

A [m2 ]

Weight [kg]

200 200 1200 700 700 1200

30 30 40 15 15 40

1000 1000 450 1000 1000 1000

0.96 0.86 0.72 0.30 0.40 0.01

8.24 8.40 60.84 12.35 16.40 12.01

192 173 144 60 80 2

Total

Fire load [MJ] 5780 5184 5760 900 1200 96 18900

Figure 26.

Stage design fire loads, local and global (red).

Figure 27.

Geometrical fire model of the stand.

Table 5. Quantity data of one seat in the stand.

Plywood PP PU

Density [kg/m3 ]

Heat value [MJ/kg]

HRR [kW/m2 ]

V [m3 ]

A [m2 ]

Weight [kg]

Fire load [MJ]

700 1200 100

15 40 25

150 1200 400

0.0072 0.0024 0.0096

0.72 0.72 0.72

5.04 2.88 0.96

75.6 115.2 24.0

Total

214.8

390

Figure 28.

Local and global design fires for the. Time in minutes.

Figure 29.

Stand fire.

The size of the burning area is 12 m2 . The attributes of the single homogeneously burning stand material are: – – – –

HRRPUA = 583.0 kW/m2 THICKNESS = 0.2 m DENSITY = 52.0 kg/m3 HEAT_OF_COMBUSTION = 24.2 MJ/kg Local and global design fires for this case are given in Figure 28. An example of a fire in the stand is given in Figure 29.

5.1.4 ESTIMATION OF ERRORS Some error estimations concerning the proposed design fires should be done before any fire simulations on the building. The selected fire scenarios meet the requirements of Finnish regulations (Ympäristöministeriö (2002), Chapter 1.3.2) and, thus, cover all fires that probably could take place in the building. They do not represent the average situation, but a rare situation which can be considered to represent 99% of the cases. This means that one fire out of 100 can be worse than expected. That is a very small number, which means that in this study the possible uncertainty of the fire scenarios will be attributed to the uncertainty of the design fires. The uncertainty of design fires consists of the uncertainty of our knowledge and our ignorance (epistemic and aleatoric uncertainty) such as: – The values used in calculations, e.g. HRRPUA values, always include noise originating from non-ideal tests arrangements, measurements and analysis models. – Possible systematic errors in the values used in calculations originating e.g. from the hypotheses made to simulate the real situation. The uncertainty of fire technical measurements is of the order of 20% as are model uncertainties. . Assuming that systematic uncertainties are of the same order (20%), the uncertainty Q of the fire 391

load can be calculated by the equation 3.

According to fire plume models, gas temperature Tg rises in proportion to ambient temperature to the power 2/3 as shown by Heskestad (1984) and Hostikka (1997).

so the uncertainty Tg of the temperature rise is

This means that the relative uncertainty of the estimations of temperatures can be described as a normal distribution with a mean of 1 and a standard deviation of 10%:

5.1.5 FIRE SIMULATIONS 5.1.5.1 The simulation environment The aim of the simulation was to estimate endurance of structures to natural fire. The structural product model was used as the basis of simulation. Beams, columns, roof and floor slabs, and concrete stairwells were incorporated in the model. The data content of the structures of that model was more complete than that of the architectural model. The building parts were not assumed to be involved the fire since all the burning material was assumed to be included in the fire packages. The material properties of structures were not needed in fire simulation. The structural model was complemented based on drawings. The airspace where the fire burned was bounded by slabs or wall panels. All doors were modelled as openings in the walls assuming that evacuated persons had left them open. Other vents were for the most part not modelled. If there were any openings, the airspace where the fire burned could also be modelled by the properties of the edge of the calculation grid. The used modelling program was Tekla Structures version 15.0. The NIST Fire Dynamics Simulator (FDS) version 5.2.5 was used for simulation. The calculation method is based on CFD (computational fluid dynamics) which uses a three-dimensional, rectilinear computation grid. All the modelled objects must be modified into cubes in some phase of the data transformation process. A special data transformation program was used to transfer the structural model data to the FDS input file. At the same time, all needed material data were stored to the same input file. The process is described more accurately by Laasonen (2010). 5.1.5.2 Selection of the grid cell size The size of a single cell of the calculation grid affects the following three important factors given in order of importance: 1) the reliability of simulation, 2) the minimum size of the objects that can be incorporate in the fire model, and 3) the computer time needed for calculations. Heinisuo et al. (2008a) have discussed the required cell size. Heskestads’s correlation is used to estimate the reliability of calculation. It uses the density of fire [kW/m2 ] and the burning area to calculate the so-called Resolution factor (R) for defining the sizes of cells. Heinisuo et al. (2008a) recommended that the sizes of cells should be selected so that the value of R is at least 10 (or inverse value r not more than 0.07). As presented in the previous chapters, the used special fires are not planar but involve threedimensional objects which may burn on many faces. Then, the acceptable limit for the Resolution 392

factor is not known. Two Resolution factors have been calculated based on simulated fires: a lower value when only the fire on the top face is included in the burning area, and the higher value when all the burning faces are included in the area. To limit calculation time, the model was divided into the several grids. A calculation environment where every grid can be calculated by a different processor was used. However, the hottest area was not divided between several grids because that could cause problems to the stability of calculation. Also, if a larger number of processors are needed, the starting of calculations could be severely delayed. Coarser grids were used for the colder parts. Alpert’s correlation was used to approximate the width of the hot area. A distance from the plume centreline where the temperature should be less than 100◦ C was calculated. This distance is always smaller than the distance to the edge of the coarse grid. In the simulation environment the co-ordinates of modelled objects were not changed in the transformation to the fire simulation program. The simulation program was allowed to locate every co-ordinate to the nearest cell corner using normal mathematical rounding rules. If all the corners of an object are rounded to the same cell corner, it will vanish from the fire simulation. Because of rounding, the thickness of some objects may be zero. As long as the rounding cause any unwanted holes in the simulation model, it should not affect the calculation. The simulation program reads the real thickness of objects from their attributes. The effect of rounding was observed by two methods. In the simulation environment the calculation grids were also added to the structural model. At least one edge of the grid could be located according to modelled structures. All the added geometry could also be located to the grid cells. For example, holes less than two cells in size were not used. The other method involved visual checking of the fire simulation model. The checking was carefully done before calculation when most of the problems could been noticed. After calculation, smoke animation could indicate unwanted air flows. To minimise calculation time, the biggest possible cell size was usually selected. Then the rounding of co-ordinates may cause structures to be lost in the fire simulation model. Profiles whose both dimensions are less than the cell size will probably be lost if not successfully located between cell corners. Profiles exactly the size of a cell can be lost if the cell corner is located exactly in the middle of the profile. That is highly improbable. Heinisuo et al. (2008a) have tested the effect of different sizes of obstacles in a fire model. They noticed that if the obstacle height versus corridor height is below 0.1 in a ceilinged space, and the obstacles are not located close to each other (less than three times their height), it is not essential to model them in a fire simulation. Consequently, slender profiles do not change substantially the flow of air. The height of the modelled spaces was typically between 4 and 10 metres. Then it can be assumed that ignoring of obstacles smaller than 400 mm has little effect on simulation. In the hot area the upper limit of cell size was 200 mm. Outside the hot area, the flow of air is even slower and bigger obstacles can be ignored in the fire model. There the upper limit of the cell size was 400 mm. The end result of the investigation of the effects of rounding was that profiles smaller than the cell size could be freely rounded off. The pictures of the fires in Chapter 5.5 show that, for example, all diagonal members of trusses have vanished from the fire models. 5.1.5.3 Modelling of fires and grids The previously presented fire packages were used in simulations. The properties and behaviour of burning materials were converted to FDS language. The HRRPUA, CONDUCTIVITY, SPECIFIC_HEAT, HEAT_OF_COMBUSTION and DENSITY values were given. The slope depicting the development of the fire as a function of time was given. The material data of the fire were linked to the model so that the name of the FDS fire was included in the name of the geometrical object describing the fire. The fire was modelled in the form of cubic geometry which follows the cells of the calculation grid. The location of the fire was selected for maximal temperatures of structures. Then the flames should reach the structure or just underneath. The other rule was that there should be enough air for the fire since the area around the opening is the severest. 393

The finest grid was located around the fire. One edge of the grid was aligned with the bearing structures. The exact location of the fire was fine-tuned accordingly. Then the other fine grids where located around the first one. Finally, the rest of the model was filled by coarser grids. 5.1.5.4 Output of temperatures Air temperatures were output at certain points during fire simulation. The location of the points must be entered by co-ordinates to the input file of fire simulation. The middle point of every steel member was selected as a control point. That allowed reading the co-ordinates automatically from the structural model. Temperatures at different locations of long and vertical rods varied sometimes. The safe solution in such instances is to assign critical members the highest calculated temperature of the surroundings. In some cases extra control points above the fire were also included in the calculation. The air flow near the flames and plumes is turbulent. The programs can simulate this when output temperatures vary a lot between successive calculation steps. In an intense fire the difference could be about 100◦ C. If we wish to know the temperature at one point at a certain time, it is not advisable to take a single value from the time-temperature curve because of the turbulence. It is better to use the so-called ‘sliding window’ with the mean of several successive calculation steps. One simulated second may involve several steps of calculations. That would make the amount of output data huge. The temperature of structures corresponds closely to the temperature of air. For these reasons, all the calculated steps are not used in post processing. Hostikka et al. (2001) have presented an equation to calculate the width of the sliding window. In the output diagrams of simulations they reduced air temperatures to 10 seconds wide time steps. That value was considered suitable in all cases. The temperature of a steel part can be calculated by integration from the time-temperature curve of air. Heinisuo et al. (2008b), among others, have presented examples of such calculations. In the following, only the air temperature curves are given. These temperatures were used by the structural engineers of the project to check the resistance of the structures in fire. 5.1.5.5 Simulated cases The calculations of the fire cases presented in Chapter 3 were done to determine the worst-case scenarios. The cases involving the highest temperatures are presented in the following. Results are presented mainly for those control points where air temperatures were over 400◦ C. That is a critical limit because the yield stress of steel decreases at temperatures above it. Table 6 lists the documented cases. The Resolution factor (R) is output as told in Chapter 5.2. An exception is the coat-rack fire where the relative area of the top faces was very small and the top of the coat-rack was closed as shown in Figure 33. The R value of the top faces in the coat-rack fire has not been output. All calculated values are at least near the minimum target value 10. The worst R value was calculated for the storage fire, but there only the top faces of the fire elements were burning. The number of grids of both used cell sizes is given. The total number of grids of the fire models was between 7 and 16. The initial simulation time was one hour. In cases where the combustible material burned away, the simulation was stopped earlier. The output temperatures should have settled down before the stopping. Table 7 shows the calculation times of simulations. The maximum numbers of cells in one grid and simulation time were output to compare different cases. As stated earlier, it is advisable to avoid dividing the grids around the fire to keep calculation times short. A long, intense fire also lengthens the calculation time in addition to the wideness of the grids. The ice hall was modelled in actual size bounded by the designed walls. The space was so large that the fire qualities of the walls did not matter in the simulation. The fire was situated near a door so as to provide enough air. The burning part of the machine was at the actual level. Figure 30 is an example of the visualisation of simulation. The door openings are white and the green points indicate where temperatures were output. Only a few bottom flanges of the trusses were included in the fire model while all other parts were rounded off. 394

Table 6. Documented simulations.

Ice hall, ice machine Ice hall, storages Restaurant, coat-rack Fun park, slide Dance hall, stage Volleyball hall, stand Climbing hall, climbing wall

Resolution factor R

Number of grids

Burning faces

Size of cells [mm]

top

all

200

400



9.6 9.2 – 13.9 13 10.4 13.4

13.9 – 11.4 30.6 23.8 12.8 25.6

3 4 4 6 6 7 8

6 6 5 10 2 0 0

9 10 9 16 8 7 8

Table 7. The simulation and the calculation times.

Ice hall, ice machine Ice hall, storages Restaurant, coat-rack Fun park, slide Dance hall, stage Volleyball hall, stand Climbing hall, climbing wall

Maximum number of cells of grids

Simulation time [min]

Calculation time of simulation [hh:mm]

109824 109824 100000 83200 72000 52000 190256

33 60 60 60 50 25 23

33:47 41:02 33:51 41:01 86:13 16:33 55:28

Figure 30. The ice resurfacing machine fire.

The time-temperature curve of Figure 31 shows that the fire was decaying rather quickly. The control points are indicated by the letter ‘B’ followed by the consecutive number of the corresponding member. In the case of the storage fire, burning objects filled the space and flames shot out of the open door. The structures most endangered by the fire were those above the door opening. In Figure 32 air temperatures are represented by a coloured slice. The other colours of the slice only visualise temperatures while the red objects are structures. The Figure 32 shows that the ceiling above the door spreads the heat so that the air at the ceiling level is not very hot. On the other hand, the temperatures at the platform just above the door and the column are rather high. 395

Figure 31. The highest air temperatures above the ice resurfacing machine fire at various control points.

Figure 32.

Storage fire in the ice hall.

The restaurant was modelled in actual size as an open space with all the doors open. Thus, the lack of air did not limit the fire. The burning coat-rack was situated according to architectural drawings. The fine tuning of its position was done by testing when the flames reached the bottom flange of a truss. The fire is visualised in Figure 33. Air temperatures at control points were not raised above 400◦ C. The plastic slide fire was situated in an open hall near an open door. No other equipment or possible separating walls were modelled. The fire is visualised in Figure 34. The plastic slide fire was very severe although the structure of the slide was thin. Thus, the combustible material was consumed quite quickly as shown by the time-temperature curve in Figure 35. In the dance hall model, the entire space was left open. The separating walls were left out of the model in order to produce simulation results on the safe side as the lack of air could not limit the fire. The fire was located near the emergency exit which was modelled open. Test calculations were made to determine the most severe situation of the stage fire. It was noticed that that tallest speaker caused the highest and longest-standing flames. At the time, the speaker was located under the truss as can be seen from Figure 36. The air temperatures caused by the stage fire were not very critical to steel structures as can be seen from Figure 37. 396

Figure 33.

Coat-rack fire in the restaurant.

Figure 34. The plastic slide fire.

Figure 35. The highest air temperatures above the slide fire at control points.

Figure 38 shows the geometrical model of the storey where the volleyball hall is located – without the walls. The calculation grids can be seen as darkened areas on the floor. The thinner grids are indicated by the darkest colour. The perimeters of the grids follow the walls of the hall. The gray doors and ventilation opening are also pictured. According to Chapter 3.8, only the upper part of the stand is assumed to burn. Therefore, only the seats of the upper part of the stand are modelled. 397

Figure 36. The stage fire in the dance hall.

Figure 37. The highest air temperatures above the stage fire at control points.

Figure 38. The geometrical model of the volleyball hall.

The highest point of the flames varied across the stand. The locations of the highest temperatures varied correspondingly. Thus, it was difficult to determine the single most critical spot of the fire. Therefore, the highest registered temperature should be used for all structures above the fire. Figure 39 also shows how the diagonal braces modelled in green in Figure 38 have been modified into cubes in the fire model. The air temperatures near the structures were high because the fire spread up towards the ceiling. The duration of the fire was, again, short as can be seen from Figure 40. 398

Figure 39. The stand fire in the volleyball hall.

Figure 40. The highest air temperatures above the stand fire at control points.

Figure 41. The climbing hall fire.

399

Figure 42. The highest air temperatures of the climbing hall fire at control points.

The climbing hall was modelled as a three-dimensional multistorey space. Three trusses supported the ceilings. The climbing equipment and plywood based climbing wall were assumed to catch fire. A temperature slice was output also for the climbing wall fire of Figure 41. Figure 42 shows that the temperatures of the climbing hall fire were not very critical to the steel structures. All simulation data were delivered to the structural engineer of the project. That allowed him to visualise the simulation results using all temperature histories of all control points. Using this information he could check the resistances of the trusses of every store. 5.1.6 FURTHER DEVELOPMENT Performance based fire engineering is increasingly used in projects not only for evacuation, smoke control and exit design, but also to determine the resistance of structures in fire. It is not used just to minimise or reduce fire protection, but to enhance the fire safety of structures. In some cases it provides better fire protection than traditional fire design. Performance based fire design is not suitable only for large projects, but for all projects. It has been frequently applied in a wide variety of projects. Lot of work will be required in Europe to bring fire design to the same level in different countries, which would make the market for products subject to the same regulations wider. Fire design has been typically incorporated in different sections of national codes as structural codes. In many countries national rules have been changed to allow applying Eurocodes to fire design as required in EU regulations. The lack of experience and confidence of authorities and design fire definitions seem to be the largest challenges to performance based fire design in projects. The checking of design calculations is a major challenge to authorities. The article presented a case study on how to define the temperatures of fire compartments. Only the fire scenarios and the definitions of the design fires were given. The structural design of the project was done by others. Fire scenarios for all the parts of the building were defined in close co-operation with the client, the authorities and other designers of the project. In this kind of performance based design cooperation between all partners to the project is essential and leads to a thorough survey of the worst-case scenarios. The authors believe that the end result is a very high level of fire safety for buildings. This kind of design requires first rate fire engineering skills and good computing facilities. The developed integrated fire engineering tool was used in the project. In this case a module was used to transfer the data between the product model (Tekla Structures) and the fire simulator (FDS). Careful grid sizing, fitting the obstacles and fire packages to the right locations, etc. require experience from the end user of the system. Similar integrated systems, in fact the same simulator, FDS, can be used e.g. in evacuation design and other design tasks. High unused potential lies in the integration of design procedures. However, the expertise of skillful engineers cannot be substituted by computers. 400

Performance based design should be incorporated in design at an early stage of the project. In the case study it was done by the steel contractor at a rather late stage. Earlier introduction could result in improved fire safety over the life cycle and bigger savings during the building phase compared to this project.

ACKNOWLEDGEMENTS The fire engineering work of this study was done by Dr. Jukka Hietaniemi, VTT, Finland. His input to this project was huge and is gratefully acknowledged.

REFERENCES Barry, T., F., 2002. Risk-Informed, Performance Based Industrial Fire Protection – An Alternative to Prescriptive Codes. Tennessee Valley Publishing: Tennessee, USA. CEA 4001: 1998-12 (fi). Sprinklerlaitteistot. Suunnittelu ja asentaminen. (In Finnish) EN 1991-1-2. 2002. Eurocode 1: Actions on structures – Part 1-2: General actions – Actions on structures exposed to fire. CEN: Bryssels. Hadjisophocleous, G., Zalok,. E., 2004. Fire loads and design fires for commercial buildings. Interflam: Scotland, UK. Heinisuo M., Laasonen M., Hyvärinen T., Berg T., 2008a. Product modeling in fire safety concept, effects of grid sizes and obstacles to steel temperatures. IABSE Conference Information and Communication Technology (ICT) for Bridges, Buildings and Construction Practice. Helsinki. pp. 82–83. ISBN 978-385748-117-8 Heinisuo M., Laasonen M., Hyvärinen T. 2008b. Product modeling in fire safety concept, calculation of steel temperatures. EG ICE08 Conference, Intelligent Computing in Engineering, edited by Rafiq, Y., de Wilde, P., Borthwick, M., University of Plymouth, Plymouth. pp. 460–469. ISBN 978-1-84102-191-1. Heinisuo, M. 2009. Fire design in Europe. In: Fire resistance, Technical sheets, Urban habitat constructions under catastrophic events. Print Prazska technical: Czech Technical University in Prague. pp. 133–139. Heskestad, G., 1984. Engineering relations for fire plumes. Fire Safety Journal, Vol. 7, nro 1. Hietaniemi, J., 2007. Palon voimakkuuden kuvaaminen toiminnallisessa paloteknisessä suunnittelussa, Internet publication, updated 15.5.2007. http://proxnet.vtt.fi/fise/simon/Fise/opetusmateriaali/ mitoituspalot/MITOITUSPALOT_15052007.pdf (In Finnish). Hietaniemi, J., 2008. Performance of fire brigades – literature study. Tutkimusraportti Nro VTT-R-01744-08, VTT: Espoo. Hietaniemi, J., 2009. Salmisaaren liikuntakeskuksen korkeiden liikuntatilojen teräsristikoiden toiminnallinen palomitoitus: Paloskenaariot ja mitoituspalot. Tutkimusraportti Nro VTT-R-01035-09, VTT: Espoo. (In Finnish). Hostikka, S., 1997. Plume models in numerical simulation of fire (Palopatsasmallit tulipalon simuloinnissa). Master’s Thesis, Helsinki University of Technology, Department of Technical Physics and Mathematics: Espoo. (in Finnish). Hostikka, S., Kokkala, M., Vaari, J. 2001. Experimental study of the localized room fires, NFSC2 Test Series. VTT Research Notes, 2104, VTT, Espoo. ICECAT, 2008. Technical data. www.icecat.fi International Fire Engineering Guidelines. 2005. National Research Council of Canada (NRC), International Code Council (ICC), United States of America, Department of Building and Housing, New Zealand (DBH) & Australian Building Codes Board (ABCB). Isaksson, S., Holmberg, L., Jakobsson, P., 1998. Sprinklersystems vattentillop – tillförlitlighet. SP Rapport 12, SP Sveriges Provnings- och Forskningsinstitut: Borås. (In Swedish). Iqbal, N., Salley, M., Weerakkody, S., 2004. Fire Dynamics Tools (FDTs): Quantitative Fire Hazard Analysis Methods for the U.S. Nuclear Regulatory Commission Fire Protection Inspection Program. Final Report. Karhula, T., Hietaniemi, J., 2008. Palokunnan operatiivisten toimien vaatimien aikojen selvittäminen – esitutkimus, Tutkimusraportti Nro VTT-R-01060-08, VTT: Espoo. (In Finnish). Korpela, J., 2002. Märkäasennuksen käytettävyysanalyysi. Teknillinen korkeakoulu, Talonrakennustekniikan laboratorio, Diplomityö, Espoo. (In Finnish). Laasonen M. 2010. Data exchange from BIM to building-use simulation. Icccbe Conference. Nottingham. (not yet published).

401

Lyon, R., Walters R., 2001. Heat Release Capacity. Fire & Materials Conference, San Francisco, CA, January. Mouritz, A., Mathys, A., 2006. Gibson. Heat release of polymer composites in fire. Composites: Part A 37. pp. 1040–1054. NFPA Fire Protection handbook, 1996. Cole, A. E. (Ed.). National Fire Protection Association. Scudamore, M., Briggs, J., Prager F., 1991. Cone Calorimetry – A Review of Tes Carried out on Plastic for Association of Plastics Manufacturers in Europe. Fire and Materials, Vol 15. pp. 65–84. Tillander, K., Oksanen, T., Kokki, E., 2009. Paloriskin arvioinnin tilastopohjaiset tiedot, VTT Working Papers, VTT: Espoo. (In Finnish). Ympäristöministeriö, 2002. Suomen rakentamismääräyskokoelman osa E1. Rakennusten paloturvallisuus. Määräykset ja ohjeet. (In Finnish).

