Shot Peening

III Shot Peening Edited by Lothar Wagner Deutsche Gesellschaft für Materialkunde e.V. 566 I Shot Peening Edite...

Author: Lothar Wagner

167 downloads 3719 Views 15MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

III

Shot Peening

Edited by Lothar Wagner

Deutsche Gesellschaft für Materialkunde e.V.

566

I

Shot Peening

Edited by Lothar Wagner

II

International Scientific Committee on Shot Peening Prof. S. Al-Hassani Dr. J. T. Cammett Mr. J. Champaigne Prof. C. L. Corum Prof. R. Clausen Dr. E. R. de los Rios Dr. B. Eigenmann

University of Manchester, Manchester, UK Naval Aviation Depot, Cherry Point, NC, USA Electronics Inc., Mishawaka, IN, USA Purdue University, West Lafayette, IN, USA Technical University of Hamburg – Harburg, Hamburg, Germany University of Sheffield, Sheffield, UK Diehl Munitionssysteme GmbH & Co. KG, Röthenbach a. d. Plegnitz, Germany Mr. J.-F. Flavenot CETIM, Senlis, France Mr. P.-O. Karlsson Volvo Aero Corporation, Trollhättan, Sweden Dr. D. Kirk Kenilworth, , Warwickshire,UK Mr. A. Levers Airbus UK Ltd., Broughton, Flintshire, UK Mr. B. Miller Boeing Company, Seattle, WA, USA Prof. E. Müller University of Applied Sciences of Bochum, Bochum, Germany Prof. A. Nakonieczny Technical University of Warsaw, Warsaw, Poland Dr. A. Niku-Lari IITT - International, Noisy - le - Grand, France Mr. M. J. Platts University of Cambridge, Cambridge, UK Prof. B. Scholtes University of Kassel, Kassel, Germany Dr. V. Schulze University of Karlsruhe, Karlsruhe, Germany Prof. M. C. Sharma Maulana Azad College of Technology, Bhopal, India Prof. K. Tosha Meiji University, Tama-ku, Tawasaki, Japan Prof. O. Vöhringer University of Karlsruhe, Karlsruhe, Germany Prof. L. Wagner Technical University of Clausthal, Clausthal-Zellerfeld, Germany Mr. Y. Watanabe Toyo Seiko, Ama-gun, Aichi, Japan Prof. H. Wohlfahrt Technical University of Braunschweig, Braunschweig, Germany Prof. K. Xu Xi’an Jiaotong University, Xian, China

III

Shot Peening

Edited by Lothar Wagner

Deutsche Gesellschaft für Materialkunde e.V.

IV Prof. Dr.-Ing. Lothar Wagner TU Clausthal Institut für Werkstoffkunde und Werkstofftechnik Agricolastr. 6 D-38678 Clausthal-Zellerfeld Germany

Proceedings of the 8th International Conference on Shot Peening (ICSP-8) in Garmisch-Partenkirchen, Germany, 16-20 September 2002

This book was carefully produced. Nevertheless, editor, authors, and publisher do not warrant the information contained therein to be free of errors. Readers are advised to keep in mind that statements, data, illustrations, procedural details or other items may inadvertently be inaccurate.

British Library Cataloguing-in-Publication Data: A catalogue record for this book is available from the British Library Bibliografic information published by Die Deutsche Bibliothek Die Deutsche Bibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliografic data is available in the Internet at . ISBN 3-527-30537-8 © 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Printed on acid-free paper All rights reserved (including those of translation in other languages). No part of this book may be reproduced in any form – by photoprinting, microfilm, or any other means – nor transmitted or translated into machine language without written permission from the publishers. Registered names, trademarks, etc. used in this book, even when not specifically marked as such, are not to be considered unprotected by law. Composition: W.G.V. Verlagsdienstleistungen GmbH, Weinheim Printing: Druckhaus Darmstadt GmbH, Darmstadt Bookbinding: Buchbinderei Schaumann GmbH, Darmstadt Printed in the Federal Republic of Germany

V

Introduction This book covers the papers which were presented either orally or as posters during the 8th International Conference on Shot Peening held in Garmisch-Partenkirchen, Germany, September 16-20, 2002. Shot peening has been proved to be a powerful instrument in enhancing the resistance of materials to various kinds of stress-induced damage, particularly against damage due to cyclic loading in air or in aggressive environments. As shot peening can be used for a wide variety of structural components irrespective of shape and dimensions, the number of shot peening applications in many industrial branches is increasing. The use of peen forming as a technique to form large metal parts into complicated shapes is also increasing, particularly in the aerospace industry. ICSP8 was the eighth in a series of international conferences (1981 Paris, 1984 Chicago, 1987 Garmisch-Partenkirchen, 1990 Tokyo, 1993 Oxford, 1996 San Francisco, 1999 Warsaw) and offered scientists and engineers a unique opportunity to update their knowledge. This highly successful conference attracted more than 200 participants from 18 countries. Of the more than 100 abstracts submitted, 71 papers were finally accepted for publication by the International Scientific Committee on Shot Peening (ISCSP). I would like to thank those members of the ISCSP who have been willing to review the submitted manuscripts in order to improve the quality of the proceedings. The contributions are arranged in the book according to the conference topics, i.e., Applications, Techniques and Controlling, Surface Layer Properties, Peen Forming, Corrosion and Fretting, Fatigue of Fe- and Ni-based Alloys, Fatigue of Light-Weight Alloys, Alternative Mechanical Surface Treatments and Modeling.

L. Wagner, Chairman of ICSP8 Clausthal-Zellerfeld, February 2003

2

VII

Table of Content I Applications ........................................................................................................................... 1 The Application of Mechanical Surface Treatment in the Passenger Car Industry P. Hutmann, BMW Group, Munich, Germany ............................................................................. 3 Life Enhancement of Aero Engine Components by Shot Peening: Opportunities and Risks G. König, MTU Aero Engines, Munich, Germany ..................................................................... 13 Study of Methodology to Increase Fatigue Limit of Gears K. Ando, Yokohama National University, Yokohama, Japan; K. Matsui, H. Eto, Isuzu Motors Ltd, Kawasaki-shi, Japan..................................................................................... 23 Searching for the Most Suitable Condition and the Suggestion of Each Application in Ultrasonic Shot Peening Y. Watanabe, K. Hattori, M. Handa, Tokyo Seiko, Japan; J.-M. Duchazeaubeneix, Sonats, France............................................................................................................................ 31 Consideration of Shot Peening Treatment Applied to a High Strength Aeronautical Steel with Different Hardnesses M. Torres, Department of Mechanics State University of São Paulo, São Paulo, Brazil; M. do Nascimento, H. Voorwald, Department of Materials and Technology State University of São Paulo, São Paulo, Brazil....................................................................... 37 Towards Peen Forming Process Automation F. Wüstefeld, W. Linnemann, S. Kittel, Kugelstrahlzentrum Aachen GmbH, Aachen, Germany....................................................................................................................... 44 Current Applications of Advanced Peen Forming Implementation A. Friese, J. Lohmar, F. Wüstefeld, Kugelstrahlzentrum Aachen GmbH, Aachen, Germany....................................................................................................................... 53 II Techniques and Controlling.............................................................................................. 63 The Unsatisfactory Situation in Residual Stress Evaluation R. Bosshard, ANVIL Developments, Volketswil, Switzerland .................................................... 65 Vacuum-Suction Peening: A Novel Method for Emission-free Shot Peening G. Pieper, S. Ruhland, GP Innovationsgesellschaft mbH, Lübbenau, Germany ....................... 73 New Developments in Cut Wire Shot for Shot Peening U. Kersching, R+K Draht GmbH, Mittweida, Germany ........................................................... 78

VIII Virtues & Limitations of Almen Round M. Sharma, H. Deo, M. A. C. T. Regional Engineering College Bhopal, India; S. Modi, MEC Jodhpur, India, R. Bosshard, Anvil Developments, Switzerland ....................... 83 Device for the Determination of Impact Velocities in Shot Peening J. Stangenberg , R. Clausen, Technical University of Hamburg-Harburg, Hamburg, Germany.................................................................................................................... 89 Effective Use of Flourescent Tracers for Peening Coverage P. Bailey, Electronics Inc., Cincinnati, OH, USA....................................................................... 96 A Theoretical and Experimental Investigation into the Development of Coverage in Shot Peening S. Karuppanan, Universiti Teknologi Petronas, Bandar Seri Iskandar, Perak Darul Ridzuan, Malaysia; J. Romero, E. de los Rios, C. Rodopoulos, SIRIUS Department of Mechanical Engineering, University of Sheffield, Sheffield, UK; A. Levers, Airbus UK, Chester, UK......................................................................................... 101 Almen Gage Calibration J. Champaigne, Electronics Inc, Mishawaka, IN, USA............................................................ 108 Performance of Almen Strips which are Straightened after Tempering J. Champaigne, Electronics Inc., Mishawaka, IN, USA........................................................... 114 Optimization of the Shot Peening Parameters F. Petit-Renaud, USF Impact Finishers (Shot Peening Division) and USF Vacu-Blast International, Slough, UK........................................................................................................ 119 Shot Peening on Pelton Wheels: Methods of Control and Results P. Marconi, 2 Effe Engineering s.u.r.l., Manerba del Garda, Italy; M. Lauro, W. Bozzolo, Enel Produzione S.p.A., Turin, Italy ........................................................................................ 130 Effect of Shot Peening on Erosion and Fatigue in Combined Bending and Torsion of the Magnesium Alloy AZ80 A. Jain, V. S. Nadkami, M. C. Sharma, Maulana Azad National Institute of Technology, Bhopal, India........................................................................................................ 137 III Surface Layer Properties............................................................................................... 143 Characteristics of Surface Layers Produced by Shot Peening V. Schulze, Institut für Werkstoffkunde I, Universität Karlsruhe (TH), Karlsruhe, Germany ................................................................................................................ 145

IX The Influence of the Velocity of a Peening Medium on the Almen Intensities and Residual Stress States of Shot Peened Specimens W. Zinn, B. Scholtes, Institute of Materials Technology, University of Kassel, Kassel, Germany; J. Schulz, R. Kopp, Institute of Metal Forming, Aachen University of Technology (RWTH Aachen), Aachen, Germany ................................................................. 161 Correlation between Mechanical and Geometrical Charasteristics of Shot and Residual Stress induced by Shot Peening P. Marconi, 2 Effe Engineering S.r.l., Manerba del Garda, Italy; G. Citran, Pometon S.p.A., Venice, Italy .................................................................................................................. 167 Influence of Pre-Annealing on Surface and Surface Layer Characteristics Produced by Shot Peening K. Tosha, Meiji Univ. Higashimita, Tama-ku, Tawasaki, Japan; J. Lu, D. Retraint, B. Guelorget, Universite de Technologie de Troyes, Troyes, France; K. Iida, Society of Shot Peening Technology of Japan, Meiji Univ. Kawasaki, Japan.......................... 173 Endurance Life of Aging Aircraft Components Predicted by Conductivity Changes in Aluminium 2024 J. Soules, Surface Stress Technology Inc. Cleveland, OH, USA .............................................. 181 Shot Peening of Ceramics: Damage or Benefit? W. Pfeiffer, T. Frey, Fraunhofer Institute for Mechanics of Materials, Freiburg, Germany................................................................................................................... 185 Influence of Retained Austenite, Strain-induced Martensite and Bending Stress upon Shot Peening-induced Residual Compressive Stress K. Ando, Yokohama National University, Yokohama, Japan; H. Eto, K. Matsui, Isuzu Motors LTD, Kawasaki, Japan ...................................................................................... 191 Lining of Metal Surface with Hard-Metal Foil using Shot Peening Y. Harada, K. Mori, S. Maki, Toyohashi University of Technology, Japan........................... 200 Relaxation of Shot Peening Residual Stresses in the 7050-T7451 Aluminium Alloy after Heat Cycles for Adhesive Bonding M. Roth, C. Wortman, Department of National Defence, Ottawa, Ontario, Canada ...................................................................................................................... 208 IV Peen Forming ................................................................................................................. 215 Peen-Forming – A Developing Technique P. O’Hara, Metal Improvement Co Inc, Newsbury, UK........................................................... 217 Optimising the Double-Sided Simultaneous Shot Peen Forming R. Kopp, J. Schulz, Institute of Metal Forming, University of Technology Aachen, Aachen, Germany..................................................................................................................... 227

X Impact Metal Forming H. Reccius, IHR, Gummersbach, Germany ............................................................................. 234 V

Corrosion and Fretting................................................................................................... 241

Mechanisms and Modelling of Cracking under Corrosion and Fretting Fatigue Conditions E. de los Rios, Department of Mechanical Engineering, University of Sheffield, UK............. 243 Influence of Shot Peening on Stress Corrosion Cracking in Stainless Steel J. Kritzler, Metal Improvement Company, Inc., Unna, Germany............................................ 255 Investigating the Benefits of Controlled Shot Peening on Corrosion Fatigue of Aluminium Alloy 2024 T351 S. Curtis, E. de los Rios, C. Rodopoulos, J. Romero, University of Sheffield, Sheffield, UK; A. Levers, Airbus UK, Chester, UK. ................................................................ 264 Influence of Shot Peening on the Fatigue and Corrosion Behavior of the Die Cast Magnesium Alloy AZ91 hp C. Müller, R. Rodríguez, Physikalische Metallkunde, Technische Universität Darmstadt, Darmstadt, Germany ........................................................................................... 271 VI Fatigue of Fe- and Ni-based Alloys ................................................................................. 279 Shot Peening and Fatigue Strength of Steels K.-H. Lang, V. Schulze, O. Vöhringer, Institut für Werkstoffkunde I, Universität Karlsruhe (TH), Karlsruhe, Germany .................................................................. 281 The Effect of Shot Peening Coverage on Residual Stress, Cold Work and Fatigue in a Ni-Cr-Mo Low Alloy Steel P. Prevey, Lambda Research Inc., Cincinnati, OH, USA; J. Cammett, U.S. Naval Aviation Depot, Cherry Point, NC, USA ................................................................................. 295 Effect of Ultrasonic Shot Peening on Fatigue Strength of High Strength Steel Y. Watanabe, K. Hattori, M. Handa, Toyo Seiko, Japan; N. Hasegawa, K. Tokaji, M. Ikeda, Gifu University, Japan; J.-M. Duchazeaubeneix, Sonats, France .......................... 305 Residual Stress Relaxation and Fatigue Strength of AISI 4140 under Torsional Loading after Conventional Shot Peening, Stress Peening and Warm Peening R. Menig, V. Schulze, O. Vöhringer, Institut für Werkstoffkunde I, Universität Karlsruhe (TH), Karlsruhe, Germany ..................................................................................... 311 Influence of Optimized Warm Peening on Residual Stress Stability and Fatigue Strength of AISI 4140 in Different Material States R. Menig, V. Schulze, O. Vöhringer , Institut für Werkstoffkunde I, University of Karlsruhe (TH), Karlsruhe, Germany ..................................................................................... 317

XI Thermal Fatigue of Shot Peened or Hard Turned Hot-work Steel AISI H11 M. Krauß, B. Scholtes, Institute of Materials Technology, University of Kassel, Kassel, Germany ...................................................................................................................... 324 Effect of Short-Time Annealing on Fatigue Strength of Shot Peened AISI 4140 in a Quenched and Tempered Material State R. Menig, V. Schulze, O. Vöhringer , Institut für Werkstoffkunde I, Universität Karlsruhe (TH), Karlsruhe, Germany ..................................................................................... 331 Effect of Shot Peening on Improvement of Fatigue Strength for Metal Bellows H. Okada, A. Tange, NHK SPRING CO.,LTD. ,Kanagawa-Pref, Japan; K. Ando, Yokohama National University, Kanagawa-Pref, Japan.......................................................... 338 VII Fatigue of Light Weight Alloys.................................................................................... 347 Property Improvement in Light Metals Using Shot Peening J. Gregory, Springfield Metallurgical Services, Inc., Springfield, VT, USA; L. Wagner, Chair of Physical Metallurgy and Materials Technology, Technical University of Brandenburg at Cottbus, Cottbus, Germany ...................................................... 349 Fatigue Strength Improvement of Welded Aluminium Alloys by Different Post Weld Treatment Methods Th. Nitschke-Pagel, H. Wohlfahrt, Institut für Schweißtechnik, Technische Univcersität Braunschweig, Braunschweig, Germany ............................................................ 360 Influence of Mechanical Surface Treatments on Notched Fatigue Strength of Magnesium Alloys B. Küster, L. Wagner, Chair of Physical Metallurgy and Materials Technology, Technical University of Brandenburg at Cottbus, Cottbus, Germany; M. Hilpert, Otto Fuchs Metallwerke, Meinerzhagen, Germany; A. Kiefer, OSK Kiefer, Oppurg, Germany .................................................................................................................... 367 Shot Peening of Cast Magnesium Alloys T. Ludian, L. Wagner, Chair of Physical Metallurgy and Materials Technology, Technical University of Brandenburg at Cottbus, Cottbus, Germany; M. Hilpert, Otto Fuchs Metallwerke, Meinerzhagen, Germany; A. Kiefer, OSK Kiefer, Oppurg, Germany .................................................................................................................... 374 Shot Peening to Enhance Fatigue Strength of TIMETAL LCB for Application as Suspension Springs J. Kiese, O. Schauerte, Volkswagen AG, Wolfsburg, Germany; J. Zhang, L. Wagner, Chair of Physical Metallurgy and Materials Technology, Technical University of Brandenburg at Cottbus, Cottbus, Germany............................................................................ 380

XII Effect of Test Temperature on Fatigue of Shot Peened Magnesium Alloys J. Wendt, A. Ketzmer , L. Wagner, Chair of Physical Metallurgy and Materials Technology, Technical University of Brandenburg at Cottbus, Cottbus, Germany.................. 386 Effect of Shot Peening on Fatigue Performance of Gamma Titanium Aluminides J. Lindemann, L. Wagner, Chair of Physical Metallurgy and Materials Technology, Technical University of Brandenburg at Cottbus, Cottbus, Germany; D. Roth-Fagaraseanu, Rolls-Royce Deutschland, Dahlewitz, Germany ................................. 392 Mechanical Surface Treatments on the High-Strength Alpha-Titanium Alloy KS 120 J. Kiese, O. Schauerte, Volkswagen AG, Wolfsburg, Germany; J. Zhang, L. Wagner, Chair of Physical Metallurgy and Materials Technology, Technical University of Brandenburg at Cottbus, Cottbus, Germany............................................................................ 399 The Effect of Cold Deformation and Surface Treatment on Fatigue Behaviour of Al2O3-Al6061 Composite Material G. Quan, W. Brocks, Institute for Material Research, GKSS Research Center, Geesthacht, Germany............................................................................................................... 406 Effect of Overloads on Fatigue of Shot Peened 2024 Al V. Šupík , L. Wagner, Chair of Physical Metallurgy and Materials Technology, Technical University of Brandenburg at Cottbus, Cottbus, Germany ..................................... 413 VIII Alternative Mechanical Surface Treatments ............................................................ 419 Alternative Mechanical Surface Treatments: Microstructures, Residual Stresses and Fatigue Behavior I. Altenberger, Department of Materials Science & Engineering, University of California, Berkeley, CA, USA, now at: Institute of Materials Technology, University of Kassel, Germany ............................................................................................... 421 Cavitation Shotless Peening for Improvement of Fatigue Strength H. Soyama, D. Odhiambo , K. Saito, Department of Machine Intelligence and Systems Engineering, Tohoku University, Sendai, Japan......................................................... 435 Comparison of the Residual Stresses Induced by Shot (Stress) Peening and Rolling in Spring Steel E. Müller, Fachhochschule Bochum (University of Applied Science), Bochum, Germany.................................................................................................................... 441 Isothermal Fatigue Behavior and Residual Stress States of Mechanically Surface Treated Ti-6Al-4V: Laser Shock Peening vs. Deep Rolling U. Noster, B. Scholtes, Institute of Materials Technology, University Kassel, Kassel, Germany; I. Altenberger, R. O. Ritchie, Department of Materials Science and Engineering, University of California, Berkeley, CA, USA ................................. 447

XIII Influence of Shot Peening and Deep Rolling on High Temperature Fatigue of the Ni-Superalloy Udimet 720 LI J. Lindemann, T. Raczek, L. Wagner, Chair of Physical Metallurgy and Materials Technology, Technical University of Brandenburg at Cottbus, Cottbus, Germany; K. Grossmann, Rolls-Royce Deutschland, Dahlewitz, Germany ............................................. 454 Shot Peening and Roller-Burnishing to Improve Fatigue Resistance of the (=+>) Titanium Alloy Ti-6Al-4V M. Kocan, L. Wagner, Chair of Physical Metallurgy and Materials Technology, Technical University of Brandenburg at Cottbus, Cottbus, Germany; A. Ostertag, Ecoroll AG, Celle, Germany .................................................................................................... 461 Fatigue Performance of the Mechanically Surface Treated Steels 42CrMo4 and 54SiCr6: Shot Peening vs. Roller-Burnishing D. Wierzchowski, L. Wagner, Chair of Physical Metallurgy and Materials Technology, Technical University of Brandenburg at Cottbus, Cottbus, Germany; A. Ostertag, Ecoroll AG, Celle, Germany .................................................................................................... 468 Process Control Techniques for Laser Peening of Metals R. Specht, F. Harris, L. Lane, Metal Improvemenet Co., Paramus, NJ, USA; W. Wübbenhorst, Metal Improvement Co., Haan-Gruitzen, Germany; D. Jones, Rolls-Royce PLC, UK; M. Hill, UC Davis, CA, USA; L. Hackel, J. Halpin, T. Zaleski, Lawrence Livermore National Laboratory (LLNL), Livermore, CA, USA................................................................................................................. 474 High Temperature Fatigue of Mechanically Surface Treated Materials I. Altenberger, R. O. Ritchie, Department of Materials Science & Engineering, University of California, Berkeley, CA, USA; U. Noster, B. Scholtes, Institute of Materials Technology, University Kassel, Kassel, Germany .............................................. 483 Roller Pressing or Shot Peening of Fir-Tree Root of LPT Blades of 500 MW Steam Turbine M. Sharma, Maulana Azad College of Technology, Bhopal, India .......................................... 490 Comparison of Surface Characteristics and Thermal Residual Stress Relaxation of Laser Peened and Shot Peened AISI 4140 R. Menig, V. Schulze , O. Vöhringer , Institut für Werkstoffkunde I, Universität Karlsruhe (TH), Karlsruhe, Germany ..................................................................................... 498 IX Modeling............................................................................................................................ 505 Finite Element Simulation of Shot Peening – A Method to Evaluate the Influence of Peening Parameters on Surface Characteristics J. Schwarzer, V. Schulze, O. Vöhringer, Institut für Werkstoffkunde I, University of Karlsruhe (TH), Karlsruhe, Germany ................................................................................. 507

XIV Shot Peening and Coverage A. Tange, H. Okada, NHK Spring Co., Ltd., Yokohama, Japan ............................................. 516 Example of the Computer Simulation of Shot Peening Process A. Nakonieczny , Institute of Precision Mechanics, Warsaw; P. Borkowski , P. Wymyslowski , Institute of Air Technical and Applied Mechanics of Warsaw Technical University, Warsaw, Poland ..................................................................................... 523 Modeling of Fatigue Behavior due to Shot Peening Conditions M. Tufft, GE Aircraft Engines, Cincinnati, OH, USA .............................................................. 530 Finite Element Impact Modelling for Shot Peen Forming T. Wang, M. J. Platts, University of Cambridge, Cambridge, UK; A. Levers, Airbus UK Ltd, Chester, UK ................................................................................................................ 540 Theoretical Analysis of Beneficial and Detrimental Effects of Controlled Shot Peening in High Strength Aluminium Alloys C. Rodopoulos , R. Edwards , S. Curtis , J. Romero , Division of Aeronautical Applications, Department of Mechanical Engineering, University of Sheffield, Sheffield, UK; J-H. Choi , Hyundai Motor Company, Korea; A. Levers, Airbus UK, Chester, UK; E. de los Rios, Division of Aeronautical Applications, Department of Mechanical Engineering, University of Sheffield, Sheffield, UK..................... 547 Finite Element Simulation of Shot Peen Forming Y. Zeng, Beijing Aeronautical Manufacturing Technology Research Institute, Beijing, P. R. China.................................................................................................................. 554 Author Index ............................................................................................................................ 563 Subject Index .......................................................................................................................... 567

1

I Applications

2

3

The Application of Mechanical Surface Treatment in the Passenger Car Industry Peter Hutmann BMW Group, Munich, Germany

1

Introduction

At the beginning of automobile manufacturing, about 100 years ago, customer expectations concerning performance and quality were relatively low. People drove cars for the purpose of getting quickly and reliably from point A to point B. Cars were built from conventional engineering materials of that time. Iron and steel constituted 80 % of vehicle weight. Synthetic materials did not exist then. Nowadays, this situation has completely changed. Modern cars must meet a large variety of additional requirements. Among these, the most important are safety, environmental effect, resource preservation, climate comfort and favorable cost of ownership. Ultimate driving performance can be achieved by fulfilling these requirements and, in addition, by providing an individual driving experience, i.e. superior agility, engine power, elasticity in speed changes and corner handling. These high demands on automotive engineering can only be met through the appropriate use of advanced materials in combination with light-weight design.

Figure 1: Materials in the BMW 520i

4 As a result, the composition of materials in passenger cars has changed significantly during the last decades. Fig. 1 shows a typical distribution of the materials of a modern passenger car, in this case the current BMW 5 series. Iron and steel only make up 51 % of the weight, whereas light metals and plastics make up 15 % and 12 % of the weight, respectively.

2

The Role of Materials Technologies in Automotive Engineering

Most materials in modern vehicle have been specially developed for the new automotive requirements. The main objective of the development of advanced materials technologies for automotive applications is to promote the desired properties of a vehicle. This is usually achieved by: • •

• •

technologies improving material properties (e.g. mechanical surface treatment of metal components) technologies where the material has an additional function (e.g. combination of reduced machining costs and enhanced strength of fracture-splitting materials for conrods or crankcases) technologies reducing the number of processing steps (e.g. bake hardening effect during the painting process of body panels which subsequently reduces the number of forming steps) technologies preparing a material for a new application and enabling it to substitute other materials (e.g. replacing cast iron by aluminum alloys for crankcases)

Neither the complexity nor the novelty of these technologies is the key factor. Of prime importance is the contribution to enhanced driving performance achieved by advanced materials technologies. In addition to the manufacturing processes, economical and ecological aspects must also be considered. In this respect, notable achievements made by advanced material technologies include: • • • • • •

reduction of fuel consumption through light-weight material design improvement of comfort and handling performance through light-weight material design for unsprung masses increased power output through light-weight material design for oscillating masses improvement of crash energy management through optimized deformation behavior reduction of cost of ownership through wear resistant material systems maintenance of the original appearance and function throughout the lifetime of the vehicle through corrosion resistant material systems

Many of the above mentioned requirements can only be achieved by increasing the fatigue strength of highly loaded car components. The optimal development of the strength of metals requires mechanical surface layer hardening.

5

3

Concept of Mechanical Surface Treatment

The mechanical surface treatment is based on the elastic-plastic cold-working of the surface. The surface layers are work-hardened and residual compressive stresses are generated. The surface resistance against fatigue crack initiation and propagation, corrosion fatigue or friction fatigue increases significantly and therefore, improves the structural performance under cyclic loading. In addition to that, a reduced surface roughness due to the flattening of roughness peaks can be expected. The increase in surface hardness depends on the hardening potential of the material as well as on the material condition before cold-working. For instance, in the case of a highly hardened initial condition, no additional increase in surface hardness is achieved by mechanical surface treatment. Moreover, mechanical surface treatment of highly hardened metals might even result in a reduction of the surface hardness. The following sections present various examples of advantages and disadvantages of mechanical surface treatment of automotive components.

3.1

Roll Strengthening

We use the example of a stub axle to demonstrate the strengthening potential of rolling. The bold line in Fig. 2 presents loading sequences for a typical consumer, extrapolated until 300,000 km. The dashed line in the same figure shows the loading sequences under racing conditions on the Nuerburgring, extrapolated until 10,000 km. The loading sequences for the Nuerburgring primarily show intermediate and high load levels and a few low load levels. It must be noted that both sequences were measured from the same car.

Figure 2: S-N-diagram for a stub axle before and after strengthening as compared to two loading sequences

6 In addition to these two loading sequences, Fig. 2 shows the Woehler curves for the stub axle before and after roll strengthening. It appears that the life span of the untreated stub axle is shortened under racing conditions. The corresponding S-N-curve and the loading sequence intersect. In other words, the load levels that intersect with the Woehler curve are already high enough to initiate partial failure of the stub axle. However, the S-N-curve of the roll strengthened stub axle is above the loading curves and thus, prevents the stub axle from fatigue failure. Roll strengthening can be used to induce high residual compressive stresses within critical areas of the stub axle. As a consequence, critical tensile stresses due to the loading sequence are successfully reduced. It is known that the residual compressive stresses increase as the rolling force increases. The optimal rolling pressure can be found from measurements of the residual stresses.

Figure 3: Rolling force optimization based on residual stress measurements

The upper half of Fig. 3 shows a diagram of the lifespan of a stub axle as a function of the rolling force. The lower half of Fig. 3 shows the corresponding measurements of the compressive residual stresses inside the critical radius. The quantitative agreement is good. It appears that the maximum compressive residual stresses are reached over a rolling force range of 12,000 N to 14,000N. Rolling forces above 14,000 N lower the residual compression. Similarly, the stub axle loses its fatigue strength in this regime. Fig. 4 shows the change in diameter as a function of the rolling force. Again, in close analogy to the residual stresses, the curve attains its maximum at 14,000 N. Further increase of the rolling force results in a reduction of the diameter. As a consequence, the dimensional accuracy and the surface roughness are no longer satisfying. Moreover, the loss of roughness indicates that very high rolling forces destroy the surface layers by means of material separation.

7

Figure 4: Dimensional change as a function of the rolling force

3.2

Thread Grooving

Another possibility to increase the lifespan of car components is the application of chipless forming processes. The following example demonstrates the increased fatigue resistance of components with internal threads within heavily loaded areas. For functionality reasons, internal threads within critical areas of the part can hardly be avoided. In the case of internal combustion engines, the positioning of the crankshaft bearing is subjected to various constraints. On the one hand, a slender design of the bearing area is highly desirable. In addition to that, due to stiffness requirements, the taphole of the main bearing screw joint should be in the close vicinity of the main bearing boring. On the other hand, the component design must guarantee sufficient fatigue strength. In particular, the taphole of the screw joint between the main bearing and the crankcase is critical. From a mechanical point of view, such a cut thread can be seen as a deep notch in a highly stressed material. During dynamic experiments on a light metal crankcase, fatigue cracks initiated within the thread (Fig. 5). The following factors contribute to the likely initiation of fatigue cracks at the lower end of the thread: • • • •

small wall thickness between the thread and the main bearing boring stiffness discontinuity due to the lower end of the screw significant notch effect due to the thread high stresses due to the mean bearing loading and the pre-stressing of the screw

A first step to improve the design was to increase the length of the critical screw by 5 mm (depth now 30.5 mm instead of 25.5 mm). Thus, the most critical spot moved to a region of thicker cross-sections. The fatigue strength increased from 52 kN to 67 kN. However, the minimum design requirement was still not reached. A significant improvement was achieved by changing the tapping method from cutting to grooving. The resulting fatigue strength of 92 kN was twice as high as the work load of the engine. It must be noted that the thread in the crankcase was no longer critical. The fatigue strength was now limited by the screw.

8

Figure 5: Dynamic experiments on aluminum crankcases; comparison of cut and grooved thread

By contrast to conventional threading, thread grooving does not imply any cutting. Instead, the material is pushed away from the thread valleys. Thus, residual compressive stresses are present in the thread valleys. In the case of dynamic loading, the acting tensile stresses are reduced by the residual compressive stresses. The load level is shifted into a less critical regime, which results in a significant increase of the fatigue strength of the design. Throughout thread cutting, chips are produced. However, in the case of thread grooving no chips are formed. Especially for blind holes a chip-free method is advantageous.

3.3

Shot Peening

The mostly used mechanical surface treatment method is shot-peening. Some major advantages of this method are • • • •

adjustability of the strengthening effect high processing quality easy surface cleaning being well established in the industry

It must be noted that shot-peening is not applied in order to compensate for deficiencies in other steps of the manufacturing process. The following examples demonstrate the use of shot-peening by the automotive industry. The design of the chassis suspension is usually based on fatigue strength. Fig. 6 shows the fatigue strength of barrel springs before and after shot-peening. For this case, shot-peening lengthens the lifespan by a factor approximately equal to 70. The springs were tested under cyclic compressive loading at a constant pre-stressing. However, it

9

Figure 6: Woehler curves for barrel springs (before and after shot-peening)

must be noted that these results were found under lab conditions, i.e. no other effects that might have reduced the lifespan were present. Corrosion has a strong impact on the lifespan of barrel springs (Fig. 7). A salt-spray test was performed (DIN 50021SS, 240 hours). After an optical observation, the barrel springs seemed to fulfill all surface requirements. But the springs did not perform well under fatigue loading. The lifespan was reduced by approximately 40%. The surface protection had to be improved. The standard surface protection is achieved by a layer of paint. The properties of the paint depend on the baking temperature that usually ranges from170 to 190oC. But tempering of mechanically strengthened components reduces the residual stresses and thus, the durability.

Figure 7: Impact of corrosion on the lifespan of barrel springs

10

Figure 8: Influence of the baking temperature on the residual stresses in barrel springs

Fig. 8 presents the reduction in compressive residual stresses as a function of the baking temperature after painting. The duration of the baking process was 30 minutes. It appears that the stress reduction increases dramatically for temperatures higher than 180oC. Therefore, temperatures above 180oC must be avoided. During vehicle operation, the temperature in the barrel springs is about 60 to 80oC. A significant impact on the compressive residual stresses is not expected.

3.4

Shot Peening under Pre-Stressing

An additional increase in strength due to shot-peening may be achieved by the pre-stressing of selected components. We will use the example of a connecting rod. The general design objective for a connecting rod is to keep the oscillating masses as low as possible. This requires very high fatigue strength of the connecting rod shank whereas the scattering must be as low as possible. This is the only way to achieve both high fatigue strength and cost efficiency. In the case of engines that are running within a high revolution range, tensile stresses are primarily due to inertia forces. The fatigue-critical sections of the connecting rod surface have to be optimized along the loading direction. An increased performance due to the use of high strength materials with a tensile strength above 1200 MPa is not possible. Without any additional surface treatment, the notch sensitivity of such alloys limits the fatigue strength. One method to optimize the strength properties is shot-peening under pre-stressing. This procedure induces residual compressive stresses oriented along the loading direction of the connecting rod. Pre-stressing induces additional elastic deformation energy in the component. During shot-peening, this energy contributes to an increase of the dislocation density at the surface of the component. The pre-stressing equipment can introduce compressive stresses up to the order of the yield stress of the material.

11

Figure 9: Shot-Peening under Pre-stressing (schematic)

3.5

Shot Peening – Special Procedures

Challenging engineering tasks often yield innovative solutions. As an example, we will show the engineering of a lighter front-axle for rear-wheel driven cars with front engines. An anti-roll bar is used to limit body roll and thus increase the cornering performance. Typically, it is formed as a torsion bar with a solid cross-section. However, changing the design to a hollow cross-section allows for additional weight savings. A longitudinally welded tube (material 34MnB5) is bent first by cold forming and brought into its desired shape by swaging. Next, to increase the fatigue resistance, shot-peening is applied to the inner surface of the tube. A flexible lance is introduced into the open end of the anti-roll bar. Shot peening media is continuously added to an air flow. The lance tip has a conical deflector so that the steel shot exiting the nozzle is hitting the part at nearly 90 degrees. And finally, after flattening both ends of the antiroll bar, the outer surface of the part is shot peened. As indicated by torsion tests throughout the procedure development, premature fatigue failure occurred unless shot peening was applied to the inner surfaces of the anti-roll bars. Typically, the fatigue crack started at the inner surface. A required 100% increase of the fatigue life could only be achieved by expensive deflector lance peening. After this treatment, the fatigue cracks initiated at the outer surface, which corresponded to the location of the highest stresses. A permanent monitoring of the relevant process parameters is required to insure high quality standards. Among these, the most important are: • • • •

shot media flow-rate peening intensity (air pressure) Initial positioning of the lance Velocity of the lance

The technology presented above is currently used for the hollow anti-roll bars of the BMW 3 and 7 series. As compared to the traditional design with a solid section, weight savings of about 45% could be achieved.

12

Figure 10: Hollow anti-roll bar of the BMW 3 series with shot peened inner surfaces

4

Summary

The development of light-weight and thus fuel efficient and environment-friendly cars is one of the most important challenges facing the automotive industry. Methods to increase the performance and strength of individual components are very important. Shot peening is competing with various other mechanical and thermo-mechanical methods for surface treatment. Requirements to the component design as well as manufacturing conditions determine the choice of the optimal method. However, shot peening is a well-established method that efficiently increases the fatigue strength of highly stressed automotive components.

13

Life Enhancement of Aero Engine Components by Shot Peening: Opportunities and Risks Gerhard W. König MTU Aero Engines, Munich, Germany

1

Abstract

In service, aero engine parts are subject to high temperatures and extreme cyclic loads. As a consequence, the initiation and propagation of surface fatigue cracks is life-limiting for many components. The application of surface treatments, such as shot peening, can cause a shift of the crack initiation site from surface to subsurface and concomitantly, a significant increase in the cyclic life. However, in practice, the actual improvement achieved by shot peening is sometimes found to be less than expected. One reason is that residual stresses produced by shot peening are not stable against high temperatures and non-elastic deformation induced by service loads. Another reason is that defects (intrinsic defects of the material, surface defects caused by handling as well as surface damage caused by peening itself) may also reduce the benefit of shot peening. The following paper discusses the possibilities to exploit the benefits of shot peening for life enhancement of aero engine components and the requirements that must be fulfilled to take full advantage of shot peening.

2

Introduction

Shot peening is a very common method for increasing the fatigue strength and has frequently been used for aero engine applications for many years. Aero engine components are characterized by high-strength materials (titanium and nickel base materials), high cyclic loads (often locally exceeding the yield strength) and high service temperatures (in the region of timedependent material behavior). In the past, shot peening has been used for safety-critical parts mainly as a measure to increase safety margins rather than for extending service lives. However, with increasing understanding of the mechanisms of shot peening and increasing predictive capabilities, the opportunities of increasing service lives by shot peening or other surface treatments are expected to grow in the future.

3

Life Enhancement by Shot Peening

The experience from laboratory tests, component tests as well as failure analysis clearly indicates that the large majority of fatigue cracks initiate from the surface. The crack origin is either directly at the surface or slightly subsurface within a distance of less than about 25 microns to the surface. As an example, Figure 1 shows the result of a component test performed on a titanium compressor disk (simulating 36000 start/stop cycles). After this test, cracks were found in the disk bore and in the cooling holes as well as in the lobes (blade fixtures). The fractographic

14 evaluation showed that all these cracks had their origin directly at the surface. There are several reasons for the surface to be the preferred location of fatigue crack initiation: For most situations, there is a negative stress gradient at the surface (Figure 2). The reduction in stress below the surface (say at a distance of 0.1 mm) is not very important for unnotched features (such as the bore) but is far from negligible for notched features. Furthermore, surfaces are prone to specific defects produced by machining (e. g. wear or ductility exhaustion) or handling (e. g. scratches). Another important aspect is that the fatigue resistance of the bulk material is generally higher compared with that of the surface (Figure 3): there are two distinct stress-life curves for surface and for internal initiation. From this diagram it is obvious that life benefits are to be expected from any surface treatment capable of shifting the crack initiation site from surface to subsurface and that the maximum possible life benefit is limited by the internal initiation curve.

Figure 1: Example of surface crack initiation at various locations of a compressor disk

Normalized stress Disk bore 1

0,95

Cooling hole 0,9

0,85

Typical depth affected by shot peening

0,8

Bolt hole 0

0,05

0,1

0,15

0,2

Depth from surface (mm)

Figure 2: Stress gradient below the surface for various disk features (bore, cooling and bolt hole)

15

4

Stress amplitude

Maximum life benefit from surface treatments

Surface initiation Parameters: Material Residual stress profile Hardness Surface integrity

Internal initiation Parameters: Material Stressed volume Microstructure Stress gradient Temperature

Number of cycles

Figure 3: Surface crack initiation vs. internal crack initiation

The capability of shot peening to suppress surface crack initiation is mainly due to two effects: The impact of the shot leads to heavy work hardening in the surface layer with a concomitant increase in fatigue strength (Figure 4). Furthermore, shot peening produces compressive residual stresses in the surface layer. Typically, the amount and depth of these compressive stresses (an example is shown in Figure 14) are significantly higher than those produced by standard manufacturing methods such as turning, drilling or broaching. The enhancement of fatigue lives through compressive residual stresses relates both to crack initiation and propagation. Crack initiation life modeling considers the lowering of the mean stress and the resulting increase in fatigue lives: Figure 5. The prediction of crack propagation life (based on fracture mechanics) assumes crack-like surface defects. Indeed, fractographic observations support the

Figure 4: Fatigue strength increases with hardness; defects disturb this relationship

16 existence of crack-like defects (e. g. scratches or metallurgical defects) at the origin of engineering surface cracks (Figure 6). According to fracture mechanics estimations, compressive residual stresses may cause arrest (of small cracks) or retardation of crack growth (Figure 7). For large cracks even an acceleration is possible, caused by tensile stresses (which balance the compressive stresses).

5

Stress

Smean

unpeened

Decrease in

Strain range De

mean stress

Stress

by peening

Strain 10 4 cycles Life benefit

peened

10 5 cycles Smean

sm

Smean Compressíve

Strain

residual stress

Figure 5: Model to predict the benefit of compressive residual stresses on fatigue life caused by a reduction of the mean stress Smean

Figure 6: Examples of crack-like surface defects found at the origin of fatigue cracks Scratch produced by handling Non-metallic inclusion (nickel base alloy)

17

Broken line: unpeened

Crack size

Solid line: peened Crack acceleration a3

Crack retardation

a2

Non-propagating crack

a1 Number of cycles

Figure 7: Effect of shot peening on the propagation behavior of surface cracks

4

Factors Reducing or Preventing Life Improvements by Shot Peening

Numerous specimen and component tests verify the capability of shot peening to improve fatigue life (Figure 8) as well as to shift the crack initiation from surface to subsurface (Figure 9). However, there is also ample evidence of shot peening resulting in life debits rather than benefits (Figure 8). Therefore, the exploitation of shot peening benefits requires detailed knowledge about the mechanisms and relevant parameters. These are addressed in the following discussion.