402

Urban Habitat Constructions under Catastrophic Events (Final Report) – Mazzolani (Ed). © 2010 Taylor & Francis Group, London, ISBN 978-0-415-60686-8

5.2 Demands and recommendations for assessment and mitigation of risk under exceptional earthquakes A. Plumier University of Liege, Belgium

R. Landolfo University of Naples “Federico II”, Naples, Italy

D. Dubina The Politehnica University, Timisoara, Romania

5.2.1 STATE OF THE ART 5.2.1.1 Introduction to the concept of exceptional earthquakes At present, seismic design is made with reference to an earthquake level which has a given probability of being exceeded over a certain period, or, which is equivalent, to a level of earthquake which on average should take place after a certain period of time (return period). In Eurocode 8, for instance, the typical value of reference peak ground acceleration for each seismic zone and for the no collapse requirement corresponds to a reference probability of exceedance PNCR in 50 years equal to 10% or to a reference return period TNCR equal to 475 years. Implicitly, this definition contains the fact that greater values of accelerations can take place: those greater values correspond to what is called “exceptional” or “catastrophic” earthquakes: those which impose abnormally large inelastic deformation demand to structures. When inelastic deformation demand is very large, complex non-linear phenomena occur: buckling and/or fracture, in case of steel structures; material crushing and/or loss of bond between concrete and steel, in case of reinforced concrete structures; crushing and/or loss of integrity, in case of masonry constructions. The effect of such phenomena is a degradation of mechanical response, such as strength and/or stiffness degradation of plastic zones in engineered frames, which ultimately may lead to complete collapse of the structure. Several comments can be made: – whatever the value of the probability chosen in the definition of the design earthquake, a certain risk of failure exists since there is always a greater level of possible earthquake than the one chosen for design; – uncertainties are present in many aspects of seismic design; some of those uncertainties are explained in a following paragraph; their practical effect is that there is an uncertainty on the exact level of probability of failure which correspond to a given design; – the designer can try by adequate choices in his design to provide a structure which has extra margins of safe behavior beyond the design earthquake (Fig. 1); the recommendations envisaged here should indicate ways to those extra margin of safety; this objective may seem unduly requiring, but it is in fact not so strange if one thinks of the uncertainties affecting seismic design; – existing structures can be rather weak constructions and raise other problems; for them, a “normal” intensity earthquake can be “exceptional” by imposing inelastic deformation demand greater than the capacity of the structure. 5.2.1.2 Uncertainties affecting seismic design The recommendations for design against “exceptional earthquakes” envisaged in COST C26 WG2 should indicate how to design structures which are “better than good”. This objective may seem 403

Figure 1. Pushover curves of 2 structures valid for a given design earthquake. Structure 2 has an ability to survive an exceptional earthquake.

unduly requiring, but it is in fact not so strange if one thinks of the uncertainties which remain in many steps of seismic design. In particular, one can think of the following problems. Talking about exceptional earthquakes certainly includes the uncertainty on the level of action. Every earthquake brings a modification to the seismic map, which is always a rise in the design action. So the “exceptional” earthquake of to day often becomes the design earthquake of tomorrow. Furthermore, besides of uncertainty on the zone seismicity, there are fundamental aspects of the seismic motion which are ignored, though there effect could be detrimental, in particular directivity effects which exist in near-fault regions and soft soil conditions. Such situation can generate ground motions with long period pulse-type form (Stratan and Dubina, in Mistikadis et al., 2007). The acceleration response spectrum of this type of motions is characterized by a large value of the control period TC (limiting value between the constant acceleration and constant velocity region of the spectrum). While modern design codes generally recognize this effect in the case of soft soil conditions, it is not considered in the case of near-fault ground motions, though structures with fundamental period of vibration smaller than the TC control period of the seismic motion are subjected to accelerations greater than foreseen. Earthquake force reduction factors valid for standard ground motions may be inappropriate in this cases. For the design of the resisting structure, many codes, including Eurocode 8, only consider the horizontal component of the seismic action. However, several recent earthquakes have given evidence of significant and damaging effects of the vertical component of the seismic action. For a given behaviour factor q, the local ductility required by codes is the same for all potential plastic zones: local µ related to q for RC structures, plastic rotations θ related to q for steel structures. However, depending on the design and on the real distribution of strength of materials in the structure, some first formed plastic zones may experience request for ductility far greater than the code prescribed ones. Differential settlements related to severe earthquakes may add strains in plastic zones. Modifications may be brought to a structure in the course of its life which can modify the action effects in comparison to the design. Though they should not, such modifications can also find their origin in the differences between the drawings of the design stage and the “as built” drawings. The code requires that the influence of modifications be computed in a re-analysis of the structure and measures be taken, but the story can be different in the real world. Some typologies of structures activate many plastic zones at the ultimate limit state. It is fine for a high capacity to dissipate energy, but some design relies on a chain of elements such that the 404

failure of one of the components may mean a complete failure of the structure. Such structures are very sensitive to the quality of the execution of the many details. If an uncertainty in the quality of the execution exists, such design can behave far under the expectations. 5.2.1.3 Features of existing seismic codes contributing to a reduction of risk to new design submitted to exceptional earthquakes There are two ways to design an earthquake resistant structure: – with structural elements large enough to remain elastic; – with smaller elements which have the ability to deform plastically. Since the 1980’s, design codes have developed the possibility of doing design which can withstand earthquakes by undergoing plastic deformations. Such design provides safety if the following conditions are realised: 1) an intended global plastic failure mechanism is defined; it avoids partial mechanisms like a single storey failure; it contains numerous or large dimensions places where local plastic deformations will exist in case of an earthquake; 2) those places called “dissipative zones” are designed to be able to undergo several plastic cycles of deformation without significant loss of resistance; 3) the other zones of the structures are designed to remain elastic and stable while the dissipative zones are yielding; the components adjacent to a dissipative mechanism have greater resistance than the dissipative mechanism; they are ‘capacity designed’. Design criteria in codes enforce the objectives of global ductility of structures and of local ductility of the components. For the global plastic mechanism, the criteria vary from one structural type to another. Typical is the “weak beam-strong column” rule for moment resisting frames present in all codes. More specific to Eurocode 8 are the rules tending to the homogenization of overstrength over the building heigth in steel frames with concentric or eccentric braces, through a newly introduced parameter  which is the ratio of the design strength of components to their minimum required strength, and the rules for ductile reinforced concrete walls and lightly reinforced large dimensions walls. The local ductility of dissipative zones is obtained by the respect of rules specific to the constitutive material. For steel structures, it is a matter of classes of sections or of connection design. For reinforced concrete sections, the condition bears on the content of longitudinal and transverse reinforcing steel. In both type of materials, the requirement for local behaviour is related to the intended level of global ductility represented by the behaviour factor. In both types of materials also, a margin of safety is taken on the conditions put to achieve the local ductility: it is not uncommon that the local ductility which is available be in fact two times greater than strictly required by the code. For that reason, the set of rules present in modern design codes usually provide some safety in case of an exceptional earthquake: if an the action and/or its effects are greater than expected, the excesses can be absorbed by the greater energy dissipation achieved by greater plastic deformations of structural components. On the contrary, a similar component designed to resist by its strength cannot provide much more strength than the design one. 5.2.1.4 Guidance for the assessment of existing structures Assessment of collapse of structures is a difficult scientific and technical issue that needs to be addressed in prevision of catastrophic earthquakes. Many progresses have been made in case of engineered structures, such as steel or reinforced concrete frames, in order to understand and define the limit state of “collapse” during earthquakes. Robust documents like FEMA 356 and Eurocode 8 Part 3 (EN1998-3:2004) are now available for practical cases, which allow evaluations of the seismic vulnerability of individual structures based on an extensive experimental basis and on background studies. However research work is still needed to improve the current seismic technical regulations with respect to the assessment of collapse conditions of new and existing constructions. 405

5.2.1.5 Measures to reduce risk under earthquakes An important task with respect to public safety concerns the existing building stock, for which it is fundamental to develop systems and techniques which can reduce damage to structures and increase life safety. Recent research work has provided significant steps in that direction. Guidelines for the seismic vulnerability reduction in the urban environment are presented in a report of the recent LESSLOSS project (Plumier 2007). They cover very different interventions, as there are many types of structures, many materials and many ways to reduce vulnerability. This Report focuses on practical applications rather than on theory. A variety of topics is treated: – The screening of buildings on an urban scale to identify which need retrofitting; – Conventional retrofitting methods; – New retrofitting techniques like the application of Fibre Reinforced Polymers (FRP) on existing structures are explained, with design methods; a user friendly design tool, experimental data on durability and fatigue, a design method considering the contribution of steel rebars and FRP to resistance and experimental studies on masonry infill which FRP can effectively reinforce against transverse move and for their in-plane strength are presented; – Rehabilitation using FRP retrofitting at an urban scale; – The use of dissipative devices to reduce the vulnerability of structures; the technique is applied to precast concrete portal frames and to steel frames with concentric bracings; – base isolation for seismic upgrading of historical buildings, with an example of displacement based method applied to a light house; – The mitigation of hammering between buildings, with a methodology to face various situations; – A displacement based methodology of analysis for underground structures in soft soils is presented at Chapter seven. Detailed information on the deliverables of the LESSLOSS project is available on the website www.lessloss.org. In the same line as LESSLOSS, the recent PROHITECH research project (Mazzolani 2007) presents an exhaustive overview of structural issues for the seismic protection of existing and/or historical buildings against the possible devastating effects of earthquakes. Innovative technologies are searched in order to minimize damage to the main structure by localizing damage into dedicated damage-tolerant structural fuses. Though the idea is clear and the technology offers many alternatives, the practical implementation may sometimes be difficult. The most delicate case is that one of historical masonry constructions, which are generally stiff and brittle. The high initial stiffness of the original masonry structure strongly reduces the efficiency of displacement-based hysteretic dissipation devices; more favorable are viscous dampers. However, the integration and collaboration of the devices with the existing construction is often difficult to realize without significantly modifying the building. Research efforts are needed to develop non-intrusive reversible techniques. In the PROHITECH Report, the results of innovative and/or advanced systems for seismic protection are investigated, with a focus on the important aspects of reversibility, low-intrusiveness, and sustainability of the protection system. Figure 2 reproduces Tables taken from the final report of the PROHITECH (WP6 activity). The results of LESSLOSS and PROHITECH projects, perfectly coherent with the activity of the COST Action C26, present recommendations for the seismic protection of new and existing constructions; they also highlight those aspects still demanding more research. 5.2.2 CONTRIBUTIONS FROM COST MEMBERS 5.2.2.1 Introduction Along the four years of the project, in full agreement with the tasks fixed by the Memorandum of Understanding, the collaborative research of the Group, developed work focused on the following topics: – Characterization and modeling of seismic action – Evaluation of structural response under exceptional seismic actions – Performance based evaluation and risk analysis 406

Figure 2. Advanced mixed reversible technologies for seismic protection as defined in the PROHITECH research project (Mazzolani 2007).

– Innovative protection technologies and study cases – Demands and recommendations for damage prevention under exceptional earthquakes” Four scientific events have been organized by or with the cooperation of COST C26 i.e. the Workshop in Prague, on march 2007, the Seminar, in Malta, October 2008, the “PROHITECH” 407

International Conference, in Rome, June 2009, and the Final Conference in Naples, September 2010. A total of 101 papers, including the Keynote Lectures, have been produced and published by the Group on Earthquake Resistance; they cover the above mentioned topics. On the following, an interpretative and selective review of these contributions, organized and labeled according with the contents and structure of the actual Final Report is presented. 5.2.2.2 Assessment of existing structures One important aspect for the assessment of the seismic performance of existing structures is the characterization of the seismic input. This could be done in different ways, i.e. with different levels of detail, according to the type of existing construction being assessed and its relevant importance. The most simple way is to use the acceleration and/or displacement spectra fixed by the code for the design of new constructions. However, a specific study of the seismicity for the site of the construction under investigation may be required for more important cases. The degree of accuracy in the definition of the seismic input may also be related to the type of analysis tool that is selected for the structure. An excessively refined analysis tool, such as non-linear dynamic time-history analysis (THA), may appear to be disproportionately complex as respect to the degree of accuracy in representing the seismic actions. In fact, the sensitivity of structural response to the ground acceleration timehistory is well proved by the large number of analysis carried out for research purposes. On the other hand, when assessing an existing building, large inelastic deformation demand may be expected, even under “normal” earthquake intensity, i.e. the intensity usually used for the design of new constructions. In this case, THA may represent the only rational tool to assess seismic risk and, therefore, an accurate representation of the seismic action is also required in such a case. The work presented by Stratan and Dubina (2008) and by Lungu et al. (2008a, b) discusses such aspect of seismic assessment, either from a general point of view or with reference to specific conditions (Bucharest soil conditions). In particular, Stratan and Dubina (2008) discuss the issue of record selection for THA, from the viewpoint of the current codified suggestions and requirements. Aspects such as the number of records, the type of record (far-fault, near-fault, recorded, artificial, generated), the scaling procedure eventually adopted are some of the aspects that are briefly reviewed with consideration of code requirements. Lungu et al. (2008a, b) report methods to assess soil conditions and to relate such geotechnical/geological information to the selection of the earthquake actions. Though those papers discuss the specific case of Bucharest (Romania), the work presented may be taken as a useful example as to how incorporate site effects in the evaluation of seismic actions. This is considered an important item in view of further developments of Eurocode 8 and for the harmonization of National seismic codes with the European code. One additional important aspect in the assessment of existing constructions is to deal with uncertainties, both in the structural model and the seismic input. The importance of considering uncertainties in the perspective of a modern performance-based evaluation methodology has been clearly addressed and clarified by a lot of researchers. The work presented by Sickert et al. (2008) deals with this subject, using the fuzzy stochastic analysis methodology. Though the results presented are still at a research stage, in the long term this work may be seen to be one contribution towards the development of comprehensive performance-based guidelines for a rational assessment and the reduction of seismic risk of existing constructions. Many different structural types can be observed for buildings throughout the world. Not all of them are completely codified, understood and studied. One example is represented by thin, lightly reinforced, structural RC walls (Fig. 3a). They are frequently used with a double function: to be partitions between rooms in buildings and to give lateral stiffness and strength. This represents one existing structural type that is not covered by the current version of Eurocode 8, as highlighted by Fishinger et al. (2008). The need to develop analytical tools for assessing and eventually mitigate the seismic risk of such structures is also emphasized by the Authors, especially with reference to the development of flexural-shear-axial interactions (Fig. 3b), which may be significant in such a case. One additional example of a very frequently adopted structural solution is the one of precast, prestressed, RC frames. There are hundreds of such structures used for industrial buildings. The need to assess the seismic risk of such structures has been addressed by Fishinger et al. (2008), who emphasize that the main source of such risk comes from weak connections. 408

Figure 3. Analysis of thin lightly reinforced RC shear walls. (Fishinger et al., 2008).

Figure 4. Experimental tests on typical European beam-columns: a) general view; b) plastic hinge. (Landolfo et al., 2008).

5.2.2.3 Assessment of seismically strengthened structures Seismic design of new constructions is currently codified on the basis of significant research efforts that have been carried out over the past 20–30 years. The large number of both experimental and theoretical research results has nowadays led to develop rather detailed design rules and guidelines. Notwithstanding several points still need to be clarified and/or improved. With reference to the design of steel structures, one point is the classification of beams and beam-columns in terms of available ductility and plastic overstrength. The current version of Eurocode 8 is based on the cross section classification reported by Eurocode 3. In principle, such a classification does not consider strength and stiffness degradation of plastic hinge regions due to repeated local buckling in compression because of the cyclic nature of the earthquake actions. Therefore, an improvement of the design rules for new constructions, that are seismically strengthened constructions, may be considered the development of a new cross section classification based on seismic test results. The work presented by Landolfo et al. (2008) may be considered as one step through this direction. As shown on Figure 4, the Authors carried out experimental tests on beam-columns under uniform moment gradient and with typical European cross section shapes. 409

Figure 5. Effect of soil-structure interaction for a RC wall structure: a) flexible soil-foundation system; b) fixed base model. (Apostolska et al., 2007)

One design aspect that is still not covered with sufficient detail by seismic codes is the design of structures when significant soil-foundation-structure interaction occurs. The work presented by Apostolska et al. (2007) deals with this subject. The Authors investigated the behavior of some typical RC wall structures by means of the non linear static (pushover) analysis and application of the capacity spectrum method to calculate the target displacement. Numerical results show that in case of soft soils the target displacement is reached with significant soil deformation and associated reduction of plastic deformation of the superstructure, which may even remain elastic (Fig. 5a) while a fixed-base analysis model would indicate significant spread of plasticity (Fig. 5b). This consequently leads to smaller values of the behavior factor. Though numerical analyses and results presented are limited in the extension of case studies, they can be considered as indicative of the importance to eventually consider soil-structure interaction for those situations with rigid super-structure and soft foundation-soil systems.

5.2.3 INNOVATIVE STRUCTURAL SOLUTIONS Several innovative structural solutions have recently been proposed and investigated by researchers. Among the contributions of COST members, presented and discussed during the development of the C26 Action, such innovations include: 1. 2. 3. 4. 5.

Innovative base isolation systems Buckling-restrained braces Steel shear panels Novel bracing types Composite fiber reinforced materials

The use of both advanced bracing systems (buckling-restrained braces and eccentric braces) and composite fiber-reinforced materials for seismic retrofitting of existing “under-designed” RC buildings has been investigated both theoretically and experimentally. One significant experimental research was carried out by Mazzolani et al. (2007a, b) and D’Aniello et al. (2008). Several full-scale tests on novel “all-steel” BRBs (Fig. 6a), eccentric bracing (Fig. 6b) and a FRP strengthening system (Fig. 6c) were investigated by means of collapse tests on the systems applied to existing RC structures. Besides to the possibility to understand typical features of each one system, the similarity of the existing bare RC frames allows also a comparison between the three alternatives to be carried out. Then, experimental results are very useful to improve knowledge and to develop design guidelines to use different retrofitting techniques according to the specific needs. For example, it clearly emerges that eccentric braces permit significant increases of stiffness and strength while limited global ductility is achieved because of the large local plastic deformation demand to the yielding shear links, while on the contrary limited improvements of stiffness and strength can be obtained using FRP materials which are 410

Figure 6. Experimental tests on advanced bracing systems (Mazzolani et al., 2007, D’Aniello et al., 2008, 2009, Della Corte and Mazzolani 2010).

Figure 7. Experimental tests on metal shear panels: a) laboratory tests; b) on-site testing of prototypes (Mazzolani et al., 2007).

instead useful to increase the ductility of existing members and, consequently, of the overall structure. Buckling-restrained braces are at an intermediate position, because they permit to increase stiffness and strength as well as the global structural ductility, though no improvement is brought to the local ductility of existing RC members. Similar experimental results, but on the use of metal (steel and aluminum) shear walls were presented by Mazzolani et al. (2007) (Fig. 7). Both component-level and full-scale global experiments were carried out, confirming that a large increase of global stiffness and strength can be obtained, as well as a significant improvement of the overall displacement-capacity but at the expense of significant local damage to the existing RC members. The development of design guidelines could/should properly account for such results, by clarifying the limits and potentialities of each different seismic upgrading system. In some cases, the designer may found effective a combination of “global” and “local” type of seismic retrofitting systems, as testified by the study carried out by Bordea et al. (2007). The Authors studied theoretically the application of buckling-restrained braces (BRBs) and composite fiber-reinforced polymer (FRP) materials for seismic retrofitting of existing gravity-load designed reinforced concrete buildings. The study was based on numerical analyses of case studies. The design of the BRB system was based on current codified rules for steel buildings, while the design of FRP confinement of columns was also based on existing design rules. Different combinations of retrofitting systems were considered: (i) only FRP; (ii) only BRBs; (iii) FRP and BRBs with a q-factor equal to 6; (iv) FRP and BRBs with a q-factor equal to 3. The case study frames were analyzed using a static pushover. Based on such numerical results, Authors conclude that the use of BRBs only was not able to meet the code requirements, while it was necessary to consider integration of “local” strengthening by FRPs and “global” strengthening by BRBs (Fig. 8). While giving a useful information about the possible need to combine the two upgrading systems, the work also highlights the need for further research to be carried out in order to develop design guidelines (e.g. the most appropriate value of the behavior factor to be assigned at the design stage). 411

Figure 8.

Comparison of simple and combined global and local strengthening (Bordea et al., 2007).

Base isolation is nowadays a rather consolidated seismic protection system. However, research is still active on the subject, aiming to both solve some unresolved design issues and to propose novel and better techniques. For example, the search for more economic systems, as respect to the standard rubber-bearing based solution, was the motivation for the study presented by Michalopoulos et al. (2007). The research may usefully be exploited to develop design guidelines of such a special isolation system. 5.2.3.1 Improvement in design methods The use of High Strength Steel HSS with Grade up to 690 MPa in seismic resistant building structure can be really effective when it is combined with conventional steels. This is demonstrated by Dubina et al. (2007), who propose a design concept of “Mixed Steel Building Technology”: HSS is used for high elastic strength and conventional steel for low yield strength and ductility. An attractive application field consists in dual frames with inverted V braces because the seismic requirements lead to a very high demand for strength in columns and beams coming from the unbalanced forces resulting from the difference between the capacity of the braces in tension and in compression. This may be a good opportunity to use HSS and the results obtained so far shown this gain. However, it is shown that more advanced static or dynamic inelastic analysis are needed to come to conclusive design. Figure 9 shows experimental work realised by Dinu et al. (2010) on two storey frames with dissipative shear walls; the aim of the tests is to obtain an experimentally calibrated seismic behaviour factor q for such structures. Different types of beam-to-column joints have been considered. It has been observed that rigid connections of the panels increase slightly the yield resistance and the ultimate capacity of the structures. Behaviour factor q is on average equal to 5, considering the contribution of the ductility only, which indicates that Steel Plate Shear Walls SPSW provide a dissipative behaviour, similar to other high dissipative structures like moment resisting frames or eccentrically braced frames. Numerical parametrical investigations on multi-storey frames are under way to calibrate the values of behaviour factors q to be used in design. A reinforced concrete building designed for gravity loads and retrofitted with Buckling Restrained Braced systems (BRB) is studied in Bordea et al. (2010, with the objective to assess the behaviour factor of the retrofitted system, providing in that potential design value of q for future retrofitting projects. A Performance Based Evaluation of the RC frame building is applied before and after retrofitting a historical building constructed during the first half of 20 century. Nonlinear static and dynamic analyses (Incremental Dynamic Analysis IDA) are performed using the relevant EC8 elastic spectrum adapted to the soil characteristics of the building location. To validate the results of IDA, two full scale replicates of a portal frame of the structure were constructed, one with and one without BRBs (Fig. 10); they were tested under monotonic and cyclic loadings. Figure 10 412

Figure 9. Tests on steel plate shear walls.

Figure 10. The experimental test set up and the pushover curves of the initial RC frame (MRF) vs. the retrofitted frame (MRF+BRB).

Figure 11. A house stabilised by cold formed steel walls.

413

shows the pushover curve and the results of the monotonic and cyclic tests. Numerical results are close to the results of the cyclic test. The failure of the BRB core took place before the failure of the concrete elements. The experimental values of q is 3.7 or 4,9 depending on hypothesis in interpretation of results, which represent a very significant increase with respect to the original situation (q = 1,5). Buildings stabilised by cold formed steel walls is a subject poorly covered until now. The paper by Iurorio et al. (2007) provides design data and a design method for such structures. The seismic design is based on the results of a parametric study performed trough an analytical method that consents to predict the nonlinear shear vs. top wall displacement relationship for sheathed cold-formed shear walls on the basis of screw connection test response.