Figure 8: Examples of benefit and debit of fatigue lives produced by shot peening

18

150180 μm

Surface initiation

Subsurface initiation

Surface: drilled & polished

Surface: shot peened

Figure 9: Examples for the shift of the crack initiation site caused by shot peening Notched specimen; titanium alloy

4.1

Local Variations of Residual Stresses

The process of shot peening of engineering components is subject to considerable statistical scatter including local variations of the number of impacts, the size and mass of individual shot particles, their angles of impingement etc. As a result, residual stresses are subject to considerable statistical variations, too. Figure 10 shows residual stress measurements (X-ray measurements at the surface) at various identical locations on a titanium disk. Evidently, increasing the coverage from 100 to 200 percent increases the mean value of the compressive stress and redu-

Probability

Coverage: 100%

Coverage: 200%

Compressive stress in MPa Figure 10: Influence of coverage on mean values and scatter of residual stresses

19 ces the scatter. In order to ascertain a specified minimum level of residual stresses, a high coverage is desirable. But this conflicts with another effect of shot peening: the damage in the surface (see below) also increases with increasing coverage. Therefore, the process has to be optimized such that sufficient values of residual stresses are ensured while keeping the peening damage low.

4.2

Residual Stress Shake-Down Caused by Thermal Exposure and Plastic Strain

After shot peening, residual stresses are typically close to the compressive yield stress of the material. However, these compressive stresses are not stable under the temperature and plastic deformation imposed under service conditions. The mere exposure to service temperatures may already reduce residual stresses by 50% or more (see Figure 11). The underlying mechanism is very similar to the well-known stress relaxation observed for creep under constant strain. In addition, further reduction of residual stresses is experienced through monotonic loading (such as that occurring in the first half of a fatigue cycle). This behavior is a consequence of the flat slope of the typical stress-strain curve for a high-strength material. Compressive residual stresses are reduced with increasing tensile strain and tend to disappear completely (Figure 11). Similarly to monotonic plastic deformation, cyclic strains may also lead to further reductions of residual stresses with increasing number of cycles. This effect is well established by specimen tests. However, for many applications, the cyclic amplitudes are low enough such that this effect is negligible compared to thermal and monotonic relaxation. Monotonic and cyclic relaxation are the reasons why residual stresses are not effective at high cyclic loads. This is illustrated in Figure 12 comparing test results of peened and unpeened specimens of a nickel base alloy. At low amplitudes, crack initiation is observed to be subsurface both for the peened and unpeened conditions, whereby peening has no effect on fatigue life. At high amplitudes, relaxation effects remove the residual stresses initially produced by shot

Residual stress in MPa

Titanium alloy IMI834; 500 deg. C 0

-200

Tensile strain

-400 -600

Thermal exposure: 24h

-800

-1000 0

0,2

0,4

0,6

0,8 0,8

11

1,2 1,2

% Tensile strain in % Figure 11: Modeling of residual stress shake-down caused by thermal exposure and subsequent tensile strain at 500 deg. C

20 peening such that the behavior of peened and unpeened specimens becomes very similar (with surface crack initiation in both cases). Only at intermediate strain amplitudes is there a significant difference between peened and unpeened specimens: The unpeened specimens show initiation at the surface while the compressive residual stresses of the peened specimens shift the initiation location to subsurface, leading to longer fatigue lives. This example shows that there is no general life enhancement to be expected from shot peening. Benefits are limited to “windows” of loading parameters. Of course, these windows are very much dependent on parameters such as material, surface conditions, temperature, loading conditions etc. The reliable prediction of these effects is a continuous challenge for testing as well as for modeling of the material behavior. Closed symbols: surface initiation

Strain amplitude (per mill)

Open symbols: subsurface initiation Residual stress

6

No peening benefit at high loads Initiation: both at the surface

“breakdown”

5

peened

4

Peening benefit at intermediate loads Initiation: surface

3

subsurface 2

unpeened

No peening benefit at low loads both subsurface initiation

1 1,E+03

1,E+04

1,E+05

1,E+06

1,E+07

Number of cycles

Figure 12: Effect of peening on LCF life (nickel base alloy)

4.3

Defects in the Material and in the Surface

The fatigue strength of a structure is determined by its weakest element. Therefore, the minimum life is often related to the behavior of defects. In the following, the role of defects on the shot peening effects is considered. 4.3.1 Volume Defects As shown in Figure 3, the fatigue life of a subsurface volume element is longer than that of an equivalent element at the surface. In principle, this is also true in the presence of defects: A given defect is less detrimental in the volume than at the surface. However, there is a marked difference with regard to defect sizes between surface and subsurface because of the statistical size effect: The density of large defects is much lower than that of small defects (Figure 13). For an unnotched feature, the volume of the surface layer typically represents less than 1% of the highly stressed volume. Therefore, the worst defects that can be found in the volume tend to be larger than those in the surface layer. In the situation of very large internal defects these may

21 become the life-limiting feature and surface treatments are no longer of any benefit (this effect can be quantified on the basis of probabilistic life models). To avoid this situation, it is necessary to control defect densities by choosing appropriate materials and manufacturing routes. In addition to the statistical size effect, technological size effects also have to be considered (e. g. local degradation of material properties due to unfavorable conditions of forging or heat treatment). Defects per volume

Stressed volume

Probabilistic Model

Defect size a

Cyclic life

Figure 13: Volume-dependent cyclic life due to the statistical size effect

4.3.2 Surface Defects The increased hardness of a peened surface improves the resistance against handling damage. But the main advantage of shot peening is the capability to reduce the harmfulness of surface defects. Crack initiation from defects can become suppressed and crack propagation stopped or delayed (see Figure 7). Therefore, shot peening is a very effective means of enhancing minimum lives. This is even true considering the fact that considerable surface damage is produced by the process of shot peening itself. An illustration of such behavior is the increase in surface roughness or the smearing of material (Figure 15). However, in most cases, the life benefit outweighs the debit due to additional surface damage. But there is an important limitation: The defect depth must not exceed the depth of the compressive residual stress profile. Figure 14 shows Nickel Nickelbase basealloy alloyatat500 500deg. deg.CC

Factor of life improvement

Stress in MPa

Strain Strainrange: range:0.65% 0.65%

10 10

100 0 -100 -200

unpeened

-300 -400 -500

peened

-600 -700 0

0,2

0,4

Depth in mm

0,6 0,6

0,8 0,8

11

11 0,01 0,01

0,1 0,1

11

Figure 14: Decreasing life benefit with increasing defect size Residual stress depth profile (left) Fracture mechanics calculation for a semi-circular surface defect (right)

22 that the factor of life improvements dramatically decreases with increasing defect depth. Similarly, the fatigue strength benefit due to surface hardening is also reduced by defects (Figure 4). Again, there exists a “window” of defect sizes where shot peening is effective. Benefits can only be achieved when surface defects are kept below a size determined by the residual stress profile.

Smearing caused by

Increase in surface roughness

Smearing caused by

Increase in surface roughness

due to steel-wire shot peening due to steel

-

Non-optimized peening of a corner

Non

-

optimized peening of a corner

wire shot peening

Figure 15: Examples of surface damage produced by shot peening

5

Summary and Conclusions

Shot peening is an effective means to enhance fatigue life of aero engine components operated at severe loading conditions. The life benefits are restricted to “windows” of boundary conditions with key elements including loading parameters, residual stress profiles, work hardening and inelastic material behavior as well as the size, location and density of defects. Important aspects of shot peening effects can be predicted on the basis of elasto-plastic material models and fracture mechanics as well as specially-designed specimen tests. Further progress in life enhancement can be obtained by development of surface treatments with the following properties: • Capability to produce deep profiles of compressive stresses and work hardening • High stability of residual stresses and work hardening with regard to thermal exposure as well as inelastic deformation • Minimizing surface damage caused by the surface treatments • Good reproducibility of surface properties • Suitability for different geometric requirements

23

Study on Methodology to Increase Fatigue Limit of Gears Kotoji Ando1), Katsuyuki Matsui2) and Hirohito Eto2) 1) 2)

1

Yokohama National University, Yokohama-shi, Japan, Isuzu Motors Ltd., Kawasaki-shi, Japan

Introduction

At present, improvement of the fatigue limit of automotive components is a top priority. The demand is especially high for automobiles due to the environmental and fuel economy expectations. To improve the fatigue limit of automotive components, the following three methods are currently in common use: (i) Minimize surface roughness, (ii) Increase the hardness of the material, (iii) Introduce a large compressive residual stress at the surface. Item (iv) is an extremely adequate method; therefore, a close attention is being paid to achieve a minimized surface roughness at this time. Although item (ii) is a reasonable method, it is difficult to apply to the automobile gears and springs with hardness as high as 600–700 HV. Fatigue limit (DIw) is proportional to hardness up to 400–500 HV. Above 500 HV, DIw does not increase with hardness, but decreases as hardness increases. Therefore, it is not appropriate to increase hardness further for components such as gears and springs. Item (iii) is a general method and shot peening is used widely. However, it is extremely difficult to introduce a large compressive residual stress by shot peening since the hardness of automotive components reaches 600–700 HV. To solve above problems and improve the fatigue limit of automotive components, the authors conducted a study focusing on the following points: (a) What is the process of fatigue fracture and what is the resistance factor in each stage? (b) What stress ratio (R) can be applied to automotive components? (c) Why does the fatigue limit start decreasing when hardness reaches some level? Is there any way by which the decrease can be inhibited? (d) How can a large compressive residual stress be introduced to a material with 700HV or more? As a result of this study, it was found that the number of components subjected to cyclic loading with positive stress ratios (R > 0) is unexpectedly high among automotive components such as gears and springs. Therefore, after working closely on the above four points concentrating on the R > 0 components, the following results for improvement of the fatigue limit were proposed: (1) Increase the hardness of materials as high as possible. (2) Introduce compressive residual stress as high and deep as possible. (3) Decrease the grain size as much as possible. (4) Grind the surface region of components to remove early stage fatigue damage such as extrusion and intrusion, and stage I fatigue crack. (5) Heal the stage I fatigue crack during service if possible [1]. In this paper, the fatigue limit range of a gear was improved considerably, simply and economically, using above methods (1), (2) and (3) together.

2

Fatigue Process and Resistance Factor to Fatigue Fracture

The process of fatigue fracture is introduced in many references [2]. According to these references, the process of fatigue fracture consists of the following seven steps: (1) Cyclic stress be-

24 low yield stress activates the dislocations. (2) The activated dislocations then create a slip band inside a grain. (3) Extrusions and intrusions are formed near the surface region. (4) A stage I fatigue crack is formed along the intrusion and then propagates through few grains. (5) The stage I crack is then transformed to a stage II fatigue crack by cyclic tensile stress. (6) Propagation of the stage II fatigue crack. (7) Final fracture. In the above seven steps, promoting and resistance factor for fatigue damage of each step were considered systematically as shown in Table 1. The promoting factor of early stage of stage I crack is a cyclic shear stress. And to resist the damage and the crack propagation, it is important to increase the hardness as high as possible and to make the grain size as fine as possible. The promoting factor of stage II crack is a cyclic tensile stress. To increase the resistance of the stage II crack propagation, it is important to introduce a large compressive residual stress. Tange et al. [3] showed that it is very important to increase the compressive residual stress at the surface in fine grained steel. From the statements above, it can be concluded that using a combination of the following methods improves the fatigue limit of the components subjected to R > 0 loading: (1) Maximizing the yield stress (the hardness). (2) Introducing a large and deep compressive residual stress. Introducing a large compressive residual stress at the surface improves the fatigue strength especially for fine grained steel. (3) Decreasing the grain size as much as possible. Table 1: Promoting and resistance parameter for fatigue damage at each stage

:contribute considerably, ×: have no relation

3

Application to Test Gears

3.1

Materials, Samples and Test Method

Two kinds of steels were used in this study. Chemical compositions (wt.%) of these steels A and B are; C: 0.19, 0.51, Si: 0.06, 0.20, Mn: 0.84, 0.74, P: 0.010, 0.02, S: 0.019, 0.02, Ni: 0.09, 0.04, Cr: 0.11, 0.11, Cu: 0.09, 0.08 and Mo: 0.4, 0.0, respectively. Both steels were quenched and tempered to a hardness level of 200 HV. Then, they were machined to a gear (module: 3, number of teeth: 36, helix angle and hand: 17° right hand, pressure angle = 14°30’, over-ball diameter: 123.6 mm). Six kinds of gears were made. Gears I-IV were made of steel A while the gears V and VI were made of steel B. The surface treatment techniques adopted are Vacuum

25 Carburizing (VC), Contour Induction Hardening (CIH) and Double Shot Peening (DSP). After machining, the gears were surface-treated with these combined treatments. Table 2 shows the combined surface treatments for each gear. For example, gear IV was first vacuum carburized, then contour induction hardened and finally double shot peened. The gears I-IV were first vacuum carburized to C = 0.8 wt.%. The vacuum carburizing conditions were: pressure in furnace = 6.67 · 10–2 kPa, temperature = 1223 K, atmosphere = C3H8 gas, carburizing time = 2.88 ks. After being carburized, the gears were cooled to 1173 K and subsequently quenched using N2 gas of 5 · 102 kPa. The gears III–VI were induction hardened. The contour induction hardening conditions were: 3 kHz frequency for pre-heat, 1000kW power for pre-heat, 150 kHz frequency for main-heat and 600 kW power for main-heat. Two kinds of shots were used: Ø0.6 mm shot for primary peening and Ø0.08 mm shot for secondary peening. Using shot sizes in this sequence is the key to introduce an appropriate compressive residual stress near the surface in the gear [4]. The hardness of both shots is 700 HV. Peening conditions were: 490 kPa and 392 kPa air pressure, 0.35 mmC and 0.26 mmN arc heights, respectively. Details of the shot peening parameters and their effects are explained in [4]. To measure the residual stress and the volume fraction of the retained austenite (CR), a micro X-ray stress measuring apparatus was used. The gear surface was masked with Ø5 mm window, and was polished to a specified depth using electrolytic polishing method. The X-ray conditions are: Cr-K= beam X-ray spectrum and 0.3 mm X-ray beam injection diameter. The fatigue testing was done using an electro-hydraulic testing machine at a stress ratio of R = 0.1, 10 Hz frequency and sine wave load cycle in air. Table 2. Gears and combined surface treatment Gear

I

II

III

IV

V

VI

Steel

A

A

A

A

B

B

Surface Treatment VC

3.2

VC+DSP VC+CIH

VC+CIH+DSP CIH

CIH+DSP

Retained Austenite

Fig. 1 shows the distribution of CR. The open diamond symbol shows CR of gear I. The CR at the surface is 11.5 % and the maximum CR is 26.8 %. The solid square symbol shows CR of gear II. The CR at the surface is very low (1.8 %) and the maximum one is 16.5 %. The CR of gear II was reduced considerably as compared to gear I. This CR reduction is attributed to double shot peening. The CR of gear III is shown by the open triangle symbols in Fig. 1. The CR at the surface and the maximum are 24.5 % and 31.3 %, respectively, showing extremely high values. CR of gear IV at the surface and the maximum values are 3.4 % and 21.2 %, respectively. Similar to gear II, CR of gear IV was reduced drastically meaning that the retained austenite was transformed to martensite by strain transformation caused by DSP.

26

Figure 1: Depth profile of retained austenite (CR) in the roots of the gear teeth.

3.3

Residual Stress– Depth Profiles

Fig. 2 shows the residual stress distribution of gears I-IV. The open diamond symbols show the residual stress distribution of gear I with the compressive residual stress at the surface (Is) of about 300 MPa and the maximum compressive residual stress (Imax) of about 400 MPa, respectively. Both values are not so high. Solid square symbols show the residual stress distribution of gear II. Imax was introduced at the surface with a value of 1838 MPa. The value of Imax is surprisingly high. Open triangles show the residual stress distribution of gear III. The Is and Imax are 801 and 1054 MPa, respectively. In gear IV, the maximum residual stress was also introduced at the surface with a value of 1862 MPa. Even 300 mm below the surface, a very high compressive residual stress of 900 MPa was measured. Extremely high residual stresses were introduced in gears II and IV for the following reason: the retained austenite was transformed to

Figure 2: Depth profiles of residual stresses (IR) in the roots of the gear teeth.

27 martensite by DSP, resulting possibly in a quite large compressive residual stress. In gear V, Is and Imax values are only 662 and 810 MPa, respectively. In gear VI, values of Is and Imax are 1159 and 1346 MPa, respectively. These compressive residual stresses are much higher than those of gear V. The only possible reason for high compressive residual stress in gear VI is that retained austenite was transformed to martensite by strain transformation caused by DSP. No retained austenite was present in gear V. On the other hand, compressive residual stresses in gear VI were not very high compared with those in gear IV. The occurrence of transformation made this difference in residual stress values.

3.4

Hardness–Depth Profiles

Fig. 3 shows the hardness–depth profiles in gears I–IV. The open diamonds and triangles symbols show the hardness distribution of gears I and III, respectively. The highest hardness of gears I and III are 799 HV and 893 HV, respectively. However, the highest hardness of gears II and IV are 1040 and 1067 HV, respectively, showing much higher hardness than those of gears I and III. These results are attributed to the strain induced martensite transformation as previously mentioned. The highest hardness of gear V is 757 HV. The highest hardness of gear VI is 792 HV. The maximum hardness of gear VI is 35 HV higher than that of gear V. This result can be attributed to the work hardening by DSP.

Figure 3: Vickers hardness (HV)–depth profiles in the roots of the gear teeth.

3.5

Fatigue Strength

Fig. 4a shows S-N curves of the gears I, II and IV. The open diamonds, solid squares, solid circles show the S-N data of gears I, II and IV, respectively. The gear III was not fatigue tested. The stress range at the fatigue limit DIw of gear I is 883 MPa, while that of gear II achieved an

28 increase of about 118 % up to 1931 MPa, and a further increase of 150 % to 2207 MPa was realized with gear IV. These surprising increase in the fatigue limit are attributed to the following two reasons: (a) The carbon content at the surface of these gears is about 0.8 % and the retained austenite near the surface was reduced considerably by strain induced martensite transformation. (b) The hardness near the surface is over 1000 HV, thus, the material has high resistances to initiation and propagation of stage I fatigue cracks. Fig. 4b shows S-N curves of gears V (solid triangles) and VI (solid circles). The stress range of the fatigue limit DIw of gear V is 1256 MPa, while that of gear VI is increased by 38 % to 1710 MPa. Both gears showed similar hardness. It can be said that the difference in the fatigue limit resulted from the different compressive residual stress distribution of both gears. The conclusion is that double shot peening played a key role in increasing the fatigue limit.

a) Gear I, II and IV

b) Gear V and VI

Figure 4: S-N diagram of gears

3.6

Regression Analysis of the Fatigue Strength

The fatigue tests were conducted on 8 kinds of gears by these authors. All tests were performed under R = 0.1 loading. To understand the important factors to increase the fatigue limit, a regression analysis was made using the above 8 data with special attention to the following three parameters: (a) Yield stress converted from Vickers hardness (HV). (b) Maximum compressive residual stress Imax. (c) Grain size dC.The information required for the regression analysis (HV, Imax and dC) of 5 gears out of 8 is included in this paper. The yield stress IY was estimated from HV: IY = 3.27 HV (MPa). Fig. 5 shows correlation between DIw and {0.478 (IY + Imax) + 1.363dC–1/2 – 894}. From Fig. 5, it can be seen that {0.478 (IY + Imax) + 1.363dC–1/2 – 894} is an important parameter to increase fatigue limit. DIw is given by the following equation DIw = 0.478 (IY + Imax) + 1.363 dC –1/2 – 894(1)

29

Figure 5: Regression analysis of experimental fatigue data.

and this parameter shows good agreement with our proposal (1) to (3) in the chapter 2 for increasing fatigue limit.

4

Conclusions

To increase the fatigue limit of car components, fatigue processes were analyzed and new methodology for increasing fatigue limit was proposed. To achieve an increase in fatigue performence, combined surface treatments were applied to the gears. These treatments were: Vacuum Carburizing (VC), Contour Induction Hardening (CIH) and Double Shot Peening (DSP). Using these treatments, the following results were obtained: (1) With VC, the carbon density at the surface was increased up to 0.8 wt.%, and grain boundary oxidation was completely prevented. (2) With CIH, the grain size was refined to about 5 mm. (3) With DSP, most of the retained austenite near the surface was transformed to martensite, resulting in increased hardness up to 1067 HV and extremely high compressive residual stresses. (4) With the above combined effects, the stress range of the fatigue limit (R = 0.1) was increased up to 2207 MPa. (5) The stress range of the fatigue limit of gears DIw is given by the equation (1) as a function of yield stress, the maximum compressive residual stress and the average grain size. This result shows good agreement with the new methodology proposed in this paper based on the fatigue process analysis.

30

5 [1] [2] [3] [4]

References K. Ando, K. Furusawa, M. C. Chu, T. Hanagata, K. Tuji and S. Sato, J. Am.Ceram.Soc. 2002, 84 (9), 2073. K. J. Miller, Materials science perspective of metal fatigue resistance, Materials Science Technology, 1993, 9, 453. A. Tange, Thesis for doctorate to Yokohama National Univ., 2001. H. Ishigami, K. Mastui, Y. Jin, and K. Ando, Fatigue Fract. Engng. Mater. Struct., 2000, 23, 959.

31

Searching for the Most Suitable Condition and the Suggestion of Each Application in Ultrasonic Shot Peening Kaneshi Hattori, Yoshihiro Watanabe, Mitsuru Handa Tokyo Seiko, Japan

Jean-Michel Duchazeaubeneix Sonats, France

1

Abstract

Abstract

Ultrasonic shot peening, named Stressonic®, developed by SONATS, is being popularized among aeronautic industries for production parts and maintenance application because of its compact equipment that allows treatment partially and it can be integrated into production line. However, Stressonic® has a big difference in processing method compared to that of classical shot peening. Therefore, the peening management, which influences the peening effect, needs to be renewed. In this report, treatment conditions as amplitude, shot media (material, diameter), injecting distance, are set as Stressonic® management items. Then the most suitable Stressonic® condition is experimented by intensity and residual stress distribution obtained from the management above. Stressonic® application is also studied using parts, such as gears and springs, and confirmed its quality after Stressonic® process. As a result of study and experiment, Stressonic® application has an outstanding peening effect. Key words: ultrasonic shot peening, bearing ball, tungsten carbide ball

2１ Introduction Introduction Ultrasonic shot peening (USP) treatment has a lot of characteristics which are different from classical methods in principle and results. The outline of the principle is shown in Figure 1. A piezoelectric transducer emits ultrasonic waves at 20kHz. The waves are amplified when they travel through an acoustic booster. Finally by way of SONOTRODE, the kinetic energy is transmitted to shots. The dimension of the vibrating part, SONOTRODE, which contacts shots allows vibration amplitudes of 50 to 200 microns to be attained. In case the amplitude is 90 microns, the injection velocity of shots is approximately 10 to 20 m/sec. The shots strike the vibrating walls and are reflected off the surface. Then they collide with one another. The balls are scattered randomly throughout their encasing, like molecules of gas. A homogeneous treatment is obtained on the injected surface. At this study, the effects of shot amount and injection distance were examined. Generally stainless balls, bearing balls or tungsten carbide balls which have high spherical accuracy and smooth surface are used as the shot media. The surface roughness after USP treatment becomes the smoothest that could never obtained in the classical methods

32 using high hardness shot to apply high compressive residual stress distribution. In USP treatment, combined (so-called multi) shot peening is not required because of the applied high compressive residual stress distribution and the smoothest surface. At this study, the housing for a carburized gear and a nitride valve spring are proposed, also the residual stress distribution and the surface condition are examined. Further the system to harden the bearing ball is introduced and the hardness distribution is measured.

・Shot media ・Housing

・Sonotrode ・Booster

・Piezo transducer

・Generator

Figure 1. Principle of USP treatment

32 Experimental Experimental Procedures procedures First of all, the effect of the shot amount on the intensity was examined. The variation of coverage was observed by pressure sensor film. And the intensity was measured by almen Table 1. Test (1) Conditions SONOTRODE Shot

Injecting condition Pressure sensor film

Diameter Amplitude Material Diameter

(mm)

(mm)

70 90 SUS304 0.6

Hardness Amount Possession rate

(HV) (g) (%)

461 2.5 , 5.0 , 10.0 13 , 26 , 52

Injecting time Injecting distance Measurable range

(sec) (mm) MPa

5 30 0.5 ∼ 2.5

(μｍ)

33 A-strip. The test conditions are shown in Table 1. Figure 2 and 3 show the shape of shots and an application to measure intensity.

Figure 2. φ0.6mm shot

Figure 3. Intensity measuring application

(2) Relation between amount of shot and intensity, and 100% coverage time were measured minutely. Also the influence of injecting distance were investigated. Test conditions are shown in Table 2. Table 2. Test (2) Conditions SONOTRODE Shot

Injecting condition

Diameter Amplitude Material Diameter Hardness Distance

(mm)

70 90

(μｍ) (mm) (HV) (mm)

SUJ2 1.2

Tungsten carbide 0.86

850

1,500 10, 20, 30

43 Results Results The result of test (1) is shown in Figure 4. The uniform color developed in peened area (70mm diameter) of each condition show that Ultrasonic shot peening was treated homogeneously. Also both of the coverage and the intensity show a constant tendency to decrease when the shot volume is increased. It seems that there is interference among the shots themselves. The result of test (2) is shown in Figure 5. Figure 5a and 5b show there are the best conditions which can satisfy both of the higher intensity and the shortest full-coverage time. Figure 5c shows that the distance is an important factor in USP, because doubled (tripled) distance requires treatment time of the same ratio (doubled or tripled).

34 Shot amount

2.5 g

5.0 g

10.0 g

7 70 mm Results

Intensity

0.058 mmA

0.041 mmA

Figure 4. Intensity and result of pressure sensor film

0.035 mmA

35

54 Suggestion Suggestions In this section, some applications for a gear and a spring are proposed. And the qualities (residual stress distribution and surface condition) after USP treatment are reported. Also the circulation system to harden the bearing ball is introduced which developed in TOYO SEIKO and the distribution of hardness is also reported. The image how to shot-peen gear in USP treatment is shown in Figure 6.

Rotaiting

Housing

SONOTRODE

Figure 6. Application for gear The residual stress distribution after USP treatment to carburized gear is shown in Figure 7. The result of SEM observation of the peened surface is also shown as Figure 8. There is few changes in the roughness value before and after.

Figure 7. The residual distribution

Figure 8. SEM observation of the surface afterUSPwithtungstenballs0.86mm

36 The image of the application for a valve spring is shown in Figure 9. The spring was pressured from both sides. In case of nitride spring, an endowed compressive residual stress value could exceed –2000MPa at the peak. The residual stress distribution after USP treatment to nitride spring is shown in Figure 10.

Figure 9. The application for a spring

Figure 10. The residual stress distribution

Bearing balls were shot peened themselves at USP treatment. To manufacture more durable bearing balls, the circulation system for balls under 0.6mm diameter was developed. The result of hardness distribution before and after is shown in Figure 11. The work hardened layer are confirmed.

Figure 11. The distribution of hardness

6 5 Conclusion Conclusions Ultrasonic shot peening has unique characteristics summarized as below. 1) Shot amount, injecting distance are important factors to obtain the highest peening effect. 2) This is a suitable method to obtain a larger residual compressive stress distribution without a deterioration of surface condition.

35

Consideration of Shot Peening Treatment Applied to a High Strength Aeronautical Steel with Different Hardnesses Marcelo A. S. Torres1, Marcelino Pereira do Nascimento2, Herman Jacobus Cornelis Voorwald2 1

Department of Mechanics/ 2Department of Materials and Technology -State University of São Paulo, 333, Ariberto Pereira da Cunha Ave., Guaratinguetá, São Paulo, 12516-410, Brazil

1

Introduction

One of the most important components in a aircraft is its landing gear, due to the high load that it is submitted to during, principally, the take off and landing. For this reason, the AISI 4340 steel is widely used in the aircraft industry for fabrication of structural components, in which strength and toughness are fundamental design requirements [1]. Fatigue is an important parameter to be considered in the behavior of mechanical components subjected to constant and variable amplitude loading. One of the known ways to improve fatigue resistance is by using the shot peening process to induce a compressive residual stress in the surface layers of the material, making the nucleation and propagation of fatigue cracks more difficult [2,3]. The shot peening results depend on various parameters. These parameters can be grouped in three different classes according to R. Fathallah et al [4]: parameters describing the treated part, parameters of stream energy produced by the process and parameters describing the contact conditions. Furthermore, relaxation of the CRSF induced by shot peening has been observed during the fatigue process [5-7]. In the present research the gain in fatigue life of AISI 4340 steel, obtained by shot peening treatment, is evaluated under the two different hardnesses used in landing gear. Rotating bending fatigue tests were conducted and the CRSF was measured by an x-ray tensometry prior and during fatigue tests. The evaluation of fatigue life due the shot peening in relation to the relaxation of CRSF, of crack sources position and roughness variation is done.

2

Experimental Work

The chemical composition of AISI 4340 used is 0.41C-0.73Mn-0.8Cr-1.74Ni-0.25Mo-0.25Si, weight percent. The mechanical properties of base material from the 53HRC are: yield strength 1511 MPa, ultimate tensile 1864 MPa. These properties were obtained by means of quenching from 815 ºC followed by double tempering in the range (230±5) ºC for 2 hours. The mechanical properties of base material from the 39HRC are: yield strength 1118 MPa, ultimate tensile 1240 MPa. These properties were obtained by means of quenching from 815ºC followed by tempering in the range (520±5) ºC for 2 hours. Shot peening treatment was performed with steel shot in 0.008A peening intensity. This Almen intensity was adopted for the following reason: in the previous studies this peening intensity when applied on surface of AISI 4340 steel prior to the hard chromium plating resulted in significant recovering of fatigue strength. [8]. In the 53HRC condition, the effect of the shot peening pre treatment of hard chromium plated AISI 4340 steel, in the same intensity, was still better. The process parameters were: outflow 3 kg/min, speed 250 mm/min, distance 200 mm and rotation 30 rpm, shot S230 (Æ 0.7mm), coverage 200 %

38 carried out with an air-blast machine according to standard MIL-S-13165. The shot peening treatment was done with high quality control, in which the shots are automatically selected and kept in perfect conditions. The specimens used were tested in rotating bending fatigue tests at frequency of 50Hz at room temperature (Figure 1). The fracture planes of the fatigued specimens were examined using a scanning electron microscopy model LEO 435 Vpi in order to identify the crack initiation points. The compressive residual stress field induced by shot peening was determined by x-ray diffraction method, using the Raystress equipment, whose characteristics are described in [9]: y goniometer geometry, Cr-k= radiation and registration of {221}diffraction lines. The accuracy in the stress measurements was Ds = ±30MPa. In order to obtain the stress distribution by depth, the layers of specimens were removed by electrolytic polishing with a non-acid solution. All surface roughness data measured in this research was obtained by Mitutoyo 301 equipment using a cut-off of 0.8 mm.

Figure 1:- Rotating bending fatigue test specimen (mm)

3

Results And Discussion

Figure 2 shows the Compressive Residual Stress Field (CRSF) for the 0.008A peening intensity for both 39 and 53HRC (solid lines). It is possible to observe, from Figure 2, that the surface stress and the depth of CRSF for 39HRC were lower than 53HRC condition. These results were expected, since the surface stress and the maximum value of the CRSF beneath the surface are a function of the yield strength and ultimate strength, respectively. However, the width of CRSF for 39HRC was a little bigger than 53HRC one. Since the width of CRSF is actually the plastic deformation depth, it is possible to suppose that the smaller hardness the larger the affected layer. The S/N curves for the base material and with shot peening treatment with 53 and 39HRC condition are shown in figures 3 and 4 respectively. It is possible to observe in figure 3 an improvement in the fatigue resistance of the specimens shot peened, compared to the base material. In relation to the base material, the shot peening influence in high stress (1370MPa) was almost null in the number of cycles until failure, for medium and high cycles an increase in fatigue life resulted from the shot peening treatment. Yet, the shot peening for the 53HRC condition represents an enhancement of the fatigue limit about 10% when comparing to the base material. For intermediate conditions (105-106 cycles) however, the fatigue gain in relation to the base material was much more expressive, about three times. Figure 4 shows that, surprisingly, there is not gain in fatigue life as a consequence of the shot peening treatment when applied

39 to specimens 39HRC, for all stress levels studied. According to previous studies [7], the effects of CRSF induced by the shot peening treatment on fatigue life gain is due to two different mechanisms, which act simultaneously. The first mechanism considers that the CRSF pushes the crack source beneath the surface of the base material, resulting in larger crack nucleation period and, consequently, larger fatigue life of a component. The second ones consider that the CRFS delays the nucleation/propagation process from superficial crack sources, increasing the fatigue life as well. The factors controlling crack origin at the surface or below the surface are a function of: CRSF dimensions, stress applied by fatigue tests and roughness generated by 100 0 -100 -200

Stress (MPa)

-300 -400 -500 -600 -700 -800

Original CRSF 0.008A 53HRc

-900

CRSF 0.008A 53HRc after 10 cycles Original CRSF 0.008A 39HRc

5

-1000 -1100

5

CRSF 0.008A 39HRc after 10 cycles

-1200 -1300 0,00

0,05

0,10

0,15

0,20

0,25

0,30

Depth (mm) Figure 2: Compressive residual stress field produced by 0.008A peening intensity in two hardnesses prior and after cyclic loading (105 cycles)

Base Material 53HRc Base Material and shot peening 0.008A

1400 1300

STRESS (MPa)

1200 1100

1007MPa

1000 900 800 700 600 500 3

10

10

4

10

5

10

6

10

7

CYCLES (NF)

Figure 3: Fatigue results of AISI 4340 steel with and without shot peening with 0.008A and 53HRC

40 1100

1000

Base material 39HRc and shot peening 0.008A Base Material 39HRc

STRESS (MPa)

900

800

730MPa 700

600

500 10

3

10

4

5

10

10

6

10

7

CYCLES (NF)

Figure 4: Fatigue results of AISI 4340 steel with and without shot peening with 0.008A and 39HRC

the shot peening treatment. Moreover, it was observed that despite both mechanisms caused an increase in fatigue life of a component, the first one is more efficient [7]. Therefore, it is desirable that the shot peening treatment is capable of pushing the fatigue life crack nucleation below the surface. It is also known that the shot peening intensity results in an increase in the maximum compressive residual stress and the width of the CRSF, but the stress at the surface is maintained almost the same [10,11]. On the other hand, the increase of the shot peening intensity causes the roughness to increase due to larger cavities created by the impact of the shot, which results in faster fatigue crack nucleation when the crack source is from the surface. Therefore, it is necessary to achieve a balance of this variable in order to obtain a good performance of the shot peening process. Due to this fact, researchers have obtained better fatigue results by using intermediate shot peening intensity [7,12]. Shot peening intensities which balance a CRSF with enough dimensions to push the crack source below the surface or to delay its nucleation/propagation in surface cracks due to the more positive surface conditions, are more adequate in obtaining a greater gain in fatigue life. Table 1 presents the increase of roughness as a consequence of the shot peening treatment used in this study. It is possible to observe that the both initial (before shot peening) and final roughness (after shot peening) were larger for specimens 39HRC than specimens 53HRC. By considering that the material is softer it is subjected to larger superficial deformations, which cause the roughness to increase. Table 1: Roughness results AISI 4340

Without shot peening

With shot peening 0.008A

39HRC

0.08 ±0.021mm

1.34 ±0.095mm

53HRC

0.22 ±0.016mm

0.92 ±0.063mm

41 It was observed, in a previous work, that the fatigue strength in the AISI 4340 steel was quite sensitive to the roughness variation. In this research a roughness variation from 0.1 to 1.5mm produced by sandpaper was made and as a result, the fatigue strength in rotating bending tests decreased approximately 60% to low cycle and 90% to high cycle [13]. Although the roughness variation created by shot peening and sandpaper were originated from different situations, the reduction in fatigue strength due to the increase of the roughness induced by the shot peening was expected. Due to the fact that there was not a reduction in fatigue strength of specimens 39HRC shows the performance of CRSF induced by the shot peening treatment. For specimens 53HRC, the fatigue strength gain allows us to speculate about a more appropriate correlation between roughness and CRSF generated by the shot peening process. However, the CRSF can suffer variations under cyclic loading. Many authors studied this subject and they showed that, usually, the CRSF suffers a decrease in the absolute stress value during the fatigue process [5,7]. This stress relaxation is directly related to the applied stress and the number of cycles to which the specimens are subjected [6,7]. For the study of the reported situation, additional tests were performed to verify the possible variation of the residual stresses induced by shot peening under cyclic loading. In these tests, the specimens were subjected to cyclic loading and removed from the rotating bending machine before failure occurred, and then, the new CRSF was measured. The measurements were carried out in 105 cycles, in which the greatest gains in fatigue life for the 53HRC condition could be observed (figure 3). Figure 2 shows the relaxation of CRSF after 105 cycles for both hardnesses studied (dot lines). It can be observed that the decrease of the CRFS was significant enough for both. This stress relaxation is justified when in the rotate bending fatigue test the compressive stress applied is added to the local compressive residual stress induced by shot peening. If the result of this superposition is big enough, there will be a plastic deformation and consequently a rearrangement of the stresses, causing relaxation of the original CRSF [5].Table 2 shows the residual stresses values at the surface produced by shot peening treatment prior and after fatigue tests (105 cycles). For the 53HRC condition, the rate between applied stress in fatigue tests and the residual stress after105 cycles was smaller than 39HRC condition (column F). Before cyclical loading, the ratio between the applied stress and residual stress at the surface was approximately the same for both hardnesses. So, it is expected that after 105 cycles the CRSF is more efficient in avoiding the crack nucleation at the surface for specimens 53HRC than specimens 39HRC. In addition, the algebric sum of the applied stress and residual stress after 105 cycles was, approximately, the same (table2, column E). Therefore, for the 53HRC specimens, the crack source from the surface would have a higher possibility to have its nucleation delayed, since with a greater hardness and smaller roughness the material has bigger fatigue strength with same applied stress.The fracture surface of fatigue specimens can prove the previous considerations. The shot peening treatment pushes the crack sources beneath the surface in most of medium and high cycles cases for the 53HRC showing the CRSF influence (figure 5). All the specimens with shot peening in low cycle or without shot peening had their cracks originated from the surface. This fact can be explained since the high applied tension stresses always significantly surpass the residual stresses induced by the shot peening process in case of low cycle conditions with and without stress relaxation. In specimens without shot peening it is natural that the crack source comes from the surface, where the maximum tension stress occurs, induced by the test characteristics. For the specimens 39HRC, with and without shot peening treatment, all the cracks originated from the surface, indicating a lower performance of the CRSF induced by the process.

42 Table 2: Residual Stress values at the surface (MPa) AISI 4340 A

B

C

D

Hardness

Applied Stress

Original Residual Stress

A/B

Residual Stress after A + D 105 cycles

A/D

53HRC

1007

-970

1,04

-480

527

2,02

39HRC

730

-697

1,05

-230

500

3,17

4

E

F

Conclusions

Some of factors that influence the fatigue life with the shot peening process are: the capacity of the CRSF of pushing the crack source to beneath the surface or not, the increase of roughness variation induced by shot peening process and the relaxation of the CRSF due to cyclic loading.Due the favorable combination of above factors, the shot peening intensity of 0.008A resulted in an increase of fatigue strength of 53HRC specimens. For other hand, the same intensity of shot peening did not produce fatigue strength variation for 39HRC.

(a)

(b)

Figure 5: Detail of crack source beneath the surface: a) Applied stress =931MPa; NF = 552900 cycles; crack source = 225mm from the surface; b) Applied stress =1007MPa; NF = 98600 cycles, crack source = 72mm from the surface.

5 [1] [2] [3] [4]

References L. B. Godefroid. Fatigue crack growth under constant and variable amplitude loading in aluminium alloys of aeronautical applications. 1993. (Thesis) - COPPE, UFRJ, RJ, Brazil. E. R. Los Rios, et al. Fatigue crack initiation and propagation on shot-peened surfaces in A316 stainless steel. International Journal of Fatigue, 1995, 17, 493–499. C. P. Diepart. Modeling of Shot Peening Residual Stresses. Practical Applications. Materials Science Forum. 1994,163–165. R. Fathallah et al. Prediction of plastic deformation and residual stresses induced in metallic parts by shot peening. M. S. and Technology,1998,14, 631–639.

43 [5]

[6] [7] [8] [9] [10]

[11] [12] [13]

A. Bignonnet. Fatigue Strength of shot-peened grade 35NCD15 steel. Varation of residual stresses introduced by shot peening according to type of loading, in: H. Wohl fahrt, R. Kopp, O. Vöhringer (Eds.), Proc. Int. Conf. Shot Peening 3,1987,659–666. S. Kodama. The behaviour of residual stress during fatigue stress cycles.1972, 111–118. M. A. S. Torres. An evaluation of shot peening, residual stress and stress relaxation on the fatigue life of AISI 4340 steel. International Journal of Fatigue. 2002, in the press. M. P. Nascimento, et al. Effects of surface treatment on the fatigue strength of AISI 4340 aeronautical steel. International Journal of Fatigue.2001, 23, 607–618. Gurova,. et al. Study of the residual stress state during plastic deformation under uniaxial tension in a 5.0Cr and 0.5Mo steel. Scripta Materialia, 1997, 36,.9,1031–1035. M. A. S. Torres, et al. Mechanical and microstructural study of residual stresses induced by shot peening on the fatigue strength of AISI 4340 steel. In: EUROMAT 2000, Tours, France: Elsevier, 2000, 1033–1038. J. K. Li, Zhang Ruiping, Yao Mei. Experimental Study on the compressive residual stress field introduced by shot-peening. In: ICRS3,1991. Proceedings... 750–757. L. Wagner. Mechanical surface treatments on titanium, aluminum and magnesium alloys. Materials Science and Engineering A, 1999, 263, 210–216. H. J. C. Voorwald. Análise sobre a influência de algumas variáveis no comportamento em fadiga por flexão rotativa em aços ABNT 4140 e ABNT 4340. 1983. 166f. Dissertation in Mechanical Engineering – ITA, Aerospace Center Technical, São José dos Campos, Brazil (in portuguese).

42

Towards Peen Forming Process Automation Frank Wüstefeld, Wolfgang Linnemann, Stefan Kittel Kugelstrahlzentrum Aachen GmbH, Aachen, Germany

1

Introduction

The KSA - Kugelstrahlzentrum Aachen GmbH (Aachen Shot Peening Centre), a spin-off from Aachen Technical University, was founded in 1994. Its management has an approximately 25% share in the company, the majority of the shares being held by RAG Aktiengesellschaft in Essen, one of the twenty largest German concerns. The company’s aim is to establish a new way of carrying out peening in the marketplace by moving away from the widespread trial and error approach, replacing it instead by really controlled shot peening with high-level process automation. Thereby, we prosecute an open policy in terms of full process data transparency and in this respect: our team of peen forming engineers, process engineers and automation specialists is available both on the spot to ensure the reliable performance of your process as well as for production work on a contract basis. The advantages of our process automation include high process reliability and quality with a high level of productivity and unlimited flexibility in design.

2

Controlled Shot Peening

Various applications have now been developed in Aachen to the point where they can be produced in series using controlled shot peening with process automation (= KSA -Controlled Shot Peening). This technology has its origins in Aachen Technical University, in particular the Institute for Metal Forming, and still draws on its potential and environment. Well over 500 side shells for the Airbus A310, about 600 structural tank segments for Ariane 4 and over 250 tank bulkhead segments for Ariane 5 have been produced successfully with shot peening to date in Aachen. We also use shot peening to form various extremely complex cone panels for the frame of the Ariane 5 power module and are currently using shot peen forming to support Airbus Deutschland GmbH to re-shape laser-welded skin panels [1, 2, 3] / (Fig. 1). Despite the large number of applications, there has not been one case in our company’s history where a component from series production has been rejected because of imperfect shot peening treatment. This indicates the quality and reliable performance of our controlled shot peening. Following successful qualification, the process runs on a controlled basis, is documented and can be reproduced exactly (= Frozen Process). Manual intervention in the production sequence of peening technology is neither possible nor necessary.

45

Figure 1: Production experiences

46

3

CNC-Controlled Shot Peening Facilities and Models

For our work we use 7-axis CNC-controlled/robot-aided shot peening facilities, our special know-how of the process with regard to the normal distribution of shot beneath the nozzle [4] as well as our models for on-line logging, evaluation and documentation of all the parameters of the peening process carried out on the component surface (Fig. 2).

Figure 2: Controlled shot peening

Thereby we are the only company worldwide which can provide visual representation not only of the amount of shot used per second and its normal distribution but also of shot velocity directly beneath the nozzle and the degree of shot coverage on the surface of the component. This has finally put an end to trial and error approach in shot peening and has facilitated clear description of the influences on the material caused by forming. A further innovation is that our customers receive complete documentation for each component, which explains what happened on the component surface itself. Our customers speak of Shot Peening – Improved Technology or even of New-Generation Shot Peening.

47

4

Process Implementation

We would like to outline our unique controlled shot peening process using the example of the current production of tank segments for the European rocket Ariane 5 (Fig. 3). This rocket has five spherically curved tank bulkheads which are currently assembled from eight individual segments for technical reasons. All the tank segments are made from the aluminium alloy 2219 T37* / T87*, the thickness of the sheet in the field varying from 1.6 mm to 3.5 mm and widening in stages towards the edge up to a maximum of 6.4 mm (= Integral Construction / High Design Flexibility). Our task is the precise and reproducible conversion of this integral component from the flat state in which it is delivered into the spherical contour R = 3004,6 mm ± 4 mm. Thereby, the maximum reduction of sheet thickness is 0,1 mm (= Customer Requirement). Depending on the thickness of the component, we use shot of varying diameter (4.75 mm – 9.575 mm) and carry out various steps in sequence as well as concave peening with high peening pressure and through-forging of the cross-section of the component (= Concave Peening / Partial Forging Process).

Figure 3: Dome tank segment of Ariane 5 during shot peening treatment

Depending on the geometry of the component, the sequential steps of the process are: five basic treatments, three homogenisation steps, 1 convex and 1 concave step for treating the edge as well as a maximum of eight field correction steps. In order to ensure high production standards for such components, a modern 7-axis CNCcontrolled shot peen forming facility was put into operation in Aachen at the Institute for Metal

48 Shaping at the beginning of 1993. This was originally a research and development facility with the aim of serving as a foundation stone for the development of a shot peening centre in Aachen. The facility features a continuous compressed air peening system with a nozzle for shot of up to 4.00 mm in diameter as well as a combined injector-gravitation peening system for shot of up to 10 mm in diameter. It is designed for forming components with maximum dimensions of 6,000 x 3,000 x 1,500 mm3 (length x breadth x height) and can be positioned to an accuracy of ± 0.1 mm for the x, y and z axes and to ± 0.1 degree for the possible rotary and slewing motion of the component (v and w axes). The components are charged using a transport wagon on rails with a possible rotary and slewing motion (c and a axes).

5

Control Units

The next Figure shows a diagram of the control PC and control unit for our on-line logging and documentation of the process parameters (Fig. 4). These enable us to carry out off-line programming on the basis of CAD data, on-line logging of readings as well as on-line documentation and evaluation with data output (Fig. 5).

Documentation Analysis

CAD

Planning Data Output

Off-line Programming

On-line Logging

7-Axis CNC/SPS-Control System

Figure 4: Control unit and control PC for on-line control and documentation of the peening process parameters

Apart from providing conventional information such as the current axis position, peening pressure, the shot dosing or the velocity of the nozzle over the component, essential data for the forming process is logged on-line and presented in visual form on our peening facilities for the first time. This includes the number of registered particles per second (= shot as a tool) and their normal distribution (= mass per surface) and, for the first time, shot velocity and shot coverage on the component surface itself as a function of the place and time of peening.

49

Figure 5: On-line logging and control of all essential peening process parameters

6

Process Automation

Each segment is accompanied by a forming programme, storaged on disk/CD-ROM, on which all the forming steps and processing times are chronologically recorded. The production process takes place with full computer control, thus eliminating any errors in processing the component. In order to determine the final field correction steps, the component is measured outside the peening chamber on a 3D measuring gage (Fig 6). The contour which was achieved is logged electronically using a measuring sensor and any deviations from the target contour are represented by contour lines with a similar deviation from the target contour R = 3004.6 mm (Fig. 7). Then the facility programs the necessary field correction steps fully automatically, as seen in the diagram of the field correction steps 1 – 4. We have had an optoelectronic, integrated 3D final component measuring feature on our new 7-axis CNC-controlled/robot-aided shot peening facility since 2001 (Fig. 8). Currently we are working on qualifying production of double size ¼ dome segments using this second, highlymodern shot peening facility (= Flexibility in Process and Design). The ¼ dome segments are peened simultaneously from both sides [5] using two articulated industrial robots, thereby considerably reducing run times and optimizing the sequential steps of the process (= Learning Curve Effects).

50

Figure 6: 3D contour measurement

Step 1

Step 2

Contour Plot with Lines of Equal Deviation from R = 3004,6 mm Step 3

Step 4

Figure 7: Computer-aided calculation and programming of the final field correction steps 1–4

51

7

Customer Benefits

Controlled shot peening with process automation takes place on CNC-controlled shot peening facilities in a way which is controlled, documented and can be reproduced precisely (= High Process Reliability). Our new-generation peening facilities feature central control units which facilitate off-line programming and on-line logging, evaluation and documentation of all the parameters essential for the peening process (= High-end Process and Model Technology). Each segment is accompanied by a forming programme. The production process is fully computer-controlled, thus eliminating any errors in processing the component. (= Process Automation). This ensures high quality (= Quality), optimum process sequencing (= Technology) and reduced run times (= Profitability).

Figure 8: Modern 7-axis CNC-controlled shot peening facility with two articulated industrial robots and integrated 3D measurement / facility 2

8

Summary

The market increasingly expects stable and automated processes with high process reliability and practical flexibility in design. Controlled and automated peening processes have put an end to the trial and error approach to the shot peening process and facilitate clear description of the influences of forming on the

52 material. Each processing step can be traced and identified at any time. Thus, the lack of clarity which characterized the shot peening process has given way to transparency, resulting in an industrial process which permits a breakdown of the individual steps. Production and quality control are process-oriented and completely fulfil the requirements of DIN ISO 9001 - 2000.

9 [1] [2] [3] [4] [5]

References F. Wüstefeld, Metal Finishing News 2000, December Issue, page 9–10, Shot Peening Process Automation for the Aircraft Industry F. Wüstefeld, Metal Finishing News 2001, March Issue, page 1–2, New Shot Peening Facility meets highest Standards in Terms of Control, Automation and Flexibility F. Wüstefeld, Metal Finishing News 2001, August Issue, page 12, Successful Shot Peening Tests for Airbus A380 Fuselage Shells F. Wüstefeld, DGM Informationsgesellschaft mbH, 1993, Modelle zur quantitativen Abschätzung der Strahlmittelwirkung beim Kugelstrahlumformen F. Wüstefeld, Metal Finishing News 2001, November Issue, page 18–19, Shot Peening Forming of new ¼ - Dome Tank Segments for Ariane 5

51

Current Applications of Advanced Peen Forming Implementation Axel Friese, Jürgen Lohmar, Frank Wüstefeld Kugelstrahlzentrum Aachen GmbH, Aachen, Germany

1

Introduction

Shot peen forming is above all carried out in the aeronautic and aerospace industry for the production of complex integral structures in fuselage and wing components (= Partial Forging Process / Shot as a Tool). Particularly in the aeronautics industry, development engineers and designers intensively work on complex, single-axis or multiple-axis curved geometries, varying panel thicknesses and/or stringer stiffenings. The aim here is to optimize weight and hence to save on aircraft operating costs, as well as to optimize production processes (= Design for Manufacturing Feasibility). A further important customer requirement comes from production managers: completely controlled, documented and reproducible processes with practical flexibility in design are a prerequisite for the industrial series production of carrier rocket structures or airplane components such as fuselage and wing panels (= State-of-the-art Industrial Processes). Appropriate documentation for today’s market should include the component data and the influences of forming on the material as well as all process data relevant to the customer (= Complete Transparency and Constant Quality Control). The KSA – Kugelstrahlzentrum Aachen GmbH (Aachen Shot Peening Centre) uses modern modelling and facility technologies which facilitate controlled shot peening with complete process control and automation (= KSA – Controlled Shot Peening). The following describes the procedure for implementing controlled shot peening. We will use examples from our current work to show that this highly-automated process can be employed both for producing components on a contract basis as well as for series production at the customer’s own plant.

2

Project Planning and Implementation

Clear project structuring is necessary before peening applications can be controlled and automated. This is the only way in which complex matters can be described, simplified and automated. The KSA – Kugelstrahlzentrum Aachen has an open policy in this respect, because we believe that the customer has the right to receive documentation covering all the information relevant to the component and process (= Maximum Transparency). First the project and customer needs are analyzed and studied with regard to their practicability. To do this we feed in CAD data, analyze the specifications of the process and the project conditions and discuss the necessary production and measuring toolings with the customer. We proceed as follows for new projects (Fig. 1):

54

computer-aided documentation analyzing and optimization

Production, Inspection and Documentation

Solution Approach

project targets

specifications

CAD data

Continuous Improvement

Concept

Expert Knowledge

Implementation

Customer

Process Introduction

Customer

planning Planning

data base intensity curves pilot tests

Forming Strategy

frozen process

reproducible quality

design flexibility

transparent prozess data

reduced runtimes and costs

programming basic and homogenisation peenings final forming steps zero defect peening

production requirements process parameters forming steps

Figure 1: Shot peening process automation by KSA

After adopting joint project planning (= Milestone Planning), we work out a possible solution, define a forming strategy and produce a first qualification specimen. For this we use our know-how regarding the process, our existing data sheets on peening process parameters and on projects (= Expertise / Data Library) and/or our evaluation of smaller previously-defined samples which have been subjected to shot peening treatment (Fig.2). The latter, which are often described as intensity curve samples having the same material and thickness, enable us to determine how to approach the project without processing or losing an original component. Programming of the peen forming strategy takes place off-line. We log, control and finally regulate production of our first specimen with our models for on-line acquisition, control and documentation of the peening process parameters. In addition to conventional machine information such as the current axis position, peening pressure, shot dosing or nozzle velocity over the component, essential process information is logged and presented in visual form on-line. This includes the number of registered particles per second and their normal distribution (= Mass per Surface) as well as shot velocity and shot coverage on the component surface itself as a function of the time and place of peening [1]. This means an end to the trial and error approach to shot peening for the first time and facilitates clear description of the influences of forming on the material. Our integrated 3D contour measuring with on-line evaluation and representation of contour lines enables our facilities to actually program themselves to attain the target contour. Our CNC-/robot-aided nozzle movements and positioning provide an additional guarantee of high precision, process reliability and control (= State-of-the-Art Shot Peening Facilities). This procedure means that regularly our qualification components can be used later as series parts (Fig. 3).

55

Figure 2: Determination of intensity curves and forming parameters

Our model and facility technology and its facilitation of the exact depiction and documentation of the process allow us to further optimize and automate the forming process (= Reduction of Sequence Steps and Run Times). Hence, after treating the fourth component we have as a rule already reached a point at which series production is stable, so that we can achieve the same peening result as with the first specimen, but with the additional advantage of optimized run times and sequence steps (= Learning Curve). Finally, the peening process runs on a controlled and automated basis and can be reproduced exactly with a high level of process reliability and precision (= Frozen Process). Manual intervention in the peening sequence is neither possible nor necessary. Despite the large number of applications, there has not been one case in our company’s history where a component from series production has been rejected because of imperfect shot peening treatment (= High Process Reliability). We would like to illustrate the procedure and quality of our controlled shot peening with various examples from our current production. Our reference [2] already describes the automated production of ? dome bulkheads for Ariane 5 with shot peening.

56

Figure 3: Forming example Airbus A380: qualification panel with precise contours

3

Re-shaping of Laser-welded Fuselage Skin Panels for Airbus A380

Our latest project deals with the re-shaping laser-welded fuselage panels for Airbus A380 [3,4] (Fig 4). The components are between 2,700 and 10,500 mm long and have varying sheet thicknesses, the stringers and skin panels also having varying geometric profiles. The thickness of the material ranges from 1.6 mm to 5.8 mm, the material used being various, aluminium alloys for panels and stringers (= High Design Flexibility). As a result of the heat applied during the previous laser welding process, the components are warped along both their length and breadth. Therefore, the components have to be re-shaped in a controlled and automated way. To achieve this, defined shot peening treatment is applied to the area of the fuselage skin panels on the inside of the component which was initially influenced by the heat. The qualification work which is necessary for the later re-shaping shot peening process is currently carried out successfully in Aachen on our two 7-axis CNC-/robot-controlled shot peening facilities for Airbus Deutschland GmbH [4]. It is intended that the re-shaping of the skin panels will finally take place at the Airbus plant itself with a new, corresponding facility for maximum dimensions of 11,600 × 3,000 × 750 mm (L × B × H), featuring high process reliability. Being the systems supplier, this will be our responsibility (= Open Policy, Shot Peening Systems Integration).

57

Shot Peening Facility No. 1

7-Axis CNC Control Unit

Left Stringer Treatment

Right Stringer Treatment

Completed Component

Process Visualization

Figure 4: Controlled re-shaping of laser-welded skin panels for Airbus A380

58

4

Production of Conical Segments of Power Module Frame for Ariane 5 using Shot Peening

We use controlled shot peening to make various extremely complex panels for the frame of the Ariane 5 power module (Fig. 5). The components consist of large flat segments measuring approximately 1 m2 (Box Cone Panel) to 2,5 m2 (Cone Panel) with varying sheet thicknesses, stringer stiffening and cut-outs. They have to be converted into a conical contour with radii of about 900 to 2,700 mm by means of controlled and automated shot peening.

Box Cone Panel

Process Visualization

Cone Panel

Process Visualization

Figure 5: Segments of power module frame of the european rocket Ariane 5

There is a total of 12 different geometric profiles within the required 24 panels per carrier rocket. The components are made of reinforced aluminium alloys for panels and stringers, the thickness of the material ranging from about 2.3 mm to 8 mm (= Design Flexibility). The panels are formed by shot peening both sides simultaneously with shot of a diameter of 8 mm. The production process can be reproduced exactly with high reliability (= KSA -Controlled Shot Peening).

59 Compared to conventional hand peening work, we carry out the forming process in a fraction of the time which would otherwise be required (= Process Automation / Reduction of Sequence Steps and Run Times). Furthermore, possible changes in design can be incorporated into our modelling and NC peening programs quickly and flexibly. A significant innovation is that our customer, Dutch Space B.V., receives complete documentation of what happened on the surface of the component (= Open Policy).

5

Production of Spherically Curved ¼ Dome Bulkheads for Ariane 5 using Shot Peen Forming

A further very recent project is the production of ¼ dome tank bulkheads for European space launcher Ariane 5 (Fig. 6).

Shot Peening Facility No. 2

¼ Dome Tank Segment

Control Stand

Process Visualization

Figure 6: ¼ dome tank bulkheads for the european space launcher Ariane 5

These components consist of segments of approximately 8 m2 which have to be converted by shot peening into a spherical contour with a constant radius of approximately 3,000 mm.

60 There are five tank floors for each Ariane 5 carrier rocket with four different geometric profiles and varying sheet thicknesses of 1.6 mm to 3.5 mm in the panel. The panel thickness widens in stages up to a maximum of 6.4 mm at the edge of the component (= Integral Structure / High Design Flexibility). The segments are peened simultaneously on both sides on our new 7-axis robot-aided shot peening facility using two articulated industrial robots. In comparison to our usual ? dome segment production [5,6], this results in a considerable reduction in run times and costs for our customer MAN Technologie AG (= Customer’s Competitive Advantages). The only sequence steps now required in the process are the basic peening and the final forming steps. The contour which was achieved is electronically logged using integrated 3D laser measurement, any deviations being represented in the form of contour lines with the same deviation from the target contour. Finally, our control software programs the remaining necessary forming steps fully automatically. Manual intervention in the production sequence of peening technology is neither possible nor necessary.

6

Customer Benefits

KSA – Controlled Shot Peening fulfils the customer requirements of design flexibility combined with the highest level of process control and automation as well as complete documentation and transparency (= State-of-the-art Industrial Processes / Open Policy). Proven process and model technologies provide the customer with shorter development times and facilitate changes in design during production (= Simultaneous Engineering, ‘Time to Market’) without losing an original component (= Reduced Pre-run and Adjustment Costs). All data on the component and the process parameters are recorded in the data library and can be accessed by the customer at any time (= Development and Production Partnership). Our modern, new-generation peening facilities feature central control units which facilitate off-line programming as well as on-line logging, evaluation and documentation of the essential peening process parameters (= High Process Reliability). The production process is fully computer-controlled, which eliminates any errors in processing the component (= Process Automation). In order to give all our customers this advantage, whether they prefer peening on a contract basis or in-house production, KSA devises a customized package for each case (= Customized Automation Solutions). Our support service for the customer during the implementation of highly-automated shot peening processes can include engineering work for the development and qualification phase, contract peening and the supply of control systems and complete facilities with process introduction at the customer’s plant (= Systems Integration). Peening work on a contract basis is carried out on our highly-automated facilities (= High Productivity / Short Run Times), providing the highest quality and reliability regarding delivery. Our high-end shot peening capacity can also be used to cover customers’ peak demand (= Back-up Production).

61

7

Summary

Controlled and automated peening processes have brought about an end to the trial and error approach to the shot peening process and facilitate clear description of the influences of forming on the material. We view ourselves as a partner for implementing actually entitled KSA – Controlled Shot Peening with the aim of supporting the customer to develop the potential for increased quality and economic efficiency. All our processes, facilities and models completely fulfil the requirements of DIN ISO 9001 - 2000. KSA’s shot peening automation technology can, in principle, be applied to all peening processes on the market and can be transferred to surface hardening peening and other applications.

8 [1] [2] [3] [4] [5] [6]

References F. Wüstefeld, Metal Finishing News 2000, December Issue, page 9–10, Shot Peening: Process Automation for the Aircraft Industry F. Wüstefeld, ICSP 8 2002, Towards Peen Forming Process Automation A. Kielies, Fertigungshandbuch Rückformen LBW 2002, page 2–3, Airbus Industry Specification 80-T-32-1025 F. Wüstefeld, Metal Finishing News 2001, August Issue, page 12, Successful Shot Peening Tests for Airbus A380 Fuselage Shells F. Wüstefeld, Metal Finishing News 2001, November Issue, page 18–19, Shot Peening Forming of new ¼ - Dome Tank Segments for Ariane 5 R. Kopp, J. Schulz, ICSP 8 2002, Optimising the Double-sided Simultaneous Shot Peen Forming

1

II Techniques and Controlling

2

63

The Unsatisfactory Situation in Residual Stress Evaluation Rudolf G. Bosshard ANVIL Developments, Volketswil, Switzerland

1

Abstract

Shot peening and its measuring tool “Almen procedure” are chained together ever since. As the process of shot peening has been improved considerably and numerous variations have been developed in the past 60 years, no real effort can be noted on the measuring side, at least not for industrial application. Other possibilities such as X-ray diffraction, Barkhausen noise principle and others have not been able to meet industrial needs so far. Statistical process control tries to solve the problem from a different point of view. The creation of new residual stress evaluation methods is still a challenge to scientists. All people involved in shot peening, regardless of origin or tradition or commercial affiliation should feel free to raise new ideas. This for the reason to make shot peening more economical, more simple, more reliable and to expand this technology into the field of further industrial applications.

2

Present Situation vs. the Engineer’s Dream

Consider todays daily procedure, e.g., in a maintenance and repair peening workshop: The qualified operator gets a work-piece and some papers such as part drawings, job instructions, photos, fill-in forms, time sheets or other documentation. He or she is in charge for a peening machine with media management, some measuring gear and special tools and is supposed to do his or her job under full responsibility. Happily, the operator will put up a dummy to simulate the work-piece to follow. He or she will put up Almen blocks, will install the strips, will do the mechanical setting and the setting for a first life run. Following that, the operator will go into the commonly known Almen procedure and after repeated test runs, will be satisfied with the results acquired with all the parameters. Then the set-up with the original part and the effective peening process are to follow. After that, rigging down, completion of paperwork and that¢s it. As experience of over half a century show, this works quite well and even at acceptable costs. But the component design engineer works, or should work, with residual stress-depth profiles and various surface conditions, at least when he or she is concerned about fatigue. This means true physical figures, that can be evaluated and also measured on the effective work-piece. And these figures must be found again in the workshop, must be recognizable on the work-piece, before or after a treatment. This is the aim that should be taken into consideration over the next few decades. Then and only then, we will have the standard of most other industrial measuring processes, in other words: 1. Step: Actual value measured 2. Step: Process performed 3. Step: Desired value measured (aim of process)

66 Certainly, the task is not simple, and was not even possible in the past. However, with the fast development in micro and electronic engineering, all involved people should support this Table 1: Possible evaluation methods for shot peening-induced residual stresses a Principle:

Destructive

b Technique: Fatigue testing (and exotic)

Non-destructive

Simulated

Physically scien- Direct Mechanical tific

Indirect resource related

c Procedure:

Lifetime test run Examination pro- Process related intensity cess and coverage set-up copied over to original partShort interval specific conformity check after time periods (4 hrs . .)

d Tools:

Materials testing X-Ray/Neutron machinery and Magnetism based instr.b) “XRD, BNA”

Metal plate with deforSensors with intellimation measuring equip- gent software procesment sing “PDT” “SPC”

e Main utilization:

Research

Research and Industry

Industrial production

high

Can be raised to accepta- Errors are minimized ble level

f

Reliability: Very high

Process control based Machine calibration and transformation to part Longtime interval standard conformity check after time periods (1 shift . .)

Industrial production

g Accuracy:

high

good

sufficient, fully human resource depending

h Hardware costsc) (EUR):

2k to 50k + part

70k to 300k instrument

1 to 3 k for equipment 5k to 30k 0.5 to 2 for 1 piece plate machine outfit

i

Costsc) per single test (EUR):

high to very high high value of part + 200–600 processing

sufficient, partly human resource depending

little does not apply 10 to 70 (intensity and coverage by Almen Strip) 5 to 10 (intensity and coverage by Almen Round = uncommon))

k Trend of Constant application:

Increasing

Fading ?

Increasing

l

high

little

high

partly (ASTM/SAE)

Almen Strip = extensive

no

Future no development potential:

m Standardization:

does not apply

Remarks:

a) to give just a fragmentary idea b) X-Ray also used partly “destructive”

(electropolishing) to compensate for the disadvantage of low penetration depth c) all monetary figures are very approximately due to the wide range of product/make variations (validity year 2002, 1 EUR ~ 1 US$)

67 idea and work on that. Then, the component design engineer’s dream becomes true and thus, the peening process can become a fine machine tool. Results can be truly measured and the time of simulation combined with various rather complex activities are past. The benefit would be an improvement of part functional safety and a reduction of quality control steps. Of course, total costs must be kept in relation. So far, the Almen test procedure is today¢s superior technique, but other quite different methods to evaluate or to inspect residual stress are known as listed in table 1.

3

The Plate Deforming Technique “PDT”

Shot peening has been introduced about 80 years ago and in 1943 has become a companion that up to now could not have been separated from any kind of peening process. It is the donation of Jo Almen who suggested to use a flat plate for testing the effect of a shot bombardment, just by measuring the curvature resulting when processing one side only. This simple plate has been blown up to a world wide success with a yearly consumption well over 5 million pieces. In standard configuration, brilliant as the idea is, even more astonishing is the fact that in the past 60 years, this technique remained nearly unchanged and has nothing left from its importance to industry and research. Henry Ford with his T-model motorcar was not as lucky as Mr. Almen was with his A/N/C strips, The T-metamorphosis is still in progress. Not so the petrified Almen [1] technique (Fig. 1).

Figure 1: Typical Almen procedure

There are various reasons for that: a) The procedure is very simple, hardware costs and evaluation time can be defined together with machine down time costs. The exact costs can be worked out (table 2). b) The process specific behavior of the specimen can be highly consistent. It can be a reliable measure for comparison with effects on actual work-piece. c) The application and standardization comes from a national high tech source that includes design, manufacturing and post treatment and is the basic of all international regulations. d) The workshop activity and paper work in all companies are well organized and routine in every respect.

68 e) Standardization is worldwide on a top level. f) The process can be done by specifically trained, reliable persons. No basic background is really required. g) Plates can vary in size, shape and thickness as well as in material. For example, for small hole application, a sectional peening treatment is even possible. On the other hand, some limitations have to be mentioned: a) A certain weakness [2] of the Almen strip system is documented in the literature and all experts dealing with this subject are aware of that. b) No direct correlation exists between an Almen evaluation testing procedure and the actual work-piece. c) No real measurements of residual stresses in surface layers of actual work-pieces are done. This is the outstanding disadvantage of the Almen principle. d) It needs quite an effort to prove the comparability of the process for a particular part with the Almen procedure run. e) As for peening processes, more and more expensive machinery becomes common. Machine down time costs, just for the time consuming Almen strip procedures, must be taken into consideration. f) The standard regular Almen strip is not ideal for an automatic handling and processing. Simple handling and simple installation is the domain of the so-called Almen Round [3].

4

X-ray Diffraction “XRD” (and Variations)

A true recognition of residual stress in MPa up to a depth of a few microns is offered by this highly scientific machine that can cost between 70 and 300 kEUR or kUS $. The function on the basis of X-rays is quite complex. For a comprehensive understanding, we would have to go into details which is not possible here. The instrument is very practical [4, 5] and widely used in universities and research centers. It has the outstanding advantage that true stress figures can get evaluated. However, there are also disadvantages that should be mentioned: a) Due to physical laws similar to optical in/out race of the rays (Fig. 2), the front end key device has a certain minimal size and geometrical shape. And this is a handicap for measure-

Figure 2. XRD-principle (STRESSTECH)

69

Figure 3. XRD-principle (BRUKER AXS)

ments on many typical structural parts, e.g., aircraft parts with small radii, small holes or extreme shapes. However, other principles do exist (Fig. 3). b) The penetration depth of normal instruments is very small - only a few microns [mm] - and also the exposure time required can be quite long (over minutes). c) Such an instrument requires a skilled operator, but X-ray diffraction scientists are rare and a short training course is definitely not sufficient to reach the necessary scientific level. A prosperous attempt, especially in the field of shot peening has been made by a US-industrial group [6], but did not pursue this concept due to the rather small potential market that was expected at that time. At present, there may be a dozen companies worldwide which offer adequate XRD machinery to be used for peened parts. However, the application is rather in research and not really in production and re-conditioning. Since improvements in this line are usually made in small steps, over the years, the peening engineer might become more familiar with XRD than he is at present and was in the past. Similar to the standard XRD is the neutron diffraction method [7]. At present, this method which is still under development is used for very special applications, but could be a highlight towards improvement in residual stress evaluation for peening processes. The penetration depth in neutron diffraction is much greater than in X-ray diffraction.

4.1

Barkhausen Noise Analysis “BNA”, Eddy Current

A very nice and not too complicated method [8] makes use of the materials magnetic properties and elastic strains/stresses. The so called Barkhausen noise signal is the answer of the material to a magnetic field which is applied on to the surface of the part by the sensor. This signal is a function of hardness and residual stress state of the part and can be correlated. The BNA can characterize the peening process applied and could be standardized. The advantage of a relatively deep penetration of up to 1 mm is worth to mention, but also the severe disadvantage that it is impossible to examine non-ferromagnetic materials. The practical measurement is very simple, various kinds of sensors even for holes down to less than 10 mm in size are possible. There are portable monitors, bigger installations and automatic devices. The BNA is well established in the automotive industry, e.g., for grinding processes, or examination of suspension springs. It should be possible to introduce this system also

70 extensively for shot peening application. Also “Eddy current” technology should be taken into consideration [9].

Figure 4. BNA-principle (STRESSTECH)

4.2

Statistic Process Control “SPC”

If you process an Almen plate with a conventional peening machine, it is similar to a blacksmith job. There, you bring a piece of steel into the desired shape. Hammering and examination and hammering again. Finally you get what you want, the same process for the next piece. If you are more advanced you will use a forging die and then, all the pieces processed by this tool are identical. From time to time you can examine a piece, also you can estimate the time of the next examination, even the expected lifetime of the tool you can work out from statistic calculations [10-12]. The same idea applies for a modern high-tech peening machine. Such normally CNC controlled machines with closed loop outfit for pressure and media flow, also multi-axes CNC movement and with various sensors e.g. for nozzle identification, air flow and shot storage

Figure 5. Almen statistics (BAIKER)

71 facilities. The software handles sensor signals, stores data, calculates functions and prepares messages and communicates with the operator. Behind all this, a software manages all the statistic procedures as illustrates in Figure 5 and informs the operator about normal operation or is involved in service recommendations or emergency actions. Following international trends, such equipment can easily be connected via standard protocols to selected networks and communicate, e.g., with the supplier in case of exotic failures. In practice, such a machine allows a set-up/programming according to specifications or stored job routines. Once created, the machine will guide the operator through the full job. It monitors all sequences and processes, will also tell when, e.g., a manual intensity check [13], say an Almen plate procedure, has to be performed. Even such a check can be done automatically, e.g., when using Almen Round technology. So here the machine is self controlling, once a parameter is stored, e.g., peening intensity, it can be repeated and even varied, or several different intensities and coverages can be stored and reproduced when ever necessary. Together with the automatically requested intensity checks or even calibration/service recalls, the machine has highest reliability and has a very high producTable 2: Time evaluation (conventional Almen versus SPC) TIME EVALUATION SHEET SHOT-PEENING for Almen-Processes (Air operated peening machines)

Datum: 2002-May Source.: (Name, Company) general example

Job description: Intensity and coverage machine set-up procedure Shot peening machine A = A comparatively simple peening machine with manual media feed and pressure control Kind of work piece

Simple

Complex

Example, designation (GE, CFM, PW, . .

2nd St LPT Disk 80C2

Time spent for set -up of measuring devices, time for actual processing, strip handling, arc high determination, clearing away and paper work [h]

up to 3 hrs

Effective process time for treating one original work piece after the set up [h]

0.5 hrs

Part specific intensity verification after . . hours (e.g. company typical)

4 hrs

Remarks

Shot peening machine B (SPC equipped) = A modern CNC machine fully controlled with closed loop technique and pre set parameters, statistic process controlled Kind of work piece

Simple

Complex

Remarks

Example, designation (GE, CFM, PW, . .

2nd St LPT Disk 80C2

Same job as for machine “A”

Time spent for set up of measuring devices, time for actual processing, strip handling, arc high determination, clearing away and paper work [h]

max 1 hr

Effective process time for treating one original work piece after the set-up [h]

0.2 hr

Machine general intensity verification after . . hours (ev. calculated by machine)

8 hrs

Remarks: Above figures may vary widely depending on company/skill of labour/equipment

72 tion availability. Practical experience shows that with the Statistical Process Control, one single intensity test per 8hrs shift is absolutely acceptable. Obviously, such a technique will reduce machine running time and overall costs considerably (table 2).

5

Conclusions

Although politics and philosophy should not be touched here, it seems to be a fact that highly civilized countries can only economically survive, if there is a continuous, even very small advancement to be noted. Shot peening as a microscopic part in the industrial field is not yet on its peak. Its subdivision “Residual Stress Evaluation” still is in a high technological “vacuum”. And this should be reduced by innovations and contributions of the present and future generations. Regarding the matter “Future development in Residual Stress Evaluation”, it seems that the Plate Deformation Technique (Almen) cannot be replaced in the near future due to the high technical and administrative level established in the key industries. Because of the high efforts in budget and personnel necessary to introduce new technology, no fast change in present situation can be expected. But then again, a development is imperative and shall be a challenge to human resource.

6

References

Only a few directly text related references are mentioned here. For more information, the internationally accessible documents can be traced, e.g., under - www.shotpeener.com [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13]

J. O. Almen, The Iron Age, 1944 (June). R. S. Simpson and S. Terry, Shot Peening (Ed.: D. Kirk) Oxford, 1993, 39. R. G. Bosshard, Shot Peening and Blast Cleaning (Eds.: M. C. Sharma and S. K. Rautaray) MACT, 1996, 37. P. S. Prevey, The Shot Peener, Vol.15, Issue 1, March 2001, page 4 and Vol.15, Issue 3, Sept 2001, page 7. A. Dugeon, SOPNATS, Rue Ile Mace, OP 2017, F-44406 Reze cedex, [email protected] K. J. Kozaczek, Shot Peening (Ed.: A. Nakonieczny), 1999, 313. H. G. Priesmeyer, Inst. Angewandte Kernphysik, Univ. Kiel, private communication S. Tiitto and A. S. Wojtas, SURFAIR IX, Cannes, June 1992 Hong Chang, Shot Peening (Ed.: J. Champaigne), 1996, 356. R. C. Wieland, Shot Peening (Ed.: D. Kirk) Oxford, 1993, 27. M. R. Vincek, Shot Peening (Ed.: D. Kirk) Oxford, 1993, 111. C. Mason, Shot Peening (Ed.: J. Champaigne), 1996, 328. S. Baiker, MFN, Vol. 3, August 2002, 16.

71

Vacuum-Suction Peening: A Novel Method for Emission-Free Shot Peening Gerard Pieper and Sigurd Ruhland GP Innovationsgesellschaft mbH, Lübbenau, Germany

1

Introduction

Conventional shot peening equipment such as injector or direct pressure blast systems require compressed air of sufficient pressure and volume flow to accelerate the shot material to desired velocities at certain mass flows onto the surface of the workpieces to be treated [1, 2]. A disadvantage of these pressurized systems is that dust which production can not be avoided during shot peening is emitted to the surroundings. Therefore, in applications where the peening treatments cannot be performed in closed chambers, breathing masks and protective clothing need to be worn. In addition, noise levels caused by conventional shot peening treatments are usually high. The goal of this presentation is to introduce to the shot peening community a novel shot peening machine, the so called vacuum-suction system whose operating principle does not rely on compressed air. This vacuum-suction system has been patented [3] and one of the authors (G. P.) has received the Grant of Innovation in Brandenburg in 1998.

2

Working Principle

Without the need of compressed air from a compressor, a vacuum-generator, e.g., a by-pass compressor, industrial exhauster or piston pump produces a negative pressure relative to the surroundings of the system. The incoming air generates a suction stream which is used to accelerate the shot medium onto the surface of the parts to be treated and also to transport the shot material within the system. Since the entire unit is under partial vacuum relative to its surroundings, there is no emission of dust. In addition, the partial vacuum keeps the noise level of the shot peening operation low. In order to design a vacuum-suction unit, the characteristic curves of the suction side of the stream producer and of the unit are optimized. Figure 1 illustrates the experimentally determined dependencies of the volume flow on negative pressure for both the exhauster and the unit. Since these characteristic curves have reversed dependencies on pressure, there is an optimum working point at the crossing of the curves. As seen in Figure 1, preferred working pressures of the vacuum-suction unit are in the range of 35.000 < pu < 45.000 Pa (350 – 450 mbar). Depending on nozzle size, the corresponding volume flows are estimated to result in flow velocities of 150 to 280 m/s.

74

Negative pressure, pu [Pa]

60000 Suction hood

50000 40000 30000

Suction unit

20000 10000 0 0

2000

4000

6000

8000

10000

12000

14000

16000

18000

Volume flow, Q [m3/h]

Figure 1: Determination of the working point of a vacuum-suction unit

3

Design of a Vacuum-Suction Unit

The vacuum-suction system was designed to enable sand-blast cleaning even in containers without any particular precaution. There is no emission of dust. Noise transport is hindered inside the vessel. Peening without using a breathing mask is possible since fresh air continuously postflows due to the sucking action inside the unit. Even in case of leakage or other defects, no dust will be emitted owing to the negative pressure principle. The system which is schematically illustrated in Figure 2 is a fairly simple technical construction. Therefore, running the unit does not require any particular qualification of the technical personnel. Maintenance consists merely of changing filters and if necessary, replacing worn out parts. Vacuum pump station

Blast section

Medium tower

Separator

Shot supply

Exhaust

Air supply

Blast lance

Suction hood

Dust collector

Reservoir

Suctioning

Almen strip Air suctioning

Dosing equipment Figure 2: Vacuum-suction system for shot peening (schematic)

75 The use of shot material is highly economical. Before the air passes through the filter, re-usable shot material returns to the cyclone and then into the process. The air is then cleaned by filters which are adapted to the particular application. Depending on the actual job, the shot material can be used for several cycles. Since its introduction to the market, the vacuum-suction system has been used successfully in a number of different applicationsnt removal on automobile parts and copper conductors • • • • • • • • •

Paint removal on automobile parts and copper conductors Layer removal on galvanized sheets Cleaning of railroad axles, metal moulds and insulation parts Matte finishing of glass and stainless steel Engraving of logos and serial numbers Surface roughening before painting Descaling of structural parts Deflashing of cast aluminum parts, blanks and milled parts Smoothening of welds or milled parts

This large variety of possible applications results from the applicability of almost all of the various clean blasting media available on the market. The vacuum-suction system is portable and therefore, can be used in the machine shop, laquering shop or in other locations where dustfree operation is required. The following summarizes the advantages of the vacuum-suction system: • • • • • • •

4

Emission-free operation even in sensitive environments Low cost operation at versatile and flexible applications A broad spectrum of possible working media No particular effort for precaution Fulfills high standards for worker safety and environmental protection Easy handling and maintenance through simple construction Easily adaptable to various application areas

Experimental

While the vacuum-suction system described above is already being used in a couple of different applications regarding mainly cleaning, descaling and roughening jobs, the applicability of this system to typical shot peening procedures has not yet been examined. Unlike the above mentioned operations, in shot peening, round shot is used to minimize erosion effects and to induce cold work and near-surface residual compressive stresses mainly to improve fatigue performance [4]. To examine the capability of the above described vacuum-suction system in typical shot peening applications, Almen intensity measurements were performed using commercially available type A Almen strips. For these measurements, the mobile vacuum-suction system and a standard suction hood were used. An industrial exhaust was used having a power of 5.5 kW, a maximum volume flow of 810 m3/h and a maximum negative pressure of 42.000 Pa (420 mbar). The following shot materials were used:

76 • • •

SCCW14 (spherically conditioned cut wire, 0.35 mm average diameter) S 230 (cast steel, 0.6 mm average diameter) S 330 (cast steel, 0.8 mm average diameter)

After positioning the suction hood, the circle is closed and the shot material is automatically sucked through an asymmetrical suction nozzle. Within the injector peening lance, the pre-accelerated shot material is further accelerated to velocities ranging from 20 to 80 m/s. After hitting the Almen strips, the shot material is sucked off immediately and is directed in the media-tower via a cyclone separator. Any re-usable shot material is separated and can be used several times. Finer particles such as dust and broken shot are removed by the exhaust and retained in a filter (19.500 cm2 filter area). The noise level of the vacuum-suction system is only 71 dB(A). The results regarding capability to shot peening are shown in Figure 3 where the arc height of type A Almen strips is plotted versus exposure time. 0,35 0,3

Arc height [mm A]

S 330 0,25

S 230

0,2

SCCW 14

0,15 0,1 0,05 0 0

20

40

60

80

100

120

Exposure Time, t [s]

140

160

180

GP Innovationsgesellschaft mbH 03222 Lübbenau, 28.06.02

Figure 3: Arc height vs. exposure time in vacuum-suction peening

As seen, there is a clear ranking among the various shot materials with lowest and highest values for SCCW14 and S 330, respectively. These results reflect the differences in kinetic energy caused by the differences in mass being lowest for SCCW14 and highest for S 330. For the given process parameters (16 mm nozzle diameter, 25 mm stand-off distance), saturation in arc height was reached after short exposure times. According to the definition of Almen intensity (arc height increase less than 10 % by doubling the exposure time), the corresponding Almen intensities amount to 0.16, 0.27 and 0.31 mmA for peening with SCCW14, S 230 and S 330, respectively. Recent results indicate that by using more powerful industrial exhausts, Almen intensities higher than 0.5 mmA can easily be obtained. Thus, it can be concluded that the vacuum-suction system described above can be used in typical shot peening applications where strengthening of the surface regions of structural components by inducing cold work and residual compressive stresses is utilized to improve fatigue performance.

77

5

Acknowledgements

The authors would like to thank Prof. L. Wagner of TU Cottbus for his support regarding Almen intensity measurements and helpful comments on the manuscript.

6 [1] [2] [3] [4]

References G. Nachman, Shot Peening (H. Wohlfahrt, R. Kopp, O. Vöhringer, eds.) DGM (1987) 37. R, D, Gillespie, Shot Peening (D. Kirk, ed.) (1993) 81. G. Pieper, Patent DE 10102924 H. O. Fuchs, ASTM No.196 (1957) 22

76

New Developments in Cut Wire Shot for Shot Peening Uwe Kersching, R+K Draht GmbH, Mittweida, Germany

1

Introduction

Todays production of dynamically stressed components in many industrial branches requires Shot Peening applications with a high level of optimized processes and materials. The media used in that processes is becoming more important to achieve best results at the parts surface and to run the application in the most economical way. This recognition was the reason for a two years research at R+K Draht GmbH to develop a Conditioned Cut Wire Shot (“Special Quality”) that corresponds to the a.m. needs much better than conventional well known Conditioned Cut Wire is doing today. The extensive research period was finished in 1999 and was continued by a period of large scale tests in industries with various applications. Special Quality showed its drastic increased lifetime and provides industries a shot peening media that allowes to run also applications where hardnesss with 700 HV (60 HRC) and above are needed in an economical way for the first time. The following presentation shows the results that could be achieved in industrial applications during the last years.