5.2.4 RECOMMENDATIONS FOR THE DESIGN OF STRUCTURES SUBMITTED TO EXCEPTIONAL EARTHQUAKES 5.2.4.1 Use only the most reliable global typologies and local details Annex B of Eurocode 0 gives an opening to the characterisation of different typologies of structure by a different reliability coefficient KFI . Typologies which are more prone to defects would be penalised by being attributed a reliability KFI coefficient greater than 1 if it is used to multiply the standard design action. The KFI coefficient of a very unreliable typology could for instance be as great as KFI = q. At present, there is to our knowledge no use in Eurocodes or in U.S. codes of a characterisation of a given structural type made of a given material by means of a reliability coefficient KFI . However the concept has been introduced in some countries. The Algerian code RPA2003, for instance, puts limits to moment resisting frames in reinforced concrete: they may not be more than respectively 4 or 5 storeys high in zones IIa and I (the high seismicity zones), no more than 3 storeys high in zones IIb and III. In addition, moment resisting frames of any height are forbidden if the architecture is such that there are infills at upper floors and no infills at ground floor. In fact, those rules mean that MRF’s in RC are not considered very reliable, as it has been demonstrated by past earthquakes, including in particular Boumerdes 2003. There may be several causes to that fact, like the effect of the vertical component of the earthquake and the uneven quality of reinforced concrete, but rather than putting additional rules in the code to improve quality control or to stress the vertical component, another choice was made: give up with tall MRF’s and favour wall structures. They are simple structures with one single plastic hinge; they are dissipative enough by the dimensions of that plastic hinge; they are easy to control because there is only one plastic zone. By that they are reliable. On the contrary, MRF’s present as many potential causes of failure as there are beam-column nodes. Those nodes are highly stressed in shear. In reinforced concrete, making and placing the stirrups necessary to the resistance of the node is very difficult due to the presence of continuous vertical and horizontal rebars of the beams and columns; it is so difficult that post earthquake survey show that in many instances, it has been a “good” reason to skip those costly and complicated stirrups. Similarly, the Northridge (1994) and Kobe (1995) earthquakes have set forward problems of connections in steel MRF’s. Better design rules have been defined since those events, but still some connection design are more reliable than other ones, even if those are allowed. The general conclusion is: in order to mitigate exceptional earthquakes, use the more reliable global typologies and the more reliable design of local zones. To put the idea into practice, a ranking of reliability of typologies and a ranking of details should be made available to designers so that they can better select the structural systems and the details for their projects. 5.2.4.2 Impose details for seismic robustness Details for seismic robustness are additional construction measures, independent of analysis and design, which are applied to improve the reliability of structures designed to the code, by bringing them the robustness required by Eurocode 1: “the ability of a structure to resist events . . . without effects disproportionate to the cause . . . in particular the ability to avoid progressive failure, a chain in which a local failure generates a global failure, effect out of proportion to the original local problem”. 414

Figure 12. The INERD concept promotes the use of a composite column on the 1st storey of reinforced concrete buildings as a measure for robustness.

An example is the INERD concept to improve the seismic resistance of moment resisting frames in concrete (Degee et al., 2010). Those structures, though designed to Eurocode 2 and 8 may need robustness measures because they are analysed as bare frames. But their behaviour can be modified by infill walls, leading to the most frequent failure mode of RC moment frame, a “soft storey” mechanism in the bottom storey. This is caused by openings present in the bottom storey for offices, shops or lobby, which weaken the 1st storey, so that bending moments and deformation required in the columns are higher there than at upper levels, where masonry infill walls are present. Bending moments combined with compression and shear results in the failure of the columns of the 1st storey, which induce the collapse of the building. Also, in many seismic design codes, there is an allowance to neglect the vertical component of earthquakes in the analysis made for design. It results that the axial force in columns considered in design is in fact approximate. As reinforced concrete columns are not prone to ductility, they can fail locally and generate a global failure, which would less be the case with steel profiles, because their bending resistance M is constant in a wide range of values of N. Another problem is the variability of the quality of concrete. It is not the case with steel sections, because they are industrial products. If the concrete strength is less than required in one column, a local failure is possible which can initiate a chain of failures. Robustness in earthquake conditions is thus needed even for moment resisting frames designed to the code. Research has demonstrated that the INERD concept (Fig. 12), which consists in placing inner steel profiles into what otherwise remain a reinforced concrete design is a way to provide robustness for earthquake situations. Figure 13 shows how much the INERD concept can improve the resistance and ductility of a given RC column. Design criteria of the encased steel profile have been defined. There are other examples of robustness measures, some of which are already in application. For instance, Eurocode 8 prescribes for bolted connections of steel structures that the design shear resistance of the bolts should be higher than 1,2 times the design bearing resistance. This is a detail for robustness, since it adds a contribution to ductility within the joint, while this one is already capacity designed to the structural element resistance and should thus normally not yield. So a first general recommendation in order to improve the behaviour of structures in case of an exceptional earthquake is: use details which improve the robustness of the structures. To put the idea into practice, a set of details for robustness should be defined, so that designers can easily put them into practice. 5.2.4.3 Use typologies with q factor greater in reality than the q indicated by the code Values of q given in the design codes are lower bound of the many values (tenth of thousands) established by researchers. In reality, there is a great scatter in the values of q as they depend on the many parameters: strength of materials, beams spans, number of spans, height of structure, seismicity level, etc. . . Obviously, in redundant structures, the energy dissipation is greater if as many dissipative zones a possible are activated as soon as the structure enters the plastic range. 415

Figure 13. Moment-Rotation curves showing the improved capacity of a composite column (right) in comparison to the original reinforced concrete column (at left).

In addition, if, at an early stage, the plastic mechanism is quite global, then the requirement for ductility is more even between plastic zones than if the plastic mechanism implies only few zones. Structures with an early global plastic mechanism are characterised by a plastic redistribution factor αu /α1 close to 1. Room has been made in Eurocode 8 for less favourable design, by introducing that plastic redistribution factor αu /α1 in the definition of q so that structures less favourably designed, having great real αu /α1 , are not too much penalised. It has been shown that the best design objective is indeed structures in which as many dissipative zones a possible are activated soon after the structure enters the plastic range (Plumier, 2007). This explains why, in the conversion phase of Eurocode 8 from the ENV (provisory version, 1994) to EN (final version, 2004), it has been a constant concern to implement design rules in favour of a relatively early development of a global plastic mechanism. This concern can be found in the “weak beams-strong columns” condition, which has been modified from MRc ≥ MRb into MRc ≥ 1,3 MRb . The same concern is present behind the “homogenisation” rule which, for steel and composite structures with concentric or eccentric bracings, requires that the ratio  of the design strength to the design action effect be kept within 25% over the building height. However, the reality can differ from the design due to the uneven distribution of the real yield stress of materials in comparison to the theoretical one. Two practical ways can be followed to realise an early formation of a global plastic mechanism. One way consists in selecting design which, by themselves, activates almost simultaneously energy dissipation in all potential plastic zones. For instance, in the wide family of possible frames with eccentric bracings, such design is achieved by the “zipper” EBF – Figure. In MRF’s, it could be achieved by imposing a stronger “weak beams-strong columns” condition, like for instance MRc ≥ 2,0MRb (factor 2,0 to be calibrated by research) or by using a different design method. Another way consists in creating conditions such that the ratio  of the real strength to the design action be very close to 1 over the building height. This can be achieved by using specific “fuse” zones which are not simply steel sections from the catalogue, but are industrial products with calibrated strength, like the BRB’s (Fig. 10) or the INERD pin connections (Fig. 2). It has for instance been shown on a particular design example of a building that the behaviour factor q of a frame with X concentric bracings was q = 3,3 for a classical design but it jumped to q = 6,4 if variable INERD pin connections were used. 5.2.4.4 Do design following concepts associated with seismic motion typology One of the crucial decisions influencing the building structure to withstand earthquakes is the basic plane shape and configuration. In some extent seismic design codes contain provisions related to building regularity, both in plane and in elevation, and configuration principles related to structural 416

Figure 14. A “zipper” EBF enforces a global plastic mechanism.

typologies. However, as recalled by Stratan and Dubina in (Mistikadis et al., 2007), there are two requests which must be achieved in order to resist severe earthquakes: – Building structure should be provided with balanced stiffness and strength between its members, connections and supports; – Overall conception and detailing should provide the structure with balanced overstrength and ductility of its members and connections in order to possess an enhanced redundancy characterized by the largest number of defense lines against seismic action. Different structures may respond differently to different type of ground motion. Some structural typologies are more sensitive to particular type of motion (pulse, repeated pulses, long duration). In the light of the two previous basic principles and in order to optimize structural response, the conceptual design of a given structure should take into account the specific features of the possible ground motion. 5.2.5 RECOMMENDATIONS FOR FURTHER DEVELOPMENTS 5.2.5.1 Improvements in seismic design codes A robust seismic resistant structure should be characterized by balanced stiffness, strength and ductility among component members and connections. According to dissipative design philosophy, a robust structure, able to face catastrophic seismic events must posses the following three key attributes i.e.: (1) Secure plastic deformation capacity in structural components, targeted as dissipative, which are key members in any seismic-resistant structure; (2) Dispose for multiple routes for transfer of forces and ensure their redistribution through yielding of other members; (3) Provide sufficient overstrength to those structural components designed to remain predominantly elastic during earthquake, in order avoid the collapse of structure, at any cost. Japanese scientists are calling this approach “Collapse Control Based Design” (Guidelines for collapse control design-construction of steel Buildings with high redundancy, JISF, Tokyo, September 2005). It can be applied trough a Performance Based Seismic Design (PBSD) philosophy and with adequate design methodology. In European Seismic Design Codification, either National or Eurocode 8 package, minimum criteria are still applied in order to protect the health and safety of occupants. However, in recent years, the limitations of design procedures based on minimum requirements and associated criteria have become more apparent as building owners, managers, and regulators have recognized that other factors besides structural safety, such as life cycle, maintenance and repair costs, the performance of non-structural components, need to be considered in designing specific facilities. This has led to the emergence of a performance-based approach to design and construction. Under this method, individual buildings or classes of structures can be designed to perform at levels commensurate with applicable hazards, risks, and risk tolerances. Performance-based 417

methods have the potential to significantly enhance many aspects of building design, including seismic protection. This potential will be realized as the approach is further developed and refined and becomes more widely used and integrated into building codes. In the United States the development of PBSD began in the mid-1990s, largely for use in evaluating and upgrading existing buildings. After FEMA 283 Performance-Based Seismic Design of Buildings – an Action Plan, prepared by the Earthquake Engineering Research Centre, University of California at Berkeley in 1996, and FEMA 349 Action Plan for Performance Based Seismic Design, which was prepared by the Earthquake Engineering Research Institute in 2000, FEMA 356 document proposes to the engineering community a Pre-standard and Commentary for the seismic rehabilitation of buildings. In 2006, the FEMA 445 document titled Next-Generation Performance-Based Seismic Design Guidelines-Program Plan for New and Existing Buildings, declared the following objectives: • To revise the discrete performance levels defined in first-generation procedures to create new Performance measures (e.g. repair costs, casualties, and time of occupancy interruption) that better elate to the decision-making needs of stakeholders, and that communicate these losses in a way that is more meaningful to stakeholders; • To create procedures for estimating probable repair costs, casualties, and time of occupancy interruption, for both new and existing buildings; • To develop a framework for performance assessment that properly accounts for, and adequately communicates to stakeholders, limitations in our ability to accurately predict response, and uncertainty in the level of earthquake hazard. In June 2000, the Interim Testing Protocols for Determining the Seismic Performance Characteristics of Structural and Non-structural Components (FEMA 461) was issued. This document FEMA 461 describes in detail laboratory testing protocols that can be used to determine fragility functions for various building systems and components. Fragility functions express in mathematical terms the likelihood that a component will sustain a specified level of damage when exposed to a specified level of demand (e.g., force, acceleration, displacement). These functions are of fundamental importance to PBSD. Such documents, either normative or for design guidance, are absolutely necessary and should be drafted or appropriated for seismic design practice in Europe, too. In 2007, the Joint Research Centre (EU Commission) of Ispra, published the Document EUR 22858 EN-2007- Pre-normative research needs to achieve improved Design Guidelines for seismic protection in the EU, as technical Support to the implementation, harmonization and further development of the Eurocode 8. The following possible objectives for future earthquake engineering research in Europe have been found as relevant to achievement of improved design guidelines i.e.: • Development of a common methodology to evaluate the earthquake hazard in Europe: the research should be at least conducted at a regional scale, because the methodology depends on the tectonic context. Hazard from 475 years to 10.000 years return period should be envisaged. • Development of assessment and strengthening methodology for more economical and safe solutions stock in European earthquake prone areas. • Development of strengthening techniques of low intrusive effect for application in monuments, historical buildings and other structures. • Seismic design and upgrading of mechanical, electric and other types of equipment used in the lifelines and industry. In particular, the following particular research topics are identified to be of importance for the next generation of Seismic Design in Europe: 1 Harmonized European Seismic map 2 Provisions for the design of irregular-in-plan buildings 3 Primary vs. secondary seismic elements: Elaboration of the implications and reevaluation of the concept 4 Seismic design rules for flat slab systems 5 Seismic design rules for prestressed concrete elements and systems 6 Design rules for masonry buildings 7 Seismic assessment and retrofitting (emphasis to masonry-infilled frame buildings) 418

8 Seismic design of the structure-foundation-soil system 9 Seismic protection of sensitive or valuable equipment and artifacts The authors of this Report are of the opinion that the revision of Eurocode 8 should be made with the aim to provide appropriate, but differentiated, design criteria and rules low-to-moderate and moderate-to-high seismic risk European Regions. In order to establish those rules and criteria, the need for Normative and Guidance documents for implementation of PBSD, for current design practice of both new and existing constructions, is once more underlined. There are also missing rules for low dissipative structures, because is not enough to specify in the code that a q factor of value ranging between 1.5 and 2.0, as it is case of steel structures, for instance, can be applied: there is a real need for specific criteria. For low to moderate seismicity regions, this subject is of particular interest. A particular attention has to be paid to development and application of new structural systems, materials and seismic protection technologies, for which the actual seismic design code like Eurocode 8 do not contain provisions, or only in an informative way. One can observe, looking at the list of contributions to COST C26, that most of these advanced systems and technologies have been, in fact, addressed in the activity of COST C26. Finally, one can observe that there are no contradiction between the research priorities specified in EUR 22858 EN and the ones which have been considered in the COST C26 activity (see V 2.2.) In fact, all the research topics addressed by COST C26 WG2 and its specifically related subjects are in line with those integrated into JCR Document. 5.2.5.2 Some specific aspects of research needs related to new design For new constructions, the current design methodologies should be improved in order to meet acceptable risk of collapse. This can be done through several ways. The simplest way is reviewing values of the behavior factor assigned to the various structural typologies. Historically, the value of the behaviour factor has been assigned mainly based on experience from the behaviour of structures during historical earthquakes or engineering judgment (besides to conventional agreement between parties involved into code development). More recently, numerical validation has become of more widespread use. However, unavailability of adequate hysteresis models has usually limited attention to the system response in its stable range of behavior. Research of the last decade has highlighted the need to consider more refined hysteresis models to correctly capture collapse conditions of structures. Refinement of hysteresis models should capture degradation phenomena, in particular strength deterioration, as degradation could be of paramount importance in accelerating collapse due to P-Delta effects; the behavior of tall buildings, which are particularly susceptible to P-Delta effects and may collapse even with small or no degradation of plastic zones, should be carefully looked at. Special care should also be addressed to masonry constructions, which are at severe risk of collapse even in case of earthquake actions having a relatively small return period, hence relatively small intensity. Numerical analyses with advanced and refined hysteresis models could usefully be employed to improve the design codes for new constructions. 5.2.5.3 Some specific aspects of research needs related to existing constructions For existing constructions, research work is needed for non-engineered masonry, because of the inherent difficulties of the problem. A probabilistic approach is necessary for a scientifically sound and rationale approach to the problem. The seismic fragility or vulnerability of a given structure, i.e. the probability of exceedance of a given damage state for a given earthquake intensity, must be combined with the rate of exceedance of that earthquake intensity, in order to calculate the probability of that damage state. This process involves the consideration of both epistemic and random uncertainties, because not only the earthquake intensity has a random component but also the structure behavior and its assessment are affected by both types of uncertainties. The problem becomes more complex when shifting from the scale of an isolated structure to grouped constructions or urban habitat. In such a case, the analyst must accept a smaller degree of confidence in the analysis results, because of the larger difficulties in the damage assessment process and consequent uncertainties. 419

Mitigation of risk for urban habitat constructions still demands significant research efforts, especially in view to develop scientifically sound methods to evaluate either monetary or life losses. REFERENCES Apostolska R., Bonev Z.P., Blagoev D., Vasseva E., Necevska-Cvetanovska G. 2008. Design seismic response evaluation for 2D frames and 3D wall systems with flexible foundation using capacity spectrum method. Proceedings of the International Symposium “Urban habitat constructions under catastrophic seismic events”, Malta, 23–25 October, pp. 152–157. Bordea S., Stratan A., Dogariu A., Dubina D. 2007. Seismic upgrade of non-seismic r.c. frames using steel dissipative braces. Proceedings of the COST C26 Workshop, 30–31 March, Prague, Czech Republic, pp. 211–220. Bordea S., Dubina, D. 2010. Numerical and experimental evaluation of q factors for RC MRF strengthened of steel BRB. Proceedings of the COST C26 Final Conference, Naples. D’Aniello M, Della Corte G., Mazzolani F. M. 2008. Response of buckling restrained braces to catastrophic seismic events. Proceedings of the International Symposium “Urban habitat constructions under catastrophic seismic events”, Malta, 23–25 October, pp. 158–163. D’Aniello M, Della Corte G., Mazzolani F. M. 2009. Seismic protection of reinforced concrete buildings by means of “all-steel” buckling-restrained braces. Proceedings of the International Conference on Protection of Historical Buildings (Prohitech 09), Rome, 21–24 June. Degee, H, Plumier, A. 2010. An innovative solution to improve the seismic robustness of reinforced concrete frames, 14th European Conference on Earthquake Engineering 14ECEE, Ohrid, Republic of Macedonia. Della Corte G., Mazzolani F.M. 2010. Response of BRBs to catastrophic seismic actions: experimental results. Proceedings of the COST C26 Final Conference, Naples. Dinu F., Dubina, D., Neagu, C. 2010, Experimental evaluation of q factor for dual steel frames with dissipative shear walls, Proceedings of the COST C26 Final Conference, Naples. Dubina D., Dinu, F., Ungureanu, V., Zaharia, R. & Grecea, D. 2007. High strength steel for seismic resistant building frames, Proceedings of the COST C26 Workshop, 30–31 March, Prague, Czech Republic, pp. 211–220. EN 1998-1: 2004. Eurocode 8: Design of structures for earthquake resistance, Part 1: General rules, seismic actions and rules for buildings. EN 1998-3, 2005. Eurocode 8: Design of structures for earthquake resistance – Part 3: Assessment and retrofitting of buildings. FEMA 356. 2000. Prestandard and Commentary for the Seismic Rehabilitation of Buildings. FEMA 445. 2006. Next-Generation Performance-Based Seismic Design Guidelines,Program Plan for New and Existing Buildings. Fischinger M., Kramar M., Isakovic T. 2008. Seismic risk of prefabricated “RC” industrial buildings with strong connections. Proceedings of the International Symposium “Urban habitat constructions under catastrophic seismic events”, Malta, 23–25 October, pp. 170–175. Fischinger M., Kramar M., Isakovic T. 2008. Seismic resistance of a thin, lightly reinforced coupled wall evaluated by large-scale shaking table test. Proceedings of the International Symposium “Urban habitat constructions under catastrophic seismic events”, Malta, 23–25 October, pp. 164–169. Iuorio O., Landolfo R., Fiorino L. 2007. Seismic design of cold-formed steel housing: a case study. Proceedings of the COST C26 Workshop, 30–31 March, Prague, Czech Republic, pp. 245–251. Landolfo R., Piluso V., Brescia M., D’Aniello M., Mammana O., Tortorelli S. 2008. Rotation capacity vs demand of steel beams under catastrophic events. Proceedings of the International Symposium “Urban habitat constructions under catastrophic seismic events”, Malta, 23–25 October, pp. 146–151. Lungu D., Arion C., Calarasu E. 2008a. Bucharest soil conditions and input ground motion for the structural performance analysis. Proceedings of the International Symposium “Urban habitat constructions under catastrophic seismic events”, Malta, 23–25 October, pp. 135–140. Lungu D., Vacareanu R., Aldea A. 2008b. Seismicity of Vrancea subcrustal source and corresponding seismic instrumentation. Proceedings of the International Symposium “Urban habitat constructions under catastrophic seismic events”, Malta, 23–25 October, pp. 129–134. Mazzolani F.M. 2007. Earthquake protection of historical buildings. Proceedings of the COST C26 Workshop, 30–31 March, Prague, Czech Republic, pp. 260–267. Mazzolani F.M., Della Corte G., Barecchia E., D’Aniello M. 2007a. Experimental tests on seismic upgrading techniques for RC buildings. Proceedings of the COST C26 Workshop, 30–31 March, Prague, Czech Republic, pp. 221–228.

420

Mazzolani F.M., Della Corte G., Fiorino L., Barecchia E. 2007b. Full-scale cyclic tests of a real masonryinfilled RC building for seismic upgrading Proceedings of the COST C26 Workshop, 30–31 March, Prague, Czech Republic, pp. 229–236. Mazzolani F.M., De Matteis G., Panico S., Formisano A., Brando G. 2007c. Shear panels for seismic upgrading of new and existing structures. Proceedings of the COST C26 Workshop, 30–31 March, Prague, Czech Republic, pp. 237–244. Michalopoulos A., Nikolaidis Th.., Baniotopoulos C.C. 2007. The MNB aseismic isolation system for the seismic protection of structures. Proceedings of the COST C26 Workshop, 30–31 March, Prague, Czech Republic, pp. 268–272. Mistakidis, E, Apostolka, R, Dubina, D, Graf, W, Necevska-Cvetanovska, G, Nogueiro, P, Pannier, S, Sickert, J-U, Simoes da Silva, L, Stratan, A, A, Terzic, U. 2007, Typology of seismic motion and seismic engineering design. Proceedings of the COST C26 Workshop, 30–31 March, Prague, Czech Republic, pp. 268–272. Plumier, A. 2007, Editor, Guidelines for Seismic Vulnerability Reduction in the Urban Environment, LESSLOSS Report 2007/04. IUSS Press. ISBN 978-88-6198-008-2. RPA 2003. 2003. Règles Parasismiques Algériennes. Sickert J.-U., Kaliske M., Graf W. 2008. Fuzzy stochastic earthquake analysis of structures. Proceedings of the International Symposium “Urban habitat constructions under catastrophic seismic events”, Malta, 23–25 October, pp. 141–145. Stratan A., Dubina D. 2008. Selection of time-history records for dynamic analysis of structures. Proceedings of the International Symposium “Urban habitat constructions under catastrophic seismic events”, Malta, 23–25 October, pp. 123–128.

421

Urban Habitat Constructions under Catastrophic Events (Final Report) – Mazzolani (Ed). © 2010 Taylor & Francis Group, London, ISBN 978-0-415-60686-8

5.3 Impact and explosion M.P. Byfield & P.P. Smith School of Civil Engineering and the Environment, University of Southampton, Southampton, UK

5.3.1 INTRODUCTION This chapter provides a critical review of case studies in an effort to identify the respective effectiveness of robustness design measures. A series of case studies from buildings subject to severe structural damage are considered. Some of the buildings featured survived the extreme loading whilst others suffered progressive collapse disproportionate to the initial structural damage. An attempt will be made to understand which types of building posses high levels of robustness and which forms of construction are potentially vulnerable to collapse subsequent to minor structural damage. Strategies for damage prevention depend on the level of threat imposed by impact and explosions and this chapter aims to assist in the selection of an appropriate design strategy by rating the different design approaches using a simple 5 star rating system. This classification system is made deliberately simple and is intended to assist engineers in their selection of appropriate robustness strategies should severe loading from blast or impact be identified as a significant possibility during the lifetime of the building. This chapter is of interest to engineers faced with the challenge of designing structures at risk from vehicle borne improvised explosive devices (VBIED’s) or potential damage caused by vehicle impact or severe accidental natural gas explosions. This chapter is intended to assist the Structural Engineer faced with the challenge of designing a building with a level of robustness suited to its intended use. At present a range of different design strategies exist to comply with the regulations related to robustness but there is at present no information as to the level of protection afforded by each design approach. This chapter will review the main methods of providing robustness to structures, along with case studies of buildings subjected to damage, in an attempt to inform as to the level of protection provided by each method. An attempt will be made to classify each design strategy using a simple 5 star rating system to differentiate between the performance of each design strategy. This paper is of most interest to engineers faced with the challenge of designing structures to survive localised structural damage from a possible attack by a vehicle borne improved explosive device (VBIED). It is also relevant to those practitioners concerned with potential damage from vehicle impact and or accidental natural gas explosion. The VBIED is extreme in its ability to cause severe structural damage to building frames. Since blast attenuates quickly with distance, VBIED’s often leave the majority of a structural frames only superficially damaged and it is often the case that buildings have sufficient redundancy in order to bridge locally damaged areas. However, some buildings lack this redundancy and VBIED’s have a proven ability to cause severe progressive collapses in modern multi-storey framed buildings. The ideal method of protecting a building against impact and a VBIED is to maintain a safe distance away from potentially hostile vehicles. This is known as the safe stand-off distance approach. If this is not possible then the designer has other options. The first is to attempt to design the frame to resist a given weight of explosives, known as the specific resistance method. Alternatively, local damage to a frame is accepted as inevitable and redundancy is built into a frame to redistribute loads away from damaged areas (the alternative load path method). The final method is to comply with basic design rules that have a proven track record of providing a reasonable level of blast protection at little cost and these are the tying force and key element methods. 423

Figure 1.