1.1

Short History of the Development of Dynamically Stressed Components

Dynamically stressed components are mainly: • springs, • gears • aircraft components The technical development of these applications is characterized by highest dynamic. The increasing request to the performance of these components caused permanent innovation: • • • • •

1.2

material alterations new constructions new raw materials increased strengths new technologies in production

Development of Shot Peening Media

In the past major use of shot peening media was cleaning of surfaces.The cognitions by WÖHLER that special tensions at the surface of parts increase the life time substantially, were applicated in large-scale production with the introduction of the shot peening process to the production of dynamically stressed components. This process without shot peening today is in-

79 conceivable. Initial shot peening media mainly were made of steel cast. Using of Cut Wire Shot became a very important alternative. Advantages of Cut Wire Shot are: • • • •

increased regularity of mechanical characteristic closer tolerances of hardness more uniform screen structure increased lifetime

The analysis of shot peening media situation shows that the development of shot peened parts is much faster than the development of the needed shot peening media. To open opportunities for Shot Peening applications which meet future demands it is necessary to find ways for a Conditioned Cut Wire Shot that will provide better technical and economical characteristics.

2

Requirements on Dynamically Stressed Components

The requirements in general are: • • •

very long lifetime smaller size of components cost reduction

Ways to get there: • •

3

on one hand changes in the raw materials and the production process of these parts and on other hand by harden the surfaces with shot peening

Requirements on Shot Peening Media

By analysing the mentioned situation we derive the recognition that there is a delay of development in Conditioned Cut Wire Shot in two directions: First is lifetime of cut wire shot, including the influence to the costs and second point is the ratio hardness / lifetime and costs of Cut Wire Shot.

4

Examinations at R+K Draht GmbH

4.1

Examination of Lifetime

The procedure of the examination of lifetime and the limited values are prescribed at VDFI 8001. The above-mentioned procedure – the Ervin Test – is a prestigious method to describe the duration of the shot peening media in practice. The increase of the Ervin-values means also a decrease of consumption of shot peening media in minimal the same equivalent percentage.

80 The increase of lifetime of media tested with the Ervin-Test machine to 140 % means in practice a decrease of consumption of 40 to 50 %. With an additional operation (heat treatment) it was reached that the lifetime was increased to 140 to 160 % as compared with conventional Cut Wire Shot. The expansion of the loss of conventional cut wire shot is nearly identical to special cut wire shot up to 2.200 cycles (point A) in the Ervin-Test machine. From this point the breaking characteristic of conventional shot changes drastic. The loss increases, the Ervin-Test curve breaks away. If using cut wire shot made of drawn steel wire there is a uniaxial tension state. The particles breaks in longitudinal axis and rounds again. This is a critical phase, because there are sharp-edged particles that could cause violations at the parts surface. In difference to that the new Special Quality with heat treatement has a normalized inner structure so that the breaking characteristic is not influenced by unfavorable tensions. The use of this new special cut wire shot guarantees a very uniform structure that also wears out like that. The rate of breaks is minimal, the danger of surface violations nearly excluded. Only occasional particles break, the most quota wears out to its spherical form more and more. The production of special cut wire shot is not cost-neutral. But cost reduction by increase of

typ ic a l c u rse o f c u rv e c o m p a rin g c o n v e n tio na l a n d sp e c ia l C u t W ire S h o t n or m a l cu t w ir e sh ot loss % sp ecia l cu t w ir e sh ot loss % 100 90 80 loss in %

70 60 50 40 30

A

20 10

75 00

70 00

65 00

60 00

55 00

50 00

45 00

40 00

35 00

30 00

25 00

20 00

15 00

10 00

50 0

0

c yc les in E rvin -T es t m ac h in e

Fig. 1: Comparison of lifetime of conventional and special Cut Wire Shot (dia.: 0,6 mm, shape: elliptic, hardness: 640 HV1)

the lifetime is proved. Further advantages like decrease of decay, reduction of efforts for handling and other are possible. In Table 1 all achieved results are shown. This list is to be continued.

81 Table 1: Achieved increase of lifetime with conditioned cut wire “special quality” Dia- Hardness meter/ Vickers mm

Lifetime/ number of cycles in Ervin test machine conditioned cut wire “ conventional”

nomi- Toleran- Shape** nal ces HV 1* HV 1* 0,30

0,40

0,50

0,60

0,70

0,80

0,90

G1

G2

Increased lifetime with conditioned cut wire ”special quality”

conditioned cut wire “special quality” Shape**

G3

G1

G2

Shape** G3

G1

G2

G3

640

610–670 4600

4500

4400

9700

9600

9500

210,87% 213,33% 215,91%

670

640– 700 4500

4400

4300

9600

9500

9400

213,33% 215,91% 218,60%

700

670–730 4600

4500

4400

9500

9400

9300

206,52% 208,89% 211,36%

640

610–670 4400

4300

4200

9400

9300

9200

213,64% 216,28% 219,05%

670

640–700 4300

4200

4100

9300

9200

9100

216,28% 219,05% 221,95%

700

670–730 4400

4300

4200

9200

9100

9000

209,09% 211,63% 214,29%

640

610–670 4200

4100

4000

8800

8700

8600

209,52% 212,20% 215,00%

670

640–700 4100

4000

3900

8700

8600

8500

214,00% 215,00% 217,95%

700

670–730 4200

4100

4000

8600

8500

8400

207,00% 207,32% 210,00%

640

610–670 4000

3900

3800

6500

6400

6300

162,50% 164,10% 165,79%

670

640–700 3900

3800

3700

6400

6300

6200

164,10% 165,79% 167,57%

700

670–730 4000

3900

3800

6300

6200

6100

157,50% 158,97% 160,53%

640

610–670 3700

3600

3500

6000

5900

5800

162,16% 163,89% 165,71%

670

640–700 3600

3500

3400

5900

5800

5700

163,89% 165,71% 167,65%

700

670–730 3700

3600

3500

5800

5700

5600

156,76% 158,33% 160,00%

640

610–670 3500

3400

3300

5500

5400

5200

157,14% 158,82% 157,58%

670

640–700 3400

3300

3200

5400

5300

5100

158,82% 160,61% 159,38%

700

670–730 3500

3400

3300

5300

5200

5000

151,43% 152,94% 151,52%

640

610–670 3300

3200

3100

4800

4700

4500

145,45% 146,88% 145,16%

670

640–700 3200

3100

3000

4700

4600

4400

146,88% 148,39% 146,67%

700

670–730 3300

3200

3100

4600

4500

4300

139,39% 140,63% 138,71%

* for evaluation 0,30 mm it is to use HV 0,5 ** shape: G1 = broken edges, G2 = elliptical, G3 = ball

4.2

Examination of Hardness

If industries are looking for Cut Wire Shot with hardnesses above 700 HV (60 HRC), a contrast between the high hardness and increased wear is to be mentioned. Very often the use of Condi-

82 tioned Cut Wire with that high hardnesses causes high costs by rapid wear of the material. Solutions could be: a) to increase the tensile strength of the drawn wire b) use of different raw materials c) to optimize the production procedure By considering these three opportunities it was the target to find the best ratio of highest hardness, highest lifetime and economical production. In Table 2 examples are shown: Table 2: Examples of cut wire shot with higher hardnesses and increased lifetime Diameter

Hardness

Lifetime

mm

HV 1

Cycles in Ervin Test

0,35

ellyptic

HV 750 (± 30)

3500

0,50

ellyptic

HV 750 (± 30)

3740

0,60

ellyptic

HV 750 (± 30)

4000

5

Shape

Benefits and Profits

The application of special cut wire shot saves media as compared with conventional cut wire shot about 30-60% and offers cost-reduction of about 10-20%. The exact amounts are to be calculated by the actual basis at the user.The actual basis is described as follows: a b c

consumption kg/shot peening hour composition of operating mixture technical specification of the applicated cut wire shot

The saving refers to the consumption kg/shot peening hour. The composition of operating mixture is keeping constant, like actual basis

83

Virtues & Limitations of Almen Round Mahesh C. Sharma & Himanshu Deo M. A. C. T. Regional Engineering College Bhopal, India

Suresh, C. Modi MEC Jodhpur, India

Rudolf G. Bosshard ANVIL Developments, Volketswil, Switzerland

1

Introduction

There are many parameters involved in shot peening. Among them are: The shot speed, the size, shape, nature and hardness of shot, projection angle and exposure time. This multiplicity of parameters makes the precise control and repeatability of a shot peening operation very problematical. For this purpose, research was carried out to develop a more comprehensive and easier method of controlling these parameters. Present investigation gives a detailed discussion of the development of various methods of controlling shot peening parameters and proposes also a promising method which is called “Interactive Almen Round Method”.

2

Chronological Development in Shot Peening Process Control

Shot peening control aims to either reproduce a previously-established intensity of peening or to induce a specified level of peening intensity. Control of shot peening operations for strain hardening is currently achieved by means of Almen gauge system or “Almen Strip” system. The “Almen Gauge” introduced by J. O. Almen in 1943 has served the shot peening industry very well. Rectangular steel strips of controlled chemical composition and thermal history are shot peened whilst being held flat and deflection of the strip on release is measured .The deflection from flatness is Almen arc height H , being measured using an Almen gauge and is presented for strips with a choice of three different thickness. These three thickness accommodate the wide range of peening intensities that need to be used for different applications. No doubt, Almen strips have proved out to be the only means to set up the shot peening parameters for different materials and applications But apart from having such extensive usage, Almen strips do have some limitations. They are work intensive like clamping and unclamping (tightening/loosening). This is a time consuming procedure and regarding details there are very many tricky mechanical process effects in respect of accuracy and application. Apart from these limitations for taking one set of reading we need at least 4-5 strips to establish a saturation curve. Until 1993 this method was the only method, but in 1993 the term “Interactive Shot Peening” (1) came into existence and at that time lot of work have been done on this particular field. An initial attempt to produce an interactive peening intensity measurement device (2) involved the use of standard Almen strips but without the standard four screwed securing procedure. This device incorporated LVDT (Linear variable displacement transducer) for sensing and measuring the deflection of strip continuously. This method was reliable and fast but the main problem with this method was that LVDT was very sensitive to long term damage.

84 Further development on LVDT based devices had given a device in which a screw down ring presses the circular test disc against a recess in the disc holder. An LVDT is secured in the disc holder by means of a grub screw and the output fed to a displacement recorder. The screw down ring and disc holder afford excellent protection for the LVDT. This device was more versatile and accommodating for wider range of peening intensities. To further improve this device masking washers were used to expose different areas to the incoming shot stream. These fit on top of the disc. This technique is in regular use for research and development purposes, but the LVDT itself is relatively bulky and too delicate and has to be carefully protected from high-velocity shot. After this LVDT method a new development in monitoring device (3) involves the use of elastically–strained arm carrying a strain gauge that is in contact with a disposable steel disc. As the disc deforms, due to peening, the arm relaxes with consequent strain gauge signal changes being continuously chart-recorded. The very high sensitivity of this strain gauge device allows very small changes in intensity to be detected. This method used test specimens which were made of simple mild steel rather than the very hard steel of original Almen Strips.

3

Alternative Method (“Almen Round” Technique)

Pertaining to above-mentioned limitations of the “Almen Strip” technique and other methods, the present presentation proposes the concept of having an alternative test piece, the “Almen Round” with an unique on-line monitoring of shot-peening process. This system (predecessor was the “Impact Sensor” (4), established in centrifugal peening / blasting applications) facilitates the use to mount Almen Rounds in a special fixture, having a sensor (for convexity measurement), which displays the values of convexity on a digital monitor. The most outstanding feature of this system (5) is that everything is online, i.e. we can see the increasing arc height, the saturation curve, on the monitor. Procedure: This method involves two steps for any set of readings which are: 1) Setting the arrangement: In this step we have to make all arrangements which include tightening of an Almen Round with swivel nut using simple hand force, plugging in the monitor and linking monitor with the sensor head arrangement. 2) On-line reading: In this step we expose the sensor head with the fixed Almen Round to the shot process and read the increasing arc height with the help of the monitor simultaneously or have the event e.g. software processed.

4

Constructional Details

This arrangement has in minimum 4 elements: The test specimen, the sensor head, the monitor and calibration disks A, N and C.

4.1

The Test Specimen (Almen Round)

The test specimen or so called Almen Round (Fig.1) is a cut out from a standard Almen strip of either A, N or C–thickness and of original quality Premium. The precision Laser-cut operation

85 followed by an extensive flatness examination guarantees a Almen strip equivalent round disk in every respect.

Figure 1: Almen rounds

4.2

The Sensor Head

A heavy, specially shaped steel bar, rectangular at one side , carries a finger like dome. On top of this dome the Almen Round specimen gets tightened with the swivel nut. The specimen is only fixed peripherally on both sides on a tiny rim. The dome bears inside a pin touching the specimen on the inner side, the other side of the pin works on a strain gauge, actually transmits the deflection of the specimen. This will induce a output signal exactly linear to the deflection.This unit (Fig. 2) is treated to highest hardness in order to withstand the hazardous conditions when exposed to shot stream. It is designed for simple installations supposed to be exposed

Figure 2: Sensor with cable

86 to the shot stream. The unit does not contain any other items then the strain gauge. A protected cable also connects to the monitor box or interface outside the peening area.

4.3

The Monitor

Common to this equipment is a hand hold battery operated monitor (Fig. 3) that allows a 3 digit reading of the Almen arc height in metric or inch system. Additionally an analog out signal 0-1 V is provided. One toggle switch ON/OFF one switch for DATA/HOLD and the ZERO setting knob allow the operation of the outfit. However the sensor out signal can be connected directly e.g to a chart recorder or to any suitable interface for further processing by an suitable software even combined with an automatic 2T calculation.

Figure 3: Hand held monitor

4.4

Saturation Curves

When a component is shot peened, it incurs a deflection, which is the result of residual stresses that have been induced by shot peening. For a given shot peening condition the variation in deflection with the shot peening duration may be plotted and these kind of curves are called saturation curves. These curves are very helpful in controlling the shot peening parameters and are used worldwide for this purpose. Some saturation curves are shown. The values of the deflections obtained after a saturation time are considered to be the characteristics of the intensity of the shot peening. Following figures show the saturation curves obtained from shot peening carried out using an air operated

87

0,2 t.

0,15

Round Arc H

0,1

Strip Arc.Ht.

0,05 0 0

20

40

60

Arc.Height(mms)

Arc.Height(mm)

standard peening machine. For such preliminary studies injection type equipment has been used and parameters such as shot, flow, pressure etc. have been set to commonly used standards. Saturation curves have been plotted by using the conventional “Almen Strip” method and the alternative “Almen Round” technique. The results obtained from both of the methods are given and found as a result, they are almost identical. It could be discussed whether discrepancies are a matter of tolerances and errors or whether a real physical difference could be noted. At least it seems that for industrial use such differences could be disregarded.

0.4 0.3

Round Arc.Ht.

0.2

Strip Arc.Ht.

0.1 0 0

80

50

0,25 0,2 0,15

Round Arc.Ht. Strip Arc.Ht.

0,1 0,05 0 20

40

60

80

Time(secs)

Curve 3 “A” specimen peening pressure 4.0 bar

5

Curve 2 “N” specimen peening pressure 2.0 bar

Arc.Height(mm)

Arc.Height(mm)

Curve 1 “N” specimen peening pressure 1.0 bar

0

100

Time(secs)

Time(secs)

0,3 0,25 0,2 0,15 0,1 0,05 0

Round Arc.Ht. Strip Arc.Ht.

0

20

40

60

80

Time(secs)

Curve 4 “A” specimen peening pressure 5.0 bar

Discussion

This interactive “Almen Round” method is very useful for determination of Almen values (intensity) at the time of setting of machines. This method is considerably faster than “Almen Strip” method as this method doesn’t involve cumbersome process of clamping and un-clamping of strips which is replaced by a simple hand clamping of the swivel nut. This method uses the test specimen (Rounds) which are made of original Almen strips (thickness, hardness, material etc). In this method the clamping practically does not affect the buckling up of the disc and the setup is rigid, and all the components of forces are in defined position. The measuring gear is mechanically linked to the test specimen therefore reducing errors and increasing accuracy. In view of the calibration provided for digital gauge giving convexity (spherical shape) accurately at par with conventional Almen arc height (as it is evident from the herewith presented research work according above given results/graphs), this interactive peening control method can allow the control of shot peening parameters more accurately and conveniently. Also this device is claimed to be industrially approved.

88 If there is an interest from the shot peening industry or from research side, certainly the method can be trimmed to meet international approval standards.

6 [1] [2] [3] [4] [5]

References D. Kirk, ICSP5, Oxford, “Interactive Shot Peening Control” p.10 to 15 (pre-prints) see page 12, fig. of reference1) D. Kirk, ICSP6, San Francisco, “Developments in Interactive . .”, page 95 to 106 R. G. Bosshard, MAT-TEC91, Paris, “Shot Intensity Monitoring . . ”, page 197 to 200 R. G. Bosshard, SP-BC1, Bhopal, 1966, “On Line Monitoring . . ”, page 37 to 44

89

Device for the Determination of Impact Velocities in Shot Peening Rolf Clausen, Jürgen Stangenberg Technical University of Hamburg-Harburg, Hamburg, Germany

1

Abstract

A device was developed for the determination of impact velocities of shot-peened grains impinging workpiece surfaces. The impact velocity of shot-peened grains can be used as explanation of the characteristics of the workpiece changes, like roughness, tension or surface coverage. Also, modifications of these changes due to variation of shot peening parameters (shot peening media, jet parameter, jet direction) can be detected and evaluated.

2

Introduction

In the department "Cutting Technology" of the Technical University of Hamburg-Harburg research was performed in the field of the shot peening, concentrating on Peen Forming and the roughness of the shot-peened surface. In this context, the velocity of shot-peened grains at the impact on the workpiece surface had to be measured. A simple measurement device was designed and will be introduced in the following: The impact velocity of the shot-peensd grains can not be determined by measuring parameters like the jet pressure, the shot-peened grain mass flow, the nozzle geometry or the jet intensity. However, the grain’s impact velocity and the grain’s mass determine the kinetic energy of a shot-peened grain. With that, the deformation of the surface (e.g. roughness) and the change of the peripheral area of the workpiece (e.g. compression stress) as well as the coverage of the surface can be identified. The knowledge of the impact velocity can contribute to the explanation of the type and the characteristics of the measured changes. The attempt to measure the velocities of shot-peened media is not new and is documented in different papers [1, 2, 3, 4]. Partially with elaborative efforts [3].

3

Description of the Measurement Principle

The principle of the measurement device designed in the department of Cutting Technology of the Technical University of Hamburg-Harburg (TUHH) is based on a so-called " two-disk method " and is shown in Figure 1. On a common shaft two disks are fastened, which turn with a constant rate. The first disk, the so-called hole disk, has a hole at a certain distance (r) of the disk’s center. The second disc is the so-called registration disk. The jet nozzle is located with the distance (a) from the hole disk at the height of its hole. Whenever the through hole of the hole disk and the nozzle exit aligned, some of the shot-peened grains pass through the hole and mark the registration disk. Depending

90

Figure 1: Measurement principle according to the “two-disk method”

on the grain velocities it takes a certain time (t) for the grains from passing through the hole disk and impinge on the registration disk. During the time span (t) the disks keep on turning about a angle (j), so that the impact on the registration disk does not align with the hole in the hole disk and the nozzle exit. The grain impact velocity can be calculated with the parameters b (distance of disks), n (RPM of disks) and j (turn angle of disks during t).

4

Description of the Measurement Device

In Figure 2 the built measurement device is shown, including necessary power supply for the drive. In the following some boundary conditions for the measurement device are listed: • • • • •

Dimensions < 600x400x400 mm (workspace of the shot peening cabin), facility for converting the input parameters (a, b, r, d, n) for optimal adjustments of the device, low weight for transportation purposes, easy handling (registration disk, modifications of input parameters), low cost.

The device frame consists of aluminium profiles, which allow easy adjustments. The two disks, the bearing blocks and engine mounting plate are also made out of aluminium, the shaft consists of steel. The drive, an DC motor and a drive belt, allows an easy variation of the disk revolutions. The two disks are wedged on the shaft and are also easy to adjust. Likewise the mounting plate of the jet nozzle can be adjusted to the height and distance necessary according

91

Figure 2: Measurement device with power supply

to the hole disk. The attachment of the graph paper on the registration disk is performed with the aid of a second recessed disk, with wedges the paper onto the base of the registration disk. Figure 3 shows the hole disk and the registration disk from an other view for better understanding. The distance (r) from the rotation axis to the disk hole can be changed by the exchange of an insert. This insert is available twice, in order to keep the spinning disks in balance.

Figure 3: Measurement device with the hole disk and registration disk

92 The actual hole is in a metal sheet with a thickness of 0.25 mm. The thin metal sheet is necessary to prevent the collusion of the shot-peened grains with the wall of the through hole.

5

Test Run

For optimal test results the different, in chapter 3 described variable parameters had to be investigated and adjusted. Figure 4 shows the setting up of the measurement device in a shot peening cabin. Following dimensions and/or settings turned out as favorable for the test run: • • • • • • •

hole diameter in the hole disk: d = 5 mm distance r (scribed circle): r = 95 mm distance between the disks: b = 110 mm distance nozzle exit to hole disk: a = 135 mm RPM of the disks: n > 1200 min–1 (2400 min–1) graph paper: Copy paper with 80 g/mm² necessary jet duration: 30 to 120 s

Figure 4: Measurement device set up in a shot peening cabin

Figure 5 shows the registration disk schematically. With selected settings (b = 110 mm, n = 2400 min-1) the grain impact velocities show up according to the angle j. The range of 10° to 70° equals grain velocity range of 158.8 m/s to 22.6 m/s. Shooting grains on the standing disk through the hole onto the registration disk result in a scattering range diameter of 8 mm. This range does not change significantly using different shot peening media with different jet pressures. The effect shown in Figure 6 results from a jet pressure of p = 4 bar using round cast steel shot peening media with the size of 0.2 to 0.4 mm (GSR 0.2-0.4).

93

Figure 5: Registration disk with indication of the impact velocity according to the angle j

Figure 6: Graph paper, shot-peened grain impacts on standing and spinning disk

Shooting grains on the spinning disks (n = 2400 min–1) the grain imprints on the graph paper appear in form of a tail, whereby an accumulation can be detected at larger angles. A test run

94 with the parameters mentioned above indicated a grain velocity rate of v = 51.1 to 77.3 m/s. The scattering range diameter is again about 8 mm. For the calculation of an average grain impact velocity the weighting for the lower speeds was taken double due to the accumulation of imprints at lower grain velocities:

v = 2 × v min + v max / 3

(1)

To prove the functionality of the designed measurement device the following four types of shot peening media were used: • •

MGL 0.250-0.420 and MGL 0.420-0.590 (glass beads), GS-R 0.2-0.4 and GS-R 0.4-0.8 (cast steel shot peening grain, round) The jet pressure was within the range of 3 to 7 bar.

6

Interpretation of Experimental Results

In Figure 7 the average grain impact velocity is shown depending on the jet pressure. The curves rise approximately linear within the examined jet pressure range. The glass beads achieves higher rates than the steel grains, as well as for smaller grains a higher velocity can be detected than for larger grains.

Figure 7: Average grain impact velocity depending on the jet pressure

For the upper and lower curve in Figure 7 the determined scattering range is show in Figure 8. The curves represent the weighted average grain impact velocity. It can be seen that the scattering range is relatively large, for the glass beads even more than for the steel grains.

95

Figure 8. Scattering range of grain impact velocities

7

Summary

In summary the following conclusions can be drawn. The presented two-disk measuring method is basically suitable for determination of shot-peened grain impact velocities on workpiece surfaces. The shot-peened grain impact velocity rises almost linear with the jet pressure within the examined pressure range from 3 to 7 bar. However, further investigations are necessary for determination the possibilities of utilization of the presented measurement device.

8 [1]

[2] [3] [4]

Literature Martin, P.: Einzelkornversuche, Bestimmung der Kugelgeschwindigkeit nach dem Laufzeitverfahren mit Hilfe von 2 Impulsgebern (in: Beitrag zur Ermittlung der Einflußgrößen beim Kugelstrahlen durch Einzelkornversuche. Dr.-Ing. Diss. Uni. Hamburg 1980). TRAVEL®: Optische, berührungslose Methode zur Messung der projizierten Geschwindigkeit der Strahlpartikel in Strahlkabinen und Schleuderradanlagen (ICSP5, 1993). Linnemann, W., Kopp, R., Kittel, S. and Wüstefeld, F.: Shot Velocity Measurement (ICSP6, 1996). Andziak, J.: A new method of measurement of the velocity of solid particles and their mass for air blasting (ICSP7, 1999).

96

Effective Use of Fluorescent Tracers for Peening Coverage Peter Bailey Electronics Inc., Cincinnati, OH, USA

1

Abstract

Fluorescent tracers for determining shot peening coverage have been in use for several decades. Their most appropriate applications are nozzle targeting and detection of coverage in difficult to peen areas. To determine actual percent coverage more sophisticated techniques are required. Frequently, tracers are not used properly or effectively. The most common misuse is to assume that full coverage always requires complete tracer removal. In fact, supplier instructions warn against this assumption and recommend the use of a standard on which tracer removal is compared with magnified visual determination of coverage. Tracer misuse often results in over-coverage by continuing peening well past full coverage until all tracer is removed. Undercoverage is also possible where excessive small or abrasive media prematurely removes the tracer. Neither is good for fatigue protection. This paper describes the removal behavior of fluorescent tracers under a range of peening media and process conditions and part material hardness. Analysis of the data suggests methods that make 100 % coverage more easily determinable by tracers. The purpose of the study was to increase the effectiveness of using fluorescent tracers and also their credibility. The author believes that if they are used with knowledge and common sense, improved production efficiency will result and over-peening will be minimized.

2

Results, Conclusions and Recommended Practice

Tests in a peening machine, described in detail later, show that tracer removal exactly corresponds with peening coverage under conditions that are moderately aggressive – for instance, S110 shot at 6A intensity with an impingement angle of 45 degrees. Larger shot at a higher intensity also fits this favorable situation. This correlation was valid for all the target materials tested – from soft aluminum to hard tool steel. Lower intensities, larger shot, hard materials or more direct impingement reduce the aggressiveness and thereby the effectiveness of tracer removal. Rough machined surfaces can also interfere with tracer removal. Less than complete tracer removal at 100 % coverage does not render the process less useful. It does however, make essential the use of a standard (part material piece) which has been peened by the part peening process to 100 % coverage. Observation of tracer removal from the standard shows what tracer removal should look like on production parts. The tracer manufacturers recommend the use of such standards. In addition to reinforcing the need for standards, the author recommends the use of more effective magnifiers, such as low cost 20´ binocular microscopes. Especially useful can be the use of both ultraviolet and white light under the microscope to determine peened and unpeened areas in situations where dimples are shallow and difficult to see.

97

3

Description of Fluorescent Tracers

Fluorescent peening tracers are proprietary mixtures of fluorescent dyes and organic compounds in quick drying solvents. Alcohol and Methyl Ethyl Ketone (MEK) are typically used solvents, with MEK usually preferred because its tracers puddle less on reapplication to spots missed on the first application. Unfortunately, MEK is considered a health risk in some societies. Actual risk is very small because of the minute quantities used. The tracer utilized in this investigation was Fluoro-Finder III produced by American Gas and Chemical. Spot comparisons with other available tracers during the tests confirmed that these other tracers behave similarly in removal behavior.

4

Experimental Equipment

The equipment used was simple in construction because it is based on a modified hand blasting cabinet using a single fixed position nozzle. Nozzles were constructed of small diameter steel pipe so that nozzle length could be changed easily. Surprisingly, they worked better than onhand commercial blasting nozzles in reaching desired intensities with available 6.5 plus 2 hp air compressors. The addition of the following made the setup quite reliable and intensity/coverage reproducible: • • • • • • • •

Pressure pot Shot flow control Cast and cutwire steel shot to AMS 2431 specifications Fixed speed turntable Stopwatch timing Intensities by computer generated saturation curves Startup off the samples until shot flow uniformity achieved Screens to separate previous shot after shot size change

Pressure Pot

compressors

98

5

Sample materials

Test samples included a range of materials and hardnesses: • Aluminum - HRB 48 • Titanium 6-4 - HRB 80 • Titanium 6-4 - HRC 36 • Titanium 6-2-4-6 - HRC 46 • Nickel-base Rene’ 88 - HRC 43 • Steel 1070 - HRC 47 • Tool steel - HRC 63 Sample preparation for each test run comprised removing the peening dimples of the previous run by (1) orbital machine sanding with 240 grit paper, (2) longitudinal hand sanding with 240 grit paper, (3) transverse hand sanding with 400 grit paper and (4) longitudinal hand sanding with 600 grit paper. After solvent cleaning, the tracer was applied with a cotton swab and allowed to dry. The samples were mounted on a 10 inch diameter turntable.

Shot utilized included cast steel S110, S230 and S330 and conditioned cutwire CW14. The cast shot was supplied by Ervin Industries and the cutwire shot by Premier Shot Company. Shot flow control was by MagnaValve from Electronics Inc.. Screens were from W. S. Tyler. Addition of the pressure pot and nozzle fabrication was by the author. The nozzles are shown below. The extra pipe fittings for the shorter nozzles were used to maintain nozzle to sample di-

99 stance. Nozzle wear was minimal during the tests, though they were made from commercial galvanized pipe – 1/8 NPT – internal diameter ¼ inch.

6

Test Results

The tests were conducted in the following sequence: Table 1: Sequence of Tests Shot

Intensity Angle [°] Nozzle [inch] Distance [inch] Mass flow [lb/min]

Air pressure [psi]

S230

15A

90

6

6

18

60

S230

11A

90

2

6

18

60

S230

14A

90

6

6

2

60

S230

6A

90

2

6

10

30

S230

5A

90

2

6

10

20

S230

3A

90

2

6

10

10

S230

12A

90

2

6

2.5

70

S230

7A

90

2

6

2.5

20

S230

7A

90

2

8

2.5

20

S230

6A

90

2

8

10

20

S230

12A

90

2

8

2.5

70

S110

9A

90

6

8

2.5

70

S110

8A

90

4

8

2.5

70

S110

9N

90

2

8

10

20

S110

9N

45

2

7

10

30

S110

6A

45

2

7

2.5

70

S330

13A

45

4

7

10

70

S330

7A

90

2

8

10

20

CW14 8A

90

6

8

2.5

70

7

Summarizing the test observations:

The most aggressive conditions of 45 degree impingement angle with S110 shot at 6A and S330 at 13A showed complete tracer removal from every dimple and complete removal at 100% coverage. (1) Reducing the aggressiveness by changing to 90 degrees (lower shot velocity to reach same intensity), (2) reducing intensity and (3) increasing shot size at same intensity (shallower dimple – same dimple diameter) incrementally reduced the correlation of tracer removal with coverage. The soft aluminum and annealed titanium maintained the correlation further down the progression to decreased aggressiveness. As levels decreased, determining coverage on the harder materials became increasingly more difficult, especially on the HRC 63 tool steel. CW14 behaved exactly the same as S110 which is the same diameter.

100 As coverage determination became more difficult with lower aggressiveness levels, a helpful technique emerged. Looking at the peened surfaces through a 20´ binocular microscope and alternating between white and ultraviolet light made it much easier to determine dimples from areas not impacted. The white light picture below may illustrate. The UV light did not generate enough light for picture taking.

Unpeened

Peened

Another coverage observation was made that may contradict folklore and some existing specifications. It was found that the Almen strips at the saturation point were usually not 100 % covered. This makes sense to the author, because even though the term “saturation” is used, the strip is not really “saturated”. It is still increasing in curvature as time is increased. To make that happen, it makes sense that the coverage is yet to be completed. In determining intensities for this study, a computer generated saturation curve program was utilized. This made intensity determination more accurate because the program picks an exact point from which arc height rises 10 % on doubling the time. It also gave the author confidence that the peening system was working consistently because the curve was very smooth without any outlying points. An example is shown below.

101

A Theoretical and Experimental Investigation into the Development of Coverage in Shot Peening Saravannan Karuppanan*, Jose Solis Romero+, Eduardo R. de los Rios+, Chris Rodopoulos+ and Andrew Levers# *

Universiti Teknologi Petronas, Bandar Seri Iskandar, Perak Darul Ridzuan, Malaysia SIRIUS, Department of Mechanical Engineering, University of Sheffield, S1 3JD Sheffield, UK # Airbus UK Chester Road, Broughton, Chester CH4 0DR, UK +

1

Introduction

Shot peening is a mechanical pre-stressing surface treatment that substantially improves the strength of metals if the process is carefully controlled. The earliest record of mechanical prestressing probably predates 2700 BC, when hammered gold helmets were found during the Crusades, as reported in reference [1]. Peening was a well-accepted technology in the early 1920’s when hand-peening with specific hammers was used in the race-car industry [2]. However, shot peening as a process of the cold working of metal surfaces, was only realised in the middle of 1920’s, as a consequence of the accidental observation that the parts which were sand-blasted for cleaning purposes showed an increased fatigue life. Since the 1960’s, the understanding of the shot peening process has increased significantly, especially in the area of fatigue life improvement. The use of shot peening to improve component fatigue life has also been standardised [3]. However, shot peening process parameters are still selected by means of empirical considerations or by experience. Determining the peening schedules required for optimum shot peening is still a grey area. In most shot peening applications, uniform residual compressive stress in the surface zone is the sole desired effect, as the stresses will resist the formation of fatigue cracks within the component during service, thereby improving significantly the life of the peened component. A few examples of the type of part which have shown a good response to shot peening, include crankshafts (900% life increase), gears (1500% life increase) and connecting rods (1000% life increase) [4]. Although the mechanism of shot peening is a simple concept, the process is complex. The effectiveness of the shot peening process is dependent upon the uniformity of the induced compressive residual stresses and the energy transfer that occurs during the impact of the shots with the target surface. In practice, the process efficiency is established by means of coverage, intensity and saturation. The scope of this study is to investigate the development of coverage and its relationship to intensity and saturation peening. Within this scope, the objectives of this research are: (i) to compare coverage results obtained experimentally with theoretical models of coverage development, (ii) to establish a relationship between coverage and intensity and (iii) to obtain an empirical relationship to predict coverage.

102

2

Coverage

Coverage is defined as a measure of the area fraction of a component surface that has been impacted in a given peening time, usually expressed as a percentage. It is a measure of the interaction between neighbouring indentations, and hence the uniformity of the residual stresses within the surface layers of the shot peened component. Complete visual coverage, (100 % coverage), is reached when the entire surface of a reference area has been indented. At this point, the residual stresses are assumed to be uniform in the surface layers of the component. Coverage of less than 100% is ineffective because of the unpeened surface which contributes to uneven distribution of residual stresses in the surface layers of the component. Coverage above 100 % is achieved by using multiples of the exposure time to 100 % coverage. Indentations are most likely to occur without overlap in the early stages of the shot peening process so that the coverage increases linearly with time. The rate of coverage decreases with time because the probability of overlap increases. The probability that an uncovered area be covered by a new indentation becomes smaller and smaller with time. Hence the approach to 100 % coverage is exponential. In practice, 100 % coverage can neither be accurately measured nor achieved with certainty after a definite exposure time. Hence, complete coverage is assumed to occur when the observed coverage reaches 98 % [5,6]. Coverage can be assessed qualitatively by visual inspection of the reference area with a magnifying glass, or quantitatively by image analysis or by the dyescan tracers technique. Theoretical models have been developed to predict the development of coverage. In this project, the development of coverage will be determined experimentally with the use of an image analysis technique. Two theoretical models, the Avrami equation and the Holdgate model, will also be used to predict the development of coverage.

2.1

The Avrami Equation

A theoretical model reported by Kirk et.al.[7] considers shot size indentation, peening rate and exposure time for the prediction of coverage. This model was based on the earlier work by Avrami and therefore is called the Avrami Equation. This equation is based on the assumption that each shot particle makes the same size of indentation and that the shot particles arrive at the surface in a statistically random manner, but at a rate which is uniform over a significant period of time. In this respect, the Avrami equation in terms of the parameters that are readily determined for a particular peening system, is written as follows: æ 3r 2 mt ö ïü ïì C (t ) = 100 í1 - exp ç 3 ÷ý 4 Ar r ø þï è îï

C(t) is the coverage at any particular time r is the average radius of the indentations A is the area of shot spread t is the time during which the indentations were being created

(1)

103

m is the mass flow rate of shots H is the density of the shot r is the average radius of the shots 2.2

The Holdgate Model

N.M.D Holdgate [8] extended existing models for describing the development of coverage, to one applicable to a general peening system involving multiple peen sources. The proposed model states that if the overall coverage C of a reference area S is known at time t, the overall coverage after the interval @t is approximately given by: ns é a ù C (t + dt ) = 1 - [1 - C (t )] Õ ê1 - j ú j =1 ë Sû

dN j

(2)

nS is the number of peen sources aj is the total area of indentation caused by the peens from the j-th peen source at time @t @Nj is the number of peens from the j-th peen source expected to impact the reference area in an interval of time @t This model can be simplified for a single peen source as:

é aù C (t + dt ) = 1 - [1 - C (t )] ê1 - ú ë Sû

2.3

(3)

Intensity and Saturation

Intensity correlates the amount of energy transferred during the impact of a typical shot with the work piece and is related to the kinetic energy of the blast stream [9]. The Almen strip test, which was originally proposed by J.O.Almen, is used to quantify the intensity level [10]. Saturation refers to the number, uniformity and relative position of the impingements caused by the shot striking the work piece during the exposure time. Saturation is a measure of the effectiveness of the shot peening process. Almen strips can be used to measure the saturation point which is defined as the earliest point on the curve of arc height versus peening time, where doubling the exposure time produces no more than a 10 % increase in arc height. An algorithm developed by one of the authors, for determining the saturation point by means of full regression analysis, has been used in this study. An equation in the following form has been adopted for the solution. Arc Hight =

A A - p p (time + b) b

(4)

Where A, b and p are fitting parameters. Figure 1 show the implementation of equation (4).

104

3

Experimental Details

3.1

The Shot Peening Machine, Media and Target Material

The shot peening machine used in the experiments was a direct-pressure air-blast type, where compressed air at a desired pressure is supplied to the pressure vessel. The pressure of the air is monitored by a pressure transducer and is indicated on a digital display in the facia control. The combined air-media flow then passes through the boost hose into the nozzle mounted at the top of the cabinet. The nozzle directs the shot to the work piece to be peened and can be set to remain stationary or move at a selected speed. An electronically controlled feed valve system (MagnaValve), located at the bottom of the pressure vessel, controls the feed rate of the shot. The shot peening media used in the experiments were S110, SCCW20, S230 and S330, and were projected at impingement angles of 30o, 45o and 90o. Aluminium 2024-T351 and aluminium 7150-T651 were used for the coverage investigation and A type Almen test strips (cold rolled spring steel SAE 1070) were used for intensity determinations.

Arc height (0.001")

saturation curve 32 28 24 20 16 12 8 4 0

data points fitted saturation point

0

10

20

30 40 time (secs)

50

60

70

Figure 1: Saturation curve for S230 shot, the saturation point is at 17A and 8 sec

3.2

Experimental Determination of Coverage

Dimensions of the specimens used for the coverage experiments were 25 mm × 19 mm × (5~7) mm. The surface of the specimens was polished to 1mm finish before peening. A microscope with magnification ×32 was used to capture images of the specimen after each shot peening pass and an image analysis program (SigmaScan), was used to determine the percentage of coverage. An image taken after the 1st pass was used to determine the indentation radius of different shots. Figure 2 show an example of coverage determination by image analysis.

105

1 pass, 37.7% coverage

4 pass, 76.64911 %

8 pass, 95.05208%

Figure 2: Development of coverage in 7150 alloy with shot S230

4

Results

For the application of the Avrami equation the average radius of the indentations was measured from photographs taken after the first pass, as in Fig. 2, while the area of the shot spread was measured from wide metal strips peened along the centre. Regarding the Holdgate equation, the values of the ratio a/S after the first pass were obtained from a regression analysis of the experimental determinations. The regression equation is: a = b0 + b1 x1 + b2 x2 + b3 x12 + b4 x22 + b5 x1 x2 S

(5)

x1 = shot diameter (mm) x2 = impingement angle (o) b0 to b5 = regression coefficients Solutions for the regression coefficients were solved using a Microsoft Excel program. The predicted expressions for both materials are as follows:

4.1

Al 2024

a = 1.43 - 1.72 x1 + 1.48 × 10 -3 x2 + 0.56 x12 - 4.40 × 10 -6 x22 + 9.71 × 10 -4 x1 x2 S

4.2

(6)

Al 7150

a = 1.43 - 1.72 x1 + 1.48 × 10 -3 x2 + 0.56 x12 - 4.40 × 10 -6 x22 + 9.71 × 10 -4 x1 x2 S

(7)

Figure 3 shows the comparison of coverage development as predicted by the two theoretical methods together with experimental results, while Figure 4 shows the effect of shot size, angle of impingement and target material on the rate of coverage.

106 o

o

S230/30º/Al 2024

S230/30 /Al 7150

S230/45 /Al 7150

100

100

80

80

80

Coverage (%)

60

40

60

Coverage (%)

Coverage (%)

100

40

60 40

20 20

20

0 0

0 0

1

2

3

4

5

6

7

8

9

10 11

1

2

3

4

5

6

7

8

0

9

0

No of Pass (1 pass = 0.2 sec)

1

2

3

4

5

6

7

8

No of Pass

No of Pass

100

100

80

80

80

60

S110 SCCW20 S230 S330

40

20

Coverage %

100

Coverage %

Coverage %

Figure 3: Comparison of coverage results obtained with the Avrami equation (fine line), Holdgate method (coarse line) and experimental (dash line and symbols)

60

30 deg 45 deg 90 deg

40

(b)

0

(c) 0

0

2

4

6

Al 2024 Al 7150

40

20

20 (a)

0

60

8

0

No of Passes

2

No of Passes

4

0

2

4

6

8

No of Passes

Figure 4: Coverage development (a) in Al 2024, 90o angle of impingement, using different shot sizes, (b) Al 2024, shot SCCW20, using different impingement angles and (c) in two different materials, shot S230 and 30o angle

5

Discussion and Conclusions

The experimental method used to determine coverage is reliable provided high quality images are obtained from previously polished samples. The SigmaScan program can be used for a faster and easier coverage determination. Application of the Avrami equation requires the determination of two parameters which are the indentation radius, r and the shot spread area, A. These two parameters are determined from simple experimental tests. Application of the Holdgate model requires the determination of the coverage ratio after an initial interval time of shot peening. This ratio is obtained directly from experimental measurements, or can be obtained by regression analysis of peening data. Coverage predictions by the Holdgate model are more accurate than those obtained using the Avrami equation. Coverage development is faster using fine size shots. At a fixed mass flow rate the number of shots impacting the sample is higher using a fine shot than a coarse shot. Coverage development is faster at an impingement angle of 90o followed by 45o and 30o. The area of shot spread at impingement angle of 90o is smaller than at 45o or 30o. Thus, the number of shots per unit area is higher at a 90o impingement angle. Coverage development is faster in Al2024 compared to Al7150, which is a reflection of the softer Al2024 material. The indentations created by shots impacting Al2024 are bigger than in Al7150, which explains the faster coverage rate in Al2024.

107 The time taken to achieve 98% coverage in Al2024 and Al7150 is faster than the time taken to achieve saturation in an Almen strip. This was expected because the hardness of the aluminium specimens is lower than the hardness of Almen strips (steel). The ratio of the time taken for 98 % coverage to the time taken for saturation, tcov/tsat was observed to be between 0.10~0.36 for Al2024 and 0.12~0.48 for Al7150. A clear relationship between this ratio and the shot type or angle of impingement could not be obtained.

6 [1]

References

P. E. Cary, The First International Conference on Shot Peening (ICSP1), Pergamon Press, pp 23-28, 1981. [2] K. H. Kloos & E. Macherauch, Shot Peening : Science-Technology-Application (The Third International Conference on Shot Peening [ICSP3]), Deutsche Gesellschaft fur Metallkunde e.V., pp 3-27, 1987. [3] Military Specification, MIL-S-13165C. Military Specification, ‘Shot peening of metal parts’, The Wheelabrator Corporation, USA 7 June 1989. [4] A. W. E. Corporation, ‘Shot Peening’, 3rd ed. USA: Mishawaka, Indiana, 1947. [5] SAE Fatigue Design and Evaluation Committee, ‘SAE Manual on Shot Peening (SAE HS-84)’, 3rd ed :Society of Automotive Engineers, Inc., 1991. [6] Airbus UK, ‘Process specification (ABP 1-2031)’, pp. 1-11, September 93. [7] Kirk, D. and Abyaneh. M.Y., The Fifth International Conference on Shot Peening (ICSP5), pp 183-190, 1993. [8] N. M. D. Holdgate, ‘Peen Mechanics in the Shot Peening Process’, PhD thesis, University of Cambridge, October 1993. [9] Kyriacou, The Sixth International Conference on Shot Peening (ICSP6), pp 505-516, 1996. [10] Kirk, D., The Fifth International Conference on Shot Peening (ICSP5), pp 9-14, 1993.

108

Almen Gage Calibration Jack M. Champaigne Electronics Inc, Mishawaka, IN, USA

1

Introduction

The Almen gage is used to measure the curvature or arc height of the Almen test strip that has been subjected to particle impacts on one side. The resulting impingement on the test strip causes it to stretch and arch. The resulting measurement is used to determine the blast stream energy or peening intensity. The Almen gage was invented by J. O. Almen of General Motors Corporation in 1942 and a

U. S. Patent was issued in June of 1944 (see Appendix A. for drawing). The original gage used two knife-edge supports for the test strip. However, in November of 1943 Engineers at General Motors revised the original design by replacing the knife-edges with four-ball support and designated the new gage as #2 Almen gage. This same basic design is in use today around the world with only minor modifications. The Society of Automotive Engineers, SAE, has developed a standard practice for the construction of the gage in document J-442 [1] The present application of the gage includes the use of a digital indicator replacing the original dial indicator and the addition of end-stops to help assure proper positioning of the strip for measurement. The SAE specification gives dimensional data necessary to construct the gage but does not mandate a calibration procedure. This article will address calibration procedures and recommended practices. Three areas are explored. • Calibration of the indicator • Measurement of ball position(s) • Matching gages with special gage block

109

2

Calibrate the Indicator

Primary importance is often assigned to the dial indicator accuracy and performance and it is often the only element of the gage that is calibrated. Most large aerospace firms and calibration laboratories are capable of calibrating the indicator.

2.1

Commercial Calibration Stand

A commercial indicator calibration stand is shown in figure 1. This is a common tool used to calibrate analog or digital dial indicators and it offers a high degree of precision. Unfortunately it is expensive and time-consuming to use this method of calibration.

Figure 1: Commercial indicator calibration stand

2.2 Calibration Step Blocks A simpler method of indicator calibration using precision step-type gage blocks can be employed. These blocks, shown in figure 2, have a flat datum reference surface and a slot of precise depth from that surface. Since the gage block is not as long as the Almen strip it is allowed to slide from side to side on the gage platform and thereby allow the indicator tip to rest upon either the datum reference or the slot. This technique precludes having to dismount the indicator

110 from the gage frame, another timesaving feature. Using a set of such blocks with convenient slot depths can provide a quick and easy indicator calibration. [3]

Figure 2a: Step block used to calibrate digital indicator

Figure 2b: Step block to left for zero

3

Figure 2c: Step block to right for calibration (inch mode)

Ball Position

Ball position is very important for proper operation of the gage. In addition to ball placement the condition of the balls, namely flatness, is another important factor in gage accuracy.

3.1

Ball Placement

Incorrect ball placement can contribute to erroneous readings. There are two common methods available to determine if ball placement is within specified limits. One method is to use computer controlled measuring equipment such as a coordinate measuring machine, CMM. These ball position measurements can then be compared to the requirements of J-442 to verify compliance.

111 A special template, shown in Figure 3 can be used to validate proper positioning of the balls. This type of test is more accurately described as a “conformity” test since no numerical data is generated. It is a go-no-go affirmation of ball placement integrity. The template is placed on the gage and visual inspection should reveal that all four balls are visible within its respective template hole. If any ball is out of position it will be readily apparent upon inspection and the gage can be rejected for non-compliance. Another feature of this particular template is its thickness. It is 2mm thick, the specified minimum ball height requirement. Each of the four balls must extend above the top plane of the template to meet the 2 mm ball height requirement and sliding a straight edge over the surface and detecting the “obstruction” presence of each ball easily affirm this.

Figure 3: Special template to verify position of 4-balls

3.2

Ball Flatness

SAE guidelines now dictate that balls shall be replaced “whenever any visible signs of flatness are observed.” The importance of this is illustrated in Figure 4 that shows a schematic of a gage with a curved test specimen that measures .024 inch deflection or arc height. With balls in a flat condition the resulting gage reading increases. The “ideal” reading of .024 inch has changed to almost .030 inch, approximately 25% increase in reading. [2]

3.3

Ball Plane Flatness

Another tool is needed to affirm ball plane flatness, which is specified in SAE J-442 to be .05mm. Although it is not difficult to use conventional dial indicators it has been found to be much easier to place a known flat surface, such as shown in Figure 5, onto the measuring position. Individual balls are adjusted for height until the template does not “rock” back and forth or

112

a) Figure 4: a) Reference zero and ideal reading .024” b) Flat ball condition results in .030” reading

b)

Figure 5: Flat block used for checking for ball plane flatness

tilt. The ball plane flatness is easily obtainable with this technique. Any non-plane flatness greater than .005mm causes a very noticeable rokking or tilting which can be easily detected.

4

Matched Calibration Gages

Manufacturing tolerances in such a complicated device can result in non-uniform readings from gage to gage even though the gage is technically “in compliance” with the construction requirements. It would be desirable to offer “matched calibrated gages”, especially when a gage might be taken out of service and its replacement might give different readings. Such an opportunity exists when using the calibrated curved block.

113 A special curved block was reviewed by the National Institute for Standards and Technology (NIST) in Washington, DC and a value with an estimate of uncertainty was assigned to the block. After placing the curved block into measuring position the gage can be re-calibrated to match the value of the curved block, thus assuring gagetogage reproducibility. [4]

Figure 6: Curved block used to calibrate matched gage sets

4 [1] [2] [3] [4]

References SAE J-442 (Dec 2001) Test Strip, Holder and Gage for Shot Peening, The Society of Automotive Engineers, Warrendale, PA, USA Champaigne, J., Almen Gage Accuracy and Repeatability, 1993 Oxford, England Proceedings of Fifth International Conference on Shot Peening 1627 Champaigne, J., U. S. Patent 5,780,714 1998 Calibration Apparatus and Method for Shot Blast Intensity Measurements Champaigne, J., U. S. Patent 6,289,713 2001 Method of Calibrating Gages used for Measuring Intensity of Shot Blasting

Appendix A Almen, J. O. U.S. Patent 2,350,440 1942

114

Performance of Almen Strips which are Straightened after Tempering Jack M. Champaigne Electronics Inc., Mishawaka, IN, USA

115

116

117

118

5

Acknowledgements

6

References

117

Optimization of the Shot Peening Parameters Franck Petit-Renaud USF Impact Finishers (Shot Peening Division) and USF Vacu-Blast International, Slough, UK

1

Abstract

The shot peening process is a complicated mechanism in materials science, which is still not fully understood. Despite a long history and a large number of investigations into the process it is still characterized by many areas of uncertainty. Notwithstanding this situation the aerospace and automotive industries have for years considered shot peening as a state-of-the-art process for the surface improvement, forming and life improvement of many parts. The work described in this paper is a study of the effect of a range of process parameters on the residual stress profiles produced by shot peening coupons of case carburized 17CrNiMo6 steel. The peening process utilized for the research was undertaken using a commercial shot-peening unit supplied by USF Vacu-Blast Limited using 0.6mm diameter shot. The process parameters investigated included air pressure, the mass flow, the impact angle, the distance between the nozzle and the specimen, the exposure time and the nozzle size. Using Minitab v12 software regression analyses were performed on the results obtained from the statistically designed experiments. It was found that the most significant parameters were air pressure, the mass flow, the impact angle and the exposure time. Further important and significant interactions were also detected between exposure time and air pressure; nozzle size and mass flow; air pressure and impact angle; nozzle size and air pressure.

2

Introduction

The study of the different parameters involved in shot peening applications is important in order to have better understanding and control of such process. The significance and influence of these parameters are not yet clearly established and most of the knowledge is based on practical experience rather than detailed research. There are only limited methods of assessing the results obtained from peening (e.g. Almen strips) and prediction of final properties is not possible yet. The investigation presented in this paper was aimed at designing and carrying out experimental procedures in order to understand the effects of shot peening on components by analyzing the changes occurred during the process. Therefore, determining the parameters involved to carry out the process and measuring residual stress in peened specimen were the two objectives of the investigation. Being able to relate the shot peening parameters directly to the result produced by the process would indeed be of great advantage as it could lead to a better and more accurate control of shot peening. It would mean predicting the result induced by peening a component and increase the reliability of such process [16][17][18].

120 It was not intended to generalize the whole process but, focusing on what seemed to be the heart of shot peening and limiting the investigation to one type of material. Measuring the change in residual stress within the specimens used was thought to be the most interesting and useful way of understanding the process. A statistical approach to the problem was used to design all the experiments and specific tools for the analysis of the results considered. It is believed that this investigation was one of the most complete in terms of relating different set of parameters to different responses; the considerable number of residual stress measurements carried out could be used as a good basis for an even wider research program into the process, aiming at building software and database, dedicated to produce known effects on components and increasing the reliability of the shot peening process.

3

Process Effects

The immediate effect of bombarding high velocity shots onto a metallic target is the creation of a thin layer of high magnitude compressive residual stress at or near the metal surface, which is balanced by a small tensile stress in the deeper core (Figure 1). The magnitude of this compressive residual stress is a function of the mechanical properties of the target material and may reach values as high as 50 to 60 % of the material’s ultimate tensile strength [1][3].

Figure 1: Effects of shot peening

Its depth is largely dependent on the peening intensity and the relative hardness of the impinging shot and target material. For a relatively soft target material (230-300 HV), it is feasible to produce a compressive layer of 800 to 1000 mm deep, whilst for a harder material (700 HV), it can be difficult to produce a compressive layer of much more than 200 to 250mm [1][2].

121 The introduction of this compressive residual stress at the metal surface layer brings one major benefit: it reduces and can negate any residual or subsequently imposed tensile stress at the metal surface [9][11]. As it is well known, most fatigue failures and stress corrosion failures normally start at or near the surface stressed in tension [2][3]. Therefore, by reducing the net tensile stresses at and near the surface of the component, fatigue crack initiation and stress corrosion can be delayed, improving the fatigue life of the component treated [10][12][13]. If the resultant surface stress can be made compressive enough, cracks could virtually be prevented from opening up at the component surface resulting in a much enhanced fatigue life [3][5][6]. This is generally true for shot-peened components subjected to low stress amplitudes.

4

Process Parameters

The shot peening process has to be a precisely controlled and repeatable process for optimum benefit. To achieve this, all its process variables must be identified and controlled [7]. There are many fundamental parameters affecting the shot peening process. The most common are as follows: • • • • • • • •

Shot density; Hardness and size of the shot; Nozzle characteristics (diameter, deflection angle, length); Air pressure: Impact angle; Distance from nozzle to work-piece; Exposure time, number of passes; Linear and rotational speed of work-piece relative to nozzle.

To specify all these variables every shot-peening job would require time consuming investigations and industrially impractical procedures. To overcome this problem, J. O. Almen [4][8] introduced the concept of peening intensity measurement based on curvature induced in a thin test strip, by which most of the previously listed process parameters can automatically be incorporated into one process variable called the Almen peening intensity [2][3][4]. With peening intensity known, one has only to define the shot type and size and peening coverage desired to fully define the peening process. As experience and various studies have demonstrated the improvements induced by the peening process, it is widely used to enhance the life of components operating in highly stressed environment and other critical parts such as in Formula 1 motor racing, aero engines and aero structures [14][15]. Despite important progress in understanding the process, some areas are not totally mastered yet and difficulties are still hard to avoid. Being able to predict the effect of the process in set conditions is indeed the key to gain complete control over the process and to make it much more reliable.

122

5

The Test Specimens

The dimensions of the test specimen are 10*10*100 mm. The material used was the steel 17CrNiMo6, chosen because of its interest for gear manufacturing. The manufactured bars were carburized, quenched, tempered and lightly ground before being peened to maximize the effects of the process. At the end of heat treatment, the bars were expected to exhibit a Vickers surface hardness of approximately 700 kgf.mm-2, which is typical of many case hardened gears. In order to limit the number of specimens to be manufactured, a masking technique was devised so that a number of different peening operations could be carried out on one specimen. Also, to ensure that the surface hardness of the bars manufactured was of the expected level, some tests have been carried out. The hardness measured on each face of the specimen varied from 54 to 57HRC. Knowing that the shot (Steel Shot S230, 0.6mm) used for this investigation had an approximate hardness of 55 to 65HRC, which is harder than the specimen, it was expected to observe some effect from the shot peening process.

6

Design of the Experimental Procedures

As this investigation had a broad spectrum of possibilities, designing the experiments was a necessary step in order to focus on the relevant information and establish the effects and significance of the process from a practical point of view. Six parameters and their significance were investigated, aiming at relating their conjugate effects to the residual stress introduced. Each parameter was tested at three different levels (Low, Medium and/or High). In the following table, the list of control variables is shown, with their respective experiment levels and assigned values: Table 1: The control variables and testing levels Testing levels Parameters

Low

Medium High

Exposure Time (s)

-1

+1

(2 Levels)

Nozzle diameter (in)

-1

+1

(2 Levels)

Air pressure (bars)

-1

0

+1

(3 Levels)

Distance nozzle-specimen (mm)

-1

0

+1

(3 Levels)

Impact angle (deg)

-1

0

+1

(3 Levels)

Mass flow adjustment (kg/min)

-1

0

+1

(3 Levels)

108 different set-ups were randomly allocated on the selected specimen, making sure that some sites were kept blank. The total number of experimental sites was 132. Each experimental site was processed with the required conditions and X-ray measurements were carried out to determine the residual stress profile introduced by the process. Figure 3 is a typical example of the profiles obtained:

123

Residual stress profile- Site 445, run 83 SPOL Depth (microns)

DRSM

0.0 0

20

40

60

80

100

120

140

160

180

200

-200.0

Residual Stres, RS (MPa)

-400.0

-600.0

-800.0

RSSf -1000.0

RSM

-1200.0

-1400.0

RSo (MPa)

RS90 (MPa)

Figure 2: A residual stress profile obtained for one site for the main experimental program (e.g. RSM=1042.0MPa, DRSM=17.5mm, SPOL=105.0mm and RSSF=-944.5MPa)

The profiles obtained were then analyzed and key values recorded. The key values were: • • • • •

The maximum compressive residual stress (RSM, in MPa); The depth at which the maximum compressive residual stress occurs (DRSM, in mm); The depth of shot peened outer layer (SPOL, in mm); The surface residual stress before peening (measured on the blank experimental sites; RSSi, in MPa); The surface residual after peening (RSSf, in MPa).

These key values were selected as being the most important final results. They will be used as the responses to be explained in terms of the 6 process parameters: their 6 main effects, the 4 quadratic effects and the 15 interactions using regression analysis.

7

Statistical Treatment: Evaluation of the Significance of the Process Parameters and Interactions

To carry out the statistical analysis, five responses were investigated in the statistical analysis: • •

The maximum residual stress; (RSM, in MPa); The depth of the maximum residual stress (DRSM, in mm);

124 • • •

The shot peened outer layer (SPOL, in mm); The surface residual stress after peening (RSSf, in MPa); The variation in the surface residual stress from un-peened to peened ([RSSf-RSSi], in MPa).

These results have been selected for various reasons. However, as it is intended to predict the residual stress distribution caused by the process, if RSM, DRSM, RSSf and SPOL are known it is then possible to define and “visualize” the residual stress profile. The last result selected ([RSSf-RSSi]) for the analysis is not considered as a very important one. However, this value can be used as an indicator, showing the change in the residual stress at the surface due to the shot peening process.

7.1

The Statistical Models

The response variables in the regression were selected from the 6 main effects (a, b, c, d, e and f)1, 15 interactions (ab, ac, ad,…, ef) and 4 quadratic effects or squared terms (c2, d2, e2 and f2). Two variables, a and b, which were only at two levels (-1 and +1), their squared terms a2 and b2 cannot be included for the analysis. Therefore, they do not appear in any of the equations found. The process parameters effects were included in the final models if they were of obvious physical importance, or if they were statistically significant at the 10 % level (p-value < 0.10). In addition, the main effects were always included if the process parameter appeared in a significant interaction or quadratic term. Observing and comparing the different stress profiles, differences can be seen, clearly showing the different influence from one set of parameters to another (Figure 3). Looking at the residual stress maximum values (RSM), it was observed that the values ranged from –600 Mpa to –1400 Mpa, with the majority of experimental sites exhibiting a maximum compressive residual stress ranging between –1000 MPa and –1300 MPa (Figure 4). The corresponding depth (DRSM) varies from 0mm up 80 mm (Figure 4), whilst the shot peened outer layer (SPOL), can be seen to range from 60mm to 390 mm (Figure 5). It can also be observed that a majority of experimental sites exhibit a maximum depth of residual stress of 10 mm to 40 mm and that the maximum shot peened outer layer varied between 80mm to 280mm. Figure 5 also shows that the residual stress at the surface (RSSf) ranged from 650MPa to –1080 MPa, with the majority of samples varying from around -800 MPa to –1000 MPa. These four responses are the most important. Indeed, a high compressive residual stress introduced deep in a component will help prevent crack initiation as well as enhancing the overhaul life expectancy of the processed part. Looking at these results, it was also interesting to check if any correlation exists between these individual responses. The next graph (Figure 6) shows a plot of the depth DRSM as a function of the residual stress RSM. From this plot, it is possible to observe that when the maximum compressive residual stress RSM increases (in the negative direction), the depth of this residual stress tends to increase. 1. a: Exposure time; b: Nozzle size; c: Air pressure; d: Distance nozzle-specimen; e: Impact Angle; f: Mass Flow

125 Re sidual stre ss profile - Site 624, run 90

Residual stress profile- Site 434, run 3

Depth (mum)

0.0 0

20

40

60

80

100

140

160

180

0

200

20

40

60

80

100

120

140

160

180

200

-200.0 -400.0

RS (MPa)

-400.0 RS (MPa)

Depth (m um )

0.0

120

-200.0

-600.0

-600.0 -800.0

-800.0 -1000.0

-1000.0 -1200.0

-1200.0 -1400.0

-1400.0

RS90 (MPa)

RSo (MPa)

RSo (MPa)

RS90 (MPa)

Figure 3: Residual stress profiles for 2 distinctive sets of parameters

20

Frequency

Frequency

20

10

0

10

0

-1500

-1000

-500

0

10

20

RSM (MPa)

30

40

50

60

70

80

90

DRSM (microns)

Figure 4: Repartition of the values obtained for RSM and DRSM over the experimental sites

15

Frequency

Frequency

20

10

10

5

0

0 0

100

200

SPOL (microns)

300

400

-1100

-1000

-900

-800

-700

-600

RSSf (MPa)

Figure 5: Repartition of the values obtained for SPOL and RSSf over the experimental sites

The p-value (p-value = 0.000) shows that this sample correlation is most unlikely to have occurred by chance if the variables RSM and DRSM are independent. Carrying out the same analysis between RSM and SPOL we obtain the plot shown in Figure 6. It can be seen that, as above, an increase of RSM leads to and an increase of SPOL. From the previous results and carrying out a similar analysis between DRSM and SPOL, the following graph was obtained (Figure 7). It is possible to see that as DRSM increases SPOL increases.

126

400

80

60

SPOL (microns)

DRSM (microns)

70

50 40 30 20

300

200

100

10 0

0 -1350 -1250 -1150 -1050 -950

-850

-750

-650

-1350 -1250 -1150 -1050 -950

-550

-850

-750

-650

-550

RSM (MPa)

RSM (MPa)

Figure 6: Correlation between RSM and DRSM (e.g. correlation of RSM and DRSM = –0.588, P-value = 0.000) and correlation between RSM and SPOL (e.g. correlation of RSM and SPOL = –0.683, P-value = 0.000)

-600

-700

300

RSSf (MPa)

SPOL (microns)

400

200

100

-800

-900

-1000

-1100

0 0

10

20

30

40

50

60

70

80

-1350 -1250 -1150 -1050 -950

DRSM (microns)

-850

-750

-650

-550

RSM (MPa)

Figure 7: Correlation between DRSM and SPOL (e.g. correlation of DRSM and SPOL = 0.820, P-value = 0.000) and correlation between RSM and RSSf (e.g. correlation of RSM and RSSf = 0.214, P-value = 0.026)

The plot also shows that there are no obvious correlation between RSM and RSSf. The correlations are desirable in as much as the high compressive residual stress is associated with an enhanced life expectancy of a component (e.g. a gear). The final statistical models established for each response were as follows:

RSM (MPa) = -1158 - 58.6 a + 9.94 b -118 c - 58.3 e + +10.2 f + 88.8 c2 + 17.5 ac + 26.3 bf DRSM (microns) = 35.8 + 3.11 a + 15.3 c - 2.14 d + 6.46 e - 5.58 f -5.34 e2 + 3.83 ac + 4.33 ce - 2.53 de

(Eq. 1)

(Eq. 2)

SPOL (microns) = 199 + 2.18 a + 65.0 c + 35.0 e -19.6 f - 20.2 e2 - 9.79 ae + 12.3 ce

(Eq. 3)

RSSf (MPa) = - 906 + 5.93 a - 1.01 b + 6.54 c + 5.97 e + +45.9 c2 - 33.2 e2 + 35.9 ac - 14.5 bc

(Eq.4)

127 [RSSf- RSSi] (MPa) = -744 - 37.1 b + 41.4 c - 77.8 d - 51.8 bc + 44.7 bd

(Eq.5)

Considering the previous equation for RSM (Eq. 1). The objective is to optimize the values for a, b, c, e and f, assuming that we want to minimize the value RSM (we want the most negative value). The most negative value for RSM with the process parameters within the ranges investigated in the experiment is obtained when: • a = +1; • b = +1: • c = +0.57; • d = -1: • e = +1; • f = -1. Substituting these values into the equation (Eq. 1) leads to the following results: Table 2: Optimum parameters and results 95% CI*

95% PI**

Parameters

Results

Exposure time(a) = +1 (3s)

RSM

–1330MPa

(–1375;–1285)MPa (–1485;–1175)MPa

Nozzle size (b) = +1 (5/16”)

DRSM

64mm

(55;70) mm

(40;90) mm

Air pressure(c) = +0.57 (3.4bar)

SPOL

270mm

(250;295) mm

(185;355) mm

Distance (d) = –1 (100 mm)

RSSf

–900MPa

(–930;–860)MPa

(–1050;–750)MPa

–820MPa

(–860;–650)MPa

(–1200;–300)MPa

Impact angle (e) = +1 (90 deg) [RSSfMass flow (f) = –1 (1 kg/min) RSSi] *

95% CI is an interval of the average response if the process is operated with these parameter values.

**

95% PI is an interval within which there is 95% chance that an individual test could lie, assuming a Weibull distribution. In this case, the interval is equal to ±3.7*s (where s is the standard deviation).

Using these results, we can determine what the optimum residual stress profile created would be (Figure 8). This model is a prediction of what may happen if we carry out an experiment using the optimum parameters determined above. Ideally we would expect to achieve similar RSM, DRSM, SPOL and RSSf. The difference [RSSf-RSSi] is not critical in the sense that it relies on the “state” of the work piece prior to the process (how it has been manufactured and how well the different treatments such as grinding and/or heat treatment have been performed).

128

Optimum Residual stress profile Depth (microns)

DRSM

0 0

50

63.65

100

150

SPOL 200

250

269.85

300

-200

Residual Stress (MPa)

-400

-600

-800

RSSf

-897.47

-1000

-1200

RSM

-1329.89

-1400

Figure 8: Optimum residual stress profile

8

Conclusions

The aim of this programme was to collect as much data as possible in order to carry out a full statistical analysis. As presented in the previous chapter, this statistical analysis led to several models corresponding to the five types of results felt to be most important in terms of the final material condition (RSM, DRSM, SPOL, RSSf and [RSSf-RSSi]). By using these relationships between the selected results and the process parameters, optimum values were calculated to achieve the optimum final results required. This was based on the following requirements • • • •

A high compressive residual stress (RSM); The greatest value for the depth of the maximum compressive residual (DRSM); The deepest shot peened outer layer (SPOL); A high compressive residual stress at the surface (RSSf);

The difference [RSSf-RSSi] was not considered so important (RSSf was more important than the level of change), although it was a good indication of the effect of the process on a component. The greatest this difference is, the better.

129

9 [1] [2] [3] [4]

[5] [6] [7]

[8]

[9] [10]

[11] [12]

[13] [14] [15] [16]

[17] [18]

References E. B. CHEE, Effect of Peening and Re-peening on the Improvement of Fatigue Life of InService Components, Cranfield University, 1982 D. KIRK, Shot Peening, Seminar, USF Vacublast Slough, 1998 METAL IMPROVEMENT COMPANY, Shot Peening Applications, 7th Ed., USA, 1980> W. CAO, R. FATHALLAH, L. CASTEX, Correlation of Almen arc height with residual stresses in shot peening process, Materials Science and Technology, Sept. 1995, Vol. 11, pp. 967–973 T. SENO, H. HORIUCHI, T. NAITO, Effect of various peening parameters on compressive residual stress for carburized steel, ICSP-3 Proceedings, pp. 505–512 K. TOSHA, K. IIDA, Residual stress and hardness distributions induced by shot peening, Meiji University, ICSP-4 Proceedings, pp. 379–388 R. FATHALLAH, G. INGLEBERT, L. CASTEX, Prediction of plastic deformation and residual stresses induced in metallic parts by shot peening, Materials Science and Technology, July. 1998, Vol. 14, pp. 631–639 R. HERZOG, W. ZINN, B. SCHOLTES, H. WOHLFAHRT, The significance of Almen intensity for the generation of shot peening residual stresses, ICSP-6 Proceedings, pp. 270–280 M. OSHAWA, T. YONEMURA, Improvement of hardened surface by shot peening, Journal of the Japan Society for Heat Treatment, 1989 M. HASHIMOTO, M. SHIRATORI, S. NAGASHIMA, The effects of shot peening on residual stresses and fatigue strength of carburized gear steels, ICSP-2 Proceedings, pp. 495–504 A. NAKONIECZNY, W. SZYRLE, Residual stress, microstructure and fatigue behaviour of carburized layers before and after shot peening, ICSP-6 Proceedings, pp. 263–269 M. HASHIMOTO, S. HOYASHITA, Improvement of surface durability of case carburized and hardened gear by shot peening and barreling processes, ICSP-6 Proceedings, pp. 34–43 K. NAITO, T. OCHI, N. SUZUKI, Effect of shot peening on the fatigue strength of carburized steels, ICSP-4 Proceedings, pp. 519–526 K. OGAWA, H. YAMADA, K. SAKURI, Influence of residual stress on fatigue strength of carburized and shot peened notched specimens, ICSP-4 Proceedings, pp. 445–454 S. ADACHI, Fatigue strength of gear steels shot peened in extremely high intensity conditions, ICSP-4 Proceedings, pp. 363–372 M. WIDMARK, A. MELANDER, Effect of material, heat treatment, grinding and shot peening on contact fatigue life of carburized steels, International Journal of Fatigue, 1999, Vol. 21, pp. 309–327 A. M. KORSUNSKY, An analysis of residual stresses and strains in shot peening, ICRS-5 Proceedings, 1999 H. AOKI, E. NAGASHIMA, T. MIURA, Effect of shot peening conditions on fatigue strength of carburized steels, ICSP-4 Proceedings, pp. 513–518

130

Shot Peening on Pelton Wheels: Methods of Control and Results Paolo Marconi 2 Effe Engineering s.u.r.l., Manerba del Garda, Italy

Marco Lauro Enel Produzione S.p.A., Turin, Italy

Walter Bozzolo Enel Produzione S.p.A., Turin, Italy

1

Abstract

Shot peening is normally used to increase fatigue resistance in rather small-sized mechanical parts (gears, transmission shafts, connecting rods). Residual stress induced by this treatment has been studied extensively in the past decade. Some automobile manufacturers have recently introduced specifications concerning the depth of residual stress in gears. Most of the research done on this topic has made use of the X-ray diffraction method to determine residual stress. Because of the configuration of Pelton wheels, shot peened areas cannot be easily accessed with portable diffraction meters. For this reason, the authors have used the Barkhausen noise method to measure residual stress and have then compared results before and after peening. The Barkhausen noise method was calibrated against Almen strips by X-ray diffraction and by hole drilling methods.

2

Shot Peening

The purpose of shot peening is to increase fatigue resistance in mechanical parts. Essentially it is a process in which the surface of the part is bombarded with spherical particles made of the appropriate material and to the correct size. The impact of these spheres on the surface of the part causes a plastic deformation and creates “microcups” which are evident in the following SEM image: This plastic deformation induces a state of compression which significantly increases resistance to fatigue. Obviously the performance of the part depends on the depth and on the value of compression. These variables (depth and value) are a function of shot peening procedure parameters, which are defined by intensity, shot size and coverage. Intensity is determined by measuring the effect of peening on an Almen strip, which is a hardened C70 steel strip tempered to 44-45 HRC; strip dimensions are indicated in the following sketch (MIL-S-13165C and UNI 5394-72).

131

Figure 1: shot peened aluminium surface

Almen strip (mm) Figure 2: Almen strip (mm)

Deformation, and therefore peening intensity, is expressed in Almen readings which are divided in N, A and C, according to the thickness of the strip. The size of the shot is chosen on the basis of the desired intensity (the higher the intensity, the bigger the shot) and depending on the geometrical aspects of application (for example, a feeding hose with a small radius).

132 Coverage is a function of peening time and is measured by visual inspection (30x magnification). The proportion between peened surface and toal surface (expressed as a percentage) gives the coverage value (MIL-S-13165C and UNI 5394-72).

2.1

Measurement of Induced Residual Stress

As previously stated, the increase in resistance to fatigue is due to induced residual stress. However it must be noted that peening parameters are determined on the basis of tests done on Almen strips, which in most cases do not have the same geometric or metallurgical characteristics of the part to be peened. In order to overcome this problem, some of the more well-known automobile manufacturers have recently established internal procedures which no longer refer to traditional parameters for shot peening of gears (intensity, shot size, coverage), but instead refer to depth and value of induced compression. The profile of residual stress is usually obtained through X-ray diffraction. Diffraction of Xrays is a non destructive method of measuring residual stress, however only at surface level, since at most the X-rays manage to penetrate about 10 microns beneath the surface. In order to obtain a reading at the desidered depth it’s necessary to strip away successive layers of material by electrochemical means; this process does not induce further residual stress [1]. Another method, which however is difficult to apply to automobile gears, is by hole drilling. Through the use of strainmeters (or extensometers), a reading is made of the deformation caused by the release of tension after drilling a small hole in the part. By drilling the hole in stages at deeper and deeper levels, it’s possible to measure residual stress in the part.

2.2

Shot Peening Control on Pelton Wheels

Controlling stress induced by shot peening (and therefore controlling the increase of resistance to fatigue as a result of peening) on Pelton hydraulic wheels is problematic. In fact, even though the invasive effects of in-depth X-ray diffraction or hole drilling are negligible on such large parts, the problem is gaining access to the area to be measured, as shown in the following photograph: This problem became evident during peening qualification procedures by ENEL Produzione. On that occasion a cooperation agreement was established between 2Effe Engineering and ENEL Produzione in order to determine residual stress both with traditional methods (hole drilling and strainmetering by ENEL Produzione) and with non invasive alternative methods such as X-ray diffraction and Barkhausen noise measurements done by 2Effe Engineering.

133

Figure 3:

2.3

Barkhausen Noise Method

The Barkhausen noise method can be applied only on ferromagnetic material (iron, cobalt, nikkel and their alloys). As is well known, magnetic domains in these materials are usually oriented in casual directions and therefore overall magnetization is nil. Application of a magnetic field or a mechanical deformation determines a change in the magnetic domains which in turn determines a change in the overall magnetization of the part. By applying an alternating magnetic field and by measuring the intensity of the movement of the domains through a coil in which this movement produces a periodic electrical noise (Barkhausen noise), it’s possible to relate the noise level to the level of mechanical pressure. The intensity of the noise is influenced by various microstructural parameters (shot size, hardness, residual austenite content, etc.) in addition to, of course, by the chemical composition of the part. Generally speaking, it can be said that higher noise levels correspond to residual tensile stress, whereas compression leads to considerably lower noise levels (obviously in the presence of the same material and microstructural conditions) [2]. In order to obtain a calibration curve, which is necessary in order to transform the values of Barkhausen noise into residual stress, shot peening with 2 different intensities (11 and 14 Almen A) was carried out on samples of the same material (steel 13 Cr 4 Ni) and the same thickness. These samples were comparable to those areas of Pelton wheels which are shot peened. Measurements of residual stress were then taken on these samples with the X-ray diffreaction and hole boring methods. The results appear in the following diagram, in which stress levels are measured in 2 directions: In-depth stress levels measured with the hole boring method are as follows:

134

intensity 11A XRD

residual stress (MPa)

0 -100 0

0,1

0,2

0,3

-200 -300

parallel perpendicular

-400 -500 -600 depth (mm)

Figure 4:

intensity 14A XRD

residual stress (MPa)

0 -100 0

0,1

0,2

0,3

-200 -300 parallel perpendicular

-400 -500 -600 -700 depth (mm)

Figure 5:

The following are Barkhausen noise levels measured with a magnetic field at 7-20 KHz frequency, corresponding to a depth of penetration of about 0.2 mm: Table 1: Barkhausen noise levels untreated surface

peened 11A

peened 14A

3000

5000

5500

As can be seen, measured values contradict the Barkhausen noise theory. In fact, peened surfaces give a much higher noise level compared to the unpeened surface (when measured with

135

deformations (microepsilon)

intensity 14A extensometer

1000 500 0 0

1

2

3

-500

sigma1 sigma2 sigma3

-1000 depth (mm)

Figure 6:

the X-ray diffraction method, the unpeened area has a surface compression of –67 Mpa in the perpendicular direction and –199 Mpa in the parallel direction). The explanation for this phenomenon lies in the residual austenite content. This phase of iron (face-centered cubic lattice) is non magnetic, in addition to being unstable at low temperatures [3]. Therefore the presence in the sample piece of this type of iron, because of the heat treatment which it underwent, lowered the Barkhausen noise level. The subsequent peening procedure transformed residual austenite into martensite and this changed the magnetic behaviour of the material and increased the Barkhausen noise level, even though peening created high levels of compression, as was measured by the X-ray diffraction and the hole boring methods. Measurements made on residual austenite by X-ray diffraction have demonstrated the validity of this thesis, showing the following transformation of austenite: Table 2: Transformation of austenite untreated surface

Peened surface 11A

Peened surface 14A

Retained austenite (%)

Retained austenite (%)

Retained austenite (%)

10.6

<1

<1

3

Conclusions

The purpose of peening hydraulic Pelton wheels is to increase fatigue resistance. The use of Almen strips does not ensure appropriate residual stress levels which are necessary to achieve the desired increase in fatigue resistance, even if the strips are placed in positions which reproduce the actual areas to be peened. Refined techniques which give reliable results, such as X-ray diffraction and hole boring, are not easy to apply because of the shape of the piece and difficulty in accessing the area involved. The Barkhausen noise method, appropriately calibrated with X-ray diffraction and hole boring methods, does not encounter the same difficulty as the other me-

136 thods, but gives results which are drastically influenced by the change of magnetic properties induced by peening in the presence of residual austenite. In fact, compression induced by peening tends to lower the noise level, whereas a reduction in austenite content tends to raise the noise level. This problem can be avoided by measuring Barkhausen noise levels before and after shot peening in order to quantify the change. In fact, low starting noise levels, with the same type of surface treatment, indicate a high residual austenite content which will considerably increase noise levels once it changes because of peening. Vice versa, high noise levels implicate low volumes of residual austenite and subsequent peening will considerably reduce these levels. In both cases the difference in noise levels induced by peening is a function of peening intensity. This allows the method to contribute significantly towards evaluation of peening results. Obviously this difference can be fully understood by calibrating with X-ray diffraction o hole boring and by considering the amount of residual austenite.

4 [1] [2] [3]

Bibliography TAIRA S.,“X-Ray studies on mechanical behavior of materals”. The Society of material science, Japan,1974, pp 22-32. BARKHAUSEN H.,“Rauschen der Ferromagnetischen Materialen”. Phys. Zeitschrift, Vol. 20, 1919, pp 401-403. NICODEMI W., ZOIA R.,“Metallurgia applicata”.Masson Italia Editori, Milano, 1980, pp 143-144.

135

Effect of Shot Peening on Erosion and Fatigue in Combined Bending and Torsion of the Magnesium Alloy AZ80 Ajit Jain, Vinod S. Nadkarni and Mahesh C. Sharma Maulana Azad National Institute of Technology, Bhopal, India

1

Abstract

Erosion is a micro-fatigue phenomenon and is generally caused by the breakdown of the protection film at the spot of impingement of entrained air bubbles on suspended particles in the liquid or turbulent flow of liquid substances and gases [1]. The positive effect of shot peening on erosion resistance of the magnesium alloy AZ80 was clearly observed during the experiment on an erosion testing machine. On average, 33 to 45 percent reduction in erosion was found in shot peened as compared to non-peened specimens. Shot peening was also beneficial in fatigue even under combined bending and torsion. The improvement in fatigue life of AZ80 was found to be more than 70 times due to shot peening with 1 mm cast steel at an intensity of 0.28 mmN at 200 percent coverage.

2

Experimental

The magnesium alloy AZ80 with the chemical composition (in wt. %): 8.5Al, 0.5Zn, 0.2Mn, balance Mg was used as work piece material. Mechanical properties are listed in Table 1 [2]. Table 1: Tensile properties of AZ80 E (GPa)

I0.2 (MPa)

UTS (MPa)

El (%)

HV5

46

195

315

15.7

69

Details of the erosion set-up as designed and fabricated by an earlier investigator (3) are shown in Figure 1. The testing machine can hold two test pieces of the same size on either side of the rotating arm. The test piece holder is shown as item no.10 in the assembly drawing (Fig. 1). Non-peened and shot peened test pieces were placed side by side on each test piece holder at the same radius from the rotating arm so as to perform the test under identical conditions in the slurry. Only one longitudinal face of the work piece was exposed to the slurry. The slurry was prepared by mixing 125 c.c. silica sand particles of roughly 1 mm average size with 5 liters of water resulting in a sand concentration of 2.5 %. After 9 hours of rotation in the slurry, the test pieces were dried and their weight loss was determined.

138

Figure 1: Erosion testing machine

R 2.5

5

40

Figure 2: Geometry of test pieces for erosion testing, exposed area = 195 mm2.

Details of the test pieces machined from the magnesium alloy AZ80 are shown in Figure 2. Machining was done along the length of the test piece which was taken in rolling direction. However, when test pieces were rotated in the agitated slurry, the direction of rotation was perpendicular to the length of the test pieces. The test piece holder was rotated at 1440 r.p.m. using 1.5 H.P. motor for 9 hours while holding the non-peened and shot peened test pieces on each side of the holder. Shot peening was performed by means of a syphonic system using 1 mm cast steel shot. The nozzle bore was 8 mm and the stand-off distance 30 mm. Surface roughness was measured by a profilometer. For fatigue testing, a combined bending and torsional testing machine based on a slider crank principle was used. Test frequency was 150 oscillations per minute. The deflection was 5 mm on either side of the neutral axis while twist was 5 degrees on either side. A similar combined

139 bending and torsional testing machine had been originally also designed by H. O. Fuchs at the Stanford University [4]. However, that design was based on a four bar principle. The geometry of the fatigue specimens used in the present study is illustrated in Figure 3.

R30

30 Figure 3: Geometry of the fatigue specimens (8 mm thick)

The effect of shot peening on the fatigue performance of AZ80 in combined bending and torsion was investigated by comparing non-peened specimens with those peened to an Almen intensity of 0.28 mmN.

3

Results and Discussion

The effect of peening pressure and exposure time on the arc height readings of standard Almen strips N and of strips made of AZ80 are listed in Table 2. Table 2: Shot peening parameters

Exposure time (s)

Arc height Pressure: 1 bar

5 10 15 20 30 60 120

Pressure: 2 bar

Almen strip mmN

AZ80 strip 1.3 mm thick

Almen strip mmN

AZ80 strip 1.3 mm thick

0.09 0.12 0.14 0.16 -

0.13 0.20 0.25 0.31 0.40 0.46

0.09 0.12 0.17 0.23 0.28 0.31 -

0.30 0.42 0.51 0.57 0.62 0.70 -

The surface roughness profiles of the various conditions of AZ80 are illustrated in Figure 4. Erosion test results are given in Table 3. The results indicate that the total weight loss due to erosion of shot peened test pieces is significantly (up to 45 %) lower than that of non-peened test pieces. Since erosion is a micro-fatigue phenomenon, this improvement can be correlated with the well known beneficial effect of shot peening on fatigue performance.

140

a) non-peened, # 6

b) shot peened (0.14 mmN), # 5

d) non-peened, # 9

e) shot peened (0.28 mmN), # 10

Figure 4: Surface topography of AZ80 after erosion testing for 9 hours (note that test pieces # 6 and # 5 were coupled during testing as well as # 9 and # 10)

141 Table 3: Erosion test results on AZ80 Test piece number and condition Weight before erosion

Weight after erosion

#5, shot peened (0.14 mmN) #6, non-peened* #10, shot-peened (0.28 mmN) #9, non-peened*

570 mg 40 mg 540 mg 70 mg 540 mg 80 mg 490 mg 100 mg

610 mg 610 mg 610 mg 590 mg

Loss of Weight

erosion in %

Loss of weight per unit area mg/ mm2

6.56 11.48 13.00 17.05

0.205 0.358 0.410 0.512

* Note that test pieces #5 and #6 as well as #9 and #10 were coupled in the experiments.