Khobar Towers, Dharan, Saudi Arabia (1996).

5.3.2 SAFE STAND-OFF DISTANCE In its most basic form this method involves guaranteeing a perimeter impermeable to vehicles around a structure (Smith P.D. 2007 & 2008 and Smith P.D. & Tyas A., 2008) for protection against vehicle impact and the threat of malicious explosive loading. In the more advanced form a design charge weight of explosive is specified and the stand-off distance determined based on the strength of the building. Recent research at Southampton University has been to develop a method and associated software to determine the safe stand-off distance for a given column size and design charge weight of TNT, Paramasivam (2008). However, whichever of these strategies is employed the provision of any stand-off distance is hugely beneficial and significantly reduces the risk of collapse. The 1996 bombings of the Khobar Towers in Saudi Arabia (Figure 1) and the Islamabad Marriot Hotel bombing in 2008 are practical demonstrations of the impressive level of protection afforded by the provision of stand-off. The device detonated outside the Khobar Towers was estimated to be in excess of 10,000 kg of TNT equivalent and resulted in a crater 26 m wide and 11m deep (House Armed Services Committee, 1996). Khobar Towers were defended by a simple arrangement of precast concrete ‘Jersey Barriers’ which provided a 32 m stand-off. The blast created a crater 26 m in diameter and 11 m deep. Despite the sever nature of the attack the survival of the building is in large part due to the provision a secure perimeter, without which the internal structure of the building would have been destroyed at ground level. A similar level of protection was achieved for the Islamabad Marriott Hotel (2008) which was subjected to an attack that yielded a crater 20 m diameter 6 m deep crater. Again, the stand-off provided proved effective with only minimal structural damage due to the immediate effects of the blast. The problem with the safe stand-off method is that suicide bombers have shown a consistent ability to penetrate security check posts, as demonstrated at the Pearl Continental Hotel in Peshawar, Pakistan on June 9th 2009. Suicide bombers stormed the security gate and were able to deliver a device close enough to cause the progressive collapse of multiple bays of the building. The most famous failure to secure a perimeter occurred in 1983 when the US Marine Corps and the French military forces were attacked simultaneously in West Beirut. The modern RC-framed building housing the US forces suffered a severe progressive collapse with a high loss of life.

5.3.3 SPECIFIC LOAD RESISTANCE This entails the design of structural components to withstand the loading from a specified impact or blast size and proximity, for example this could be 5 kg of TNT detonated in a post room. The result is the design of a hardened structure which can be a relatively expensive option. The authors were not able to find any case studies of buildings subjected to blast that were in fact subjected to 424

Figure 2.

Collapse sustained by Murrah Federal Building, Oklahoma (1996) following VBIED attack.

attack, although by the standards of non-seismic regions of the world the HSBC Head Quarters in Istanbul was a building designed for exceptional circumstances. Detailed to withstand the dynamic effects of an earthquake, rather than blast or impact, the structure featured a high level of strength and ductility and performed extremely well when, in 2003, the building was subject to the blast effects of a close proximity VBIED. Inspection of photographs taken in the aftermath show that the face of the building sustained the most significant damage but that damage to the structure was largely superficial, with none of the columns lost. Another example of a building able to withstand the severe blast loading from a close proximity explosion is that of the Club El Nogal Building in Bogota, Columbia (2003). A VBIED was detonated next to a column in the car park below the building (Garcia et al., 2005) with the specific intension to cause the collapse of the building. However, due to the very high shear strength of the columns which were seismically designed, the attack was unsuccessful and the column remained practically undamaged. These incidents illustrate that it is indeed feasible to design structural elements to withstand very intense loading. In fact, the additional shear strength provided to columns to resist seismic actions may well be sufficient to resist high-explosive blast loads. Specific load resistant design is achievable and may be conducted, where concerned with resilience against the effects of high explosive blast, with the guidance of the US Department of the Army (TM5-1300), Smith and Hetherington (2007), Byfield et al. (2007) and Byfield M.P. & Paramasivam S. (2008). Any problems encountered with the method are generally due to high cost. The method is best implemented as an offset with stand-off measures. If this cannot be achieved the structure will need to be designed for blast loads throughout regions of close proximity to vehicles, practice that is costly in terms of design, materials and construction labour. Furthermore, the method is highly dependent upon an accurate estimate of the blast and a design that exhausts all possible load scenarios for the full design life of the building.

5.3.4 ALTERNATIVE LOAD PATHS The use of ‘alternative load paths’ is endorsed by the British Standards, European Codes of Practice and the US General Services Administration design regulations (GSA 2003). Rather than design structural components to resist accidental or malicious loads, this method accepts localised damage to load-bearing members. Provided a viable alternative load path exists to redistribute load, the sensitivity of the building to local damage is significantly reduced thus presenting an economical substitute to ‘specific load design’ whilst reducing the risk associated with the prediction of likely threats. Whilst the level of protection afforded by this method is potentially quite high, guidance as to the design and implementation of such mitigation systems permits some variation. As a result several alternatives are available, a number of which are reviewed below. 425

5.3.4.1 Double span method This approach requires a full-moment resistant frame designed to accommodate the loss of any individual key element from the frame grid. Each bay is thus designed (under an accidental load such as 1.05DL + 0.33IL, as specified in BS 5950) for double the service span, making the building capable of bridging a damaged area, redistributing the emergency load by means of bending in the adjacent joints and structure. Many rigidly jointed structures can achieve this requirement with little difficulty since the ultimate limit state load, as used for the design of the service span, is significantly greater than the accidental limit state load used in consideration of the double span. The importance of incorporating full moment connections throughout a structural frame has been acknowledged by many engineers since as early as the Second World War. D.G. Christopherson (1945), J.F. Baker (1948) and Rhodes (1978) each stipulate the benefits of introducing connections capable of adopting the full moment capacity of the members which they connect. Significant gains in local and structural robustness were found in such practice, when compared equivalent structures of simple construction, which supports the use of double-span design in collapse mitigation and resilient design, demonstrating its potential in providing a high level of protection. Guidance for the implementation of this scheme is provided by the US GSA (2003). The code provides detailed requirements for design which go as far as to include dynamic load factors (to account for sudden column loss) that are not found in other codified forms of direct robust design. However, the Alfred P. Murrah Federal Building, Oklahoma, demonstrates the primary concern regarding the method; by designing for singular column loss, the structure may be more vulnerable in the event of multiple column failures than if it were of simple construction. In 1996 the building was subject to an estimated blast equivalent to 1,800 kg TNT. The vehicle bourn explosion produced a crater 9.15 m in diameter and 2.45 m deep and, initiated at a distance of only 4.8 m from the near stanchion, resulted in the failure of three columns at the front of the building and a fourth at the centre line. The subsequent progressive failure resulted in the loss of almost 50% of the of the total floor area of the building making the event a significant disproportionate collapse. This case supports the argument that design for the loss of singular key element may be insufficient and, furthermore, that design under the double-span method may, in the likely case of multiple column loss, exasperate progressive failure by causing ‘drag-down’. Further detriment can be found in the economy of the scheme as material and construction costs will be substantially greater than alternatives found in structures of simple construction. 5.3.4.2 Truss systems A favoured method in the robust design of Class 3 buildings, this scheme involves the integration of truss systems at intervals throughout the height of a building. The truss systems are designed such that an unsupported section of building can be effectively suspended and the emergency load transferred safely to the ground. This is conventionally achieved by means of out-rigger trusses designed to cantilever from a protected core structure. Compliance with the majority of codes available dictates that the alternative load path need only consider the notional removal of a single column and, as with the double span method, there is a significant risk associated with underestimating the potential emergency load. Should the core be subject to an emergency load superior to the design accidental load, a significant collapse is possible. A good example of truss system performance is that of the World Trade Centre towers. Outrigger trusses were located at the top of both buildings, between the 106th and 110th floors. Having suffered aircraft impact and subsequent blast, both buildings survived for more than 50 minutes before their eventual collapse. Figure 3 shows an image of the damage scenario experienced during the 2001 September 11 attack and demonstrates how the load was likely redistributed. The splice connections of the perimeter columns were capable of full axial continuity and load reversal, allowing the transfer of the sections’ tensile capacity throughout the height of the building. The ‘hat-truss’, incorporated to increase structural stiffness under wind loading and aid support of a transmission mast on the roof of each structure, provided the buildings with significant redundancy and the ability of suspend the unsupported region. The towers were each serviced by a central core which acted with the remaining perimeter stanchions to enable the truss systems to cantilever and redistributed to the ground. The collapse of the two towers in 2001 is regarded as one of the worst building disasters in history but, in retrospect, the performance of the building was remarkable. The 426

Figure 3. Schematic illustration of the damage scenario and subsequent load redistribution during Sept 11 attacks on the World Trade Centre.

extended period of collapse aided the escape of a significant number of the buildings’occupants and can be attributed directly to the marked resilience provided by the truss systems within the structure. 5.3.4.3 Shear panels – masonry & reinforced concrete infill The use of infill panelling was strongly advocated during the 1940’s. It was found that buildings with infill panels possessed a high global stiffness and numerous alternative load paths which, in practice, proved remarkably tenacious against the onset of collapse and greatly increased the capacity of the structure to bridge a damaged area. Concerned with the resilience of structures against aerial bombardment, Christopherson (1945) and Baker (1948) commented on the added robustness afforded by the inclusion of shear panels. They remarked that the presence of masonry panels would tend to compartmentalise a building, helping to localise direct damage, and reduce sensitivity to loss of support. It was noted that the likelihood of direct damage to columns, and possible primary collapse, was generally increased by inclusion of such panels but the advantage of its presence significantly outweighed the detrimental effects. The photograph provided by Figure 4 shows direct blast damage and primary collapse subsequent to the direct hit by a medium capacity bomb during the Second World War. The building featured was a seven storey steel frame construction with masonry infill panelling throughout, which sustained the loss of the four peripheral columns at its face. Primary collapse was, in this case, limited to the unseating of the transfer plate-girder at the first floor. It can be seen that the remaining five stories were left unsupported and that the masonry infill panels assisted in sustaining the building by corbelling from the undamaged structure. There is, as yet, no established form of guidance for the design and application of masonry infill panels as an alternative in emergency load path provision and thus its use is limited as an indirect form of robust design. And whilst further disadvantage may be found in the weight of the material, the potential debris, its architectural inflexibility and vulnerability to out-of-plane buckling, the potential redundancy provided by such panels presents a cheap and rudimentary 427

Figure 4.

1940’s steel frame having sustained local damage under near blast effects (J.F. Baker, 1948).

method for increasing structural robustness that would be ideal in retrofitting if not new design. Infill panels constructed during the 1940’s were typically of solid 230 mm thick brickwork, and whilst bay-widths were smaller than in modern design, this is a good minimum. The use of shear panels is not confined to masonry infill. Alternative materials, such as reinforced concrete, should be considered together with proprietary bracing systems and prefabricated modular units which, as demonstrated by the work of Lawson et al. (2007), possess a good ability in providing alternative load paths. Whilst more expensive, these alternatives are viable in direct robust design as their behaviour is more reliable and their bracing capacities are easily quantified. Further testimony to the performance of shear panels may be found in the work of Loizeaux and Osborn (2006). The potential of this system and the testimony of its performance indicate that the provision affords a high degree of resilience against collapse. 5.3.4.4 Arching action Arching action, otherwise known as compressive membrane action (CMA), is a phenomena known to enhance the strength of RC elements. The beneficial effects of this internal compression arch have been seen in beams, in-situ slabs, composite metal decking and bridge elements. Though dependant on the span-depth ratio and concrete compressive strength, tests have shown (Peel-Cross, 2001) that the ultimate strength of an element may be enhanced by some 50–300% over standard theoretical predictions established by yield line theory and design practice. The fundamental limitations regarding this form of emergency response are the guarantee of sufficient restraint to sustain CMA and the lack of practicable design methods. The former is a case of providing an appropriate degree of stiffness in the adjacent structure. This is easily achieved mid-structure but uncertain along the perimeter of buildings. Whilst design methods have been produced based upon rigid-plastic and elastic-plastic yield line theory, application in design is laborious. Current work, being conducted at the University of Southampton, aims to consolidate and simplify existing methods for an accurate and effective design solution. 5.3.4.5 Catenary action Catenary action or tensile membrane action (TMA) is generated once deflection of an element has reached half its own depth, following CMA. Indications of the performance of tying force provision as a form of collapse mitigation are relatively mixed. Recent studies, such as those by Byfield & Paramasivam (2007, 2010), Tyas (2010), Kuhlmann et al. (2007), Demonceau & Jaspart (2008) indicate that it is unlikely that catenary action can be achieved by the consideration of tensile requirements alone. The study shows, explicitly for steel buildings, that a significant rotation is required of connections for the design tying force, stipulated by BS 5950 or Eurocode 1, to be mobilised and that standard simple connects are not ductile enough to support this degree of deformation. Similarly, the ductility of RC beam-column connections has been reviewed by Merola 428

Figure 5.

Ronan Point collapse (1968).

(2009). The study has determined that the required rotation can be acquired but that the uppermost layer of reinforcement would fracture leaving a single layer of reinforcement to adopt the tensile force required to sustain the catenary. 5.3.5 KEY ELEMENT DESIGN The provision of key element design and notional element removal was introduced in 1970 following the Ronan Point collapse (1968). The 23-storey residential high-rise, constructed of load-bearing precast units, was subject to a significant progressive collapse initiated by an internal gas explosion on the 18th floor. The blast resulted in the loss of an external wall panel at the incident storey leaving the four over-bearing apartments, which were reliant on direct transfer of gravity loads, un-supported. As can be seen from the image provided in Fig. 5, the subsequent collapse resulted in the loss of the entire South-East corner of the building; debris from the overbearing apartments successively overloading each of the underlying units. Under Key Element design, elements found to cause disproportionate collapse under notional removal – loss of the lesser of 70 m2 or 15% of the immediate floor area, in accordance with the British Building Regulations (ODPM, 2004) – are designed to sustain a 34 kPa static load applied at each available face in turn. This statutory accidental load of 34 kPa was an estimate of the blast pressure that initiated the Ronan Point collapse. Key element design is unlike specific load design in that it does not require an accurate prediction of accidental or malicious loads and the dynamic response of the element is not considered. The statutory 34 kPa static accidental load, whilst specific to the incident gas explosion of Ronan Point, may not be sufficient under certain circumstances. High explosive blasts yield substantially greater peak over-pressures respective to the detonation of fuel air mixtures, and whilst their duration is much shorter, their damage potential is significant. The collapse of the Alfred P. Murrah Federal building (1996), as detailed above, is a key example of this possible shortfall. The three columns lost to the blast were subjected to a peak-over pressure to the order of 10,000 kPa (Paramasivam, 2008). This load may have been applied for only a few milli-seconds but the columns in question failed by brisance, in the near case, and shear in those two further from the blast origin. The structure was not designed for seismic action and whilst treating each lost column as a key element would have resulted in an increased shear capacity, it is likely that they would have been unable to sustain such an extreme load, certainly the near column would still have been devastated. This case raises serious concerns with regard to the ability of key element design in providing a comprehensive form of protection. Whilst a reasonable resilience in the event of a gas explosion and perhaps collision might be expected, attack by high explosive blast is unlikely to be accommodated 429

by this scheme of robust design – specific load resistance would be more appropriate. Furthermore, given its application to crucial elements, the inclusion of a key element effectively increases the sensitivity of a building to local damage and disproportionate collapse. It is only when an accidental load can be guaranteed to be less than a distributed 34 kN/m2 , or equivalent, that this scheme is effective. 5.3.6 EFFECTIVE TYING The provision of effective tying, as stipulated by Eurocode 1, originates from the British Building Regulations (ODPM, 2004) and was intended to provide a degree of continuity throughout a structure and a minimum level of robustness. Effective ties need only be applied in a horizontal direction for most structures whilst larger and more municipal buildings may be designed with effective horizontal and vertical ties as an alternative to notional column removal (alternative load path and key element design). Horizontal ties are incorporated about the periphery and at internal structural intersections between the walls, columns, beams and slabs to provide a continuous diaphragm at each storey level. Vertical ties are applied to vertical load bearing components and are detailed to provide a degree of continuity between each other and intersecting horizontal components. Effective ties for both the horizontal and vertical directions are designed in accordance with minimum requirements specified in the structural codes of individual construction materials and provide a form of indirect robust design whereby an intrinsic level of continuity reduces structural sensitivity to accidental loading and local damage and promotes collapse resistance by catenary action in the event of key element loss. 5.3.6.1 Minimum horizontal ties The importance of effectively tying structural systems was demonstrated by research conducted during the Second World War. Surveys of bomb damaged buildings, as documented by D.G. Christopherson (1945) and J.F. Baker (1948), showed that insufficiently tied systems were the route of the majority of primary and secondary collapses as inadequate continuity, especially between floor and wall systems, would tend to exasperate direct blast damage. These findings are emphasised by the collapse of the Droppin Well Bar, Ballykelly (1982). The low-rise construction featured precast hollow-core floor units that spanned from an external load-bearing wall to an internal steel frame. A device, estimated at 5 kg of Semtex high-explosive, was placed on the lower level beside one of the steel uprights within the building. Whilst the stanchion was relatively undamaged, the explosion resulted in the collapse of a significant proportion of overbearing floor as the internal blast displaced the external masonry wall, unseating the precast units from their bearing at each end. Forensic investigations indicated that the floor units were discontinuous through the structure. Should the floor units have been tied together (continuous over one or more supports) and effectively anchored to the external masonry, it is likely that the collapse would have been reduced. Indications of the performance of tying force provision as a form of collapse mitigation are relatively mixed. Recent studies, such as that by Byfield and Paramasivam (2007), indicate that it is unlikely that catenary action can be achieved by the consideration of tensile requirements alone. Their study shows, explicitly for steel buildings, that a significant rotation capacity is required of connections before the design tying force stipulated by EC1 or BS 5950 may be effective. Whilst the use of horizontal ties unquestionably provides a good level of basic robustness, research suggests that it cannot be relied upon to provide any form of comprehensive collapse resistance. 5.3.6.2 Horizontal & vertical tying The benefits of effective horizontal tying are enhanced by the addition of vertical ties. By providing a degree of continuity between components in each direction in both the vertical and horizontal plane, a fundamental level of robustness is provided that permits conveyance of accidental and emergency loads to adjacent structural elements thus increasing the resilience of a structure against accidental damage and primary collapse. However, the ability of this system to preventing disproportionate or progressive collapse is questionable. As with the ‘effective horizontal tie’ method, this is a form of indirect design that encourages the development of a catenary to prevent collapse in the event of support loss. However, the benefit 430

Table 1. Protection strategy rating. Protection method

Rating


Safe Stand off Distance

Highly secured Lightly secured
Specific local resistance
Alternate Load Paths
Truss-systems
Masonry infill panel
Double-span method
Arching action
Catenary action
Ket Elements
Minimum Tying



           

is not that pronounced from the application of horizontal ties alone. The use of vertical ties does, in the event of column loss, mobilise the overbearing structure and encourage catenary action throughout the building height. Whilst this is likely to enhance the resilience of the unsupported height, the lateral stiffness of the adjacent structure is called into question and the risk of dragdown is increased – a concern that is reflected by the work of implosion engineers Loizeaux and Osborn (2006) – though less so than buildings designed for double span. It seems that the only clear advantage from including this measure would be if it is tied into an ancillary alternative load path that would reliably redistribute the emergency load. 5.3.7 DISCUSSION AND CONCLUSIONS FOR DAMAGE PREVENTION FROM BLAST AND IMPACT It is important that the structural robustness level achieved by a building is appropriate for the intended use and this review of case studies reveals a wide variation in the ability of buildings to survive localised damage from blast or impact without subsequent collapse. It is however possible to rank the different robustness design strategies, from those with a high level of robustness to those which provide only a nominal level of protection. As a general guide this chapter uses the review of case-studies to develop a ranking system with regards to the level of protection achieved and this is shown in Table 1. It is hoped that this can be used as a guide when selecting appropriate robustness strategies for design. In terms of the protection against VBIED’s, the Khobar Towers demonstrated that simple barrier systems can provide an impressive level of protection. For this reason the provision of a safe standoff distance is regarded as the best method of protecting buildings and is awarded 5 stars. Since suicide bombers have a history of penetrating lightly secured check posts this rating only applies when gates are highly secured. It is not always possible to provide adequate safe stand-off distances for buildings in congested urban environments. If VBIED’s are considered a threat then the engineer and client may wish to consider the option of designing a hardened structure and past experience has shown that it is indeed possible to design structural frames to survive severe bomb blasts. This method has been ranked with 4½ stars because casualties are inevitable due to the close proximity. The evidence from buildings designed to survive earthquakes, but were subsequently subjected to explosions from bomb blast suggests that seismic design achieves a high degree of robustness against blast and explosions. Therefore seismic design could arguable be ranked highly. The provision of emergency bracing systems in the form of out-rigger trusses is also ranked with 4½ stars due to the proven ability of these systems to accommodate multiple column loss. The use of masonry infill panelling along column gridlines is ranked highly with 4 stars. This is due to the well proven ability of internal walls to inhibit the passage of blast waves inside buildings, as well as to redistribute loads following damage to multiple columns. In 1996, the Murrah Building was subject to a blast that caused the failure of three columns along the front face of the building. This illustrates the potential to lose support to multiple columns 431

and it is for this reason that the double span method (or notional column removal) is ranked with 3 stars, due to the inability of the method to cope with multiple column loss. The adverse effect of building continuity into a frame to accommodate notional column removal introduces the danger of drag down, whereby the weight of the damaged structure overloads surviving columns and thereby enables the collapse to spread sideways, rather than being localised to the bays previously supported by damaged columns. Further detriment can be found in the economy of the method as fabrication costs can be higher than that of an alternative structure featuring simple connections. Catenary action as applied to steel framed structures with industry standard “pinned” connections has been largely discredited, with joints shown to rip apart due to insufficient ductility when subjected to the demands from catenary action. This problem of joint rotation capacity will also occur with partial-strength connections since beams will remain elastic and therefore place all rotation demands onto the connections. Steel framed structures incorporating full moment joints would survive the demands of catenary action although the continuity may introduce dangers from drag-down. The mechanics of catenary action in r.c. framed structures are not yet understood and therefore, for this combination of reasons catenary action is ranked with 1 star. The ability of arching action to redistribute loads from columns is not presently understood and for this reason it is ranked with 1 star, although research may demonstrate that many r.c. frames can indeed support individually damaged columns through arching action. The key element method load of 34 kN/m2 is not close to blast loads from high explosives and therefore it will not provide resistance to VBIED’s. For this reason it is ranked with 2 stars. The use of horizontal ties unquestionably provides a good level of basic robustness against blast loading, although for the reasons stated earlier it will not provide support through catenary action to damaged columns. Tying members together can also create risks from drag down, as illustrated by a recent paper on progressive collapse from a demolition contractors perspective, Loizeaux & Osborn (2006), which commented that precast concrete structures with weakly tied together joints can be difficult to demolish because damage tends to be localised to bays in which columns have been removed. From a demolition perspective it is preferably to have precast elements well tied together so that collapsing members drag down the ones in the adjacent bays. The final combination of robustness measures is the requirement to tie members both horizontally and vertically. This is again given 2 stars, because this method is only effective if the columns are anchored to a stiff bearing somewhere higher up in the structure, as illustrated in Fig. 10. The Building Regulations include no requirement for this anchorage, although if one were included then this method could arguable be raised to 4½ stars.