The fatigue results in combined bending and torsion can be summarized as follows: At the same loading conditions, the non-peened test pieces failed after only 20 – 25 minutes (3000 – 3750 oscillations) while the shot peened test pieces performed splendidly well and ran for more than 24 hours (216000 oscillations) without failure. These tests clearly show that shot peening to an Almen intensity of 0.28 mmN of the magnesium alloy AZ80 using 1 mm cast steel shot is very beneficial to the materials performance in combined bending and torsion.

4

Summary

It is evident from both experiments on erosion resistance and fatigue performance in combined bending and torsion that shot peening of the magnesium alloy AZ80 gave very encouraging results even at intensities higher than 0.05 mmN reported as an optimum intensity by other investigators. Presumably, this is due to the different loading conditions used being rotating beam loading in [2] and combined bending and torsion in the present study.

5

Acknowledgements

The authors are thankful to Prof. Dr. S. Purohit and co-workers of GSITS, Indore, India for providing the erosion testing set-up. We are also thankful to Prof. L. Wagner of the Technical University of Brandenburg at Cottbus, Germany for providing the magnesium alloy AZ80 used in this investigation.

6 [1] [2] [3] [4]

References K. J. Pascoe: An Introduction to the Properties of Engineering Materials, Pergamon Press London L. Wagner and M. Hilpert: Shot Peening and Blast Cleaning (Ed. M.C. Sharma) MACT (2001) 49. Ph. D. thesis on erosion testing machine under the guidance of Dr. S. Purohit GSITS Indore, India. H. O. Fuchs: Private communication at Stanford University, CA, USA (1984).

1

III Surface Layer Properties

2

145

Characteristics of Surface Layers Produced by Shot Peening Volker Schulze Institut für Werkstoffkunde I, Universität Karlsruhe (TH), Karlsruhe, Germany

1

Introduction

Production processes, especially mechanical surface treatments like shot peening, lead to changes in the materials state close to the surface, which severely affect the success of the treatment, especially the resulting fatigue properties. Formerly these effects on fatigue life were controversely discussed as effects of mechanical workhardening, which first were postulated to be dominating by Föppl and his group [1,2], and effects of compressive residual stresses, which first were assumed to increase the fatigue properties by Thum and his group [3,4]. Additionally, effects of topography on fatigue properties were studied by Houdremont and Mailänder [5] and Siebel and Gaier [6]. Today it is well known that most of the changes of surface characteristics induced by shot peening which are listed in Fig. 1 and the stability of these changes may affect the fatigue properties of components and that these effects can be described in the so called concept of local fatigue properties [7,8,9]. Therefore the influence of process parameters of the shot peening treatments listed in Fig. 2 on the surface characteristics has to be well known. Besides the parameters concerning the peening device or the shot, the parameters concerning the workpiece are of high interest. Especially the workpiece temperature and the prestress are altered in modifications of the peening process named warm peening and stress peening which will be discussed separately. In some cases additional annealing treatments are used to achieve further improvements of the material state close to the surface. In the present paper a systematic overview of up to date knowledge about the changes in the surface state due to shot peening is given by discussing characteristic examples concerning changes of the topography, residual stress state, workhardening state and microstructure of components due to shot peening. A special focus will be drawn to the previously mentioned modifications of the conventional shot peening process like stress peening, warm peening and peening plus subsequent annealing, which show improvements in the surface properties or at least improvements of the stability of the induced surface state. Dislocation density, twin density

Half widths of x-ray interference lines

Hardness Type and distribution of lattice defects

Roughness

Mass flow rate

Coherency lengths and lattice distortions

Peening time

Coverage Nozzle diameter

Impact angle Shifts of x-ray interference lines

Device

Shot velocity

Nozzle clearance

Macro residual stresses

Topography

Deformations Size Shape (distortion)

Shot peening treatment

Phase fractions

Shot peening

Relations of intensities of x-ray interference lines

Geometry

Shape

Elastic-plastic deformation behaviour

Hardness

Curvature Texture

Density

Pole figures

Mass

Workpiece

Shot

Hardness

Size Porosity

Cracks Crack length

Orientation distribution functions Crack density

Crack depth

Figure 1: Properties of the workpiece influenced by shot peening and measures for them [acc.51]

Material Size distribution

Wear state

Chemical composition Prestress

Crystal structure Temperature

Figure 2: Parameters influencing the results of shot peening treatments

146

2

Surface Characteristics

2.1

Topography

In [10], an overview of roughness values after different shot peening treatments of the steel AISI 4140 is given. Fig. 3 shows that increasing peening pressures lead to growing values of roughness. While they increase only slightly at the quenched state, these increases are the more pronounced the lower the workpiece hardness is. Increasing size and hardness of the shot also cause higher values of surface roughness [10]. Even at values of shot hardness which are lower than that of the workpiece distinct effects on topography could be found [10,11]. These effects are caused first by increasing strengths and therefore decreasing deformations if the workpiece hardness is increased, second by increasing kinetic energy of the shot if the shot size or the peening pressure are increased and third by decreasing tendency to plastic deformations of the shot if the shot hardness is increased. At high mass flow rates and peening pressures especially at ductile materials states the roughness may decrease due to effects of mutual impacts of shots in the nozzle and in the supplies which lead to decreasing kinetic energy [10]. While stress peening causes no significant changes in the effects on topography besides elliptic impact craters found by [12], warm peening slightly increases roughness with increasing peening temperature due to decreasing warm strength [13]. 60

normalized q&t650° q&t450° q&t300°

roughness [mm]

50

q&t180° quenched

AISI 4140

40

30

20

10

0 1

2

3

4

5

6

7

8

peening pressure [bar]

Figure 3: Influence of the workpiece state on roughness vs. peening pressure at AISI 4140 (p = 1.6 bar, m = 1.5 kg/min) [10]

2.2

Residual Stresses

Fig. 4 shows typical residual stress distributions in differently heat treated and shot peened steel AISI 4140 [10]. With increasing hardness of the steel the surface residual stresses increase from the normalized state up to the state quenched and tempered at 450 °C. Further increases of the workpiece hardness lead to surface residual stress decreases of up to 100 MPa. The depth of the zone bearing compressive residual stresses decreases with increasing workpiece hardness. Only the state quenched and tempered at 180 °C shows a lower depth than the quenched state. A maximum value of compressive residual stresses below surface is characteristic for high strength

147 material states and occurs at depths of up to 0.05 mm with values of up to 800 MPa. Fig. 5a shows the same results again as the dependence of characteristics of the residual stress distribution on the workpiece hardness. According to [14] the results found are due to the concurrent processes of plastic deformation during shot peening. At low workpiece hardness the plastic stretching of regions directly at the surface is dominating and the compressive residual stresses show their maximum directly at the surface. Only at extreme peening conditions, a maximum of the residual stresses below the surface may occur. At medium workpiece hardness plastic stretching of regions directly at the surface and Hertzian pressure are almost equivalent so that high residual stresses at the surface and a distinct maximum of residual stresses below the surface are combined. At high workpiece hardness the Hertzian pressure is dominating so that relatively low compressive residual stresses at the surface are combined with a very clear maximum of residual stresses below the surface. In accordance with [15] at the material state with the lowest tempering temperature higher compressive residual stresses than at the quenched state were found. This is typical for low shot hardnesses [16] and due to increases of strength in that tempering stage caused by the formation of energetically more favorable dislocation structures and carbon clusters which pin the dislocations [see e.g. 17]. 400

AISI 4140

residual stresses [MPa]

200

0

normalized q&t650° q&t450° q&t300° q&t180° quenched

-200

-400

-600

-800

-1000 0,0

0,1

0,2

0,3

0,4

Figure 4: Influence of workpiece state on the residual stress distribution at AISI 4140p = 1.6 bar, m = 1.5 kg/min)[10]

In Fig. 5b the characteristics of the residual stress state of quenched AISI 4140 after shot peening are drawn for different shot sizes and hardnesses. The maximum and the surface values of compressive residual stresses are determined more by the hardness than by the size of the shot. In contrast to this, the depths of compressive residual stresses increase with increasing shot size and are almost not affected by changes of shot hardness. This can also be described by the model of [14] which indicates only slight changes of the positions of the maximum residual stresses induced by small changes of the size of the contact zone due to variation of the shot size at high workpiece hardness. Even the influence of shot hardness on the surface and maximum residual stress values is expected by the model in the same manner as found experimentally. According to Fig. 5c increasing peening pressures which determine the shot velocity in air blasting devices lead to no significant changes of the surface and maximum values of the residual stresses but increase the depth of compressive residual stresses and – at low pressures - the position of the maximum of the residual stresses. At high pressure the position of the maximum decrea-

148 S110

0,3

46 HRC 56 HRC

0,2 600 rs

|I surf |

0,1

400 0,0

residual stresses [MPa]

0,3 rs

|I max |

z [mm]

residual stresses [MPa]

z0

800

S330

S170

1200

0,4

AISI 4140

rs

|I max|

1000 0,2

z0

z [mm]

1000

800 ES

|I surf| 0,1 600

zmax

zmax

AISI 4140, quenched 200

300

400

500

600

0,2

700

workpiece hardness [HV]

(a)

0,0

400

-0,1

200

0,4

0,3

0,8

1,0

shot diameter [mm]

(b)

1200

0,6

0,3

800

AISI 4140, quenched

z0

rs

800 rs

|I surf|

0,1

600

zmax

0,2

700 rs

|I max | 0,1

600

z [mm]

0,2

z0

residual stresses [MPa]

1000

z [mm]

residual stresses [MPa]

|I max|

rs

|I surf | 0,0

500

zmax AISI 4140, q&t450°

0,0

400 1

(c)

2

3

4

5

6

peening pressure [bar]

7

8

-0,1

400

9

0

(d)

2

4

6

8

10

12

mass flow rate [kg/min]

Figure 5: Influence of workpiece hardness (a), shot size and hardness (b), peening pressure (c) and mass flow rate (d) on the characteristics of the residual stress distribution at AISI 4140 [acc. 10]

ses again. This may be an effect of saturation, which could be found also for the depth of compressive residual stresses in the normalized state for peening pressures higher than 5 bar. Fig. 5d shows for AISI 4140 quenched and tempered at 450 °C that with increasing media flow rate first the surface residual stresses increase to a value similar to the maximum value, so that the plateau of residual stresses typical for quenched and tempered steels forms. This plateau value seems to be not affected by the flow rate. According to mutual impacts of the particles - as described before at the effects on topography - the depth of compressive residual stresses decreases with growing flow rate. This effect could be proofed at a normalized state of AISI 4140, where coordinated changes of flow rate and peening time lead to a lower depth of the compressive residual stresses in the case of the higher flow rate at the same coverages [10]. According to [18], the impact angle of the shot leads to decreasing depths of compressive residual stresses if the component of the velocity normal to surface decreases. The residual stress state is no more rotational symmetric and shows decreasing compressive residual stresses in the direction with a tangential component of velocity. Perpendicular to this, the changes are not that pronounced because the plastic deformation in this this direction is not that affected by the angle of impact. These effects could be found also in finite-elemente-simulations of the shot peening process by [19] presented in these proceedings. Fig. 6 shows measurements of macro residual stresses and phase specific micro residual stresses at the surface of iron-copper-titaniumcarbide sintered alloys in the shot peened state with different iron or copper contents [20]. While the macroscopic residual stresses are in the compressive region as expected and decrease with increasing copper content, the micro residual stresses are in the compressive region only for the two metallic phases and are equilibrated by tensile micro residual stresses in the TiC-phase. Therefore, at least at the copper content of

149 about 10 % the sum of the two residual stresses in the TiC-phase is positive and may promote decohesion of the interfaces and therefore crack initiation at tensile loadings of the material. This means that in bi- or multiphase materials the effect of shot peening on the residual stress state has to be observed precisely as it can be detrimental to the fatigue properties.

Figure 6: Macro and micro residual stresses in sintered Fe-Cu-TiC sintered alloys [20]

At stress peening, the changes of the residual stress state with increasing prestress severely depend on the type of prestress, which may be uni- or bidirectional and homogeneous or inhomogeneous [21-28]. In general, the residual stress component in the direction of a positive prestress will increase, the residual stress component perpendicular to it will decrease and the depth of compressive residual stresses will increase due to stress peening. Fig. 7 shows this using the characteristics of the residual stress distributions for quenched and tempered AISI 4140 after stress peening with different torsional prestresses, which were tensile in the 45°- and compressive in the 135°-direction [28]. Only at the highest prestress tpre / te = 0.8 the residual stresses in 45°-direction decrease due to reverse plastifications during unloading. The effects of stress peening are the more pronounced the higher the strength of the workpiece is [29]. According to a model of [21-23] the residual stress changes are proportional to the prestress because the stress state directly after peening but before unloading is assumed to be independent of the prestress. During unloading from the prestress the stress state is shifted in the same way for the prestressed stress component and eventually in the opposite direction in the transverse direction. This model is valid up to prestresses of about half the yield stress as afterwards the residual stresses change only degressively due to the reverse plastifications during unloading mentioned before. The type of the prestress directly affects the changes of the two stress components on the prestress value. While at tensile or bending prestresses, which are uniaxial, the relation of the changes of the residual stresses perpendicular and parallel to the prestress is about one third, this is about two at torsional prestresses, which are biaxial with a positive and a negative stress component. At very high prestresses this may even lead to tensile residual stresses after torsional stress peening [30].

150 1000

0,25

0,24 rs

II surf, 45 I

800

0,22

0,21

400

z0 [mm]

0,23 600

0,20

rs

rs

II surf, 45 I or II surf, 135 I [MPa]

AISI 4140, q&t450°

z0

200

rs

II surf, 135 I

0,0

0,2

0,4

0,6

0,8

0,19

0,18 1,2

0 1,0

Jpre / Je

Figure 7: Influence of torsion prestress related to the yield stress on the characteristics of the residual stress distribution at quenched and tempered AISI 4140 [28]

Other important, but no more than recently upcoming modifications of shot peening are warm peening and stress peening at elevated temperatures. Fig. 8 shows the residual stress distributions in quenched and tempered AISI 4140 after these treatments in comparison to conventional peening [31]. While at warm peening the residual stresses in AISI 4140 initially increase only slightly with growing temperature and decrease above 330 °C due to thermal residual stress relaxation, [32,33] found clear increases of the compressive residual stresses at the surface in a spring steel after warm peening. The residual stresses and the depth of compressive residual stresses after stress peening at elevated temperatures are increased again compared to stress peening and therefore can be explained by the superposition of the influences of prestress and peening temperature [31].

longitudinal residual stresses [MPa]

400

AISI 4140, q&t450°

200 0 -200 -400 -600

Tpeen = 20°C Tpeen = 290°C Ipre =

-800

0 MPa

Ipre = 500 MPa

-1000 0,0

0,1

0,2

0,3

0,4

0,5

distance to surface [mm]

Figure 8: Residual stress distributions at quenched and tempered AISI 4140 after different modified shot peening treatments [27]

The last possible modifications of the shot peening process which should be mentioned in this paper are shot peening plus subsequent annealing treatments, which can be used to optimize the microstructure at the surface of steels and Al- as well as Ti-alloys. While at steels static strain aging effects are intended and reached at relatively low temperatures, which cause only

151 slight relaxation of the residual stresses [34], at Al- and Ti-alloys the temperatures necessary to cause recrystallization or age hardening in the surface region are that high, that they lead to a severe relaxation of the residual stresses. This is shown exemplarily in Fig. 9 for the b-Ti-alloy Timetal 21s [35]. 1.0

Tim etal 21s-PR 350°C

0.8

400°C 450°C

/ I0

rs

0.6

I

rs

500°C 0.4

0.2

0.0 1

10

100

1000

10000

aging tim e [m in]

Figure 9: Influence of aging temperature and time on the relaxation of surface residual stresses at the shot peened >-Ti-alloy Timetal 21S [35]

2.3

Workhardening State

The changes of the workhardening state at the surface are also very important for the mechanical properties of shot peened material states. Fig. 10 shows the influence of the heat treatment state of shot peened steel AISI 4140 on the distributions of the half widths of x-ray interference lines which are a measure of the micro residual stresses. While they increase at the normalized state and the state tempered at 650 °C in the region affected by the shot peening treatment due to workhardening effects and show their highest values directly at the surface, they are not changed significantly at the state tempered at 450 °C. In contrast to this, the states with the highest initial half width caused by high dislocation densities show decreasing half widths beneath the surface which increase slightly again directly at the surface. In contrast to this, the microhardness shown in Fig. 11 increases towards the surface not only at the soft materials states but also at the high strength states. As the half widths are a direct measure of the lattice distortions and therefore of the micro residual stresses induced by the dislocations [36] their increase at soft material states is due to increases of the dislocation density. In contrast to this, at hard material states decreases of the dislocation density occur as the initial dislocation density is very high and leads to dislocation rearrangements into energetically more favorable structures and to dislocation annihilations. Additionally, dissolved carbon atoms diffuse supported by the stored mechanical energy to octahedral sites where they cause lower lattice distortions due to the so called Snoek-effect [37] and decreasing half widths as found also by [36]. In contrast to this, microhardness measurements need plastic deformations and therefore are affected by the residual stress state in a way that compressive residual stresses lead to apparently increased hardness values [38]. Therefore, the microhardness is increased at hard materials states though the changes in the dislocation structure cause microstructural worksoftening. Changes of peening parameters like peening pressure, mass flow rate and velocity lead primarily to changes of the

152 depth where the workhardening state of the bulk material is reached and not to severe changes in the surface value of half width or microhardness. As the measurement of this depth is very difficult, there are no statements known which are different from those already given when discussing the distributions of the residual stresses. 8

800

AISI 4140

AISI 4140 600

hardness [HV0.3]

half widths [°23]

6

4

2

quenched

q&t300°

q&t650°

q&t180°

q&t450°

normalized

0

400

200

0

0,0

0,1

0,2

0,3

distance to surface [mm]

Fig. 10: Influence of the workpiece state on the half width distribution at AISI 4140 (S170 46 HRC, p = 1.6 bar m = 1.5 kg/min) [10]

0,4

0.0

quenched

q&t650°

q&t450°

norm alized

0.2

0.4

0.6

distance to surface [m m ]

Fig. 11: Influence of the workpiece state on the microhardness distribution at AISI 4140 (S170 46 HRC, p = 1.6 bar, m = 1.5 kg/min) [10]

While stress peening does not change the workhardening state significantly compared to conventional peening, after warm peening of steels increased half widths could be observed, which slightly decrease again at temperatures higher than 310 °C. According to [31] this is due to dynamic and static strain aging effects occurring during the peening process which lead to a lower mobility of the dislocations and therefore to multiplication of dislocations during further deformation. This causes increased values of the half widths. At very high temperatures thermal residual stress relaxation and dislocation rearrangements into energetically more favorable structures lead to decreasing half widths. After shot peening at different temperatures the microhardness of a high strength spring steel was investigated by [39-41]. As given in Fig. 12 it is increased at peening temperatures of 300 °C and above compared to conventionally peened samples. This effect is the higher, the higher the bulk hardness or the lower the tempering temperature is, resp., because the dissolved carbon content increases then. This is again a hint for the dynamic and static strain aging effects occuring during warm peening, which lead to diffuse dislocation structures pinned by clouds of carbon atoms or finest carbides [42,43]. Stress peening at elevated temperatures leads to surface half widths similar to those found in warm peened states but increased in larger regions below the surface. Conventional peening and subsequent annealing of steels leads also to static strain aging effects and therefore dislocation structures pinned by carbon atom clouds and finest carbides [34]. Similar to warm peened states this leads to increasing microhardness values in the surface region [44]. At light metal alloys based on Titanium or Aluminum annealing treatments after shot peening can also be used to increase the hardness in the surface region. Fig. 13 shows that in the b-Ti-alloy Timetal 21S annealing treatments after shot peening induce hardness increases with increasing annealing time due to age hardening [35]. These are the more rapid the higher the an-

153 600

spring steel

hardness [HV]

550

500

RT 200°C 350°C

450

100°C 300°C

400 0,0

0,1

0,2

0,3

0,4

distance to surface [mm]

Figure 12: Influence of the peening temperature at the depth distribution of the microhardness in a spring steel

nealing temperature is and may be decreasing again at the highest annealing temperatures, if the annealing time is too large and overaging occurs. 250

450°C 500°C

cahnge in hardness [HV0.05]

350°C 400°C

Timetal 21s-PR

200

150

100

50

0 0

10

100

1000

aging time [min]

Figure 13: Influence of aging temperature and time on the increase of microhardness at the shot peened >-Ti-alloy Timetal 21S [35]

2.4

Microstructure

The microstructure is altered due to shot peening in thin surface layers, which are difficult to analyze microscopically. Therefore only few investigations exist, which in some cases use special preparation techniques like the so called cross sectional preparation to get informations about the microstructure in regions very close to the surface [45,46]. Using conventional transmission electron microscopic (TEM) preparation techniques, [47] studied the changes in the dislocation structure in normalized AISI 4140 due to shot peening given in Fig. 14. The dislocation structure, which is initially very low, is changed to a uniformly distributed structure with severely increased dislocation density. This is typical for deformations at high strain rate, which may reach up to 5·10+3 1/s during shot peening [47]. Similar results were obtained by [43,48],

154 who determined depth profiles of the dislocation structure found in TEM-studies and found tangled dislocations with varying densities of up to 8·10+11 1/cm2. Another hint for high strain rate deformations was found by [49] who reports about twinning during TEM-studies of the Mg-alloy AZ31. [50] shows for normalized SAE1045 that the deformation is not restricted to the ferritic regions but affects also the pearlitic regions, where high densities of statistically distributed dislocations are observed between the cementite lamellae. For aluminum-alloys it could be shown by [51] that in AlMg5 the dislocations are uniformly distributed with increased dislocation density and only few dislocation tangles, whereas in AlMg1 the dislocations form sharp cells due to the stacking fault energy, which is three times higher than in AlMg5 and therefore allows much easier cross slip of screw dislocations.

AISI 4140, normalized

unpeeed

shotpeened

Figure 14: Dislocation structure close to the surface of unpeened and peened normalized AISI 4140 (S170, 54-58HRC, p = 1.0 bar, 1-fold coverage) [47]

In the surface region also other changes in the microstructure due to shot peening are observed. While [51] found a density of 7.72 g/cm3 close to the surface in sintered iron with a bulk density of 7.37 g/cm3, [52] describes decreases of the local porosity in unalloyed sintered steels due to ultrasonic peening, which are present in the larger depths, the higher the initial porosity is. Phase changes due to shot peening are well known especially in previously case hardened steels [53-57], hardened high carbon tool steels [58,59] and metastabile austenitic steels [5962], where austenite transforms to martensite. [63] observed transformations of retained austenite in the tool steel AISI D3 (German grade: X210Cr12) shown in Fig. 15 which bears the higher retained austenite after quenching, the higher the austenitizing temperature was. Due to shot peening the smaller transformations related to the initial value occur the higher the austenitizing temperature was. This is caused by increasing chemical stabilization of the austenite. Close to the surface the transformation to martensite is decreased significantly at the states with high retained austenite contents, so that the maximum of transformation is shifted to higher distances from surface with increasing initial retained austenite content. This may be explained by local temperature increases at the region close to the surface, which are increasing with decreasing di-

155

content of retained austenite [vol-%]

stance to surface, as they are caused by local deformations which increase in that direction. This causes temperatures close to the surface higher than the critical temperature MD, above which no martensite formation is possible.

AISI D3 shot peened

80

TA = 1100°C 60

TA = 1060°C 40

TA = 1030°C 20

TA = 940°C 0 0,0

0,1

0,2

0,3

0,4

0,5

distance to surface [mm]

Figure 15: Influence of austenitizing temperature of the tool steel AISI D3 on the depth distribution of retained austenite after shot peening [63]

In TEM-studies at the austenitic steel AISI 304 [43,48] observed strain induced martensite with contents of up to 40 %. Additionally they found a two phase nanocrystalline surface layer with a thickness of about 1 – 2 μm consisting of austenite and martensite, which has grain sizes down to 20 nm directly at the surface. This layer is a hint for static recrystallization during shot peening. Similar effects were found by [64] after ultrasonic peening of AISI 316L, by [65] at airon and by [66-69] at Nickel-base alloys. The severe plastic deformations during shot peening lead to changes in the grain orientation so that characteristic peening textures occur in states initially free of textures or preexisting textures may be changed by shot peening [see e.g. 70]. Annealing treatments after shot peening of aluminum and titanium alloys, used as modifications of the shot peening process, may lead to changes of the microstructure, which can be separated into four different mechanisms. In a-Ti-alloys like Ti-8Al, e.g., local recrystallization causes fine grained zones close to the surface because of the high nucleation rate for new grains in the deformed surface regions [71]. In (a+b)-Ti-alloys like Ti-6Al-4V, e.g., local recrystallization may change the fine grained lamellar structure of the bulk material to an equiaxed fine grained surface region [72]. In b-Ti-alloys like BetaC or Timetal 21S selective surface aging can be obtained by local precipitation of semi- or incoherent a-particles. This is possible due to the high dislocation density and therefore increased density of nuclei which decrease the precipitation times in the TTT-diagram and allow aging treatments which do not affect the bulk material but cause fine precipitations in the surface region. As they are finer than obtained by an aging treatment of the full material, the microhardness in the surface regions is increased to much higher values [35]. At age hardening Al-alloys like 2017 or 2024, e.g., preferred surface aging treatments can be performed, because shot peening leads also to increased dislocation densities and therefore increased densities of the nuclei for precipitation. In contrast to the a-Ti-alloys, in these materials the precipitation in the bulk material cannot be suppressed, so that in the surface

156 region only a higher number of smaller precipitates occurs, leading to increased hardness values in those areas [73].

3

Concluding Remarks

It is shown that shot peening causes multiple changes in the state of the material close to the surface of components. These changes which cannot be adjusted independently but can be separated into the categories topography, residual stress state, workhardening state and microstructure. It could be shown that there are numerous process parameters, whose influence on the state of the material close to the surface can only be understood because there are sufficiently systematic investigations in that area. The interactions of the surface state with the mechanical properties especially at cyclic loading are of particular interest. In this context, the stability of the changes in the surface state induced by shot peening, especially the residual stress state, is very important [31]. Though both aspects are not directly subject of this paper they lead to modifications of the shot peening treatments which get more and more important in technical practice and therefore are included in the paper. One of these is stress peening, which yields higher and deeper residual stress states and consequently leads to higher fatigue limits if the material strength is high enough to impede the relaxation of these increased residual stresses [21-28]. Another one is warm peening, which leads to dynamic and static strain aging effects, therefore to increased microhardness values, stabilized residual stresses and severely increased fatigue properties at steels [31,42,74]. At other materials which are able to strain aging similar effects should be expected. Combinations of stress and warm peening may use the advantages of both types of modifications but lead to less pronounced effects than would be necessary to vindicate the efforts for their technical application. Last but not least annealing treatments subsequent to conventional shot peening may be of benefit at steels and some light weight alloys. While they may cause static stain aging effects in steels and lead to slightly weakened increases of the fatigue behaviour compared to warm peening, they may be utilized for several effects in light weight alloys. This is first the formation of fine grained surface regions by recrystallization, which leads to optimized fatigue properties, because fine grained surface states may impede crack initiation and the coarse grained bulk material shows relatively low crack propagation rates [71]. Second, changes of grain morphology due to recrystallization can be obtained. This may lead to optimized properties at combined creep-fatigue loadings because creep resistant lamellar bulk materials can be combined with fatigue resistant fine and equiaxially grained surface, e.g. [72]. Third, selective or preferred surface aging can be performed causing increased hardness values at the surface region and fatigue limits enhanced compared to the states aged before shot peening [35,73]. All of these modified shot peening treatments are a fine example for the feedbacks of the systematic studies on the correlations between process parameters and surface characteristics mentioned before.

157

4 [1] [2] [3] [4]

[5] [6] [7]

[8]

[9]

[10] [11] [12] [13]

[14]

[15] [16] [17]

[18] [19]

References O. Föppl, Stahl und Eisen, 49 (1929), 775. O. Föppl, G. v. Heydekampf, Dauerfestigkeit und Konstruktion, Metallwirtschaft Wissenschaft und Technik 45 (1929), 1087–1094. A. Thum, H. Wiegand, Die Dauerhaltbarkeit von Schraubenverbindungen und Mittel zu ihrer Steigerung, VDI Z 39 (1933), 1061–1063. H. Oschatz, Gesetzmäßigkeiten des Dauerbruches und Wege zur Steigerung der Dauerhaltbarkeit, Mitteilungen der Materialprüfungsanstalt an der Technischen Hochschule Darmstadt, 1933. E. Houdremont, R. Mailänder, Archiv Eisenhüttenwesen 49 (1929), 833. E. Siebel, M. Gaier, Untersuchungen über den Einfluß der Oberflächenbeschaffenheit auf die Dauerschwingfestigkeit metallischer Bauteile, VDI Z 30 (1956), 1715–1723. E. Macherauch, H. Wohlfahrt, Eigenspannungen und Ermüdung, In: D. Munz (ed.) Ermüdungsverhalten metallischer Werkstoffe, DGM-Informationsgesellschaft, Oberursel, 1985, 237–283. P. Starker, H. Wohlfahrt, E. Macherauch, Surface crack initiation during fatigue as a result of residual stresses, Fatigue and Fracture of Engineering Materials & Structures 1 (1979), 319–327. D. Löhe, K.-H. Lang, O. Vöhringer, Residual stresses and fatigue behavior, In: G. Totten, M. Howes, T. Inoue (eds.), Handbook of residual stress and deformation of steel, ASM International, 2002, 27–53. H. Holzapfel, A. Wick, V. Schulze, O. Vöhringer, Einfluß der Kugelstrahlparameter auf die Randschichteigenschaften von 42 CrMo 4, Härterei-Techn. Mitt.53 (1998), 155–163. P. Martin, Dr.-Ing. Thesis, Universität Hamburg, 1980. R. Kopp, In: H. O. Fuchs (ed.) Proc. ICSP 2, Paramus, N.J., 1984. U. Martin, B. Scholtes, K. Kremmer, H. Oettel, Characterization of the defect depth profile of shot peened steels by transmission electron microscopy, In: J. Champaigne (ed.) Proc. ICSP 6, San Francisco, 1996, 142–153. 1996. H. Wohlfahrt, Ein Modell zur Vorhersage kugelstrahlbedingter Eigenspannungszustände, In: E. Macherauch, V. Hauk (eds.) Eigenspannungen: Enstehung-Messung-Bewertung, DGM-Informationsgesellschaft, Oberursel, 1983, 301–319. H. Wohlfahrt, Residual stress and stress relaxation, In: E. Kula, V. Weiss (eds.) Proc. Sagamore Army Mat. Res. Conference, 1982, 71. B. Scholtes, Eigenspannungen in mechanisch randschichtverformten Werkstoffzuständen, Ursachen-Ermittlung-Bewertung. DGM-Informationsgesellschaft, Oberursel, 1990. B. Hoffmann, O. Vöhringer, E. Macherauch, Effect of compressive plastic deformation on mean lattice strains, dislocation densities and flow stresses of martensitically hardened steels, Mat.Sci. & Eng. A319–321(2001), 299–303. A. Ebenau, O. Vöhringer, E. Macherauch, In: H. Wohlfahrt, R. Kopp, O. Vöhringer (eds.) Prof. ICSP 3, DGM-Informationsgesellschaft, Oberursel, 1987, 253., J. Schwarzer, V. Schulze, O. Vöhringer, Finite element simulation of shot peening - A Method to evaluate the influence of peening parameters on surface characteristics, In: L. Wagner (ed.) Proc. ICSP 8, Wiley-VCH, Weinheim, 2002.

158 [20] V. Hauk, B. Krüger, Residual stresses in Cu-FE-TiC sintered materials after grinding and after shot peening, Z. Metallkde 93(2002), 76–80 [21] R. Zeller, Kugelstrahlen unter Zugvorspannung, Materialprüfung 35 (1993), 218–221. [22] L. Bonus, Auswirkungen des Spannungsstrahlens auf die Eigenschaften von hoch vergüteten Bremsspeicher- und Torsionsfedern, Dr.-Ing. Thesis, RWTH Aachen, 1994. [23] J. F. Flavenot, A. Niku-Lari, La Mesure des Contraintes Résiduelles: Methode de la Flêche - Methode de la Source des Contraintes, Les mémoires techniques du CETIM, 1977, 31. [24] J. C. Xu, D. Zhang, B. Shen, In: A. Niku-Lari (ed.) Proc. ICSP 1, Oxford, 1982, 367. [25] E. Müller, Die Ausbildung von Eigenspannungen an Torsionsproben beim Spannungsstrahlen, Mat.wiss. u. Werkstofftechnik 27 (1996), 354–358. [26] F. Engelmohr, B. Fiedler, Erhöhung der Dauerfestigkeit geschmiedeter Pleuel durch Kugelstrahlen unter Vorspannung, Mat.wiss. u. Werkstofftechnik 22 (1991), 211–216. [27] A. Wick, V. Schulze, O. Vöhringer, Effects of stress- and/or warm peening of AISI 4140 on fatigue life, Steel Research 71(2000), 316–321. [28] V. Schulze, Warm- und Spannungsstrahlen - Wege zur Erzeugung stabilisierter und erhöhter Druckeigenspannungen, Proc. DVM-Tag "Federn im Fahrzeugbau", DVM, Berlin, 2002, 45–54. [29] A. Wick, Randschichtzustand und Schwingfestigkeit von 42 CrMo 4 nach Kugelstrahlen unter Vorspannung und bei erhöhter Temperatur, Dr.-Ing. Thesis, Universität Karlsruhe (TH), 1999. [30] J. Krobb, H. Weiß, Stress peening of ultra high strength steels, In: T.S. Sudarshan, M. Jeandin, K. A. Khor (eds.) Proc. SMT XI, London, 1997, 423–429. [31] A. Wick, V. Schulze, O. Vöhringer, Effects of warm peening on fatigue life and relaxation behaviour of residual stresses of AISI 4140, Mat. Sci. & Eng. A 293 (2000), 191–197. [32] M. Schilling-Praetzel, F. Hegemann, P. Gome, G. Gottstein, Influence of Temperature of Shot Peening on Fatigue Life, In: D. Kirk (ed.) Proc. ICSP 5, 1993, 227–238. [33] M. Schilling-Praetzel, Einfluß der Werkstücktemperatur beim Kugelstrahlen auf die Schwingfestigkeit von Drehstabfedern, Dr.-Ing. Thesis, RWTH Aachen, 1995. [34] R. Menig, V. Schulze, O. Vöhringer, Effects of static strain aging on residual stress stability and alternating bending strength of shot peened AISI 4140, Z. Metallkde 93 (2002), 635–640. [35] M.-C. Berger, J. K. Gregory, Selective hardening and residual stress relaxation in shot peened Timetal 21S, In: C.A. Brebbia, J.M. Kenny, (eds.) Proc. Surface Treatment IV, 1999. [36] F. Burgahn, O. Vöhringer, E. Macherauch, Mikroeigenspannungszustände kugelgestrahlter Randschichten von 42 CrMo 4, Z. Metallkde 84 (1984), 224–229. [37] J. L. Snoek, Effect of small quantities of carbon and nitrogen on the elastic and plastic properties of iron, Physica 8 (1941), 711–733. [38] H. H. Racké, T. Fett, Die Bestimmung biaxialer Eigenspannungen in Kunststoff-Oberflächen durch Knoop-Härtemessung, Mat.prüfg. 13(1971), 37–42. [39] A. Tange, H. Koyama, H. Tsuji, Study on warm shot peening for suspension coil spring, SAE Technical Paper Series 1999-01-0415, 1999, 1–5. [40] A. Tange, H. Koyama, H. Tsuji, The effect of warm shot peening on the fatigue strength of springs, Springs, 2000, 58–63.

159 [41] Y. Harada, K. Mori, Y. Fukuoka, S. Maki, Effect of processing temperature on warm shot peening of spring steel, J. Jap. Soc. Technology and Plasticity 40 (2000), 260–264. [42] R. Menig, V. Schulze, O. Vöhringer, Optimized warm peening of the quenched and tempered steel AISI 4140, Mat. Sci. & Eng. A 335 (2002), 198–206. [43] I. Altenberger, Mikrostrukturelle Untersuchungen mechanisch randverfestigter Bereiche schwingend beanspruchter metallischer Werkstoffe, Dr.-Ing. Thesis, Universität Gesamthochschule Kassel, 2000. [44] R. Menig, Randschichtzustand, Eigenspannungsstabilität und Schwingfestigkeit von unterschiedlich wärmebehandeltem 42 CrMo 4 nach modifizierten Kugelstrahlbehandlungen, Dr.-Ing. Thesis, Universität Karlsruhe (TH), 2002. [45] U. Martin, I. Altenberger, K. Kremmer, B. Scholtes, H. Oettel, Characterization of the microsturcture-depth profile of shot peened steels, Praktische Metallographie 35 (1998), 327–334. [46] F. Y. Hu H.-J. Klaar, Praktische Metallographie 32 (1995), 603. [47] B. Scholtes, E. Macherauch, Randschichtzustände von normalisiertem und vergütetem 42 CrMo 4 nach konsekutiven Kugelstrahlen- und Festwalzbehandlungen, Mat.wiss. u. Werkstofftechnik 23 (1992), 388–394. [48] U. Martin, I. Altenberger, B. Scholtes, K. Kremmer, H. Oettel, Cyclic deformation and near surface microstructures of normalized shot peened steel SAE 1045, Mat. Sci. & Eng. A 246 (1998), 69–80. [49] I. Altenberger, U. Martin, B. Scholtes, H. Oettel, Near surface microstructures in mechanically surface treated materials and their consequences on cyclic deformation behaviour, In: A. Nakonieczny (ed.) Proc. ICSP 7, Warschau, 1999, 79–87. [50] B. Scholtes, unpublished results, Universität Gesamthochschule Kassel., 1993. [51] B. Scholtes, O. Vöhringer, Ursachen, Ermittlung und Bewertung von Randschichtveränderungen durch Kugelstrahlen, Mat.wiss. und Werkstofftechnik 24 (1993), 421–432. [52] H. Chardin, M. Jeandin, Ultrasonic shot-peening of a low-carbon porous sintered steel, In: J. Lu (ed.) Proc. Mat-Tec 96, Paris, 1996, 195–204. [53] S. Pakrasi, J. Betzold, In: A. Niku-Lari (ed.) Proc. ICSP 1, 1982, 193. [54] D. Kirk, Proc. A. Niku-Lari (ed.) ICSP 1, 1982, 271. [55] M. Kikuchi, H. Ueda, K. Hanai, T. Naito, In: H. O. Fuchs (ed.) Proc. ICSP 2. Paramus, NJ, USA, 1984, 208. [56] S. Hisamatsu, T. Kanazawa, In: H. Wohlfahrt, R. Kopp, O. Vöhringer (eds.) Prof. ICSP 3, DGM-Informationsgesellschaft, Oberursel, 1987, 517. [57] Th. Hirsch, H. Wohlfahrt, E .Macherauch, Fatigue Strength of Case Hardened and Shot Peened Gears, In: H. Wohlfahrt, R. Kopp, O. Vöhringer (eds.) Prof. ICSP 3, DGM-Informationsgesellschaft, Oberursel, 1987, 547–560. [58] O. Vöhringer: Changes in the state of the material by shot peening, In: H. Wohlfahrt, R. Kopp, O. Vöhringer (eds.) Prof. ICSP 3, DGM-Informationsgesellschaft, Oberursel, 1987, 185–204. [59] D. Wehrle, A. Ebenau, unpublished results, Universität Karlsruhe (TH), 1987. [60] H. Wohlfahrt, Festigkeitsstrahlen, In: D. Aurich (ed.) Bauteile 86 - Die Bauteiloberfläche, 1986, 23–35.

160 [61] R. Herzog, W. Zinn, B. Scholtes, H. Wohlfahrt, The significance of almen intensity for the generation of shot peening residual stresses, In: J. Champaigne (ed.) Proc. ICSP 6, San Francisco, 1996, 270–281. [62] D. Kirk, N. J. Payne, Transformation induced in austenitic stainless steels by shot peening, In: A. Nakonieczny (ed.) Proc. ICSP 7, Warschau, 1999, 15–22. [63] A. Ebenau, Verhalten von kugelgestrahltem 42 CrMo 4 im normalisierten und vergüteten Zustand unter einachsig homogener und inhomogener Beanspruchung, Dr.-Ing. Thesis, Universität Karlrsruhe (TH), 1989. [64] G. Liu, J. Lu, K. Lu, Surface nanocrystallization of 316L stainless steel induced by ultrasonic shot peening, Mat. Sci. & Eng. A 286 (2000), 91–95. [65] K. Lu, J. Lu, J. Mater, Mat. Sci & Techn. 15 (1999), 193. [66] M. K. Tufft, Instrumented single particle impact tests using production shot: the role of velocity, incidence angle and shot size on impact response, induced plastic strain and life behavior, In: J. Champaigne (ed.) Proc. ICSP 6, San Francisco, 1996. [67] M. K. Tufft, Development of a fracture mechanics / Treshold behavior model to assess the effects of competing mechanisms induced by shot peening on cyclic life of a nickel-base superalloy, René 88DT, Dr.-Ing. Thesis, University of Dayton, Ohio, 1997 [68] R. Wang, X. Zhang, D. Song, Y. Yin, In: A. Niku-Lari (ed.) Proc. ICSP 1, 1982, 395. [69] P. Li, G. Spur, Berichte aus dem Produktionstechnischen Zentrum, Berlin, 2002. [70] G. Maurer, H. Neff, B. Scholtes, E. Macherauch, Texture and lattice deformation pole figures of machined surfaces, Textures and Microstructures 8 (1988), 639–678. [71] L. Wagner, J. K. Gregory, Thermomechanical surface treatment of titanium alloy, Proc. 2nd ASM Heat Treatment and Surface Engineering Conference, Dortmund, 1993, 1–24. [72] H. Gray, L. Wagner, G. Lütjering, Effect of modified surface layer microsturctures through shot peening and subsequent heat treatment on the elevated temperature fatigue behavior of Titanium-Alloys, In: H. Wohlfahrt, R. Kopp, O. Vöhringer (eds.), Proc. ICSP 3, 1987, 467–475. [73] J. K. Gregory, C. Müller, L. Wagner, Bevorzugte Randschichtaushärtung: Neue Verfahren zur Verbesserung des Dauerschwingverhaltens mechanisch belasteter Bauteile, Metall 47 (1993), 915–919. [74] A. Tange, K. Ando, Study on the shot peening processes of coil spring, In: Proc. ICRS 6, Oxford, IOM Communications, 2000, 897–904.

159

The Influence of the Velocity of a Peening Medium on the Almen Intensities and Residual Stress States of Shot Peened Specimens Wolfgang Zinn*, Jörgen Schulz**, Reiner Kopp**, Berthold Scholtes* *Institute of Materials Technology, University of Kassel, Kassel, Germany **Institute of Metal Forming, Aachen University of Technology, Aachen, Germany

1

Introduction

Up to now, the properties of shot peened components and the reproducibility of the peening process have been assessed by the aid of the Almen test. The bending height of the Almen strip is a measure of the peening intensity, taking all peening parameters into account in an integral way. The Almen test, however, can only be applied offline and it is not unambiguously correlated with the resulting properties like, e. g. residual stress. Therefore, a system for direct measurement of peening medium velocity was used to determine the correlation between peening parameters including the mean shot velocity and the Almen intensity. For that purpose, specimens, made of the German steel grade 42CrMo4 were shot peened under different conditions and velocity distributions as well as resulting residual stress distributions were analyzed.

2

Description of the Measuring Procedures

2.1

Principle of Shot Velocity Measurement

A system of measuring the velocity of shot particles in the particle flow was developed by /1/. Moving particles set off a signal at each of two light barriers in a time interval dt. Theoretically, the shot velocity can be calculated directly from the path it has taken ds and the time measured dt. In the real peening process a large number of shot particles are found simultaneously between the two measuring points, so that series of signals are registered at both light barriers. Evaluating them statistically provides us with an accumulation point at the time interval corresponding to the actual mean velocity of the shot particles. This principle of velocity measurement allows us to make statements also about the quantitative fluctuations around this mean value. The algorithm used in the measuring program subdivides each measuring cycle into a sequence of equidistant time windows. The absolute number of particles registered for each measuring cycle is then inserted in the corresponding time windows. This enables us to match the signals in the individual time windows with individual velocities, which allows us to make statements regarding the distributions of velocities.

2.2

Measuring of the Residual Stress Depth Distributions

Specimens made of 42CrMo4 of the dimension 50 mm × 50 mm × 10 mm in a normalized as well as a quenched and tempered state were shot peened. The quenched and tempered speci-

162 mens were heat treated in such a way, that they exhibited the same hardness of 500 – 520 HV as Almen strips. Residual stresses were measured always in the center of the specimens using standard X-ray diffraction methods. Crka-radiation was used to measure lattice strain distributions at {211}-lattice planes of ferrite or martensite respectively. The area irradiated by the X-ray beam was collimated to a diameter of 1 mm. For the calculation of residual stress values from the measured lattice strain distributions, the elastic constants E=210000 MPa and n=0.285 were used. Depth distributions of residual stresses were determined by successive electrochemical layer removal. For the shot peening tests a conventional NC-controlled air pressure peening machine was used and for the process the following parameters were varied: Table 1: Parameters for shot peening tests materials state:

normalised and quenched and tempered

shot diameter:

S110 and S170

nozzle diameter:

10 mm and 15 mm

nozzle distance:

100 mm and 150 mm

mean shot velocity:

(20), 25, 30, 35, 40, 45 m/sec.

mass flow:

200, 600 g/min. (S110) and 2000, 4000 g/min. (S170)

3

The Influence of the Peening Parameters on the Mean Shot Velocity and the Velocity Distribution

In order to investigate the effect of various peening parameters on shot velocity and the distribution of velocity, we used rounded off wire shot of two different sizes (S110 with Æ 0,3 mm and S170 with Æ 0,4 mm). For each Almen saturation curve we employed 8 to 10 Almen strips of type A (thickness 1,295 mm). Furthermore, the mass flow dm/dt and the velocity of the peening medium vB were varied, and two different nozzle geometries (internal jet diameter 10 mm in the case of Nozzle A and 15 mm in the case of Nozzle B) were used. The influence of the height of the nozzles was also examined in selected series of measurements (hN = 100 or 150 mm).

3.1

Influence of the Peening Parameters on the Mean Shot Velocity

Figure 1 shows the mean shot velocities related to the peening pressures for the shot media S110 (left) and S170 (right) It becomes clear that the shot velocity does not increase linearly relative to the peening pressure, but rather the relation is described by a degressive curve. A greater mass flow at the same pressure brings about a decrease in velocity. The highest velocities were attained with the narrower peening nozzle. Overall the velocities attained for S170 are lower than those for the finer S110.

163

60

Mean shot velocity in m/s

Mean shot velocity in m/s

60

Shot medium: S110

50 40 30

Nozzle A, 200 g/min

20

Nozzle A, 600 g/min

10

Nozzle B, 200 g/min

Nozzle B, 600 g/min

0 0

1

2

3

4

5

6

7

Shot medium: S170

50 40 30

Nozzle A, 2000 g/min

20

Nozzle A, 4000 g/min

10

Nozzle B, 2000 g/min

Nozzle B, 4000 g/min 0 0

8

1

2

3

4

5

6

7

8

Peening pressure in bar

Peening pressure in bar

Figure 1: Mean shot velocity related to peening pressure

3.2

The Influence of the Peening Parameters on the Shot Velocity Distribution

In what follows we have plotted the velocity distributions resulting for various types of peening medium and for mean shot velocities. Figure 2 shows typical standard deviations for Nozzles A and B, in each case for both peening media. The values for the standard deviation vary for the S170 much more than they do for the S110. It is not only that their characteristics depend on the geometry of the nozzle, the size of the shot also plays an important role. Their respective behavior was reproduced for varying mass flows. Accordingly, the reasons for these ways of behaving must have to do with the flow conditions specific to each of the nozzles, and the various peening parameters have a great influence on these. 3

Shot medium: S110

Standard deviation I

Standard deviation I

3 2.5 2 1.5 1

Nozzle A, S110

0.5

Nozzle B, S110

0

2,5 2 1,5 1

Nozzle A, S170

0,5

Nozzle B, S170

Shot medium: S170

0

20

25

30

35

40

Mean shot velocity in m/s

dm/dt = 600 g/min

45

50

15

20

25

30

35

40

Mean shot velocity in m/s

dm /dt = 2000 g/min

Figure 2: The effect of the peening medium and the nozzle geometry on the standard deviation of shot velocities

4

Influence of the Shot Velocity on the Shot Peening Results

4.1

Influence of the Mean Shot Velocity and the Velocity Distribution on the Almen Intensity

It is generally supposed that there is a linear relation between the mean shot velocity and the Almen intensity. In the investigations presented here, deviations from the linear progression can also be seen, as is shown in Figure 3 for the peening media S110 and S170. By way of example we have illustrated the correlations at a nozzle height of 100 mm. Qualitatively the result will be similar at a nozzle height of 150 mm.

164 The highest Almen intensities are attained by Nozzle B, although the kinetic energy of the shot deployed is the same for both nozzles because of the identical mean velocities and mass flows. Accordingly, the Almen intensities depend greatly on the geometry of the nozzle. The intensities are respectively higher if the mass flow is greater. Since quite radically different Almen intensities were ascertained for different mass flows, the reasons for this must have to do with the hit effects of the shot, given that the masses deployed per area mA remain mathematically speaking the same. 0,4 Nozzle A, 200 g/min

Almen intensity in mmA

Almen intensity in mmA

0,4 0,35

Nozzle A, 600 g/min

0,3

Nozzle B, 200 g/min

0,25

Nozzle B, 600 g/min

0,2 0,15 0,1

Shot Medium: S110

0,05 0

Nozzle A, 2000 g/min

0,35

Nozzle A, 4000 g/min

0,3

Nozzle B, 2000 g/min

0,25

Nozzle B, 4000 g/min

0,2 0,15 0,1

Shot Medium: S170

0,05 0

0

10

20

30

Mean shot velocity in m/s

40

50

0

10

20 30 Mean shot velocity in m/s

40

50

Figure 3: Almen intensities related to the mean velocity for the peening media S110 (left) and S170 (right)

In what follows, oversaturations of 200, 300 and 400 % were induced at the saturation point by means of repeated peening of Almen strips and the arc heights in relation to mean shot velocity were illustrated in Figure 4 by way of example for the peening medium S170, a mass flow of 2000 g/min and a nozzle distance of 100 mm. Nozzle A displays a somewhat degressive course of the arc heights at higher velocities, Nozzle B a progressive one. We also found this behavior with peening medium S110 as well as at a nozzle distance of 150 mm. The distributions of velocity outlined above cannot be the reason for the higher peening intensities of Nozzle B, since for the peening media S110 and S170 we determined opposite curvatures of the distributions, whereas the nozzle characteristics in relation to the Almen intensity remain the same for both peening media. Where the geometry of the nozzle has a small internal diameter, the jet of shot will be tightly bundled, and as a result the cone of dispersion will be slight. Consequently, a larger amount of the shot than is the case with other nozzles will strike the Almen strip almost vertically and will then bounce off again vertically, which could constitute a greater hindrance for the shot moving towards the Almen strip, or cause it to deflect or reduce in intensity. This could be substantiated by trials with different impact angles. The resulting arc height differences between the nozzles decreased substantially. Due to the complex flow conditions inside the nozzle the flow of particles will attain a mean shot velocity and a distribution of velocity. Both can be measured at the nozzle outlet online. Along their path to the component or the Almen strip interactions will take place amongst the particles, which will be significantly influenced by the nozzle geometry. On the surface double hits play a role, which are also affected by mass flow and cone of dispersion. Once these influences are known for a certain peening configuration by pre-tests, the shot velocity measurement systems supplies a good possibility of online controlling of the peening intensity.

165 0,45

0,45

0,40

S170, A*=100%

0,35

Curvature in mm

Curvature in mm

0,40

S170, A*=200%

0,30 0,25

S170, A*=300%

0,20 0,15 0,10

Nozzle A

0,05

S170, A*=100%

0,35 S170, A*=200%

0,30 0,25

S170, A*=300%

0,20 0,15 0,10

Nozzle B

0,05

0,00

0,00

0

5

10

15

20

25

30

35

40

45

0

5

10

Mean shot velocity in m/s

15

20

25

30

35

40

Mean shot velocity in m/s

Figure 4:. Arc heights in cases of oversaturation

4.2

The Influence of the Mean Shot Velocity and the Velocity Distribution on the Residual Stress Depth Distribution

As an example for the influence of the mean shot velocity on the residual stress depth distribution Figure 5 shows results for the quenched and tempered state, the S110 shot and for the nozzle parameters diameter 15 mm and distance 100 mm as well as the mass flow of 200 and 600 g/min. 200 V=25 m/s / 0.146 mmA V=30 m/s / 0.167 mmA V=35 m/s / 0.219 mmA V=40 m/s / 0.237 mmA

0

V=20 m/s / 0.149 mmA V=25 m/s / 0.170 mmA V=30 m/s / 0.210 mmA V=35 m/s / 0.236 mmA V=40 m/s / 0.267 mmA

0

-200 -400 -600 42CrMo4 quenched and tempered S110 m=200g/min nozzle: 15/100

-800 -1000

residual stresses [MPa]

residual stresses [MPa]

200

-200 -400 -600 42CrMo4 quenched and tempered S110 m=600g/min nozzle: 15/100

-800 -1000

0

20

40

60

80

100 120 140 160 180 200 220

distance from surface [μm]

0

20

40

60

80

100 120 140 160 180 200 220

distance from surface [μm]

Figure 5:. Residual stress depth distribution with shot S110; nozzle diameter 15 mm; distance 100 mm; mass flow 200 g/min. (left) and 600 g/min. (right)

As can be seen, the thickness of the layer with compressive residual stress is considerably influenced and increases with increasing shot velocity, whereas the surface value and the amount of maximum residual stress below the surface do not depend on the mean shot velocity. To compare quantitatively the influence of different process parameters on the resulting residual stress depth distributions, the surface distance, where a compressive residual stress value of –200 MPa for the normalized and –400 MPa for the quenched and tempered specimens respectively was reached, was used as a measure. These stress values correspond roughly with 50 % of the maximum compressive residual stress amount. In Figure 6, the correlation between these depths and the applied mean shot velocities is plotted. Except for the low mass flow (200 g/min) and a nozzle diameter of 10 mm in all cases a more or less linear trend with a scatter band of up to 50 μm at a given velocity can be seen. A wider spread of particles, caused by a larger nozzle distance or a larger nozzle diameter, tends to result in a deeper position of the defined stress value.

166 42CrMo4 quenched and tempered

42CrMo4 normalised 240

200/10/100 200/10/150 200/15/100 200/15/150 600/10/100 600/10/150 600/15/100 600/15/150

200

160

depth at -400 MPa [μm]

depth at -200 MPa [μm]

240

120

200/10/100 200/10/150 200/15/100 200/15/150 600/10/100 600/10/150 600/15/100 600/15/150

200

160

120

80

80 10

20

30

40

mean shot velocity [m/s]

50

10

20

30

40

50

mean shot velocity [m/s]

Figure 6: Residual stress depth distribution with S110 shot; nozzle diameter 15 mm; distance 100 mm; mass flow 200 g/min. (left) and 600 g/min. (right)

5

Conclusion

In this study peening medium velocities have been measured online and have been correlated with the results of Almen tests. In practice, the influence of shot velocity on Almen intensity has hitherto not been taken into consideration. Moreover, peening pressure was used as a process parameter. The resulting Almen intensities are substantially affected by the geometry of the nozzle. The greater the scattering range of the nozzle, the higher will be the Almen intensities measured. The reasons for this are to be found in the interactions taking place in the particle flow and in the probability of multiple hits. Different mean shot velocities do not affect the surface value and the amount of maximum residual stress below the surface but have a significant influence on the thickness of the affected surface layer. By means of online-measurement of peening medium velocity it is possible to measure and adjust the stationary state present at the nozzle outlet. More influence factors are to be located in the particle flow as well as in the type and number of strikes. As long as their influence can be ascertained through prior experiments, measuring shot velocity can be employed as an online control of the peening process. This verification of the Almen intensities resulting for various shot velocities is valid for the respective configuration of apparatus and nozzles. Cyclical interruption of production so that Almen tests can be carried out to control the process can be greatly reduced or dispensed with altogether in this case.

6

Acknowledgements

This study was carried out with the support of Deutsche Forschungsgemeinschaft. We would also like to thank the Kugelstrahlzentrum Aachen GmbH for their co-operation on the shot velocity measurement system.

7 [1]

References R. Kopp, W. Linnemann, F. Wüstefeld: Shot Velocity Measurement, , Proc. 6th International Conference on Shot Peening, ed. J. Champaigne, San Francisco, 1996, 118ff

Correlation between Mechanical and Geometrical Characteristics of Shot and Residual Stress Induced by Shot Peening Paolo Marconi 2Effe Engineering S.r.l., Manerba del Garda, Italy

Giorgio Citran and Giovanni Gregorat Pometon S.p.A., Venice, Italy

1

Abstract

Accurate measurement of induced residual stress is necessary in order to evaluate the validity of peening process parameters. Among these parameters, a fundamental role is played by size and hardness of shot. Both of these characteristics are governed by specifications. Another important aspect is the use of “conditioned” shot. The scope of our analysis is to verify if it is possible to achieve greater process repeatability and reliability by using shot with a narrower hardness range and with more precise geometrical characteristics, compared to values found in current specifications. We have therefore measured surface and in-depth residual stress by X-ray diffraction and have analyzed the shape of the shot with a scanning electron microscope (SEM). We have also evaluated the effect on both conditioned and non conditioned shot following impact on the peening surface by SEM analysis.

2

Introduction

Increase in resistance to fatigue induced by shot peening is a result of surface compression. This surface compression is caused by the transfer of part of the kinetic energy of the shot to plastic deformation energy in the peened surface. Obviously, the percentage of energy transferred from kinetic to plastic deformation is a function of the geometric parameters of the impact which takes place between the shot and the peened surface. The most important geometric parameters are the angle of impact and obviously, the shot size. However, the exchange of energy between shot and surface is also influenced by other parameters, such as for example, the hardness and the shape of the shot. Peening shot is characterized by a nominal diameter (or size). The distribution of size around a nominal value is determined by SAE AMS-S-13165. Hardness values are also specified in SAE AMS 2431/1 and 2431/2. For this study, we have used grain size distributions and hardness ranges which are narrower than those contemplated by the abovementioned specifications. Furthermore, we have used shot which has been “conditioned”, i.e. shot which has been through a sort of preliminary peening process which eliminates shot with structural defects (hollows or excessive porosity) even though they are within limits imposed by the specifications. The conditioning of three test batches (nos. 1, 3 and 4) was done at 40 cycles. The fourth batch (no. 2) was unconditioned.

168

3

Test Parameters

Four different types of shot were used for the peening tests, as follows: Table 1: Type 1 ASH 330 conditioned

Table 2: Type 2 ASH 330 unconditioned

Characteristics

Characteristics Test results (%)

Grain Sizes (mm)

>1.180

0

>1.400

0

>1.000

0.3

>1.180

0.1

>0.850

91

>1.000

27

>0.710

100

>0.850

97

Chemical Analysis

>0.710

100

Elements

Percentage

Chemical Analysis

C

0.93

Elements

Mn

0.65

C

0.97

Si

0.61

Mn

0.68

P

0.025

Si

0.66

S

0.024

P

0.016

S

0.024

Grain Sizes (mm)

Hardness HRC

Test Results (%)

Percentage

Average

61.1

Hardness HRC

Minimum

60.0

Average

59.6

Maximum

62.0

Minimum

58.0

In range

100

Maximum

62.0

In range

100

169 Table 3: Type 3 ASH 330 conditioned

Table 4: Type 4 ASH 330 conditioned

Characteristics

Characteristics Test Results (%)

Grain Sizes (mm)

Test Results (%)

>1.400

0

>1.400

0

>1.180

0.1

>1.180

0.1

>1.000

27

>1.000

24

>0.850

96

>0.850

96

100

>0.710

100

Grain Sizes (mm)

>0.710

Chemical Analysis

Chemical Analysis Elements

Percentage

C

0.95

Mn

0.65

Si

0.65

P

0.021

S

0.025

Hardness HRC Average

61.3

Minimum

60.0

Maximum

63.0

In range

95

Elements

Percentage

C

0.96

Mn

0.63

Si

0.60

P

0.0185

S

0.023

Hardness HRC Average

60.3

Minimum

58.3

Maximum

62.0

In range

100

170 The peening test pieces are made of casehardened steel type 18NiCrMo7 (depth of casehardening 1 mm) and hardened to 850 HV. The following are images of unconditioned and conditioned shot taken with an SEM:

Unconditioned Shot

Conditioned Shot

Figure 1: Unconditioned shot and conditioned shot

Figure 2: Vickers microhardness indentations on a conditioned shot particle

171

4

Results

Peening was done by a robot system with the same parameters for all four shot types: • •

intensity 14 Almen A coverage 150%

Residual stress was measured by X-ray diffraction [1] and gave the following results: 0 0

0,02

0,04

0,06

0,08

0,1

0,12

0,14

0,16

0,18

0,2

-200

type1

type2

type3

type4

residual stress (MPa)

-400

-600

-800

-1000

-1200 depth (mm)

Figure 3: Residual stress results

5

Conclusions

SEM images taken before the peening application show how conditioning eliminates imperfections (e.g. cracks) which would otherwise lead to premature breaking up of shot. This ensures a more consistent level of grain size distribution. It is interesting to note in these images the plastic deformation of shot particles after impact. Our conclusions must necessarily take into account the type of shot and the conditioning values used for this test. The fact that the hardness level before conditioning was already quite

172 high (almost 60 HRC) means that conditioning produces less of a hardening effect than if the initial hardness level had been, say, 45 or 50 HRC. In addition, the number of cycles, speed and mass of shot have a direct influence on the hardening effect of conditioning. For these reasons the conditioning process carried out for this test at the abovementioned parameters and on shot with the abovementioned characteristics had no significant effect on hardness levels, as was verified by microindentation hardness tests. Separate tests have been carried out to verify the effect on hardening and on transmitted energy with test parameters different from the ones used for this test. Induced residual stress was not influenced in any significant way by conditioning of shot with a wide hardness range. In fact, shot types 2 and 4 induced approximately the same in-depth residual stress profile, even though they generated different residual stresses at the surface. On the other hand, conditioning appears to have had a greater effect on shot with narrower hardness ranges and with narrower grain size distributions than those required by the specifications (types 1 and 3).

6 [1]

References TAIRA S., ”X-ray studies on mechanical behavior of materials”. The Society of Material Science, Japan, 1974, pp. 22–32.

173

Influence of Pre-Annealing on Surface and Surface Layer Characteristics Produced by Shot Peening Katsuji Tosha Meiji Univ. Higashimita, Tama-ku, Tawasaki, Japan

Jian Lu, Delphine Retraint, Bruno Guelorget Universite de Technologie de Troyes, Rue Marie Curie, Troyes, France

Kisuke Iida Society of Shot Peening Technology of Japan, c/o Meiji Univ. Kawasaki, Japan

1

Introduction

Several effects produced by shot peening for fatigue strength, stress corrossion cracking, wear resistance, heat transfer and flow resistance are reported in many papers. Their factors relative to those effects are surface roughness, residual stresses, hardness and texture. In general, peening effects for fatigue strength are influenced by compressive residual stresses and work hardening [1, 2, 3] , and for heat transfer or radiation is influenced by surface roughness and affected layer [4, 5], but their relations with each factors are not always simple. The most important factor of the effect for fatigue strength is residual stress, and residual stress value changes with the texture of work material and the measuring ways. In this experiment, in order to clarify the influences of shot peening on the surface roughness, hardness distribution, half width and residual stress, work materials with several different hardness produced by different annealing conditions were peened, and the influence of pre-annealing on surface and surface layer characteristics was studied.

2

Experimental Conditions and Procedures

Experimental conditions on shot peening and residual stress measurement are shown on Table 1 and Table 2 respectively. Three surface profiles for one specimen were recorded without cut- off and surface roughness Ry was calculated from the records. Hardness distribution was obtained from perpendicular section to the peened surface with Vickers hardness tester, and averaged from the data at the same depth on three positions. Residual stress was measured using X-ray diffractometer. The pre-annealing temperature and time were changed from 620 degrees to 900 degrees centigrade and from 1 hour to 8 hours respectively.

174

3

Experimental Results

3.1

Influence of Annealing

3.1.1 Number of Grains Figure 1 shows microstructures etched by 3% aqueous solution of nitric acid. Ferrite and cementite become much clear after annealing, and in the case of 900 °C, 6h and 8h of 800 °C annealing, grain growth were observed clearly.

Figure 1: Microstructures

Figure 2 shows the influence of annealing temperature and time on the numbers of grains per unit area. As shown in Fig. 2(a), the grain numbers is the most in the case of 720 °C, because the grain size of as-received material is relatively large. Figure 2 (b) shows the influence of annealing time in the case of 820 °C, and the number is inversely proportional to the time over 2 hours annealing.

175

Figure 2: Influence of annealing on the numbers of grains per unit area

3.1.2 Hardness of Work Material Figure 3 shows the influence of annealing on the hardness of the work material. As the hardness is 290HV, the hardness after annealing of 620 °C decreases largely, but annealing is not complete. After 723 °C annealing, those hardnesses are almost similar to each other. As shown in Fig. 3(b), the hardness decreases gradually with the annealing time.

Figure 3: Influence of pre-annealing on the hardness

3.2

Shot Peened Material

3.2.1 Surface Roughness Surface roughness produced by shot peening is shown in Fig. 4(a), (b). Surface roughness increases slightly with the pre-annealing temperature and time.

176

Figure 4: Influence of pre-annealing on the surface roughness

3.2.2 Hardness Distribution In this experiment, hardness distribution are all work hardening types. The influences of pre-annealing on the maximum hardness (Hmax) and the depth of work hardened layer (d) are shown in Fig. 5 and Fig. 6 respectively. Their influence of the temperature are relatively small, but the depth of the work hardened layer increases gradually with pre-annealing temperature and time.

Figure 5: Influence of pre-annealing on the maximum hardness (Hmax)

3.2.3 Half Width Half width means the micro strain of crystals and the change is similar to the hardness one. As shown in Fig. 7, the half width of 600 and 723 °C annealed materials are larger than those of other cases and this means that the micro strain before shot peening affects the values after shot peening. In the case of 820 °C pre-annealing, the influence of the time is shown in Fig. 7(b).

177

Figure 6: Influence of pre-annealing on the depth of work hardened layer

Figure 7: Influence of pre-annealing on half width

Figure 8: Influence of pre-annealing on surface residual stresses

178 3.2.4 Surface Residual Stress As shown in Fig. 8, the influences of shot peening on the surface residual stresses is very small except for the case of 620 °C pre-annealing.

3.3

Grain Size and Change of Surface Residual Stresses

As the grain size is relative to the micro residual stresses, at first, the influence of X-ray projection area on the surface residual stresses is discussed. 3.3.1 Influence of Projection Area The projection area is changes with the diameter of a collimator. Figure 9 shows the results on the scattering of surface residual stresses and Figure 10 shows the relation between the projection area and the difference of residual stress. Its difference is inversely proportional to the projection area. These differences suggest that micro residual stresses are also affected by the grain size.

Figure 9: Difference of surface residual stresses by the X-ray projection area

Figure 10: Relation between the projection area and the difference of residual stresses

3.3.2 Influence of Grain Size on the Difference of Micro Residual Stresses Figure 11 shows the difference of residual stresses measured using collimators of Æ 2 mm and 0.15 mm in diameter. As shown in Fig. 11(b), in the case of small projection area, the differences increases with the grain size.

179

Figure 11: Influence of the grain size on the differences of surface residual stresses

4

Conclusions

In order to clarify the influence of pre-annealing on surface characteristics produced by shot peening, shot peening was performed for a plain carbon steel (C:0.45%) with conditioned cut wire-shot by an air blasting machine. The influence of grain size and hardness of the work material on the several characteristics such as surface roughness, hardness and residual stresses and half width were discussed, and the following results are obtained. The grain size increases but the hardness of work material decreases with the increase of preannealing temperature. The depth of work hardened layer and the maximum hardness in the affected layer are affected by the increase of pre-annealing temperature. The influence of pre-annealing on residual stress values is not very large, but the scattering of residual stress values increase with the pre-annealing temperature. The maximum difference of surface residual stresses is 584 MPa in this experiment

5

Acknowledgement

The authors wish to acknowledge the expert assistance provided by Mr Hayato Ishii and Ryo Sasaki of the School of Science and Technology, Meiji University, in connection with the measurements.

6 [1] [2]

References J. C. Straub: Special Performance of Transmission Parts by Shot Peening, SAE730800, 1973, 1–21. G. Wigmore , L. Miles: The Use of Shot Peening to Delay Stress Corrosion Crack Initiation in Austenitic 8Mn 8Ni 4Cr Generator End Ring Steels. Proc. of the 1st International Conference on Shot Peening (ICSP-1), 1981, 61–69.

180 [3] [4] [5]

S. Hisamatsu, T. Kanazawa, T. Toyoda: Influence of the Surface Properties on the Bending Strength of Shot Peened Carburized Steel. Proc. of ICSP-4, 1990, 477–484. M. C. Sharma: Shot Peened Surfaces and Boilling Heat Transfer. Proc. of ICSP-5, 1993, 199–206. O. Vöhringer: Changes in the State of the Material by Shot Peening, ICSP3, 1987, 185–204.

179

Endurance Life of Aging Aircraft Components Predicted by Conductivity Changes in Aluminium 2024 Jack A. Soules Surface Stress Technology Inc. Cleveland, OH, USA

182

183

184

185

Shot Peening of Ceramics: Damage or Benefit? Tobias Frey, Wulf Pfeiffer Fraunhofer Institute for Mechanics of Materials, Freiburg, Germany

1

Abstract

Non-transformation toughened ceramics show the typical brittle material behavior of failure before deformation at room temperature. Thus, strengthening of ceramics due to deformation induced compressive residual stresses has been thought to be not possible. Nevertheless, preliminary investigations had shown that, using ceramic-specific parameters, shot peening can introduce high compressive residual stresses into the near-surface of silicon nitride and improve the load capacity. The aim of the presented investigation was to improve the shot peening conditions in order to extend the increase of load capacity while maintaining the surface integrity. The materials investigated where alumina and silicon nitride, the properties determined where residual stresses, load capacity and topography. For the assessment of the surface strengthening the X-Ray diffraction analysis (XRD) and the ball-on-plate strength test were used. Due to the low penetration depth of X-rays XRD allows to evaluate the peening-induced residual stresses. In the ball-on-plate test, the sample is loaded with a spherical silicon nitride indenter up to failure of the sample which is detected by a high frequency ultrasonic detector. The results show that high compressive residual stresses in the GPa-range can be introduced in silicon nitride and alumina which may boost the load capacity of the near surface layers by a factor of up to 9. Only little effect on the surface integrity could be obtained.

2

Experimental Details

2.1

Materials Investigated

The materials investigated were a commercially available silicon nitride and a commercially available fine-grained alumina. The most important material characteristics are given in Table1. Table 1: Materials investigated.

Material

Specification Company

Young’s modulus

Characteristic strength

Fracture toughness

Ref.

Silicon Nitride

SN-N3208

Ceramics For Industry, CFI

300 GPa

877 MPa

4.2 MPa m1/2

[1]

Alumina

A61

Kennametal Hertel AG

390 GPa

400 MPa

4.0 MPa m1/2

[2]

186 2.2

Shot Peening

Shot peening was performed with an injection system. Cemented carbide beads with a diameter of (650 ± 40) μm were shot on the surface of the ceramics using different pressures between 2 bar and 4 bar and treatment times between 280 and 840 seconds. The distance between the sample and the nozzle were 20 cm. All samples were polished using a 1 μm diamond abrasive prior to the shot peening. Because of the lower hardness of the Almen strip in relationship to the ceramics samples the treatment time were chosen 8.5 and 25.5 times longer than reaching the 98 % coverage of the Almen Strip.

2.3

Determination of Residual Stresses and Load Capacity

The load capacity of the shot peened and the polished reference samples was determined using the ball-on-plate test. The load on a 10 mm silicon nitride ball is increased stepwise until a typical cone-crack appears which follows exactly the maximum tensile stresses. Because of the statistical behavior of ceramics, the load which causes fracture varies from test to test within a certain scatter-band. Typically 17 samples with equal surface conditions were tested and the average fracture loads and the corresponding standard deviations were calculated. The stress field of this static ball-on-plate contact is typical for stress fields occurring in contact situation e.g. in roller-bearings. The contact induced stress fields show strong gradients and the tensile stresses, which lead to cracks in brittle materials, are restricted to a very thin surface layer. The residual stress states of the near-surface layers of the samples were measured by X-ray diffraction. The mean stress values within the penetration depth in the range of 10 μm were evaluated using the so-called sin2y-method [3]. The most important measurement parameters are given in Table 2. Table 2: Parameters of residual stress evaluations

Material

Lattice plane

Radiation

y-range

X-ray elastic constant ½ s2

Silicon Nitride

{411}

CrKa

-64° £ y £ 64°

3.89 GPa-1

Alumina

{220}

CrKa

-64° £ y £ 64°

3.15 GPa-1

3

Results

3.1

Residual Stresses and Load Capacities

Fig. 1 correlates the near-surface stress states and the load capacities of polished and differently shot peened silicon nitride and alumina samples. The polished silicon nitride reference samples – representing the near-surface condition prior to shot peening – show small compressive residual stresses in the range of 100 MPa and a fracture load of about 3 kN in the ball-on-plate test.

187

10000

?

?

?

9000

fracture load in N

8000 7000 6000 5000 4000

Alumina

Silicon Nitride 3 bar / 840 s 4 bar / 280 s 3 bar / 280 s 2 bar / 280 s polished

3000 2000 1000 0 -1400

-1200

-1000

-800

-600

-400

-200

0

residual stress in MPa

Figure 1: Fracture load versus residual stress of silicon nitride and alumina samples in polished and different shot peened conditions

Shot peening allowed introducing up to 1.25 GPa compressive residual stresses. These compressive stresses shifted the load needed to fracture the surface layers from 3 kN to more than 9 kN. In the case of the samples with the highest compressive residual stresses the increase of near-surface strength was so high, that the load limit (9.2 kN) of the ball-on-plate testing device was reached in some cases and the samples passed the test without any crack. Thus, for the most effective shot peening process no error bar could be calculated for the fracture load. The results for similar experiments on alumina are also shown in Fig. 1. As obtained for silicon nitride high compressive residual stresses up to 1.1 GPa were created to the shot peening. These compressive stresses boost the fracture load from 0.7 kN up to more than 9 kN. Again, the load range of the ball-on-plate testing device was not high enough to introduce Hertzian cone cracks in all samples treated with the two most effective shot peening conditions.

3.2

Surface Roughness

Shot peening may influence the surface integrity in two ways. First, each hit by a bead will produce a localized macroscopic deformation accompanied by the creation of dislocations in the near surface crystallites. These effects are needed to create the strength increasing compressive residual stresses. The superposition of many localized deformations will result in an overall roughness of the surface. The surface topography resulting from a shot peening condition which comprises a more or less complete overlay of dips is shown in Fig. 2. In the optical micrograph nearly no difference between the polished reference surface and the shot peened area can be obtained. In addition no significant transfer of debris to the surface is obtained which can be concluded from the natural voids of the material still being visible at the peened surface. Fig. 3 shows the average roughness of the surface for the polished reference sample and after different peening treatments respectively. The shot peening of the surface leads to a small increase of the roughness up to 0.09 μm for the highest peening pressure of silicon nitride sam-

188

shot peened polished

Figure 2: Topographical map (top) and optical micrograph near the boarder between a polished and a shot peened (2 bar) area of a silicon nitride sample

Silicon Nitride

0.12 0.10 0.08 0.06 0.04 0.02 0.00

2 bar 280 s

3 bar 280 s

3 bar 4 bar polished 840 s 280 s

0.14

average roughness Ra in μm

average roughness Ra in μm

0.14

Alumina

0.12 0.10 0.08 0.06 0.04 0.02 0.00

2 bar 280 s

3 bar 280 s

3 bar 840 s

4 bar polished 280 s

Figure 3: Average roughness Ra of the polished and shot peened silicon nitride and alumina samples

ple and 0.14 μm for the alumina sample. One reason for the higher roughness of the alumina samples could be the lower hardness compared to the silicon nitride ceramic.

3.3

Evaluation of Single Hits

Silicon nitride and alumina samples showed similar results concerning the surface integrity. The area and depth of the localized deformation can be evaluated by inspecting single hits on the

189

Figure 4: Topography of a polished silicon nitride surface with dips produced by single hits of cemented carbide beads at maximum peening pressure of 4 bar

surface. Fig. 4 shows the topography of the most intense shot peening condition with 4 bar peening pressure. Fig. 5 shows the mean values of the diameter and depth of single dips as a function of the peening pressure. The mean values were calculated from the measurement of typically 25 dips per peening pressure. Higher pressures leads to a higher velocity of the beads and to a larger deformation. The plastic deformation is somewhat higher in alumina than in silicon nitride. 1.1

220 200

Silicon Nitride Alumina

180

1.0

160

0.8

140

depth in μm

diameter in μm

Silicon Nitride Alumina

0.9

120 100 80

0.7 0.6 0.5 0.4

60

0.3

40 0.2

20 0.1

0 2

3

4

5

peening pressure in bar

6

2

3

4

5

6

peening pressure in bar

Figure 5: Diameter and depth of single hits by different peening pressures for silicon nitride and alumina

In order to detect a possibly introduced damage due to shot peening the alumina and silicon nitride samples where carefully inspected for cracks. In case of the peening pressures with 3 and 4 bar very few cone cracks could be obtained near some single hits. Fig. 6 shows one of the rarely detectable, partially developed cone cracks near a single hit of an alumina sample.

4

Conclusions

Shot peening is a common technique to improve the strength of metal components. Up to now, it has not been successfully applied to ceramics, as these brittle materials have been assumed to show no significant plastic deformation due to mechanical loading and hence would not de-

190

crack

Figure 6: Partially developed cone crack in alumina due to a single hit

velop any residual stresses improving the strength. The presented results show however that under specific shot peening conditions also in brittle materials like silicon nitride and alumina high compressive stresses up to more than 1 GPa can be introduced near the surface. No significant damage is obtained. Exposing these strengthened surfaces to loading situations, which are characterized by a steep near-surface stress gradient, a boost of load capacity and strength can be evaluated. Such loading situation exists in, e.g., roller and sliding bearings or cutting tools. Further investigations will concentrate on increasing the depth of the compressive residual stress field, on the integrity of edges and on the possibility of restoring the strength of »roughly« machined ceramics. The patented shot peening procedure is now going into the first application which is the increase of the load capacity of full ceramic roller bearings.

5

Acknowledgement

Part of the investigation was sponsored by the Deutsche Forschungsgemeinschaft (DFG) and by the Bundesministerium für Bildung und Forschung (BMBF).

6 [1]

[2] [3]

References Rombach, M.: Experimentelle Untersuchungen und bruchmechanische Modellierung zum Versagensverhalten einer Siliciumnitridkeramik unter Kontaktbeanspruchungen, Dissertation Universität Karlsruhe, 1995. Op de Hipt, M.: Randzonencharakterisierung hartbearbeiteter und tribologisch belasteter Al2O3 und SiC-Keramiken, Fortschritt-Berichte VDI Reihe 5 Nr. 521, VDI-Verlag, 1998. Müller, P., Macherauch, E.: Das sin2ψ-Verfahren der röntgenographischen Spannungsmessung, Z. ang. Phys. 13, 1961, 305-312.

191

Influence of Retained Austenite, Strain-induced Martensite and Bending Stress upon Shot Peening-induced Residual Compressive Stresses Kotoji Ando1), Hirohito Eto2) and Katsuyuki Matsui2) 1)

1

Yokohama National University, Yokohama, Japan, 2) Isuzu Motors LTD, Kawasaki, Japan

Abstract

To introduce large residual compressive stresses by means of shot peening, the authors conducted an experiment using vacuum carburized and shot peened or stress peened test pieces. The influence of arc height (Ah), shot radius (R), retained austenite (gR), strain-induced martensite (MQ) and bending stress (spre) upon residual compressive stresses was investigated systematically. Residual compressive stresses were found to be quite dependent on Ah/R, only slightly influenced by gR and hardly affected by MQ content. Approximately, 50 % of the applied spre was introduced as residual compressive stress. Finally, the amount of induced residual compressive stresses linearly depended on Ah/R, gR and spre.

2

Introduction

It is generally recognized that the introduction of large and deep residual compressive stresses is highly effective in improving the fatigue limit of mechanical parts which are subjected to cyclic loads at stress ratios R = 0 [1–3]. An economical method for inducing such beneficial residual stress distributions is to use shot peening which has been studied by various researchers [4-6]. For a given shot peening intensity, the following two conditions will help in inducing large and deep residual compressive stresses: 1) Decreasing the deformation resistance of a material during shot peening 2) Increasing the yield stress (hardness) of a material. To achieve item 1) warm peening and stress peening were developed [7, 8]. There is carburizing and quenching for item 2) as a simple and easy to apply method. However, this method can lead to significant amounts of gR which will negatively affect hardness. On the other hand, it is known that this gR can transform to MQ during shot peening. Further, the transformation from gR to MQ involves volume expansion. Therefore, the authors assumed that by utilizing this phenomenon large and deep residual compressive stresses could be obtained. An experiment was conducted using vacuum carburized and shot peened or stress peened test pieces with various contents of gR. The influence of Ah, R, gR , MQ and spre on the residual compressive stress (sr) was analyzed systematically.

192

3

Experimental Methods

Material for the test pieces was JIS SCM822H with the following chemical composition in wt. %: 0.22 C, 0.25 Si, 0.86 Mn, 1.13 Cr, 0.36 Mo. Smooth rectangular test pieces were machined having the dimensions 30 × 90 × 5 mm. To investigate the influence of gR content on sr, the above test pieces were given three different vacuum carburizing or nitrocarburizing treatments. Hereafter, these test pieces are denoted as VC1, VC2 and VCN. The maximum residual stress (srmax) among these test pieces was found to be only –412 MPa, while the maximum content of retained austenite of VC1, VC2 and VCN amounted to 15.2, 22.9 and 47.4 %, respectively. Shot peening was performed on the various test pieces being subjected to 4-point bending with spre of –1000, 0, +1000 and +1400 MPa. Peening was done either by means of an impeller type (ISP) using Ø 0.8 mm shot of 560 HV hardness to 0.55 mmA intensity or by means of a direct pressure blast system using double peening (DSP). In primary shot peening (PSP), peening was performed with Ø 0.7 mm shot of 700 HV hardness to 0.30 mmC (conversion value of A strip: 1.05 mmA) intensity and in secondary shot peening (SSP), peening was done with Ø 0.08 mm shot of 700 HV hardness to 0.35 mmN (conversion value of A strip: 0.117 mmA) intensity. Table 1 gives an overview of the various test piece conditions. Table 1: Material and shot peening conditions

I

II

III

IV

Vacuum carburizing

VC1

VC2

VCN

Shot peening

ISP

ISP

IPre (MPa)

0

0

V

VI

VII

VIII

IX

X

XI

VC1 VC2

VCN

Vc2

VC2

VC1

VC2

VCN

ISP

DSP DSP

DSP

DSP

DSP

DSP

DSP

DSP

0

0

0

1000

–1000 1400

1400

1400

0

Measurements of gR contents and of sr values were performed by means of a micro X-ray stress analyzer using Cr-K= radiation. Stresses were calculated by the sin2O-method.

4 4.1

Results and Analysis r-Depth

Profiles and

R-Depth

Profiles in Shot Peened Test Pieces

Among the 11 test piece conditions, characteristic differences in sr-depth profiles were evaluated as illustrated for test pieces III, IV, VIII and IX in Figure 1. Based on these experimental results, characteristic values within the sr-depth profiles and gR-depth profiles are defined in Figure 2. Table 2: sr and gR values in ISP test pieces

No.

Irs (MPa)

CRs (%)

Irmax (MPa)

CRD (%)

I

–819

2,8

–1067

8,7

II

–757

6,8

–963

16,1

III

–628

22,2

–869

39,1

193

Figure 1: Typical examples of theIr-depth profiles

Figure 2: Characteristic values of Ir- and CR-depth profiles (schematic)

These values as measured for the conditions ISP, DSP and SDSP are listed in tables 2, 3 and 4, respectively.

194 Table 3: sr and gR values in DSP test pieces

No.

Irs (MPa)

CRs (%)

Irmaxs (Mpa)

CRDS (%)

Irmaxp (MPa)

CRDp (%)

IV

–1249

1,7

–1448

2,0

–1388

1,8

V

–1285

1,5

–1442

2,2

–1346

3,6

VI

–1192

10,5

–1325

22,8

–1232

25,6

Table 4: sr and gR values in SDSP test pieces

No. VII

Before*1

Irs (MPa)

CRs (%)

Irmaxs (MPa)

CRDS (%)

Irmaxp (MPa)

CRDp (%)

–1352

–

–1539

–

–1439

–

*2

–1819

1.2

–2038

1.3

–1907

7.4

VIII

Before

–1281

–

–1446

–

–1266

–

After

–576

1.5

–589

2.5

–580

7.7

IX

Before

–1188

–

–1356

–

–1308

–

After

–2049

1.1

–2180

0.8

–2109

3.0

After

–2032

0.8

–2224

1.9

–2071

5.8

Before

–1223

–

–1125

–

–1119

–

After

–2092

14.1

–2140

29.9

–1871

30.4

After

X XI

*1

before removing bending stress, *2 after removing bending stress

In addition, the effect of spre on sr values can be seen in table 4. ”Before” means under bending stress, ”after” means bending stress removed. The surface residual stress srs and the maximum residual stress values srmax, srmaxp and srmaxs varied considerably from –576 up to –2092 MPa and from –580 up to –2224 MPa, respectively. Furthermore, the gR content which is one of the objects of this analysis varied extensively from 0.8 % up to 30.4 %.