REFERENCES Baker J.F., et al. (1948) The Design of Framed Buildings Against High-Explosive Bombs. The Civil Engineer In War, Vol. III – Properties of Materials, Structures, Hydraulics, Tunnelling and Surveying. London: Institution of Civil Engineers, (80–113). British Standards Institution (BSI) (2000) BS 5950-1:2000: Structural use of Steelwork in Building – Part 1: Code of Practice for Design – Rolled & Welded Sections. London: BSI. British Standards Institution (BSI) (2005) BS 5628-1:2005: Code of Practice for the use of Masonry – Part 1: Structural use of Unreinforced Masonry. London: BSI. Byfield M. & Paramasivam S. (2007). The prevention of disproportionate collapse using catenary action. Urban Habitat Constructions under Catastrophic Events, Proceedings of Workshop in Prague. Byfield M., De Matteis G., Italy, Dinu F. (2007). Robust design of steel framed buildings against extreme loading. Urban Habitat Constructions under Catastrophic Events, Proceedings of Workshop in Prague. Byfield M.P. & Paramasivam S. (2007) Catenary Action in Steel-framed Buildings. Structures & Buildings, Vol. 160(SB5), 247–257. Byfield M.P. & Paramasivam S. (2008). Byfield M.P., Paramasivam S. Protective design of RC framed building against blast loading through safe stand off distance approach. Urban Habitat Constructions under Catastrophic Events, Proceedings of the International Symposium in Malta. Byfield M.P. & Paramasivam S. (2010). Limitations of the tying force method for providing robustness in steel framed buildings. Urban Habitat Constructions under Catastrophic Events, Proceedings of the Final Conference in Naples.

432

Byfield, M.P. (2006). Behavior and Design of Commercial Multistory Buildings Subjected to Blast. J. Perf. Constr. Fac. Vol. 20, Issue 4, pp. 324–329. Christopherson D.G. (1945) Structural Defence. London: Ministry of Home Security Research & Experiments Department. Demonceau J.F. & Jaspart J.P. (2008). Experimental and analytical investigations on the behaviour of building frames further to a column loss. Urban Habitat Constructions under Catastrophic Events, Proceedings of the International Symposium in Malta. Department of Army (1990) Structures to Resist the Effects of Accidental Explosions. Publication No. TM51300., Washington, D.C. Garcia et al. (2005) Effects of the Terrorist Attack on the El Nogal Building in Bogota, Columbia. IABSE Symposium, Lisbon 2005. House Armed Services Committee, 1996 (I’ll sort out the rest of the reference). Kuhlmann U., Rölle L., Jaspart J.-P., Demonceau J.-F. (2007). Robustness – robust structures by joint ductility. Urban Habitat Constructions under Catastrophic Events, Proceedings of Workshop in Prague. Kuhlmann U. et al. (2008). Ductile and partial-strength steel & composite joints as basis for redundant and robust frame structures. Urban Habitat Constructions under Catastrophic Events, Proceedings of the International Symposium in Malta. Lawson P.M. et al. (2008) Robustness of Light Steel Frames & Modular Construction. Structures & Buildings, Vol. 161(SB1), 3–16. Loizeaux M. & Osborn A. (2006) Progressive Collapse – An Implosion Contractor’s Stock in Trade. Journal of Performance of Constructed Facilities, Vol. 20(04), 391–402. Merola R. (2009) Ductility and robustness of concrete structures under accidental and malicious load cases. PhD thesis, University of Birmingham. Office of Deputy Prime Minister (2004) The Building Regulations 2000. London: ODPM. Paramasivam, S. (2008). Protective Design against Disproportionate Collapse of RC and Steel Framed Structures. Thesis (PhD), Univ. of Southampton, Southampton, UK. Peel-Cross J., Rankin G., Gilbert S. & Long A. (2001) Compressive membrane action in composite floor slabs in Cardington LBTF. Structures & Buildings, vol. 146(02), p217–226 Smith P.D. & Hetherington J.G. (1994) Blast and Ballistic Loading of Structures. Elsevier Science & Technology. Smith P.D. & Tyas A. (2008). Blast load assessment by simplified and advanced methods. Urban Habitat Constructions under Catastrophic Events, Proceedings of the International Symposium in Malta. Smith P.D. (2008). The effectiveness of blast walls. Urban Habitat Constructions under Catastrophic Events, Proceedings of the International Symposium in Malta. Smith P.D. (2007). State of the art in Europe and activity developed within WG3 Impact and Explosion. Urban Habitat Constructions under Catastrophic Events, Proceedings of Workshop in Prague. Tyas A. (2010). Experimental studies of semirigid steel connections subjected to impulsive loading. Urban Habitat Constructions under Catastrophic Events, Proceedings of the Final Conference in Naples. U.S. General Services Administration (GSA) (2003) Progressive Collapse Analysis & Design Guidelines for New Federal Office Buildings & Major Modernisation Projects. Washington, D.C.: GSA.

433

Urban Habitat Constructions under Catastrophic Events (Final Report) – Mazzolani (Ed). © 2010 Taylor & Francis Group, London, ISBN 978-0-415-60686-8

5.4 Multi-hazard risk assessment methodology M.H. Faber & H. Narasimhan ETH Zurich, Zurich, Switzerland

5.4.1 INTRODUCTION The development and management of societal infrastructure is a central task for the continued success of society. The decision processes involved in this task concern all aspects of managing and performing the planning, investigations, designing, manufacturing, execution, operations, maintenance and decommissioning of objects of societal infrastructure, such as traffic infra-structure, housing, power generation, power distribution systems and water distribution systems. During the planning and execution of engineering facilities and activities, the management of risks should ideally be based on a holistic perspective considering all possible events which may lead to and/or influence consequences of any sort. In reality such a seamless assessment and management of risks is difficult to realize due to the way in which engineering facilities and activities are planned and organized. The basic premises required for the utilization of risk assessment to establish rational decisions for the benefit of society and other stakeholders and consistent with societal preferences are described here. This can be used by decision makers and professionals responsible for or involved in establishing engineering decision support. The broad objective from a societal perspective by such activities is to improve the quality of life of the individuals of society both for the present and the future generations.

5.4.2 FRAMEWORK FOR RISK ASSESSMENT 5.4.2.1 Risk assessment and decision making If all aspects of a decision problem were known with certainty, the identification of optimal decisions would be straightforward by means of traditional cost-benefit analysis. However, due to the fact that our understanding of the aspects involved in the decision problems often is far less than perfect and that we are only able to model the involved physical processes as well as human interactions in rather uncertain terms, the decision problems in engineering are subject to significant uncertainty. Due to this it is not possible to assess the result of decisions and consequences in certain terms. However, what can be assessed is the risk associated with the different decision alternatives. If the concept of risk as the simple product between probability of occurrence of an event with consequences and the consequence of the event is widened to include also the aspects of the benefit achieved from the decisions, risk may be related directly to the concept of utility (von Neumann and Morgenstern 1944, Raiffa and Schlaifer 1961) from the economic decision theory. A whole methodical framework is thus made available for the consistent identification of optimal decisions. This framework is considered to comprise the theoretical basis for risk based decision making. Based on these principles, a guideline document (JCSS 2008) describing the framework and principles for risk based engineering decision making has been recently developed by the Joint Committee on Structural Safety (JCSS). The following sections of this chapter describe the main features of this framework. 435

5.4.2.2 Decisions and decision makers A decision may be understood as a committed allocation of resources made by a decision maker. The decision maker is an authority or person who has authority over the resources being allocated and responsibility for the consequences of the decision to third parties. The intention of the decision maker is to meet some objective, of a value to the decision maker which at least is in balance with the resources allocated by the decision. The decision maker is faced with the problem of choosing between a set of decision alternatives which may lead to different consequences in terms of losses and benefits. The objective aimed for by the decision making represents the preference of the decision maker in weighing the different attributes which may be associated with the possible consequences of the decision alternatives. It is thus clear that the formulation of the decision problem will depend very much on the decision maker. Who are the stakeholders, the beneficiaries and the responsible parties? Each possible decision maker will have different viewpoints in regard to preferences, attributes and objectives. It is important to identify the decision maker since the selection and weighting of attributes must be made on behalf of the decision maker. 5.4.2.3 Attributes of decision outcomes There are essentially three types of attributes - natural, constructed and proxy. Natural attributes are those having a common interpretation to everyone (cost in dollars, number of fatalities and other measurable quantities). For many important objectives, such as improving image and increasing international prestige, it is difficult or impossible to come up with natural attributes. The attributes to be used must essentially define what is meant by the objective. Constructed attributes may be used for this, these are made up of verbal descriptions of several distinct levels of impact that directly indicate the degree to which the associated objective is achieved and a numerical indicator is assigned to these levels. Examples of constructed attributes turning into natural attributes with time and use are gross national product GNP (aggregate of several factors to indicate economic activity of a country), Dow Jones industrial average etc. Finally, there are cases where it is difficult to identify either type of attribute for a given objective. In these cases indirect measurements may be used. The attributes used to indicate the degree to which the objective is achieved is called proxy attributes. When an attribute is used as proxy attributes for a fundamental objective, levels of that attribute are valued only for their perceived relationship to the achievement of that fundamental objective. The decision maker will make decisions consistent with her/his values, which are those things that are important to her/him, especially those that are relevant to her/his decision. A common value is economic, according to which the decision maker will attempt to increase his wealth. Others might be personal, such as happiness or security, or social, such as fairness. 5.4.2.4 Preferences among attributes – evaluation of utility Having determined the set of attributes, the objectives must be quantified with a value/utility model. This is done by means of converting the attribute values to a value scale by means of judgment of relative value or preference strength. The value scale is often referred to as a utility function. In some cases it may not appear obvious how to directly transfer different attribute values into one common value scale. To overcome this apparent problem it is possible to consider multi-attribute decision problems. However, it is emphasized that the solution to a multi-attribute will imply a weighing of the different attributes against each other and more transparency in the decision process is thus achieved by making this weighing directly. The multi-attribute value problem is a problem of value trade-offs. These trade-offs can be systematically structured in utility functions. These are scalar valued functions defined on the consequence space, which serve to compare various levels of the different attributes indirectly. Given the utility function the decision maker’s problem is to choose that alternative from the space of feasible alternatives that maximizes the expected utility. The expected utility is used as a relative measure making it possible to choose between various actions. The action with the largest expected utility will be chosen from among the possible actions. Thus, no absolute criterion for the acceptability of the considered action is given from decision theory. 436

5.4.3 SYSTEM MODELING 5.4.3.1 Introduction In a societal context, risk based decision making needs to be understood from an intergenerational perspective. Within each generation decisions have to be made which will not only affect the concerned generation but all subsequent generations. At an intra-generational level, the characteristics of the system consist of the knowledge about the considered engineered facility and the surrounding world, the available decision alternatives and criteria (preferences) for assessing the utility associated with the different decision alternatives. A very significant part of risk based decision making in practice is concerned about the identification of the characteristics of the facility and the interrelations with the surrounding world as well as the identification of acceptance criteria, possible consequences and their probabilities of occurrence. Managing risks is done by “buying” physical changes of the considered facility or “buying” knowledge about the facility and the surrounding world such that the objectives of the decision making are optimized. A system representation can be performed in terms of logically interrelated constituents at various levels of detail or scale in time and space. Constituents may be physical components, procedural processes and human activities. The appropriate level of detail or scale depends on the physical or procedural characteristics or any other logical entity of the considered problem as well as the spatial and temporal characteristics of consequences. The important issue when a system model is developed is that it facilitates a risk assessment and risk ranking of decision alternatives which is consistent with available knowledge about the system and which facilitates that risks may be updated according to knowledge which may be available at future times. 5.4.3.2 Knowledge and uncertainty Knowledge about the considered decision context is a main success factor for optimal decision making. In real world decision making lack of knowledge (or uncertainty) characterizes the normal situation and it is thus necessary to be able to represent and deal with this uncertainty in a consistent manner. The Bayesian statistics provides a basis for the consistent representation of uncertainty independent of their source and readily facilitates for the joint consideration of purely subjectively assessed uncertainties, analytically assessed uncertainties and evidence as obtained through observations. There exist a large number of propositions for the characterization of different types of uncertainties. It has become standard to differentiate between uncertainties due to inherent natural variability, model un-certainties and statistical uncertainties. Whereas the first mentioned type of uncertainty is often denoted aleatory (or Type 1) uncertainty, the two latter are referred to as epistemic (or Type 2) uncertainties. However, this differentiation is introduced for the purpose of setting focus on how uncertainty may be reduced, rather than calling for a differentiated treatment in the decision analysis. 5.4.3.3 System representation The risk assessment for a given system is facilitated by considering the generic representation of the development of consequences shown in Figures 1 and 2. Following Faber and Maes (2005), the exposure to an engineering facility is represented as a set of different exposure events acting on the constituents of the facility. The constituents of the facility can be considered as the facility’s first defense in regard to the exposures. The damages of the constituents are considered to be associated with direct consequences. Direct consequences may include monetary losses, loss of lives, damages to the qualities of the environment or just changed characteristics of the constituents. Based on the combination of events of constituent failures and the corresponding consequences, indirect consequences may occur. Indirect consequences may be caused by e.g. the sum of monetary losses associated with the constituent failures and the physical changes of the facility as a whole caused by the combined effect of constituent failures. The indirect consequences in risk assessment play a major role and their modeling (Faber and Maes 2004). Typically the indirect consequences evolve spatially beyond the boundaries of the facility and also have a certain sometimes even postponed development in time. 437

Figure 1. Generic representation used for the risk assessment of a system.

Figure 2. Logical representation of interrelation between exposures, constituent failures, sequences of constituent failures and consequences.

Figure 3. Generic system characterizations at different scales in terms of exposure, vulnerability and robustness.

The vulnerability of a give system (facility and the rest of the world) characterizes the risk associated with the direct consequences and the robustness characterizes the degree to which the total risk is increased beyond the direct consequences. Often the constituents in a facility can be modeled as a logical system comprised by its own constituents. For instance, a highway network facility could be modeled with constituents being bridges; this is shown in Figure 3. The bridges in turn could be modeled by logical systems with constituents being structural members. Depending on the level of detail in the risk assessment, the system definition, the exposure, constituents and consequences would be different. The hierarchical risk assessment framework is applicable at any level of scale for the assessment of a given system. It may be applied to components, sub-systems and the system as a whole; thereby the framework also facilitates a hierarchical approach to risk assessment. The definition of the system in this context becomes of tremendous significance in the definition of exposure, vulnerability and robustness. 438

5.4.4 EXPOSURES AND HAZARDS The exposures and hazards for a system are defined as all possible endogenous and exogenous effects with potential consequences for the considered system. A probabilistic characterization of the exposure to a system requires a joint probabilistic model for all relevant effects relative to time and space.

5.4.5 CONSEQUENCES 5.4.5.1 Vulnerability The vulnerability of a system is related to the direct consequences caused by the damages of the constituents of a system for a given exposure event. The damage of the constituents of a system represents the damage state of the system. In risk terms, the vulnerability of a system is defined through the risk associated with all possible direct consequences integrated (or summed up) over all possible exposure events. 5.4.5.2 Robustness The robustness of a system is related to the ability of a considered system to sustain a given damage state subject to the prevailing exposure conditions and thereby limit the consequences of exposure events to the direct consequences. It is of importance to note that the indirect consequences for a system not only depend on the damage state, but also the exposure of the damaged system. When the robustness of a system is assessed, it is thus necessary to assess the probability of indirect consequences as an expected value over all possible damage states and exposure events. A conditional robustness may be defined through the robustness conditional on a given exposure and or a given damage state.

5.4.6 RISK ASSESSMENT 5.4.6.1 Indicators of risk The presented risk assessment framework facilitates a Bayesian approach to risk assessment and full utilization of risk indicators. Risk indicators may be understood as any observable or measurable characteristic of the system or its constituents containing information about the risk. If the system representation has been performed appropriately, risk indicators will in general be available for what concerns the exposure to the system, the vulnerability of the system and the robustness of the system. In a Bayesian framework for risk based decision making such indicators play an important role. Considering the risk assessment of a load bearing structure risk indicators are e.g. any observable quantity which can be related to the loading of the structure (exposure), the strength of the components of the structure (vulnerability) and the redundancy, ductility, effectiveness of condition control and maintenance (robustness). 5.4.6.2 Analysis and quantification of risk Following the assessment and evaluation of the exposures/hazards, vulnerability and consequences associated with the system considered for risk assessment, the ensuing risks then need to be quantified and evaluated. For this purpose, the system considered for the risk assessment is assumed to be exposed to hazardous events (EX ) with probabilistic characterization p(EXk ), k = 1, nEXP , where nEXP denotes the number of exposures and hazards. It is assumed that there are nCON individual constituents of the system, each with a discrete set (can easily be generalized to the continuous case) of damage states Cij , i = 1, 2..nCON , j = 1, 2..nCi . The probability of direct consequences cD (Cl ) associated with the lth of nCSTA possible different state of damage of all constituents of the facility Cl , conditional on the exposure event EXk is described by p(Cl |EXk ) and the associated conditional risk is p(Cl |EX k )cD (Cl ). The vulnerability of the system is defined as the risk due to 439

all direct consequences (for all nCON constituents) and may be assessed through the expected value of the conditional risk due to direct consequences over all nEXP possible exposure events and all constituent damage states nCSTA :

The state of the facility as a system depends on the state of the constituents. It is assumed that there is nSSTA possible different system states Sm associated with indirect consequencescID (Sm , cD (Cl )). The probability of indirect consequences conditional on a given state of the constituents Cl , the direct consequences cd (Cl ) and the exposure EXk , is described by p(Sm |Cl , EXk ). The corresponding conditional risk is p(Sm |Cl , EXk )cID (Sm , cD (Cl )). The risk due to indirect consequences is assessed through the expected value of the indirect consequences in regard to all possible exposures and constituent states, as:

The robustness of a system is defined as the ability of a system to limit total consequences to direct consequences. This characteristic may readily be quantified though the index of robustness IR (Baker et al. 2008):

which allows for a ranking of decisions in regard to their effect on robustness. 5.4.6.3 Treatment of risk The various possibilities for collecting additional in-formation in regard to the uncertainties associated with the understanding of the system performance as well as for changes the characteristics of the system can be considered to comprise the total set of options for risk treatment. The risk treatment options may, in the context of risk based decision making, be considered as the available decision alternatives. Risk treatment is decided upon for the purpose of optimize the expected utility to be achieved by the decision making. Risk treatment can be implemented at different levels in the system representation, namely in regard to the exposure, the vulnerability and the robustness, as shown in Figure 1. Considering the risk assessment of a load carrying structure, risk treatment by means of knowledge improvement may be performed by collecting information about the statistical characteristics of the loading (exposure), the strength characteristics of the individual components of the structures (vulnerability) and by systems reliability of the structural system (robustness). Risk treatment through changes of the system characteristics may be achieved by restricting the use of the structure (exposure), by strengthening the components of the structure (vulnerability) and by increasing the redundancy of the structural system (robustness). 5.4.6.4 Acceptance of risk and life quality index In addition to risks due to economic losses, the decision maker has to take into account also the risks to individuals as well as potential damages to qualities of the environment. It is hence useful to differentiate between tangible and intangible risks, i.e. risks which may easily be expressed in monetary terms and risks which are not. Intangible values especially concern loss of lives and injuries and also qualities of the environment. The Life Quality Index (LQI) is a measure that facilitates the development of risk acceptance criteria for intangible risks (Nathwani et al., 1997). It is based on demographical indicators that include the incremental increase in life expectancy through risk reduction, the corresponding loss of economic resources, measured through the Gross National Product (GNP) together with the time 440

used for work, all assessed for a statistical life in a given society. The underlying idea of the LQI is to model the preferences of a society quantitatively as a scalar valued social indicator, comprised by a relationship between the GDP per capita, the expected life at birth and the proportion of life spend for earning at living. Based on the theory of socio-economics, the Life Quality Index can be expressed in the following principal form:

where the parameter r is a measure of the trade-off between the resources available for consumption and the value of the time of healthy life. 5.4.6.5 Sustainable discounting in decision making Discounting of investments can have a rather significant effect on decision making. Especially in the context of planning of societal infrastructure for which relative long life times are desired and for which also the costs of maintenance and decommissioning must be taken into account, the assumptions in regard to discounting are of importance. Considering time horizons of 20 to 100 years (i.e. over several generations), discounting should be based on long term average values, free of taxes and inflation. In the private sector, the long term real rate of interest is approximately equal to the return which may be expected from an investment. In the public sector, the discounting rate, in the context of life saving investments, should correspond to the real rate of economic growth per capita. 5.4.6.6 Perception and communication of risk Different individuals in society perceive risks differently, depending on their own situation in terms of to what degree they may be affected by the exposures, to what degree they are able to influence the risks and to what degree the risks are voluntary. Generally risks are perceived more negatively when stake holders feel more exposed, when they feel they have no influence and they feel they are exposed to risks involuntary. Another aspect is related to how adverse events are perceived by individuals and groups of individuals in society when and after such events take place. Again, this depends on the perspective of the affected individuals and groups of individuals. Furthermore, the occurrence of adverse events and the way the information about such events is made available will affect the perception of risks in general but also in many cases trigger actions which have no rational basis and only adds to the societal consequences of such event. Due to the effects of the perception of risk, it is generally observed that different individuals and groups of individuals have different attitudes in regard to what risks can be accepted. Risk averse and risk prone attitudes are observed, which simply refers to the effect that risks are assigned different tastes depending on these characteristics. Whereas such behavior is a private matter for individuals of society, it leads to an uneven distribution of risks, if exercised in the context of societal decision making and this is clearly unethical and not rational. The perception of risks may be significantly influenced by information about the risks themselves. Information can and should be used as a targeted means of reducing potential losses caused by reactions to events beyond what is rational, seen in the perspective of normative decision making. Being provided with transparent information in regard to the nature of exposures, possible precautionary actions, information on how risks are being managed and the societal consequences of irrational behavior reduces uncertainties associated with the understanding of risks of individuals. This, in turn, adds to rational behavior and thereby reduces follow-up consequences.

5.4.7 CONCLUSIONS Engineering decision making is a complex issue often due to very significant potential consequences and substantial uncertainties. The continued successful development of society as well as the general competitiveness in engineering depends on the efficiency of identified options for the management 441

of risks as well as for the communication of the basis for decision making to all stakeholders. This situation calls for the development of a unified framework for risk based decision making which is general enough to accommodate for the special needs of different application areas but at the same time specific enough to ensure a sufficient degree of consistency in modeling and theoretical basis. This chapter presents an outline of such a framework, recently developed by the Joint Committee on Structural Safety (JCSS 2008). The presented framework should be seen as a general philosophy and a set of principles for risk based decision making rather that an operational tool box. It is implicitly understood that the user will appreciate the need for engaging experts and or appropriate tools in the implementation of the proposed framework. REFERENCES Baker, J.W., Schubert, M. & Faber, M.H. 2008. On the Assessment of Robustness. Structural Safety 30: 253–267. Faber, M.H. & Maes, M.A. 2004. Modeling of Risk Perception in Engineering Decision Analysis. In: M.A. Maes & L. Huyse (eds.) Reliability and Optimization of Structural Systems; Proceedings of the 11th IFIP WG7.5 Working Conference, Banff, Canada, 2–5 November 2003. Rotterdam: Balkema. Faber, M.H. & Maes, M.A. 2005. On Applied Engineering Decision Making for Society. In J.D. Sorensen & D. Frangopol (eds.) Reliability and Optimization of Structural Systems; Proceedings of the 12th IFIP WG7.5 Working Conference, Aalborg, Denmark, 22–25 May 2005. Rotterdam: Balkema. JCSS 2008. RiskAssessment in Engineering. Internet Publication: http://www.jcss.ethz.ch/publications/JCSS_ RiskAssessment.pdf Nathwani, J.S., Lind, N.C. & Pandey, M.D. 1997. Affordable Safety by Choice: The Life Quality Method. Waterloo: University of Waterloo. Raiffa, H. & Schlaifer, R. 1961. Applied Statistical Decision Theory. Boston: Harvard University. von Neumann, J. & Morgenstern, O. 1944. Theory of Games and Economic Behavior. Princeton: Princeton University Press.

442

Urban Habitat Constructions under Catastrophic Events (Final Report) – Mazzolani (Ed). © 2010 Taylor & Francis Group, London, ISBN 978-0-415-60686-8

5.5 A framework and guidelines for volcanic risk assessment H. Narasimhan ETH Zurich, Zurich, Switzerland

R.P. Borg Faculty for the Built Environment, University of Malta, Malta

G. Zuccaro PLINIVS Centre and University of Naples “Federico II”, Naples, Italy

M.H. Faber ETH Zurich, Zurich, Switzerland

D. De Gregorio, B. Faggiano, A. Formisano & F.M. Mazzolani University of Naples “Federico II”, Naples, Italy

M. Indirli ENEA, Bologna, Italy

5.5.1 BACKGROUND Natural hazards such as earthquakes, floods, volcanoes and tsunamis constitute a significant source of risk in several regions of the world and are often associated with widespread loss of human lives, damage to the qualities of the environment as well as to property and infrastructure. It is hence a great challenge for the engineering profession to provide methods and tools enhancing decision making for the purpose of efficient management of natural hazards. Since our understanding of the aspects involved in such decision problems is often less than perfect and that we are only able to model the involved physical processes as well as human interactions in rather uncertain terms, the decision problems in engineering are subject to significant uncertainty. Due to this, it is not possible to assess the result of decisions and consequences in certain terms. However, what can be assessed is the risk associated with the different decision alternatives. If the concept of risk as the simple product between probability of occurrence of an event with consequences and the consequence of the event is widened to include also the aspects of the benefit achieved from the decisions, risk may be related directly to the concept of utility (von Neumann and Morgenstern 1944, Raiffa and Schlaifer 1961) from the economic decision theory. A whole methodological framework is thus made available for the consistent identification of optimal decisions. This framework is considered to comprise the theoretical basis for risk based decision making. Based on these principles, a document (JCSS 2008) describing the framework and principles for risk based engineering decision making has been recently developed by the Joint Committee on Structural Safety (JCSS). In this paper, the main features of this framework are first briefly described. Then, aspects related to the modelling of the hazard process due to volcanoes are discussed. A system of classification of structures and identification of different building characteristics that could be used for volcanic vulnerability and risk assessment is then proposed. This is followed by a discussion on the fragility and vulnerability modeling of structures relevant for seismic analysis. Finally, general issues dealing with the quantification of risk and their treatment and communication are covered. 443

5.5.2 SYSTEM MODELLING IN RISK ASSESSMENT 5.5.2.1 System identification and representation In a societal context, risk based decision making needs to be understood from an intergenerational perspective. Within each generation decisions have to be made which will not only affect the concerned generation but all subsequent generations. At an intra-generational level, the characteristics of the system consist of the knowledge about the considered engineered facility and the surrounding world, the available decision alternatives and criteria (preferences) for assessing the utility associated with the different decision alternatives. A very significant part of risk based decision making in practice is concerned about the identification of the characteristics of the facility and the interrelations with the surrounding world as well as the identification of acceptance criteria, possible consequences and their probabilities of occurrence. Managing risks is done by “buying” physical changes of the considered facility or “buying” knowledge about the facility and the surrounding world such that the objectives of the decision making are optimized. A system representation can be performed in terms of logically interrelated constituents at various levels of detail or scale in time and space. Constituents may be physical components, procedural processes and human activities. The appropriate level of detail or scale depends on the physical or procedural characteristics or any other logical entity of the considered problem as well as the spatial and temporal characteristics of consequences. The important issue when a system model is developed is that it facilitates a risk assessment and risk ranking of decision alternatives which is consistent with available knowledge about the system and which facilitates that risks may be updated according to knowledge which may be available at future times. 5.5.2.2 Knowledge and uncertainty Knowledge about the considered decision context is a main success factor for optimal decision making. In real world decision making lack of knowledge (or uncertainty) characterizes the normal situation and it is thus necessary to be able to represent and deal with this uncertainty in a consistent manner. The Bayesian statistics provides a basis for the consistent representation of uncertainty independent of their source and readily facilitates for the joint consideration of purely subjectively assessed uncertainties, analytically assessed uncertainties and evidence as obtained through observations. There exist a large number of propositions for the characterization of different types of uncertainties. It has become standard to differentiate between uncertainties due to inherent natural variability, model uncertainties and statistical uncertainties. Whereas the first mentioned type of uncertainty is often denoted aleatory (or Type 1) uncertainty, the two latter are referred to as epistemic (or Type 2) uncertainties. However, this differentiation is introduced for the purpose of setting focus on how uncertainty may be reduced, rather than calling for a differentiated treatment in the decision analysis. 5.5.2.3 System representation The risk assessment of a given system is facilitated by considering the generic representation illustrated in Figure 1. The exposure to the system is represented as different exposure events acting on the constituents of the system. The constituents of the system can be considered as the system’s first defence in regard to the exposures. The damages of the system caused by failures of the constituents are considered to be associated with direct consequences. Direct consequences may comprise different attributes of the system such as monetary losses, loss of lives, damages to the qualities of the environment or just changed characteristics of the constituents. Based on the combination of events of constituent failures and the corresponding consequences follow-up consequences may occur. Follow-up consequences could be caused by e.g. the sum of monetary losses associated with the constituent failures and the physical changes of the system as a whole caused by the combined effect of constituent failures. The follow-up consequences in systems risk assessment play a major role, and the modelling of these should be given great emphasis. It should be noted that any constituent in a system can be modelled as a system itself. In the context of volcanic risk assessment, a system could be an urban area exposed to the effects of a volcano with its constituents being buildings, structures and lifelines. The buildings and structures, in turn, could also be systems with 444

Figure 1.

Generic representation used for the risk assessment of a system.

structural members as their constituents. Depending on the level of detail in the risk assessment, the system definition, the exposure, constituents and consequences would be different. 5.5.2.4 Exposures and hazards The exposure to a system is defined as all possible endogenous and exogenous effects with potential consequences for the considered system. A probabilistic characterization of the exposure to a system requires a joint probabilistic model for all relevant effects relative to time and space.

5.5.3 CONSEQUENCES 5.5.3.1 Vulnerability The vulnerability of a system is related to the direct consequences caused by the damages of the constituents of a system for a given exposure event. The damage of the constituents of a system represents the damage state of the system. In risk terms, the vulnerability of a system is defined through the risk associated with all possible direct consequences integrated (or summed up) over all possible exposure events. 5.5.3.2 Robustness The robustness of a system is related to the ability of a considered system to sustain a given damage state subject to the prevailing exposure conditions and thereby limit the consequences of exposure events to the direct consequences. It is of importance to note that the indirect consequences for a system not only depend on the damage state, but also the exposure of the damaged system. When the robustness of a system is assessed, it is thus necessary to assess the probability of indirect consequences as an expected value over all possible damage states and exposure events. A conditional robustness may be defined through the robustness conditional on a given exposure and or a given damage state. 5.5.3.3 Large scale indicator based risk modeling The risk management of large scale natural hazards requires a systematic and consistent representation and management of information for a typically complex system with a large number of constituents or sub-systems. Such representation must enable a rational treatment and quantification of the various uncertainties associated with the constituents as well as the system. The consistent handling of new knowledge about the system and its constituents as and when it becomes available and its use in the risk assessment and decision making process is also essential. Further, 445

the numerous dependencies and linkages that exist between different constituents of the system need to be systematically considered. The above requirements and considerations necessitate the use of generic indicator based risk models for the assessment and management of risks due to natural hazards. Risk indicators can be understood as any observable or measurable characteristic of the system or its constituents containing information about the risk. If the system representation has been performed appropriately, risk indicators will in general be available for what concerns the exposure to the system, the vulnerability of the system and the robustness of the system. In a Bayesian framework for risk based decision making such indicators play an important role. In the context of volcanic hazards, the exposure can be related to the triggering factors for the volcanic eruption, the vulnerability represents the physical process of the volcanic eruption and flow, damages to infrastructure and loss of lives and the robustness is associated with the follow-up consequences and the socio-economic or political impact. The use of Bayesian Probabilistic Networks (BPNs) has proven to be efficient for such large scale risk assessment applications (Faber et al. 2007, Bayraktarli et al. 2005, Straub 2005). Generally, the exposures relating to natural hazards as well as the possible ensuing consequences can be considered to depend strongly on the specific geographical location of the occurrence of the event. For this reason, the use of Geographical Information Systems (GIS) is also important in the context of natural hazards risk management. 5.5.4 MODELING OF THE HAZARD PROCESS In this section, the modeling of the hazard process for volcanoes is discussed for two phases – the Plinian phase (where typically material is ejected in a tall column, spreads in the atmosphere and falls to earth like rain) and the Peléan phase (where material flows down the sides of the volcano as fast-moving avalanches of gas and dust). Then, the modeling of two other scenarios – the occurrence of earthquakes and tsunamis, each together with a volcanic eruption are considered. 5.5.4.1 The Plinian phase 5.5.4.1.1 Tephra intensity modeling Physical phenomenon – The deposits of pyroclastics (or materials that have been blown into the atmosphere by volcanic activity) are generically called tephra and divided in three basic types: air fall, pyroclastic flows and surges. The air fall deposits are formed by the accretion of clasts which fall by gravity from the eruptive column or which are thrown directly in area from crater, according to ballistic trajectories. The deposits of pyroclastic flows and surges are those released by gas-solid dispersions with high or low concentration of particles respectively, which move along the surface under action of gravity. The fall of pyroclasts, from the eruptive column, can have different speeds depending on the pyroclasts size, density and launch height, and the deposit on the ground, happen at various distances mainly depending on the stratospheric wind pressure. The pyroclasts are supported in the column until the upward thrust exceeds the gravity force; after they fall down accelerating until the force of gravity is not counterbalanced by the air friction, when the particles fall with a constant speed. Actions on the constructions – The tephra deposits produce on the constructions a gravitational load qV on the roofs, even if the pyroclastic flows and the surges, before transforming into deposits, act through a horizontal pressure qH on the affected structure. The static load qV can be considered a gravitational distributed load and can be estimated as follows:

where g is the acceleration due to gravity (9.81 ms−2 ), h is the deposit thickness (m), ρ is the deposit density (kgm−3 ). The deposit density depends on the composition of pyroclasts, their compactness, the deposit moisture and the subsequent rains. 5.5.4.1.2 Bombs, missiles and impact modeling Physical phenomenon – The explosive eruptions can also produce flying fragments of pyroclasts called bombs and missiles. The largest clasts are exploded directly from the crater according to 446

pure ballistic trajectories. On the contrary, the smaller clasts can be sustained by convection in the eruptive column. They are then thrown in the atmosphere from the main flow to fall or be transported along the mountainside in gravitational currents. The word missile can relate to flying debris, not involved in the eruption, set in motion by pyroclastic flows. The movement of a volcanic fragment in vacuum with a ballistic trajectory has been studied and reported in Dobran (2006). During a Plinian eruption, large clasts follow a pure ballistic trajectory. The smaller ones are transported not by the wind, but by another dynamic mechanism such as the lateral and vertical expansions of the eruption column, which reduces the drag force on the particulates; such kind of clasts one called sustained ballistics. Actions on the constructions – The damage caused by bombs and missiles depends on the kinetic energy and the vulnerability of the affected object. A flying fragment can impact the roofing or the walls of a building, but, in particular, it can hit the most vulnerable parts of the building like the openings. A key factor which governs the vulnerability of buildings is the resistance of openings, especially the glass panes or the shutters which can prevent the hot ash from entering. On the contrary, possible consequent fires and/or breathing difficulties for people inside can arise. Several studies have looked at the evaluation of the speed of bombs and missiles, produced by explosive volcanic eruption, but the analysis of the effects of these flying objects buildings is not very much developed. Spence et al. (2005) have examined the window failure produced by missiles generated by pyroclastic flows. The probability of impact of flying debris on windows depends on the flow velocity, the flow density, the density of potential missiles in the area surrounding the volcano, as well as the surface and the orientation of windows. Missile impact causes failure when a fragment has a sufficient kinetic energy to break the window. 5.5.4.1.3 Lava flow, temperature and thickness/height modeling Physical phenomenon – A volcano is defined as effusive if the magma is emitted in the form of a lava flow characterized by gas bubbles dispersed in a continuous liquid: the Etna volcano in Sicily (Italy), for example, belongs to this category. The lava flows are made of totally or partially fused magma emerging on the surface. Lava can form broad flows or immediately get cold above the volcanic conduit giving rise to domed structures called lava domes. Actions on the constructions – The lava flow produces a lateral horizontal pressure which can cause the collapse of the affected buildings. The damage is also caused by the degradation of the materials produced by high temperatures of the magma. For example, during the Etna eruption of 2001, the temperature of lava flow, measured with the infrared radiometer, was 1075◦ C. Generally, the advancing speed of the lava flows is sufficiently low to allow the evacuation and the safeguarding of human lives. 5.5.4.2 The Pelèan phase 5.5.4.2.1 Pyroclastic flow and impact modeling Physical phenomenon – Pyroclastic flows can be generated by the collapse of the eruptive column (as during the eruption of the Soufrière volcano, St. Vincent, Caribbeans, 7 May 1902), by a directional explosion for the slipping of a part of the volcano (as during the eruption of the St. Helens volcano, United States of America, 18 May 1989) or by a lateral explosion at the base of a lava dome (as during the eruption of the Pelée volcano, Martinique, 8 May 1902). They are the most dangerous events of an explosive eruption. Therefore, the estimate of the main physical parameters that characterize the dynamics of transportation and deposition is extremely important. A pyroclastic flow is made of a mixture of gases, within which solid particles of various sizes are dispersed. A multi-phase physical model for the evolution of pyroclastic flows can be found in Todesco et al. (2002). Actions on the constructions – In the structural analyses, it is possible to schematize the action of the pyroclastic flows as a uniformly distributed static pressure (Petrazzuoli and Zuccaro 2004), with temperature ranges between 200 and 350◦ C (Giurioli et al., 2008). In general, the first elements to reach the collapse are the glass windows and the shutters. However, they can be easily protected by more resistant panels. Nevertheless, the lateral resistance of a building to pyroclastic flow strongly depends on the design criteria applied to resist ordinary load conditions: of course, an earthquake resistant building presents relatively larger strength and stiffness capabilities. 447

5.5.4.2.2 Lahar flow and impact modeling Physical phenomenon – After explosive eruptive events, the thermal change in the proximity of the volcano often produces rain. Combined with the pyroclasts of poor coherence, with the volcanic high slope of (20–30◦ ) and the distinctive seismicity of the eruptive phase, the rain can cause the mobilization of the volcanic deposits and the consequent formation of mudslide and lahar. The term lahar indicates any type of muddy flow containing volcanic material. Lahar and mudslide are extremely dangerous because of their high kinetic energy, they being generally characterized by speed of the order of some tens kilometers per hour up to above 100 km/h (Carlino 2001). The lahar flows are influenced by the same mechanisms of transportation and sedimentation of the non volcanic material landslides. Indeed, the lahar flows move under gravity with the influence of the shear stress, concentration of the flow and slope gradient. Actions on the constructions – The effects of lahar flows on the constructions are comparable to those ones produced by the debris flows. However, the lahar flows present the additional variable in the form of the temperature, which causes substantial degradation of mechanical properties of construction materials. The temperature of lahars is widely variable. It depends on the typology and the quantity of the erupted materials and on the time between the deposit and the mobilization. 5.5.4.3 Description and modeling of other possible scenarios 5.5.4.3.1 Eruption related earthquake Physical phenomenon – All volcanic eruptions are accompanied by local seismic activity. While tectonic earthquakes are generally related to a shear-faulting mechanism, volcanic earthquakes may involve tensile, isotropic, and/or shear rock fractures, driven by the percolation of high-temperature fluids/gases or directly by the magma-ascent mechanism. Earthquakes caused by volcanic activity are generally classified into four categories: • • • •

volcano-tectonic (VT) earthquakes, long-period (LP) earthquakes, harmonic tremor (T), surface events (SEs).

From the point of view of seismic-hazard analysis in the pre-eruptive phase, only the VT earthquakes need to be considered. Both SEs and T generally appear during an eruption, and they have very low amplitudes beforehand. Although LP earthquakes could be present in the pre-eruptive phase, highmagnitude events of such a class are rarely observed before an eruption. Moreover, LP earthquakes involve only low-frequency signals, and they are not associated with a well-understood source mechanism. Actions on the constructions – The intensity of a volcanic earthquake is a function of the entity of the eruptive event. During the evolution of a volcanic system from a quiescent state to an eruptive state, a large number of small- to moderate-sized earthquakes occur. Thus, the cumulative effects of these numerous and small magnitude earthquakes can also cause structural damage from the low-cycle fatigue phenomena. Therefore, very stiff structures such as masonry buildings or low rise reinforced-concrete-frame structures are expected to be affected during the pre-eruptive earthquake occurrence. 5.5.4.3.2 Tsunami Physical phenomenon – Tsunami refers to the phenomena of the rogue waves which produce devastating effects on the coast. It is characterized with an initial and temporary withdrawal of the waters, or with a flood which can show like a tide which rapidly comes in, like a waves trains or like a water wall. A tsunami can be produced by any cause able to vertically perturb a sufficiently big column, moving them to its equilibrium position. So, its origin is not only connected with a seismic phenomenon, but could also be volcanic eruptions, explosions, landslides, submarine tectonic displacements and impact with cosmic objects. The normal waves are caused by the wind, which produces the movement of the sea surface only while the tsunami waves move the whole water column from seabed to surface. In the context of a volcanic eruption, the anomalous waves leading to a tsunami can be produced by massive pyroclastic flows which reach the sea. This happened during the explosive eruption of the Krakatua volcano in 1883 in the Sunda Straits between Sumatra and Java that produced a large tsunami killing more than 30,000 people in the coastal villages of the Straits. 448

Actions on the constructions – According to Palermo et al. (2007), the actions produced by a tsunami on a construction can be grouped into two loading combinations: initial impact and postimpact flow. The initial impact includes surges and debris impact force components. The surge force is produced by the impact of the flood waves on the structures. The debris force is related to impact structures due to significant debris (such as vehicles, components of buildings and drift wood) which the waves can transport. After the initial impact, a proposed second loading combination results, namely the post-impact flow. During this phase, hydrodynamic (drag) forces are exerted on structures due to continuous flow of water around the structure. In addition, the inundation gives rise to hydrostatic forces. The hydrostatic forces can occur on both the exterior and interior of the structure. The latter depends on the degree of damage sustained during the initial impact. Further, the structure is subject to debris from floating objects being transported by the moving body of water. Therefore, the second phase of loading includes hydrodynamic and hydrostatic forces, debris impact forces, and buoyancy forces that result from the structure being submerged after the initial impact.

5.5.5 CLASSIFICATION OF STRUCTURES AND STRUCTURAL VULNERABILITY 5.5.5.1 Volcanic vulnerability of structures As discussed in the previous section, a volcanic eruption is characterized by a series of subsequent physical phenomena, including volcanic earthquakes, ash-fall, pyroclastic flows, lahars, landslides, volcanic missiles and tsunami. As a consequence, the damage impact due to a volcanic eruption depends upon several disastrous events, different from each other, but tightly connected to each other. Each event contributes in different ways to the final scenario. The evaluation of the possible effects due to a volcanic eruption in an urban area is therefore very complex. The damage impact scenario in fact can vary, depending on the type of eruption, and also depends on the development over time of the different phenomena characterizing it. It is also related to the location considered and the typological-structural characteristics of buildings and infrastructures in the area. Therefore, the classification of structures and the identification of the different building characteristics is an important step in the assessment of the vulnerability of structures to volcanic events. 5.5.5.2 The Vesuvius area structural vulnerability assessment The assessment of volcanic vulnerability of structures was considered in the Vesuvius area. The assessment referred to the volcanic vulnerability assessment methodology proposed within the EXPLORIS European project (EXPLORIS 2006) and developed by the PLINIVS Centre. It refers to a dynamic model which simulates the whole eruptive process and refers to the potential eruption scenarios for the volcanic activity of Vesuvius and the possible associated hazards which may develop (Neri et al., 2008). The EXPLORIS project considers three volcanic phenomena; earthquakes (EQ), ash-falls (AF) and pyroclastic flows (PF). The assessment included a data collection exercise, based also on an extensive field investigation activity conducted during the period 2009–2010 (COST C26 2009b). The surveys were necessary to collect information based on various parameters influencing the volcanic vulnerability for each construction. The methodology was applied with respect to the Torre del Greco historic urban centre, the Residential Area, and the School Buildings. (COST C26 2009b). In addition, a detailed survey was carried out for various historic 19th century villas in the Vesuvius area, namely the Vesuvian Villas along the Golden Mile. In this case, additional parameters relating to monuments and historic cultural heritage were considered. 5.5.5.3 Classification of structures and parameters for vulnerability assessment In the assessment of the vulnerability of buildings to a volcanic eruption, various relevant parameters need to be evaluated. The methodology, adopted for the volcanic vulnerability assessment of the structures, includes the collection of data, which also requires field surveys and site investigations. The data collection surveys are necessary in order to build up a database of information based on the relevant parameters influencing the volcanic vulnerability. The data collected for an area or region based on these parameters is organized and the building vulnerability can be analyzed. The 449

parameters considered in the Vesuvius field investigations conducted through COST C26 activity, include the following (COST C26, 2009a): • the Identification section is intended to locate the building with reference to the geographical parameters of the region; • the General Information Section refers to the building type (ordinary building, warehouse, electrical station, etc.), destination (residence, hospital, school, etc.), use (fully used, partially used, not used and abandoned) and exposure (ordinary, strategic, exposed to special risk) of the construction; • the Condition Section refers to age and state of conservation of the structure (poor, mediocre, good and excellent) and typology of the finishes (economic, ordinary, luxury); • the Descriptive Characteristics Section refers to the number of total storeys starting from the lowest ground level, the number of floors above the ground, including the penthouse, the number of residential apartments, the presence of occupied or not basement, the height of the first storey, minimum and maximum heights up to the roof, the presence of barriers with height >2 m, the orientation (angle between the longest or the main façade and the North) and the position of the unit in the block; • the Structural Characteristics Section refers to the principal typology (reinforced concrete, masonry, wood, steel and mixed), primary vertical structures (sack masonry with or without reinforcements, hewn stones masonry, masonry or tuff blocks, RC frames with weak or resistant cladding, etc.), primary horizontal structures (timber floor, floor with steel beams, concretetile structures, vaults, etc.), geometry of the roofing (plane, single pitched, multi pitched and vaults), thickness of the walls and the curtain walls and typology of the curtain walls (tuff blocks or squared stones, concrete blocks, etc); • the Openings Section refers to the percentage of openings on the façade, the number of small, typical and large windows, their material (timber, PVC, aluminium or timber-aluminium, light steel and steel of security anti-intrusion type), their protection and their conditions (perfect, efficient, poor, bad or lack of windows); • the Interventions Section refers to the age and type of repairs (extraordinary maintenance, upgrading and retrofitting); • the Regularity Section refers to the regularity and distribution of curtain walls in plan and along the height, the type of the structure (single or two-directional frames, single or two-directional walls and walls with frames), soft floor (pilotis on a part of the ground floor, totally open ground floor and intermediate soft storey) and possible presence of stocky beams or columns. These parameters define each building in terms of geometry, typology and importance and mainly measure the volcanic vulnerability of the construction itself. In particular, these parameters can be divided into main sections. The first section provides information on the main vertical and horizontal structures, the regularity in plan and in elevation, the age and conservation of the construction, the number of storeys. These aspects are associated with the evaluation of the seismic vulnerability of buildings. The second section is specific to the building behaviour under the effect of an explosive eruption, referring to the roof structure typology, and the openings. The information on the type of the roof structure is associated with the collapse due to ash-fall deposits during an eruption. The information on openings, including opening shape, the size and the protection of the openings, is associated with the pyroclastic flows. The structural classification is carried out, with reference to the structural vulnerability and the phenomenon/phenomena considered.

5.5.6 FRAGILITY AND VULNERABILITY MODELLING OF STRUCTURES 5.5.6.1 Seismic vulnerability assessment methodologies for building aggregates 5.5.6.1.1 General principles The seismic vulnerability analysis has the purpose to evaluate the consistency of a structure in a certain area in order to estimate its propensity to undergo a certain level of damage against an earthquake of a given intensity. To this purpose, there are several methods having a level of detail which generally changes with the scale of application. However, for building belonging to urban aggregates, only few provisions are found in literature for their vulnerability assessment. 450

One possible methodology for the structural analysis of such a building type can be carried out according to the following steps (Avorio & Borri 2001): • to perform a structural survey appropriate to the peculiarities of the group of buildings investigated; • to evaluate the influence of the masonry quality on the safety check; • to build a series of charts in order to identify the foreseeable disruptions. In the first step, it is important to understand that the study of the whole building complex is not due to the simple sum of the vulnerability of single constitutive structural units. The difficulty to have information about the buildings adjacent to the examined one suggested to adopt a quick survey method oriented to highlight the constructive typologies and their mechanical characterization. The result is the implementation of a structural survey guide composed of both a base legend and thematic forms on the different structural types. In general terms, within such a phase the following information can be achieved: • Geometrical survey of urban aggregates, which serves as basis for the structural analysis; • Structural reading, which allows to obtain information on the different structural elements as well as on the connections among them; • Analysis of disruptions, which inspect the crack pattern visible on the building. As a second step, the detection of the masonry quality, even if in a quick manner, is investigated. This is justified from the fact that many of the observed partial collapses of buildings were due to the unsatisfactory condition of the masonry apparatus. The following reference parameters, which are easily achieved, can influence the behavior of masonry buildings and therefore have to be surveyed: • • • • •

Horizontality of blocks; Offset of vertical mortar joints; Shape and dimensions of elements; Elements located orthogonally to the masonry wall plane; Quality of mortar.

The proposed procedure is based on the comparison between the features of the study masonry walls and a series of categories reported into appropriate charts. This comparison provides a given score (s) for each parameter of the study masonry. If the wall apparatus respect the “rules of art”, a score of 2 is assigned. Instead, the partial presence of each of the reference parameters listed above gives rise to a score in the range from 0.5 to 1.5. Finally, the absence of these parameters provides a score of zero. The sum of different scores, which qualitatively defines as the masonry wall respects the “rules of art”, is defined as the masonry quality. The range of the possible achievable scores is comprised between 0 (very poor quality) and 10 (very high quality). Further, three different categories about the masonry quality have been identified, namely category A (8 to 10), category B (3 to 8) and category C (0 to 3). Later on, the study of the static behavior of either an isolated building or a group of buildings within a historical centre can be conducted by the analysis of a series of standard buildings. In this way, the analysis of large scale portion of buildings is performed quickly, considering the prerequisites of the study masonry walls. Finally, once the masonry quality is attributed, the static analysis by means of comparison charts can be performed following two different directions depending on the masonry type. For category C masonry, the examination of the static disruptions is not carried out, since the poor quality of masonry does not allow for the creation of a valid resistant mechanism. Therefore, the structure is not eligible from the structural point of view. For the other two masonry categories (A and B), the evaluation of the structural integrity is made with reference to appropriate comparison charts by considering different parameters, namely the type of floors, the vertical and horizontal slenderness of walls, the presence of pushing roof, the presence of pushing arches and vaults and the mechanisms involving either one storey or more than one storey masonry walls. All these cases can be considered both for not damaged buildings and damaged ones. In the first case, from the detected boundary conditions and geometry, the structural safety of walls can be assessed by the comparison with the conditions reported in the chart. Instead, for damaged buildings, the chart represents a 451

guide to detect disruptions only. From the Italian normative point of view (M.C., 2009), no detailed rules on the global check of urban aggregates are given, but only some indications about simplified methods to be performed for this check are provided. In particular, in the case of sufficiently rigid floors, formal verification at both the Life Safety Limit State and the Serviceability Limit State of a structural unit belonging to an aggregate is carried out, even for buildings with more than two levels, through static nonlinear analysis by both checking separately each building storey and neglecting the variation of axial force in the masonry piers due to the effect of seismic actions. With the exclusion of structural units placed either at the corner or at the end of an aggregate, as well as parts of buildings not restrained along any side from other structural units (e.g. upper floors of a building having height greater than the one of all adjacent structural units), the analysis can also be done neglecting torsional effects, assuming that the floors can only translate in the direction of the seismic action considered. If the building floors are flexible, the analysis of either single walls or systems of coplanar walls of the building is done; each of them analysed as independent structure subjected to both the vertical loads and the seismic action along the direction parallel to the wall. In this case the analysis and check of each wall are made on the basis of the references given by the new technical Italian code (M. D. 08) for new ordinary masonry buildings. The lack of study on the behaviour of structural units into urban aggregates, as well as the code deficiencies on this topic, have suggested to develop a new simplified methodology, reported in the next Section, to assess the global vulnerability of such building complexes. 5.5.6.1.2 The proposed seismic assessment form The proposed procedure is applied on a regional scale and, therefore, a speedy procedure for the vulnerability estimation based on the compilation of special forms is indicated as appropriate. The seismic vulnerability evaluation of historical aggregates arises essentially from a critical review of the detection form originally introduced by Benedetti and Petrini (1984), who proposed the estimation of the susceptibility to damage under earthquakes of isolated buildings. In particular, through this type of analysis, it is possible to classify the building stock of a given area on a vulnerability scale. The procedure consists of assigning one of four vulnerability classes (A, B, C or D), as defined in order of increasing danger, to ten parameters representing the geometrical and mechanical characteristics of the building. A score is assigned to each class, whereas a weight is correlated to each parameter of the form, which represents the influence that the same parameter has on the global vulnerability of the structure. Finally, the vulnerability index is obtained as the sum of all scores related to the attribution to classes multiplied by the respective weights. The original methodology proposed by Benedetti and Petrini (1984) is, however, inappropriate for buildings placed in aggregate, because the procedure does not take into account the structural interaction among adjacent buildings. To overcome this limit, Formisano et al. (2009) have proposed a new form, achieved from the original one with the insertion of new five parameters indicative of the aggregate condition of buildings, which may increase or decrease, depending on the case, the vulnerability of a structural unit inserted within a building stock. These factors, in part derived from previous studies found in literature are: • • • • •

Interaction in height with adjacent buildings; In plan position of the building in the aggregate; Number of staggered floors between the building under investigation and those adjacent; Typological and structural heterogeneity among adjacent buildings; Difference among the opening area on the facades of adjacent buildings.

Four vulnerability classes have been assigned to each of the above parameters. According to the layout of the original survey form, scores and weights have been assigned to the introduced classes and parameters respectively. In the following, the meaning of the four vulnerability classes defined for each additional parameter of the form, as well as the scores and weights attributed to classes and parameters, respectively, and obtained from the above study, are presented and discussed into detail. 1) Interaction in height with adjacent buildings (Figure 2): • Class A: −20 points. The building is located between two buildings of equal height. • Class B: 0 points. The building is located either between higher buildings or between a building with major height and a building with the same height. 452

Figure 2. Possible in elevation configurations of a structural unit inserted into a building aggregate.

Figure 3. Possible plan configurations of a structural unit inserted into a building aggregate.

Figure 4.

Possible positions of staggered floors.

• Class C: 15 points. The building is located either between a building with minor height and a building with the same height or between a higher building and a lower one. • Class D: 45 points. The building is adjacent to lower buildings. The weight assigned to this parameter is equal to 1. 2) In plan position of the building in the aggregate (Figure 3): • Class A: −45 points. The building is restrained on three sides from adjacent buildings. In this case the nearby buildings operate a confinement function on the building under question, limiting its possible displacements and deformations. • Class B: −25 points. The building is restrained on two sides from adjacent buildings. Therefore, the adjacent buildings operate a confinement function less significant than the previous one. • Class C: −15 points. The building occupies a corner position in the aggregate. In this case the containment action is not exercised on two walls of the building and is less effective than before. • Class D: 0 points. The building occupies a leading position in the aggregate. No containment action is detected and, therefore, the building is more prone to suffer displacements and deformations. To this parameter a weight of 1.5 is assigned. 3) Number of staggered floors among adjacent buildings (Figure 4): • Class A: 0 points. Total absence of staggered floors. • Class B: 15 points. Presence of a pair of staggered floors. • Class C: 25 points. Presence of two pairs of staggered floors. • Class D: 45 points. Presence of more than two pairs of staggered floors. The weight is less than that given to the two previous parameters, it being equal to 0.5. 4) Typological and structural heterogeneity among adjacent buildings: • Class A: −15 points. The building is homogeneous with the adjacent buildings from both the typological and the structural point of view. • Class B: −10 points. The building is adjacent to buildings made of the same material but erected with a different construction technique (e.g. a sack tuff blocks building close to a squared tuff blocks one). 453

Figure 5. The new vulnerability assessment form proposed for buildings in aggregate.

• Class C: 0 points. The adjacent buildings are made of different materials which have the same structural heterogeneity (e.g. a tuff masonry building next to a brick masonry one). • Class D: 45 points. The building has structural heterogeneity with respect to adjacent buildings (e.g. a brick building adjacent to a reinforced concrete one). The assigned weight is equal to 1.2. 5) Difference among the opening area on the facades of adjacent buildings: • Class A: −20 points. Difference less than 5%. • Class B: 0 points. Difference between 5% and 10%. • Class C: 25 points. Difference between 10% and 20%. • Class D: 45 points. Difference greater than 20%. The weight assigned to this parameter is 1. Based on the above considerations, a new type of form based on fifteen parameters giving rise to a maximum vulnerability score equal to 515.25 has been therefore developed as shown in Figure 5.

5.5.7 EVALUATION OF RISKS AND THEIR TREATMENT AND COMMUNICATION 5.5.7.1 Quantification of direct and indirect risks Following the assessment and evaluation of the exposures/hazards, vulnerability and consequences associated with the system considered for risk assessment, the ensuing risks then need to be quantified and evaluated. For this purpose, the system considered for the risk assessment is assumed to be exposed to hazardous events (EX ) with probabilistic characterization p(EXk ), k = 1, nEXP , where nEXP denotes the number of exposures. It is assumed that there are nCON individual constituents of the system, each with a discrete set (can easily be generalized to the continuous case) of damage states Cij , i = 1, 2..nCON , j = 1, 2..nCi . The probability of direct consequences cD (Cl ) associated with the lth of nCSTA possible different state of damage of all constituents of the facility Cl , conditional on the exposure event EXk is described by p(Cl |EXk ) and the associated conditional risk is p(Cl |EXk )cD (Cl ). The vulnerability of the system is defined as the risk due to all direct consequences (for all nCON constituents) and may be assessed through the expected value of the conditional risk due to direct consequences over all nEXP possible exposure events and all constituent damage states nCSTA :

The state of the facility as a system depends on the state of the constituents. It is assumed that there is nSSTA possible different system states Sm associated with indirect consequences cID (Sm , cD (Cl )). 454

The probability of indirect consequences conditional on a given state of the constituents Cl , the direct consequences cd (Cl ) and the exposure EXk , is described by p(Sm |Cl , EXk ). The corresponding conditional risk is p(Sm |Cl , EXk )cID (Sm , cD (Cl )). The risk due to indirect consequences is assessed through the expected value of the indirect consequences in regard to all possible exposures and constituent states, as:

The robustness of a system is defined as its ability to limit total consequences to direct consequences, that is to restrict the development of indirect ones. This characteristic may readily be quantified though the index of robustness IR (Baker et al. 2008):

which allows for a ranking of decisions in regard to their effect on robustness. 5.5.7.2 Risk treatment The various possibilities for collecting additional information in regard to the uncertainties associated with the understanding of the system performance as well as for changes the characteristics of the system can be considered to comprise the total set of options for risk treatment. The risk treatment options may, in the context of risk based decision making, be considered as the available decision alternatives. Risk treatment is decided upon for the purpose of optimize the expected utility to be achieved by the decision making. Risk treatment can be implemented at different levels in the system representation, namely in regard to the exposure, the vulnerability and the robustness, as shown in Figure 1. Considering the risk assessment of a load carrying structure, risk treatment by means of knowledge improvement may be performed by collecting information about the statistical characteristics of the loading (exposure), the strength characteristics of the individual components of the structures (vulnerability) and by systems reliability of the structural system (robustness). Risk treatment through changes of the system characteristics may be achieved by restricting the use of the structure (exposure), by strengthening the components of the structure (vulnerability) and by increasing the redundancy of the structural system (robustness) 5.5.7.3 Risk acceptance and life quality index In addition to risks due to economic losses, the decision maker has to take into account also the risks to individuals as well as potential damages to qualities of the environment. It is hence useful to differentiate between tangible and intangible risks, i.e. risks which may easily be expressed in monetary terms and risks which are not. Intangible values especially concern loss of lives and injuries and also qualities of the environment. The Life Quality Index (LQI) is a measure that facilitates the development of risk acceptance criteria for intangible risks (Nathwani et al. 1997). It is based on demographical indicators that include the incremental increase in life expectancy through risk reduction, the corresponding loss of economic resources, measured through the Gross National Product (GNP) together with the time used for work, all assessed for a statistical life in a given society. The underlying idea of the LQI is to model the preferences of a society quantitatively as a scalar valued social indicator, comprised by a relationship between the GDP per capita, the expected life at birth and the proportion of life spend for earning at living. 5.5.7.4 Risk perception and communication Different individuals in society perceive risks differently, depending on their own situation in terms of to what degree they may be affected by the exposures, to what degree they are able to influence 455

the risks and to what degree the risks are voluntary. Generally risks are perceived more negatively when stake holders feel more exposed, when they feel they have no influence and they feel they are exposed to risks involuntary. Another aspect is related to how adverse events are perceived by individuals and groups of individuals in society when and after such events take place. Again, this depends on the perspective of the affected individuals and groups of individuals. Furthermore, the occurrence of adverse events and the way the information about such events is made available will affect the perception of risks in general but also in many cases trigger actions which have no rational basis and only adds to the societal consequences of such event. Due to the effects of the perception of risk, it is generally observed that different individuals and groups of individuals have different attitudes in regard to what risks can be accepted. Risk averse and risk prone attitudes are observed, which simply refers to the effect that risks are assigned different tastes depending on these characteristics. Whereas such behavior is a private matter for individuals of society, it leads to an uneven distribution of risks, if exercised in the context of societal decision making and this is clearly unethical and not rational. The perception of risks may be significantly influenced by information about the risks themselves. Information can and should be used as a targeted means of reducing potential losses caused by reactions to events beyond what is rational, seen in the perspective of normative decision making. Being provided with transparent information in regard to the nature of exposures, possible precautionary actions, information on how risks are being managed and the societal consequences of irrational behavior reduces uncertainties associated with the understanding of risks of individuals. This, in turn, adds to rational behavior and thereby reduces follow-up consequences.

5.5.8 CONCLUSIONS A framework for risk based decision making in the field of engineering is first described. This framework is general enough to accommodate for the special needs of different application areas but at the same time specific enough to ensure a sufficient degree of consistency in modeling and theoretical basis. Towards application of this framework for the risk assessment of volcanic hazards, aspects related to the modeling of the hazard process due to volcanoes are then discussed. A system of classification of structures and identification of different building characteristics that could be used for volcanic vulnerability and risk assessment is then proposed. This is followed by a discussion on the fragility and vulnerability modeling of structures relevant for seismic analysis. This paper can be hence seen as a preliminary version of a guideline document for the assessment and management of risks due to volcanic hazards. Further work is required, especially in the assessment and evaluation of consequences and risk treatment measures for volcanic hazards. REFERENCES Avorio, A. & Borri, A. 2001. Structural safety analysis of building aggregates in the historical centres (in Italian). In: Proceedings of the Conference “Structural Failure and Reliability of Civil Structures”; Venice, 6–7 December 2006. Baker, J.W., Schubert, M. & Faber, M.H. 2008. On the Assessment of Robustness. Structural Safety 30: 253–267. Bayraktarli, Y.Y., Ulf kjaer, J.P., Yazgan, U. & Faber, M.H. 2005. On the Application of Bayesian Probabilistic Networks for Earthquake Risk Management. In: Augusti et al., (eds.) Safety and Reliability of Engineering Systems and Structures; Proceedings of the ICOSSAR 2005, Rome. Rotterdam: Millpress, 3505–3512. Benedetti, D. & Petrini, V. 1984. On the seismic vulnerability of masonry buildings: an assessment method (in Italian). L’Industria delle Costruzioni 149: 66–74. Carlino S. 2001. The floods and the mudslides after the Vesuvius eruption. In: History and risk. Interventi di ingegneria naturalistica nel Parco Nazionale del Vesuvio. Ente Parco nazionale del Vesuvio; Napoli, 43–69 (in Italian). COST C26. 2009a. Field Investigation Form, Vesuvius Field Investigation. Urban habitat Constructions under catastrophic Events. 2009 & 2010. COST C26. 2009b. Vesuvius Field Investigation Report. Urban Habitat Constructions Under Catastrophic Events, 2009.

456

Dobran, F. 2006. VESUVIUS: Education, security and prosperity. Development in Volcanology, n. 8. F. Dobran Eds. Amsterdam: Elsevier. ISBN-13: 978-0-444-52104-0; ISBN-10: 0-444-52104-6. EXPLORIS 2006. Explosive Eruption Risk and Decision Support for EU Populations Threatened by Volcanoes. European Union 5th Framework Programme Research Project. Faber, M.H., Bayraktarli, Y.Y. & Nishijima, K. 2007. Recent developments in the management of risks due to large scale natural hazards. In: Proceedings of SMIS XVI. Mexican National Conference on Earthquake Engineering; Ixtapa, Guerrero, Mexico, 1–4 November, 2007. Formisano, A., Mazzolani, F.M., Florio, G. & Landolfo, R. 2009. Seismic vulnerability of a masonry aggregate in Sessa Aurunca (CE) (in Italian). In: Proceedings of the 13th Italian Conference ANIDIS 2009 “L’Ingegneria Sismica in Italia”, Bologna, 28 June – 2 July, CD-ROM, paper S14.15. Gurioli L., Zanella E., Cioni R. & Lanza R. 2008. Paleomagnetic determination of the pyroclastic flows temperatures produced by 79AD Vesuvian eruption. In: Proceedings of XVII GNGTS National Conference; Italy, 12 December 2008 (in Italian). JCSS 2008. Risk Assessment in Engineering. Internet Publication: http://www.jcss.ethz.ch/publications/ JCSS_RiskAssessment.pdf Nathwani, J.S., Lind, N.C. & Pandey, M.D. 1997. Affordable Safety by Choice: The Life Quality Method. Waterloo: University of Waterloo. A. Neri, W.P. Aspinall, R. Cioni, A. Bertagnini, P.J. Baxter, G. Zuccaro, D. Andronico, S. Barsotti, P.D. Cole, T. Esposti Ongaro, T.K. Hincks, G. Macedonio, P. Papale, M. Rosi, R. Santacroce & G. Woo2008. Developing an Event Tree for probabilistic hazard and risk assessment at Vesuvius. Journal of Volcanology and Geothermal Research 178(3) 397–415. Palermo, D., Nistor, I., Nouri, Y. & Cornett, A. 2007. Tsunami-Induced Impact and Hydrodynamic Loading of Near-Shoreline Structures. In: Protect; First International Workshop on Performance, Protection & Strengthening of Structures Under Extreme Loading, Whistler, Canada, 2007. Petrazzuoli S. M. & Zuccaro G. 2004. Structural resistance of reinforced concrete buildings under pyroclastic flows: a study of the Vesuvian area. Journal of Volcanology and Geothermal Research 133: 353–367. Raiffa, H. & Schlaifer, R. 1961. Applied Statistical Decision Theory. Boston: Harvard University. Spence R., Kelman I., Baxter P.J., Zuccaro G. & Petrazzuoli S. 2005. Residential building and occupant vulnerability to tephra fall. Natural Hazard and Earth System Sciences 5 477–494. Straub, D. 2005. Natural hazards risk assessment using Bayesian networks. In: Augusti et al., (eds.) Safety and Reliability of Engineering Systems and Structures; Proceedings of the ICOSSAR 2005, Rome. Rotterdam: Millpress, 2535–2542. Todesco, M., Neri, A., Esposti Ongaro, T., Papale, P., Macedonio, G., Santacroce, R. & Longo, A. 2002. Pyroclastic flow hazard assessment at Vesuvius (Italy) by using numerical modelling. I. Large-scale dynamics. Bulletin of Volcanology 64: 155–177. von Neumann, J. & Morgenstern, O. 1944. Theory of Games and Economic Behavior. Princeton: Princeton University Press.

457

Urban Habitat Constructions under Catastrophic Events (Final Report) – Mazzolani (Ed). © 2010 Taylor & Francis Group, London, ISBN 978-0-415-60686-8

List of COST papers

Proceedings of the Workshop in Prague Wald et al. eds, Print Prazska Technica, ISBN 978-80-01-03583-2 (30–31 March 2007) Consolidation, rebuilding and strengthening of the St. Panteleymon church – Ohrid R. Apostolska & G. Necevska-Cvetanovska Seismic upgrade of non-seismic r.c. frames using steel dissipative braces S. Bordea, A. Stratan, A. Dogariu & D. Dubina Connection modelling in fire I. Burgess Robust design of steel framed buildings against extreme loading M. Byfield, G. De Matteis & F. Dinu, Italy The prevention of disproportionate collapse using catenary action M. Byfield & S. Paramasivam Strengthening of masonry walls by innovative metal based techniques A. Dogariu, A. Stratan, D. Dubina, T. Nagy-Gyorgy, C. Daescu & V. Stoian High strength steel for seismic resistant building frames D. Dubina, F. Dinu, V. Ungureanu, R. Zaharia & D. Grecea Framework for risk assessment of structural systems M.H. Faber Fire analysis on steel portal frames damaged after earthquake according to performance based design B. Faggiano, M. Esposto, F.M. Mazzolani & R. Landolfo Optimization of flood protection policy P. Fošumpaur & L. Satrapa Precious and cossfire: two RFCS projects on joints subjected to fire J.M. Franssen & F. Hanus Behaviour of a cast in-situ concrete structure during a compartment fire M. Gillie & T. Stratford Impact loading of pressurized steel pipelines A.M. Gresnigt, S.A. Karamanos & K.P. Andreadakis Non-linear modelling of reinforced concrete beams subjected to fire V. Gribniak, D. Bacinskas & G. Kaklauskas Behaviour and modelling of composite columns and beams under fire conditions Hai Tan Kang, Z.F. Huang & R.B. Dharma Seismic design of cold-formed steel housing: a case study O. Iuorio, R. Landolfo & L. Fiorino Seismic vulnerability and risk assessment of urban habitat in Southern European cities A. Kappos 459

Analysis of reinforced concrete structures subjected to blast loading S. Karapinar, I. Sanri & G. Altay Aircraft impact on reinforced concrete structures S.A. Kilic & G. Altay Protecting critical infrastructure systems T. Krauthammer Experimental and numerical investigations on the Mustafa Pasha Mosque large scale model L. Krstevska, Lj. Taskov, K. Gramatikov, R. Landolfo, O. Mammana, F. Portioli & F.M. Mazzolani Robustness – robust structures by joint ductility U. Kuhlmann, L. Rölle, J.-P. Jaspart & J.-F. Demonceau Numerical analysis of beam to column connection at elevated temperatures L. Kwasniewski Peak pressure in flats due to gas explosion I. Langone, G. De Matteis, V. Rebecchi & F.M. Mazzolani Stainless steel structural elements in case of fire N. Lopes, P.M.M. Vila Real, L. Simões da Silva, Portugal & J.-M. Franssen Innovative materials and technologies for existing and new buildings in seismic areas A. Mandara Performance-based seismic retrofit of r.c. and masonry buildings A. Mandara, A.M. Avossa, M. Ferraioli, F. Ramundo & G. Spina Earthquake protection of historical buildings F.M. Mazzolani Shear panels for seismic upgrading of new and existing structures F.M. Mazzolani, G. De Matteis, S. Panico, A. Formisano & G. Brando Experimental tests on seismic upgrading techniques for RC buildings F.M. Mazzolani, G. Della Corte, E. Barecchia & M. D’Aniello Full-scale cyclic tests of a real masonry-infilled RC building for seismic upgrading F.M. Mazzolani, G. Della Corte, L. Fiorino & E. Barecchia Behaviour of microreinforced soil under impact loads of small magnitude A.I.A. Mendes, C.A.S. Rebelo, M.I.M. Pinto & I.M.C.F.G. Falorca The MNB aseismic isolation system for the seismic protection of structures A. Michalopoulos, T. Nikolaidis & C.C. Baniotopoulos Typology of seismic motion and seismic engineering design E. Mistakidis, R. Apostolska-Petrusevska, D. Dubina, W. Graf, G. Necevska-Cvetanovska, P. Nogueiro, S. Pannier, J.-U. Sickert, L. Simões da Silva, A. Stratan & U. Terzic Identification, classification of exposure events, exceptional or infrequent event scenarios J.P. Muzeau, A. Bouchaïr, V. Sesov & C. Coelho Some remarks on the simplified design methods for steel and concrete composite beams E. Nigro & G. Cefarelli Fire design of composite steel-concrete columns under natural fire D. Pintea & R. Zaharia Reconstruction and seismic strengthening of St. Athanasius church damaged by explosion V. Sendova, B. Stojanoski & L. Tashkov 460

State of the art in Europe and activity developed within WG3 Impact and Explosion P.D. Smith Atmospheric loads in considerations on extreme events and structural norms B. Snarskis & R. Simkus Structural integrity at fire Z. Sokol & F. Wald Class 4 stainless steel box columns in fire B. Uppfeldt & M. Veljkovic Analytical model for the web post buckling in cellular beams under fire O. Vassart, Luxemburg, H. Bouchaïr & J.-P. Muzeau State of art of structural fire design F. Wald Temperature of the header plate connection subject to a natural fire F. Wald, J. Chlouba & P. Kallerová Temperatures in unprotected steel connections in fire Y. Wang, J. Ding, X.H. Dai & C.G. Bailey Heat transfer in fire safety engineering U. Wickström Identification and classification of response of urban habitat constructions subjected to extraordinary loading conditions S. Wolinski

Proceedings of the Symposium in Malta Mazzolani et al. eds, Malta University Publishing, ISBN 978-99909-44-40-2 (23,24,25, October 2008) Design seismic response evaluation for 2D frames and 3D wall systems with flexible foundation using capacity spectrum method R. Apostolska, Z.P. Bonev, D. Blagoev, E. Vasseva & G. Necevska-Cvetanovska Avalanches D. Boisseir & J.P. Muzeau Performance based evaluation of a non-seismic RC framestrengthened with Buckling Restrained Braces S. Bordea, A. Stratan & D. Dubina The Seismic Risk of Buildings in Malta R.P. Borg, R.C. Borg & G. Borg Axisa Behaviour of steel and composite joints in fire I. Burgess, A. Santiago & F. Wald Component-based approaches of steel and composite joints in fire I. Burgess, A. Santiago & F. Wald Design procedures for steel and composite joints in fire I. Burgess, A. Santiago & F. Wald Introduction to Impact and Explosion Resistance M.P. Byfield & G. DeMatteis Protective design of RC framed building against blast loading through safe stand off distance approach M.P. Byfield & S. Paramasivam 461

Waves and Tsunamis C. Coelho, T. Rossetto & W. Allsop Response of Buckling Restrained Braces to Catastrophic Seismic Events M. D’Aniello, G. Della Corte & F.M. Mazzolani Blast effects on the Archimede Bridge G. De Matteis, A. Eboli & F.M. Mazzolani Seismic analysis and strengthening intervention of the Fossanova gothic church: numerical and experimental activity G. De Matteis, F.M. Mazzolani, L. Krstevska & Lj. Tashkov Hydrocarbon explosion loading G. De Matteis & P. Smith Experimental and analytical investigations on the behaviour of building frames further to a column loss J.F. Demonceau & J.P. Jaspart Response of high rise steel buildings as a result of column loss F. Dinu & D. Dubina Performance of masonry shear walls strengthened with Steel and Aluminum sheeting A. Dogariu & D. Dubina FEM modeling masonry shear walls strengthened with metal sheathing A. Dogariu, D. Dubina, F. Campitiello & G. De Matteis Seismic resistant dual steel /dual frames: Performance based analysis of structural effectiveness D. Dubina, D. Dinu & A. Stratan Volcanic hazard C. Dumaisnil, J.C. Thouret & J.P. Muzeau General Methodology for Risk Assessment M. Faber Advances in the Evaluation of Robustness of Structures M. Faber & H. Narasimhan Structural analysis and design in case of fire after earthquake B. Faggiano, M. Esposto, R. Zaharia & D. Pintea Risk management in case of fire after earthquake B. Faggiano, M. Esposto, R. Zaharia & D. Pintea Seismic resistance of thin, lightly reinforced coupled wall evaluated by large-scale shaking table test M. Fischinger & T. Isakovi´c Seismic risk of prefabricated “RC” industrial buildings with strong connections M. Fischinger, M. Kramar & T. Isakovi´c The Vesuvius risk Scenario: the Seismic Vulnerability of the “palazzo di Citta in torre del Greco (NA) A. Formisano, G. Florio, R. Landolfo & F.M. Mazzolani Thin fibre-reinforced concrete jackets for improving the seismic response of reinforced concrete members: experimental and numerical results K. Georgiadi-Stefanidi, E. Mistakidis & P.C. Perdikaris Global modelling of structures in fire M. Gillie, I. Burgess, J.-M. Franssen, L. Kwasniewski & Y. Wang 462

Design of high rise buildings to survive terrorist attack D. Hadden Fire design in Europe M. Heinisuo A proposal for a Vesuvius GIS database M. Indirli Seismic protection of buildings using innovative isolators based on magnetically controlled elastomer T. Isakovi´c & M. Fischinger Life Cycle Assessment (LCA) of an Existing Building H. Koukkari Dust Loading on Indsutrial Building Roofs A. Kozlowski Verification of effectiveness of seismic protection and retrofit techniques by experimental testing L. Krstevska & Lj. Taskov Ductile and partial-strength steel & composite joints as basis for redundant and robust frame structures U. Kuhlmann & L. Rölle Rotation capacity vs. demand of steel beams under catastrophic events R. Landolfo, V. Piluso, M. Brescia, M. D’Aniello, O. Mammana & S. Tortorelli Progressive collapse of buildings with key elements subjected to gas explosion I. Langone, G. De Matteis & F.M. Mazzolani Aircraft impact testing A. Lastunen, J. Kuutti & K. Kolari Structural member design in case of fire N. Lopes, P. Vila Real & A. Bouchaïr Structural member behaviour and analysis in case of fire N. Lopes, P. Vila Real, B. Uppfeldt, M. Veljkovic, L. Simões da Silva, J.M. Franssen, A. Bouchaïr, J.-P. Muzeau, O. Vassart, D. Bacinskas, G. Kaklauskas, V. Gribniak, M. Cvetkovska, L. Lazarov, E. Nigro & G. Cefarelli Bucharest soil conditions and input ground motion for the structural performance analysis D. Lungu, C. Arion & E. Calarasu Seismicity of Vrancea subcrustal source and corresponding seismic instrumentation D. Lungu, R. Vacareanu & A. Aldea Smart technologies in the seismic protection of existing buildings Part 1: General concepts A. Mandara, F. Ramundo & G. Spina Smart technologies in the seismic protection of existing buildings Part 2: Applications A. Mandara, F. Ramundo & G. Spina Actions in the catastrophic scenarios of a volcanic eruption F.M. Mazzolani, B. Faggiano & D. De Gregorio Analysis methodology for the evaluation of the eruption effects on buildings F.M. Mazzolani, B. Faggiano & D. De Gregorio Seismic protection of new and existing buildings using an innovative isolation system A. Michalopoulos, Th. Nikolaidis & C.C. Baniotopoulos 463

Floods J.P. Muzeau Effects of Debris Flow on Buildings E. Nigro & C. Faella The performance based approach in seismic rehabilitation of buildings. R. Pascu Mechanical properties of materials D. Pintea, R. Zaharia, M. Kaliske, G. Kaklauskas, D. Bacinskas, V. Gribniak, A. Török & M. Hajpál Lessons from the 2008 Sichang earthquake about modelling of seismic input D. Pintea, R. Zaharia, M. Kaliske, G. Kaklauskas, D. Bacinskas, V. Gribniak, A. Török, M. Hajpál & F. Romanelli Assessment, design and testing of guardrails due to vehicle impacts C. Seiler Effectiveness of seismic strengthening of monuments for blast resistance V. Sendova & G. Jekic Landslides – State of the Art V. Sesov & Z. Zugic Fuzzy stochastic earthquake analysis of structures J.-U. Sickert, M. Kaliske & W. Graf Strengthening of earthquake damaged structures by means of textile reinforced concrete J.-U. Sickert, M. Kaliske & W. Graf The effectiveness of blast walls P.D. Smith Blast load assessment by simplified and advanced methods P.D. Smith & A. Tyas Lessons from catastrophic events in the evolution of bridge and wind engineering G. Solari Structural failure and prevention during catastrophic wind events T. Stathopoulos, C.C. Baniotopoulos & I. Zisis Removable bolted links for eccentrically braced frames A. Stratan & D. Dubina Selection of time-history records for dynamic analysis of structures A. Stratan & D. Dubina Seismic effects produced by explosions Thermal properties of materials L. Tashkov & G. Mirakovski Small-scale blast load and structural response measurement A. Tyas Introduction to fire design F. Wald & I. Burgess Overview of fire design Y.C. Wang, J.M. Fransen, M. Heinisuo, F. Wald & I. Burgess Heat transfer analysis Y.C. Wang & F. Wald 464

Robustness under fire Y.C. Wang & F. Wald Fire damaged structures Y.C. Wang, F. Wald, A. Török & M. Hajpál Extreme Snow Loading Sz. Woliñski A Probabilistic Model for the Evaluation of the Impact of Explosive Eruption Scenarios at Vesuvius G. Zuccaro

Proceedings of the Final Conference in Naples Mazzolani et al. eds, CRC Press Taylor & Francis Group, A. Balkema Book (16,17,18 September 2010) Characterization of catastrophic actions on construction Characterization and modelling of seismic action Problems of seismic behaviour of buildings situated in urban habitats V. Gioncu & F.M. Mazzolani The M w 6.3, 2009 L’Aquila earthquake: linear and nonlinear site effects C. Nunziata, F. Vaccari & G.F. Panza Scenarios based earthquake hazard assessment F. Romanelli, A. Peresan, F. Vaccari & G.F. Panza Microzonation of landslide hazard triggered by earthquakes-Application of GIS methodology V. Sesov & J. Cvetanovska Modelling of explosion A feasibility study on modeling blast loading using ALE formulation L. Kwasniewski, M. Balcerzak & J. Wojciechowski Windows and glazing systems exposed to explosion loads: Part 1 – Lethality and hazard assessment A. Lööf, A. van Doormaal & M. Teich Windows and glazing systems exposed to explosion loads: Part 2 – Safety improvement strategies M. Teich, N. Gebbeken, A. Lööf & A. van Doormaal Dynamic testing of semi-rigid steel beam-column connections A. Tyas, J.A. Warren, J.B. Davison, E.P. Stoddart & A. Hindle Actions due to other natural catastrophes Landslides as a Secondary Event of Earthquakes and Eruptions E. Nigro & C. Faella Measurements of shear wave velocities for seismic and volcanic hazard assessment in urban areas C. Nunziata, G. De Nisco, M.R. Costanzo, F. Vaccari & G.F. Panza Tsunami impact evaluation for coastal areas T. Rossetto, T.O. Lloyd, C. Coelho, J.P. Carlier & W. Allsop Avalanche actions on constructions A. Talon, J.P. Muzeau, J.P. Carlier 465

Analysis of behaviour of constructions under catastrophic events Analyses of structures under fire Numerical assessment of structural assembly between steel beam and CFT columns under fire F.R. Carvalho Lopes, A. Santiago, L. Simões da Silva & J.G. Santos da Silva Application of high strength steel to seismic resistant multi-storey buildings D. Dubina, A. Stratan, F. Dinu, D. Grecea, N. Muntean & C. Vulcu Assessment of the robustness of structures subjected to fire following earthquake through a Performance-Based approach B. Faggiano, D. De Gregorio & F.M. Mazzolani The effect of different plaster types on critical limit for fire resistance used by horizontal load-bearing structures M. Hajpál & A. Somorjai Numerical modeling of steel columns in fire L. Kwa´sniewski, P.A. Król & K. Ła˛cki Application of FSE approach to the structural fire safety assessment of steel-concrete composite structures E. Nigro, A. Ferraro & G. Cefarelli Member, substructure and global structural fire analyses of steelconcrete composite frames E. Nigro, A. Ferraro & G. Cefarelli Thermo-mechanical analysis of composite slabs under fire conditions D. Pantousa & E. Mistakidis Evaluation of the steel structure fire resistance of a shopping centre using structural fire engineering P. Vila Real & N. Lopes To the composite slab on beams with corrugated web exposed to fire F. Wald, T. Jána, P. Kyzlík & M. Strejˇcek Fire after earthquake R. Zaharia, D. Pintea & D. Dubina Evaluation of structural response under exceptional seismic actions Lateral capacity of steel structures designed according to EC8 under catastrophic seismic events S. Tortorelli, M. D’Aniello & R. Landolfo Seismic assessment of historical mosques under exceptional earthquakes: A case study in Skopje F. Portioli, O. Mammana, R. Landolfo & F.M. Mazzolani Finite element modelling of lap shear riveted connections in fire R. Marmo, M. D’Aniello, F. Portioli & R. Landolfo Micro and macro-finite element modeling of brick masonry panels subject to lateral loadings M. Annecchiarico, F. Portioli & R. Landolfo Experimental methodology for verification of effectiveness of innovative seismic strengthening techniques of historical monuments L. Krstevska, Lj. Tashkov, K. Gramatikov, F.M. Mazzolani & G. De Matteis Response of BRBs to catastrophic seismic actions: experimental results G. Della Corte & F.M. Mazzolani Seismic performance of RC frame buildings with masonry infill R. Apostolska, G. Necevska-Cvetanovska & J. Cvetanovska L’Aquila earthquake: a survey in the historical centre of Castelvecchio Subequo A. Formisano, P. Di Feo, M.R. Grippa & G. Florio 466

A model for limit state analysis of wooden structures F. Ramundo & M.R. Migliore Seismic response of sheathed cold-formed steel structures under catastrophic events L. Fiorino, O. Iuorio, V. Macillo & R. Landolfo First considerations on the February 27, 2010 Chilean earthquake M. Indirli, F.M. Mazzolani & A. Tralli In-situ experimental testing of four historical buildings damaged during the 2009 L’Aquila earthquake L. Krstevska, Lj. Tashkov, N. Naumovski, G. Florio, A. Formisano, A. Fornaro & R. Landolfo Ambient vibration tests on three religious buildings in Goriano Sicoli damaged during the 2009 L’Aquila earthquake L. Tashkov, L. Krstevska, N. Naumovski, G. De Matteis & G. Brando

Analysis of structures under impact and explosion Mechanical behavior evaluation of pure aluminum by static and dynamic tests G. De Matteis, G. Brando & F.M. Mazzolani Direct design approach for seismic resistant steel frame buildings under extreme loading F. Dinu, D. Dubina & G. De Matteis Behaviour of deformable blast walls for protective structural design S.A. Kilic & P.D. Smith Simulation of pressures behind rigid blast walls S.A. Kilic & P.D. Smith Robust design – Alternate Load Path Method as design strategy L. Roelle & U. Kuhlmann Tests and simplified numerical simulation of vehicle impact on guardrails with respect to normative regulations C. Seiler Behaviour and structural evaluation procedure of a precast R.C. multi-storey building subjected to blast loading D.V. Stoian, I.S. Pescari, S.C. Florut & V.A. Stoian Consequence of natural disasters on construction Air fall deposits due to explosive eruptions: action model and robustness assessment of the Vesuvian roofs D. De Gregorio, B. Faggiano, A. Formisano & F.M. Mazzolani Review of the impacts of volcanic ash fall on urban environments V. Sword-Daniels, T. Rossetto, J. Twigg, D. Johnston, T. Wilson, J. Cole, S. Loughlin & S. Sargeant Structural design of urban habitat construction against catastrophic wind actions T. Stathopoulos, I. Zisis & C.C. Baniotopoulos

Evaluation of vulnerability of constructions Performance based evaluation and seismic risk analysis L’Aquila earthquake April 6th, 2009: the damage assessment methodologies R.P. Borg, M. Indirli, T. Rossetto & L.A. Kouris Seismic vulnerability analysis of a masonry school in the Vesuvius area A. Formisano, F.M. Mazzolani & M. Indirli 467

A quick methodology for seismic vulnerability assessment of historical masonry aggregates A. Formisano, F.M. Mazzolani, G. Florio & R. Landolfo Seismic vulnerability analysis of historical centres: a GIS application in Torre del Greco A. Formisano, F.M. Mazzolani, G. Florio, R. Landolfo, G. De Masi, G. Delli Priscoli & M. Indirli Robust design of seismic up-grading of R.C. structures with innovative bracing systems J.-U. Sickert, M. Kaliske, W. Graf & A. Mandara Vulnerability and damageability of structures under impact and explosion Limitations of the tying force method for providing robustness in steel framed buildings M.P. Byfield & S. Paramasivam Effect of column loss on the robustness of a high rise steel building F. Dinu & D. Dubina On the catenary effect of steel buildings A. Formisano & F.M. Mazzolani Performance assessment under multiple hazards Extreme flood effects as a combination of river and sea action C. Coelho & J.P. Carlier Survey activity for the seismic and volcanic vulnerability assessment in the Vesuvian area the Golden Mile Villas L. Alterio, D. De Gregorio, B. Faggiano, P. Di Feo, G. Florio, A. Formisano, F.M. Mazzolani, F. Cacace, G. Zuccaro, R.P. Borg, C. Coelho, M. Indirli, L. Kouris & V. Sword-Daniels Survey activity for the seismic and volcanic vulnerability assessment in the vesuvian area the historical centre and a residential area of Torre del Greco D. De gregorio, B. Faggiano, G. Florio, A. Formisano, T. De lucia, G. Terracciano, F.M. Mazzolani, F. Cacace, G. Conti, D. De luca, D. Fiorentino, C. Pennone, G. Zuccaro, R.P. Borg, C. Coelho, S. Gerasimidis & M. Indirli Volcanic actions and their consequences on structures B. Faggiano, E. Nigro, D. De Gregorio, G. Zuccaro & F. Cacace Survey activity for the seismic and volcanic vulnerability assessment in the Vesuvian area: relevant masonry and reinforced concrete school buildings in Torre del Greco G. Florio, D. De Gregorio, A. Formisano, B. Faggiano, T. De Lucia, G. Terracciano, F.M. Mazzolani, F. Cacace, G. Conti, G. De Luca, G. Fiorentino, C. Pennone, G. Zuccaro, R.P. Borg, C. Coelho, S. Gerasimidis & M. Indirli The L’Aquila Earthquake, April 6th, 2009: a review of seismic damage mechanisms L. Kouris, R.P. Borg & M. Indirli The Vesuvius case study in the framework of the EU COST Action C26 activity F.M. Mazzolani & M. Indirli Survey activity for the volcanic vulnerability in the Vesuvian area: the “quick” methodology and the survey form F.M. Mazzolani, B. Faggiano, A. Formisano, D. De Gregorio, G. Zuccaro, M. Indirli, & R.P. Borg A framework and guidelines for volcanic risk assessment H. Narasimhan, R.P. Borg, F. Cacace, G. Zuccaro, M.H. Faber, D. De Gregorio, B. Faggiano, A. Formisano, F.M. Mazzolani & M. Indirli Structural vulnerability assessment under natural hazards: A review D. Vamvatsikos, L.A. Kouris, G. Panagopoulos, A.J. Kappos, E. Nigro, T. Rossetto, T.O. Lloyd & T. Stathopoulos 468

Seismic impact scenarios in the volcanic areas in Campania G. Zuccaro & F. Cacace Structural damage and vulnerability assessment for service life estimation through MEDEA tool G. Zuccaro & M.F. Leone Human and structural damage consequent to a sub-plinian like eruption at mont Vesuvius G. Zuccaro, F. Cacace & S. Nardone MEDEA: a multimedia and didactic handbook for structural damage and vulnerability assessment – L’Aquila Case Study G. Zuccaro, F. Cacace & M. Raucci

Protecting, strengthening and repairing Examination, assessment and repair of RC structure damaged by fire M. Cvetkovska & Lj. Lazarov Concrete members reinforced with FRP bars in fire situation E. Nigro, G. Cefarelli, A. Bilotta, G. Manfredi & E. Cosenza Seismic protection of high-rise buildings by aluminum shear panels: a design application A. Basile, G. Brando, G. De Matteis & F.M. Mazzolani Numerical and experimental evaluation of q factors for RC MRF strengthened of steel BRB S. Bordea & D. Dubina Evaluation of re-centring capability of dual frames with removable dissipative members: case study for eccentrically braced frames with bolted links F. Dinu, D. Dubina & A. Stratan Behaviour model for post-tensioned bolted RC frame – steel brace connection A. Dogariu, S. Bordea & D. Dubina Numerical investigation of old RC frames strengthened against earthquakes by high dissipation steel link elements K.A. Georgiadi-Stefanidi & E.S. Mistakidis Application of base isolation techniques for seismic protection of structures-study cases L. Krstevska, Lj. Tashkov, M. Garevski & V. Shendova Application of smart strategies against severe dynamic actions A. Mandara, F. Ramundo & G. Spina Effectiveness of seismic strengthening of monuments for their blast resistance V. Sendova, G. Jekic & L. Tashkov

Strategy and guidelines for damage prevention Modelling issues of steel braces under extreme cyclic actions M. D’Aniello, F. Portioli & R. Landolfo Experimental evaluation of q factor for dual steel frames with dissipative shear walls F. Dinu, D. Dubina & C. Neagu Multi-hazard maps for the Valparaiso area (Chile) M. Indirli, C. Puglisi, A. Screpanti & F. Romanelli Comparison of damage assessment methodologies for different natural hazards T. Rossetto, A.J. Kappos, L.A. Kouris, M. Indirli, R.P. Borg, T.O. Lloyd & V. Sword-Daniels Avalanche risk assessment in populated areas A. Talon, J.P. Muzeau & J.P. Carlier

469