4.2

r-Depth

Profiles in SDSP Test Pieces

Figure 3 shows sr-depth profiles in SDSP test pieces before and after removing spre amounting to +1000 and –1000 MPa for test pieces VII and VIII, respectively. Comparing the data of these test pieces under bending stress, little differences were found in sr-depth profiles up to the depth of 150 ìm while in greater depths, differences were quite marked. However, after removing spre, both sr-depth profiles differed considerably. For example, srs and srmax of test piece VII were –1819 and –2038 MPa, respectively while those of test piece VIII were srs = -576 MPa and srmax = –580 MPa. This indicates that for introducing high values of residual compressive stresses, stress peening under tensile stresses should be done.

195

Figure 3: Ir--depth profiles in SDSP test pieces

4.3

Influence of gR Content on Shot Peening-Induced Residual Compressive Stresses

The amount of shot peening-induced residual compressive stresses in carburized, quenched and tempered test pieces increases by 1) 2) 3) 4)

Plastic deformation of tempered martensite (Mt) before shot peening Volume expansion by transformation of gR to MQ during shot peening Plastic deformation of MQ Plastic deformation of gR

The effect of gR content as determined after shot peening on srmax values of the test pieces I– VI (no bending stress applied) is shown in Figure 4.

Figure 4: Irmax vs. CRD content (no bending stress applied)

196 As gR content increases, srmax values decrease. Presumably, this is caused by the the yield stress of gR being lower than that of Mt and MQ. While the dependency of srmax values on gR content is similar for ISP and DSP test pieces, DSP resulted in residual compressive stresses about 400 MPa higher than those in ISP. This 400 MPa difference is attributed to plastic deformation of different quantities of Mt and MQ caused by the differences in shot peening intensity between ISP and DSP. 4.4

Influence of Ah on Residual Stresses

Although Ah values were widely varied, srmax values stayed almost the same while the corresponding depths of these values differed considerably. For example, these depths were 55–65 μm for PSP and only 8–16 μm for SSP. Thus, srmax can not be evaluated quantitatively by Ah only. Obviously, this is due to the fact that Ah is an integral value of the residual compressive stress field (unit: MPa mm). For evaluating the compressive residual stress quantitatively, relations given by Al-Hassani et al. [9, 10] are used which show that the maximum size of the plastic zone of an indentation is proportional to R. Thus, by using the ratio Ah/R, the influence of R becomes nil. Ah/R is thought to include factors a) to d) of section 3). 4.5

Influence of Ah/R on

R

and Residual Stresses

The effect of Ah/R on gR and srmax is illustrated for test piece VC2 in Figure 5.

Figure 5: CRD, CRdump and Irmax vs. Ah/R of VC2 test piece

The gR values shown were determined before and after shot peening at depths where srmax values (after shot peening) were found. The difference between gR contents before and after shot peening (Figure 5) corresponds to the amount of MQ. From Figure 5, the following can be stated: a) Increasing Ah/R leads to higher srmax. b) Increasing Ah/R decreases gR and thus, increases MQ.

197 4.6

Influence of MQ on Residual Stresses

The relation between MQ and srmax for a given ratio Ah/R = 3.0 is shown in Figure 6. Obviously, there is no simple correlation between MQ and srmax which is attributed to the fact that in shot peening both transformation-induced and plasticity-induced stresses interfere.

Figure 6: Irmax vs. MQ (Ah/R = 3.0)

4.7

Influence of

pre

in Stress Peening on

R

and Residual Stresses

The influence of spre in stress peening on gR as well as on sr is illustrated for test piece VC2 in Figure 7. From these results, the following can be concluded: a) sr is directly proportional to spre b) gR content or MQ content are independent of spre

Figure 7: CR, CRunp, and Irmaxp vs. Ipre of VC2 test piece

198 4.8

Influence of Ah/R,

R

and

pre

on Residual Stresses

Although an assessment of the individual contributions of the various parameters in shot peening to the development of residual compressive stresses is quite difficult mainly because of the complex nature and interference of transformation-induced and plasticity-induced residual compressive stresses, from Figures 4–7, the following conclusions can be made: 1) 2) 3) 4)

With an increase in gR content, sr decreases. As Ah/R increases, gR content decreases and thus, MQ content increases. No correlation exists between MQ and srmax. sr is directly proportional to spre.

Based on the conclusions a) to d) and the data from tables 2–4, a regression analysis was performed leading to the following equations I and II (Figure 8): srs = –276 Ah/R + 7.1 gR – 0.59 spre – 451 srmax = –172 Ah/R + 7.1 gRD – 0.54 spre – 822

(I) (II)

where the experimental values Ah/R = 1.4 – 3.0, gR (after shot peening) = 0.8 – 29.9 % and spre = –1000 – +1400 MPa were used. From the above equations it is seen that increasing Ah/R and spre increase the residual compressive stresses sr while increasing gR decreases sr. Again, this is attributed to the yield stress of gR being lower than that of both Mt (before shot peening) and MQ (after shot peening).

Figure 8: Ir vs. Ah/R, CR and Ipre

5

Conclusions

The authors conducted an experiment using JIS SCM8222H test pieces and investigated the influence of shot peening and material parameters on residual compressive stress formation.

199 Arc height Ah, shot radius R, retained austenite content gR, strain-induced martensite MQ and bending tensile stress spre were varied systematically. The following conclusions can be drawn: 1 2

3

4 5

6 [1] [2]

Both the surface and maximum residual stress after shot peening are proportional to the new parameter Ah/R. The higher the retained austenite content as present after shot peening, the smaller the surface and maximum residual stresses. However, there was little influence of retained austenite content as determined before shot peening. This is attributed to the low yield stress of retained austenite. Strain-induced martensite as formed during shot peening does not influence the surface and maximum residual compressive stresses. This is attributed to the interference of transformation-induced and plasticity-induced residual compressive stresses. Approximately 50 % of the amount of the tensile bending stress during stress peening is effective in increasing the surface and maximum residual compressive stresses. From the above results, equations I and II were obtained where residual compressive stresses are seen to increase with increasing arc height/shot radius and amount of tensile bending stress and to decrease with increasing amounts of retained austenite measured after shot peening.

References

K. Matsui, Doctor’s thesis of Yokohama National University, 2000. H. Ishigami, K. Matsui, A. Tange and K. Ando, Journal of High Pressure Institute of Japan, 2000, Vol.38, No. 4, p. 205. [3] K. Matsui, H. Eto, K. Yukitake, Y. Misaka and K. Ando, Trans. of Japan Society of Mechanical Engineers, 2000, Vol.66, No. 650, p. 1878. [4] A. Tange and K. Ando, Society of Materials Science, Japan, Proceeding of the 10th Symposium on Fracture and Fracture Mechanics, 1999, p. 6. [5] K. Matsui, H. Eto, K. Kawasaki, Y. Misaka and K. Ando, Trans. of Japan Society of Mechanical Engineers, 1999, Vol.65, No. 637, p. 1942. [6] H. Ishigami, K. Matsui, Y. Jin and K. Ando, Trans. of Japan Society of Mechanical Engineers, 2000, Vol.66, No. 648, p. 1547. [7] A. Tange, Doctor’s thesis of Yokohama National University, 2001. [8] H. Ishigami, Doctor’s thesis of Yokohama National University, 2001. [9] Al-Hassani. S. T. S., SAE821452, 1982. [10] Al-Obaid. Y. F., Trans. ASME, J. Appl. Mech., 57, 1990, p. 307.

200

Lining of Metal Surface with Hard-Metal Foil using Shot Peening Yasunori Harada, Ken-ichiro Mori and Seijiro Maki Toyohashi University of Technology, Japan

1

Introduction

Light metals such as the aluminum and magnesium alloys are widely used for automotive components and electronic consumer products nowadays. Both alloys have a high ratio of strength to the weight and the low densities. Magnesium alloys have a good recyclable potential[1], but the application is still limited. Light metals, because they are not nearly as wear resistance as steel in ambient and high temperatures, are not commonly used in applications where wear resistance and strength are important. Therefore, there is strong demand to improve surface treatments that could guarantee the wear resistance of the parts in aggressive environments. For the improvement, the lining processes with hard metals such as steel and nickel are useful. The surface treatments are usually used to improve the surface properties such as wear resistance and corrosion resistance. Plating, PVD and CVD are generally employed as the lining processes. The wear resistance of aluminum alloys was improved by dispersing the silicon-carbide particle to the surface [2] and by mixing PVD with ion implantation [3]. These lining processes, however, are inadequate for the use under severe conditions because of thin plating layers. Although a thick layer is formed by the thermal splaying process [4], the loss of thermal splaying material becomes large [5]. For the purpose of improving the surface properties, the bonding processes in metal forming are very effective. If hard materials are bonded to the surface of light metals, the surface wear resistance will be improved. Two metals are bonded by applying large pressure and plastic deformation in rolling and extrusion processes [6]. The pressure and plastic deformation break up the oxide film and contaminants at the interface between the two metals, and new and clean surfaces suitable for the bonding are generated. In these processes, however, the bonding becomes difficult in the case of a large difference between flow stresses of two metals, because the deformation is concentrated at the metal with a small flow stress. The authors have proposed a lining process of metals with thin foils using shot peening [7]. In this method, the foil is bonded to the surface of the workpiece bringing about large plastic deformation and the pressure generated by the hit of many shots are utilized for the bonding. The lining process using shot peening is suitable for the bonding of thin and dissimilar foils required for the improvement of surface properties. By means of peening with many shots, the aluminum foil was successfully bonded over the surface of the carbon steel workpiece. In the present study, a method for lining light metals with hard materials using shot peening is proposed. The hard materials are bonded to the surface of light metal workpieces by the collision of many shots. The effects of shot speed and the processing temperature on the bondability were examined. To evaluate the wear resistance for the bonded surface of workpiece, wear test was also examined.

201

2

Experimental Procedures

2.1

Method of Lining

In the shot peening process, a metal workpiece undergoes plastic deformation near the surface due to the hit of many shots at a high speed. This plastic deformation is utilized for lining metal workpieces with dissimilar foils as shown in Figure 1. Since plastic deformation caused by the shot peening is concentrated near the surface, the present method is suitable for the lining with the thin foil. In addition, this method is useful in bonding dissimilar metals because of the utilization of plastic deformation. The foil and workpiece are heated in order to make the bonding easy.

Shot

shot

Foil thin foil

Bonding

Bonding

Plastic deformation Workpiece Figure 1: Lining method using shot peening.

2.2

masking plate workpiece Figure 2: Lining of workpiece with thin foil using masking plate

Lining Using Masking Plate

Since the shots collide on the limited parts, shot peening is one of the partial processing, and thus this process is applicable for the partial lining. However, it is not easy to control the range of the collision of shots for the desired one. The range is limited by the masking plate as shown in Figure 2. The foil is slightly larger than the masking plate, and is fixed for the collision by the plate. The margin of the foil for the fixation is about 5mm, and the foil covered with the masking plate is removed by tearing after the shot peening. To improve the bondability, a pure aluminum foil was used as insert metal, because pure aluminum has high bondability. To examine the shot lining, a centrifugal shot peening machine was employed in the experiment. To make the bonding easy, the metal foil and workpiece were heated in air. The shots used for the experiment are made of the high carbon cast steel (HV500), and the masking plate was made of the heat-treated tool steel. The conditions used for the experiment are summarized in Table 1, where t is the thickness of the foil and insert metal. The experiment was performed between room temperature and 400 °C in air.

202 2.3

Materials used for Experiment

The workpieces were aluminum alloys (A2017, A5052, A6061, A7075) and magnesium alloys (AZ31B, AZ91D), and the foils were commercially pure nickel, commercially pure titanium and stainless steel SUS304. The surface of the workpiece was cleaned with emery papers prior to the shot lining. The dimensions of the workpieces and foils used for the experiment are summarized in Table 1. Table 1: Working conditions used for shot peening experiment. Equipment

Centrifugal peening

Shot material

High carbon cast steel

Shot diameter d / mm

1.0

Impact speed v / m/s

40, 80

Coverage %

100

Heating temp. T / °C

20 – 400

Workpiece

A2017, A5052, A6061, A7075, AZ91D, AZ31B

Foil (t = 0.01 – 0.1 mm)

Nickel, Titanium, Stainless steel 304

Insert metal (t = 0.015 – 0.020 mm)

A1050

Surface finish (Emery paper)

# 120 (Ra = 3.3 μm)

Atmosphere

Air

3

Lining with Hard Materials

3.1

Critical Heating Temperature

The critical heating temperature for the bondability of foils was investigated. The variations of the critical heating temperature with the thickness of foil are given in Figure 3. The workpieces are aluminum alloys (a)A2017 and (b)A5052. Since pure nickel has a high oxidation resistance, the critical heating temperature of nickel foil is lower than that of the other foils. The critical heating temperature increases with the thickness of sheet. On the other hand, the lining for aluminum alloy A5052 was performed in order to examine the effect of the material of the workpiece on the critical heating temperature (Figure 3(b)). In comparison with the lining of the aluminum alloy A2017 workpiece with the metal foils shown in Figure 3(a), the critical heating temperature is higher. Since the magnesium content of the aluminium alloy A5052 is larger than that of the aluminium alloy A2017, the bondability of A5052 workpiece is smaller. In addition, the metal foils were successfully bonded to the aluminum alloys A6061 and A7075 workpieces. In the present study, the lining experiment was performed for the determined heating temperature.

400 300 200 ¬ F SUS304 ¢ F Ti › F Ni

100

0

0.01 0.02 0.03 0.04 0.05 0.06

Thickness of foil t / m m

(a) A2017(Al-Cu) workpiece

Critical heating tem perature Tc / °C

Critical heating tem perature Tc / °C

203

400 300 200 100

0

¬ F SUS304 ¢ F Ti › F Ni

0.01 0.02 0.03 0.04 0.05 0.06

Thickness of foil t / m m

(b) A5052(Al-Mg) workpiece

Figure 3: Relationship between thickness of foil and critical heating temperature for aluminum alloys

3.2

Lining Surface

The lining of the aluminum alloy A2017 workpiece with the pure titanium foil using a rectangular masking plate at 300C was performed. The surface of the lined workpiece is given in Figure 4. The surface of the workpiece is uniformly hit with many shots. The exfoliation of the foil from the surface of the workpiece was not observed. The surface of the magnesium alloy AZ91D workpiece is given in Figure 5. Although the magnesium alloy has low bondability, the metal foils are successfully bonded to the magnesium alloy workpiece by increasing the processing temperature and using the pure aluminum insert. The lined shape is nearly the same as the masking shape. In addition, the lined shapes for the pure titanium and the stainless steel SUS304 foils are nearly the same.

Figure 4: SEM photogaph of surface of bonded workpiece at 300 C for A2017 workpiece and titanium foil (t=0.05mm)

Figure 5: Surface of bonded workpiece at 300 C for AZ91D workpiece and nickel foil

204 3.3

Partial Lined Surface

To examine the accuracy of the lined shapes for the masking shapes, the partial lining using some masking plates was also carried out. The surfaces of the partially lined workpices using the various masking plates are shown in Figure 6. It is clearly that the lined shape is nearly the same as the masking shape.

(a) Circles (A6061 workpiece, nickel foil)

(b) Polygons (A5052 workpiece, titanium foil)

Figure 6: Surface of bonded workpieces for aluminum alloy workpieces and hard foils

3.4

Bond Strength

The microscopic photograph of the cross-section at 300 °C is shown in Figure 7. The bonding near the interface between the workpiece and foil is good. Since the foil is very thin, it is difficult to measure the bond strength between the workpiece and foil. To evaluate the bond strength between the workpiece and foil at the boundary, the lined workpiece was bent until cracks occur. The SEM photograph of the bonded surface at boundary after bending is given in Figure 8. The workpiece is A2017, and the foil is SUS304. Although the foil tore, the exfoliation from the surface was not observed. The bonding of the foil was sufficient.

SUS304 foil

Al foil A2017

0.3 mm

Figure 7: Microscopic photograph of crosssection for A2017 workpiece and SUS304 foil (t=0.05mm)

Figure 8: SEM photograph of surface of bonded A2017 workpiece after bending

205

4

Lining of Hard Powders

4.1

Method of Lining Using Hard Powders

The hard foils such as stainless steel, titanium and nickel were successfully bonded to the surfaces of the aluminum and magnesium alloys in Section 3.2. However, it is very difficult to bond in case of lining between light metals and hard materials such as ceramic and cemented carbide. Since many shots directly collide with the hard materials, plastic deformation in light metals is very small. In the present study, by using the powder of hard material, the lining of hard materials to light metals is tried. The hard powders are bonded to the surface of light metal workpieces by the collision of many shots. The hard powders set on the workpiece are impacted with many shots in the experiment. However, on the smooth surface, the powders are moved by the impact of shots in the shot peening. Since it is not easy to fix the hard powders on the surface of workpiece in the shot peening, the lining process of workpiece with sandwich foil shown in Figure 9 was tried. In the experiment, the powders are sandwiched in between two aluminum foils by a press. The foil used for the sandwich is a pure aluminum foil of 0.015 mm in thickness. The impact speed of shot and the processing temperature are 80 m/s and 300 °C, respectively.

Sandwich foil

Aluminum foil

Hard Powder Workpiece

Surface Workpiece Figure 9: Schematic illustration of shot lining of workpiece with hard powders using sandwich foil

4.2

Lining with Hard Powders

The microscopic photographs of cross section for the magnesium alloys and hard powders are shown in Figure 10. Although the magnesium alloys have low bondability, the hard powders are successfully bonded to the magnesium alloy workpieces using sandwich foil. In addition, the bondability for the aluminum alloys and hard powders is nearly the same. The bonding for the workpieces and powders is sufficient in Figure 10. However, it is very difficult to measure the bond strength between the workpiece and powders. Thus, the bonded workpiece was bent until cracks occur. The SEM photograph of the bonded surface after bending is given in Figure 11. The workpieces are aluminum alloys A2017, and the powders are cemented carbide and zirconia. Although cracks were generated on the surface, the separation from the surface was not observed. The bonding of the powders was sufficient.

206

Cemented carbide Al Foil

0.3 mm

AZ91D

(a) Cemented carbide

Zirconia Al Foil

0.3 mm

AZ31B (b) Zirconia

Figure 10: Microscopic photographs of cross-section for magnesium workpieces and hard powders

0.3mm (a) Cemented carbide

0.3mm (b) Zirconia

Figure 11: SEM photographs of surface of bonded workpieces after bending

4.3

Wear Resistance

To evaluate the wear resistance for the bonded surface, wear test was examined. A cylindrical grinding wheel is removing a layer of material under a certain load. The workpiece is held in place with a chuck. The size of abrasive grain is about 0.2 mm. The experiment was performed in a wet atmosphere in order to prevent the friction heat. The equipment used for wear test imposes a load of 2 kg on a 10 mm diameter wheel. The surfaces of the bonded AZ31B workpieces after the wear test are given in Figure 12. The surface of the workpiece with the chromium powder becomes a flat by the grinding. In the case of the lining with the cemented carbide powders, the wear is smaller. Namely, the bonded surface has a fairly good resistance to wear. In addition, the wear test for the bonded aluminum alloy workpieces was also performed. The wear resistance for the bonded workpieces is nearly the same as the magnesium alloy workpieces. Since the hardness number on cemented carbide and alumina is higher, the wear on the surface is small. It is clearly that the lining with hard powders is effective in the wear.

207

(a) Cemented carbide

(b) Chromium

Figure 12: Surfaces of bonded AZ31B workpieces after grinding

5

Conclusions

A lining method with hard metals using shot peening was carried out. The hard metals were successfully bonded to the surface of the metal workpieces by the hit of many shots. The accuracy of the lined shapes was sufficient. The bond strength of the lined workpiece was sufficiently confirmed by a bending test. The lining with sandwich foil was also tried. The hard powders were successfully bonded to the surface of the metal workpieces. The wear resistance of the lined workpiece was confirmed to be sufficient by wear test of lined workpieces. It was found that the present method using shot peening is effective in wear resistance of the metal products.

6

Acknowledgement

This research was supported in part by a Grant-in Aid for Developmental Scientific Research by Japanese Ministry of Education, Culture, Sports, Science and Technology.

7 [1] [2] [3] [4] [5] [6] [7]

References Sasaki, G.; Hara, S.; Yoshida, M.; Pan, J.; Fuyama, N.; Fujii, T.; Fukunaga, H., Magnesium Alloys and their Applications, WILEY-VCH 2000, p.257-262. Ostermann, A. E., Society of Automotive Engineers, No. 790843, 1979, p.1-8. Asahi, N.; M. Haginoya, M.; Satou, T.; Hashimoto, I., J. of the Metal Finishing Society, 1987, 38-8, 329-333. Sasada, I.; Takamoto, Y.; Harada, K.; Izumi, T.; Ujimoto, Y.; Kameyama, R., IEEE Transaction on Magnetics, MAG-23, No. 5, 1987, p.2188-2190. Ito, H.; Fukunaga, H.; Otsuki, Y.; Matsumoto, H.; 8th Int. Thermal Splaying Conf., USA, 1976, p.38-44. Garrett, M. J. P.; Giles, J. L., Sheet Metal Indust., 1973, 50, 514-520. Harada, Y.; Mori, K.; Maki, S., J. of Materials Processing Technology, 1998, 80-81, 309314.

206

Relaxation of Shot Peening Residual Stresses in the 7050-T7451 Aluminium Alloy after Heat Cycles for Adhesive Bonding Martin Roth, C. Wortman Department of National Defence, Ottawa, Ontario, Canada

1

Introduction

Shot peening, cold working of holes, and adhesive bonding of metallic or composite doublers are often used singly or in combination in order to increase the fatigue life of critical aerospace components. A doubler, applied to a highly loaded area, provides an alternate load path and can reduce the propensity for fatigue cracking. Adhesive bonding of a doubler requires a number of steps involving heat and pressure cycles where the temperature of the part might exceed the maximum temperature to which shot peened aluminium alloys should be subjected to, because of the possibility of thermal stress relaxation. The maximum exposure temperature ranges from 90 °C to 121 °C for aluminium alloys, depending on the shot peening specification used. [1] Slow relaxation of shot peening stresses in the 7075-T6 aluminium alloy heated to 225 °F (107 °C) has been reported. [2] The following steps have been followed by the Canadian Forces to successfully adhesively bond a doubler to an aluminium alloy part, starting from a coating free surface, whether shot peened or not: • • • •

grit blasting using aluminium oxide (220 grit) to produce a chemically active surface for the silane treatment and a rougher surface with greater surface area for the bond, application, and curing of a silane coupling agent (30 minutes at 175°F (79 °C)), application and curing of a corrosion inhibiting primer (i.e. Cytec BR127) (1 hour at 250°F (121 °C) preferred, or 30 minutes at 250°F (121 °C)), application of the adhesive (i.e. Cytec FM73M) and doubler, and curing of adhesive (2 hours at 250°F (121 °C) preferred, or 8 hours at 185°F (85 °C) at a pressure of 14.7 psi (101 kPa)). [3] The preferred cycles are preferred for bonding reasons only and not based on the effects on a shot peened article.

The Aerospace Materials Specification AMS-S-13165 states in section 3.3.10: "When peened parts are heated after shot peening … the temperature employed shall be limited as follows: Aluminum alloy parts 200°F (90 °C) maximum". [4] This is qualified by section 6.13: "Processing or service temperatures of shot peened parts shall be limited to the temperatures in 3.3.10 unless test data for specific applications support the satisfactory use of higher temperatures". An experimental program was set up by the Canadian Forces to investigate: • •

the effects of the low and high temperature cycles used for adhesive bonding on the relaxation of the shot peening residual stresses in the 7050-T7451 aluminium alloy, the effects of the grit blasting,

209 • •

the effects of a 2 hour exposure at 163 °C, as some areas of a part can reach that temperature in order to achieve the desired bondline temperature of 121 °C, and the effects of the low and high temperature cycles on fatigue crack initiation of shot peened 7050-T7451 aluminium alloy under a Canadian CF188 (F/A-18) spectrum loading.

2

Experimental Procedure

2.1

Specimen Preparation

Specimens, 12,5 or 25 mm thick, were machined from 7050-T7451 aluminium alloy plate. The design properties of the plate were: 440 MPa yield stress, 510 MPa ultimate tensile strength, and 10% elongation. [5] One surface was shot peened to an Almen intensity of 0,008 ± 0,001 A (inch) (0,200 ± 0,025 A (mm)) using Z425 ceramic beads per SAE J1830 and a minimum of 200% coverage, with saturation at 100% coverage. The grit blasting was performed using 220 grit aluminium oxide.

2.2

Residual Stress Measurements

The residual stresses were measured using a X-ray stress analyzer model AST X2001 over an area 3 mm in diameter using a chromium X-ray tube operating at 30kV. The Al 311 family of diffraction planes was used and the material X-ray elastic constant E/(1 + n) = 57.1 GPa was selected, where E is the modulus of elasticity and n is Poisson’s ratio. [6] To obtain a distribution of the residual stresses as a function of depth, material was electropolished away in increments of 0,025 mm. All the measurements were performed along two perpendicular directions coinciding with the plate longitudinal, transverse or short transverse orientations. The measured residual stresses were not corrected for the relaxation occurring when layers of material are removed, because the correction would have been less than 5% for the peak stress, and the thickness of the compressive layer would not be significantly affected. [7] Direct comparison of the measured values was deemed satisfactory for this investigation.

2.3

Fatigue Specimens and Testing

The fatigue specimens, manufactured from 50 mm thick 7050-T7451 plate, were 254 mm long, 50.8 mm wide, 12,7 mm high and had a convex upper surface (Figure 1). The upper surface was shot peened as described in section 2.1. The specimens, designed for four point bending fatigue testing, provided a defined zone of inspection for crack formation, which was the primary concern of this study. The specimens were fatigue tested in a computer controlled Instron 831 servo-hydraulic fatigue testing machine, using a spectrum gathered from flight data. One spectrum block consisted of 7719 cycles, representing 325 simulated flight hours (Figure 2). The loads were scaled up so that the specimens would fail within 80 blocks. Under these conditions, the peak stress, occurring once per block, was 513.3 MPa. The blocks were repeatedly applied. The specimens were

210 inspected every 5 blocks for crack formation using acetate replicas and ultrasonic testing. Some tests were also run to final failure.

Normalized Stress

Figure 1: Fatigue specimen with a convex upper surface. The dimensions are in inches. 1 .0 0 .9 0 .8 0 .7 0 .6 0 .5 0 .4 0 .3 0 .2 0 .1 0 .0 -1 0 0 0

0

1000

2000

3000

4000

5000

6000

7000

8000

9000

10000

11000

12000

13000

14000

N u m b e r o f P o i n ts

Figure 2: Normalized spectrum block representing 325 simulated flight hours. The maximum applied stress on the specimens was 513.3 MPa.

3

Results

3.1

Effects of 12 Hours Exposure at 85 °C

There was no relaxation of the surface and subsurface residual stresses after the following heating cycles of 30 minutes at 79°C, 4 hours at 85°C plus 8 hours at 85°C, initially considered for adhesive bonding.

3.2

Effects of Grit Blasting and a 2 Hour Exposure at 121 °C

The distribution of the residual stresses with depth was measured in two areas, one shot peened and one shot peened and grit blasted, in the as prepared conditions, and after one heating cycle at 121°C for 2 hours. The compressive residual stresses increased from 248 MPa at the surface to a maximum of around 365 MPa, 0,025 to 0,050 mm below the surface. The layer with significant compressive stresses was 0,100 to 0,125 mm deep (Figure 3). The only effect of the grit blasting was a small increase of the compressive residual stresses at the surface, of the order of 10%. The heating cycle induced a relaxation of the compressive stresses. The effect was the most pronounced at the surface where the reduction was between 19 and 25% for the shot peened surface and between

211 21 and 29% for the peened and grit blasted surface. The relaxation of the compressive residual stresses was progressively smaller at greater depths (Figure 3). The peak compressive stress was reduced by about 10% from 365 to 330 MPa. 0

0.04

Depth (mm) 0.06 0.08

0.1

0.12

Peened Peened + Grit Blasted Peened + 1 h at 121 °C Peened + Grit Blasted + 1 h at 121°C

-10 Residual Stress (ksi)

0.02

-20

0 -50 -100 -150 -200

-30

-250 -40

Residual Stress (MPa)

0

-300 -50 -60

-350 -400 0

0.001

0.002 0.003 Depth (inch)

0.004

0.005

Figure 3: Residual stress distribution from a shot peened area and a shot peened and grit blasted area, in the as prepared conditions, and after one heating cycle at 121 °C for 2 hours

3.3

Effects of a 2 Hour Eposure at 163 °C

The distribution of the shot peening residual stresses with depth was measured in the as peened condition, and after one heating cycle at 163°C for 2 hours. The relaxation of the residual stresses was the most pronounced at the surface, where the reduction was between 30 and 40%. The residual stress relaxation was progressively smaller at greater depths (Figure 4). The peak compressive stress was reduced by about 20% from 330 MPa to 260 MPa.

0

0.02

0.04

Depth (mm) 0.06 0.08

0.1

Shot Peened Shot Peened + 1 h at 163 °C

-10

0.12

0 -50

Residual Stress (ksi)

-100 -20

-150 -200

-30

-250 -40

Residual Stress (MPa)

0

-300 -50 -60

-350

0

0.001

0.002 0.003 Depth (inch)

0.004

-400 0.005

Figure 4: Residual stress distribution from a shot peened area, in the as prepared condition, and after one heating cycle at 163 °C for 2 hours

212 3.4

Effects of two Bonding Heat Cycles on Residual Stress Relaxation and Fatigue Life

3.4.1 Relaxation of Residual Stresses The distribution of the shot peening residual stresses with depth was measured in fatigue specimens, in the as peened condition, and after heating cycles simulating all the adhesive bonding steps, one with a total time of one hour at 121 °C, and the other with a total time of 3 hours at 121 °C. In the as peened specimen, the compressive stresses increased from 180 MPa at the surface to a maximum of 375 MPa, 0,100 mm below the surface. The layer with significant compressive stresses was approximaley 0,250 mm thick (Figure 5). That layer was deeper than those in the sections 3.1 and 3.2 specimens, also peened to the same nominal Almen intensity. The heating cycles induced a relaxation of the residual stresses, which was slightly larger after the longer exposure at 121 °C (Figure 5). The maximum relative relaxation was at the surface where the reduction was between 28 and 31 % after 1 hour at 121°C and between 31 to 39 % after 3 hours at 121°C. The peak compressive stress was reduced by about 16 % from 375 MPa to 315 MPa after 1 hour at 121 °C, and 23 % to 290 MPa after 3 hours at 121 °C. 3.4.2 Effects on Fatigue Sixteen shot peened specimens were fatigue tested, 5 in the as peened condition, 5 after 1 hour at 121 °C, and 6 after 3 hours at 121 °C. Some tests were run to final failure. The results of the fatigue tests are summarized graphically in Figure 6. The average crack formation life (defined as the detection of a 1 mm long crack) for the baseline shot peened specimens was 69 blocks or 22425 simulated flight hours. Specimens exposed for 1 hour at 121 °C showed an average crack formation life of 43 blocks, a 38 % reduction compared to the baseline series. The average crack formation life of the specimens exposed for 3 hours at 121 °C was slightly lower at 40 blocks, a 42% reduction compared to the baseline series. 0

0.05

Depth (mm) 0.1 0.15

0.2

0.25

10 Shot Peened Shot Peened + 1 h at 121°C Shot Peened + 3 h at 121°C

0

-10

-80

-20 -160 -30 -240

Residual Stress (MPa)

Residual Stress (ksi)

0

-40 -320 -50 -400

-60 0

0.002

0.004 0.006 Depth (inch)

0.008

0.01

Figure 5: Residual stress distribution from a shot peened area, in the as prepared condition, and after 1 and 3 hours at 121 °C

Blocks

213 90 80 70 60 50 40 30 20 10 0

Crack formation

1

2

3

4

5

« As shot peened »

6

7

8

9

13

Total life

10

11

12

15

16

« 1 hour at 121ºC »

«

3 hours at 121ºC

14

»

Figure 6: Crack formation and total lives of shot peened fatigue specimens tested as shot peened and after 1 and 3 hours at 121 °C

4

Discussion

Relaxation of residual stresses by thermal means is a function of temperature and time. There was no relaxation after an exposure of up to 12 hours at 85 °C. This is not unexpected as 85 °C was significantly below the lowest maximum permissible exposure temperature after shot peening. [1] After 1 hour at 121 °C, the maximum temperature to which shot peening parts should be subjected according to some specifications, the reduction of the surface stress was around 30 %. After 3 hours at 121 °C, the reduction of the surface stresses was between 31 to 39 %. A rapid decrease of the residual stresses with time followed by a more gradual one is typical of isothermal stress relief. [8] The relaxation of the shot peening residual stresses had increased to 30 to 40 % after 2 hours at 163 °C, but the maximum compressive stresses were still in the 255 to 275 MPa range. That temperature corresponds to the final aging treatment for the 7050-T7451 aluminium alloy. If the alloy is exposed to 163 °C for longer times or to higher temperatures, there will be a gradual decrease of the mechanical properties as well as a decrease of the shot peening residual stresses. The only effect of the grit blasting was a small increase of the compressive residual stresses at the surface of the as shot peened specimen. Adhesive bonding at the lowest possible temperature (i.e. 85 °C) would be the preferred choice if the prime objective was to minimize or avoid the relaxation of the shot peening residual stresses. A heat cycle of at least 1 hour at 121 °C to cure the primer is required to produce a durable bond, but this will lead to a relaxation of the shot peening residual stresses. The curing of the adhesive can be performed either for 2 hours at 121 °C or 8 hours at 85 °C. The 2 hour exposure at 121 °C could be selected based on ease of processing as the further relaxation of the stresses by this heat cycle would be small after the part has been exposed to the 1 hour at 121 °C primer cure. The 1 hour exposure at 121 °C caused a significant reduction of the crack formation lives when the shot peened specimens were fatigue tested using a spectrum with a high peak stress of 513 MPa. It was reported that there was no significant changes in the fatigue life when specimens, shot peened to the same intensity and heated for 1,5 hours at 110 °C and 1 hour at 121 °C, were fatigue tested using a spectrum with a lower peak stress of 420 MPa. This is in accordance with the observation that shot peening is more beneficial to fatigue life at low stress levels. [9]

214

5

Conclusions

There was no noticeable relaxation of the shot peening residual stresses after an exposure of up to 12 hours at 85 °C. The preferred heat cycle used for adhesive bonding, with a exposure of at least 1 hour at 121 °C, induced a relaxation of the shot peening residual stresses, which was maximum at the surface and of the order of 30%. There was a concomitant reduction of the fatigue crack formation life of the order of 40% when the specimens were fatigue tested under spectrum loading with a high peak stress. Subsequent sustained exposure at 121 °C had only small unappreciable effects on the residual stress relaxation and crack formation time.

6 [1] [2] [3] [4] [5] [6] [7] [8] [9]

References Meguid, S.A., in Impact Surface treatment, Meguid, S.A ed., Elsevier, 1986, p. 242–249. Potter, J.M., Millard, R.A, in Advances in X-Ray Analysis Vol. 20, McMurdie, H.G. ed., 1977, p. 309–319. Raizenne, M.D., Heath, J.B.R., Sova, M., Benak, T., IAR/NRC Report LTR-ST-2066, Ottawa, 1996. Aerospace Material Specification AMS-S-13165, SAE International, Warrendale, 1997. Military Handbook, MIL-HDBK-5F, Department of Defence, 1990, p. 3-248. Metals Handbook, 9th ed, Vol. 10Am. Society for Metals, Metals Park, 1986, p. 382. Residual Stress Measurement by X-Ray Diffraction - SAE J784a, Society of Automotive Engineers, Warrendale, 1977, p. 65. James, M.R., in Advances in Surface Treatments, Vol. 4 Residual Stresses, Niku-Lari, A. ed., Pergamon Press, 1987, p. 349-365. Sharp, P.K., Clark, G., Report DSTO-RR-0208, Aeronautical and Maritime Research Laboratory, Fishermans Bend, Australia, 2001.

1

IV Peen Forming

2

217

Peen-Forming – A Developing Technique Peter O’Hara Metal Improvement Co Inc, Newsbury, UK

1

Abstract

The application of Controlled Shot Peening to form or correct the shape of components has been in process for over 50 years. The extension of the principles of the Almen strip has been well applied through a range of industries although Aerospace has been the greatest user. Shot Peen-Forming uses the compressive stresses induced by shot peening to alter the stress pattern, magnitude and depth, within a structure to deliberately create a change in product shape. Gentle curves, within the elastic range of the material, are regularly formed to consistent tolerances in to-day’s peen forming facilities. Changes to those induced stresses in magnitude or depth will alter the products profile. Consequently, the introduction of compressive stresses significantly deeper than can be achieved by conventional shot peening will extend the potential of using this forming method on a range of component parts. Lasershotsm Peening achieves that by introducing residual compressive stresses 3 to 4 times deeper than conventional shot peening. Thus opening up the Peen-Forming method to section thicknesses beyond present capability and producing tighter radii more consistently.

2

Introduction

The use of shot peening to forming the shape of components is not a new process but one that even today is finding new areas of application. For over 50 years large and small components using the basic principle that applies to the Almen strip have been formed successfully. An Almen strip builds up a compressively stressed layer at the surface being shot peened. As this material is stretched, the Almen strip curves towards the direction of shot peening to a degree that varies with the intensity of shot peening, coverage and ball size used. Consequently with peen forming the same method applies and with the use of different parameters various shapes can be achieved. Generally cast steel shot is used to achieve the high strain required but other media, glass, ceramic etc, can be used for less demanding forming. Lasershotsm Peening is an alternative technique for the introduction of residual compressive stresses. It has the advantage over conventional peening in that it induces these stresses to greater depths, consequently a tighter radius of curvature can be achieved.

3

History

Shot peen forming was discovered and patented by Lockheed Aircraft Corporation in Burbank, California in the late 1940’s. An engineer on their staff by the name of Jim Boerger was wor-

218 king on the development of a new aircraft. He recognised that with the desire to employ integrally stiffened wing panels as a weight saving measure for this aircraft that this presented some real challenges in creating the aerodynamic curvature. Boerger working with an Almen strip realized that if a curvature could be induced in an Almen strip then it could be induced in a wing panel. Several sample panels were processed and it was proved that by peening on one side you could certainly induce a reasonable curvature. The first full size wing panels were selected but were quite distorted due to lack of machining knowledge on integrally stiffened wing panels. Fortunately, it was possible to bring them back to a reasonably flat shape and then further work proved that the required aerodynamic curvature could be produced. Lockheed Corporation granted a royalty free licence for the process and ultimately, once the patent expired, Boeing, McDonnell Douglas, British Aerospace and subsequently Airbus, as well as many of the smaller aircraft manufacturers employed this process as the most cost effective manufacturing technique for inducing curvatures in fully machined complex aerodynamic panels. Consequently the first aircraft to incorporate the peen forming technique was the Constellation shown in Figure 1, followed many others over the years.

Figure 1: Super Constellation, Concord, Airbus and Boeing

4

Shot Peen Forming

Techniques for the forming of metals are many and varied, each having key features that endear them to specific materials and applications. Shot Peen Forming is a dieless forming process generally performed at room temperature, although certain applications require “warming” of the substrate for maximum benefit. During the process, the surface of the workpiece is impacted by pressure from small, round steel shot. Every piece of shot impacting the surface acts as a tiny hammer, producing elastic stretching of the upper surface. The impact pressure of the peening shot causes local plastic deformation that manifests itself as a residual compressive stress. The

219 surface force of the residual compressive stress combined with the stretching causes the material to develop a compound, convex curvature on the peened side. When curvatures are being formed within the elastic range of the metal, the core of that metal remains elastic with a small, balancing, residual tensile stress. Other mechanical forming processes that require overforming with subsequent springback induce high tensile stress. Although high tensile stress can be minimized by stretch forming techniques, stretch forming is usually not performed on tapered or sculptured sections. Figure 2 demonstrates the type of section most suitable for the peen forming technique. The size, velocity, and angle of impingement of the shot as well as the distance of the wheels or nozzles (the wheels or nozzles propel the shot) from the workpiece are automatically controlled in specially designed machines. Peen forming can be performed with or without an external load applied on the workpiece.

Figure 2: Complex multi-thickness ribbed structure

The non equal compressive residual stresses from shot peening on one side of the Almen strip, causes a degree of curvature, which is measured to give the intensity of controlled shot peening. When the Almen strip is shot peened, the top surface on which the peening has taken place is in a high degree of compression, stretching that surface to cause a change of shape. In so doing, there is also a slight compressive stress induced in the lower unpeened surface, of the Almen strip. Therefore, forming has taken place without the introduction of any tensile stresses at either surface. In order to gain strength, that type of design required a multiplicity of thicknesses of material at the locations where additional strength was necessary, the final assembly being riveted/bolted together. We are now beginning to realise, there was also another benefit in reducing doublers and treblers in panel design. Far less corrosion sites are created when fully machined wing and fuselage panels are used in an airframe construction. Previously corrosion

220 attack was very difficult to detect between the many layers of this material jointed/bonded together. Shot peen formed fully machined panels do not suffer this problem. There was not only the weight penalty to consider with the number of fasteners and joints required, but it was also a very time consuming and expensive assembly process. This was probably acceptable in the days when aircraft were manufactured at the rate of one or possibly two a month, but at today’s rates then the production engineering technology has to be improved in order to reduce the quantity of building fixtures and expense of assembly time required to produce the wings and other major structural components. In the Lockheed application, the reduction of weight from using fully machined panels was the driving force. Single curvature panels were required, and these had integrally machined stringers on them. These stringers then precluded forming by the traditional pressing, rolling, stretch, creep forming or wheeling processes at that time. The Lockheed panels had fully machined integral stringers that overcame one of the characteristics of shot peen forming that a piece of single thickness material does tend to go barrel shaped if it is peened on one side only. This meant it was a relatively simple peen forming process to give the Lockheed wing panels the shape required within the laid down straightness tolerance as the stringers tended to hold the panel true in the spanwise direction. Today, however, whilst there are still peen forming operations carried out on panels that have fully machined stringers an them, it is equally typical to see panels that are fully machined without stringers but still incorporating such features as manhole reinforcements, fuel pump locations, structural thickness differences, where for example there are engines and under-carriage loads to be carried. The limiting factor being the size of material available. This can give a very complex shaped wing panel that does not actually have integral stringers, but the variation in thickness can be as much as a ratio of 1–10 from the thickest to the

Figure 3: Wing skin panel on checking fixture

221 thinnest area of a shot peen formed panel. Typical of this complex peening process are the panels for the modern aircraft shown in Figure 3. When it was decided to use the super-critical wing design, in order to improve efficiency and consequently save fuel, the design usually requires double curvature in the lower wing surface. This is double curvature in a fully machined panel, which means that neither the previous methods of pulling to shape in the build fixture, nor more conventional forming methods can be considered. Shot peen forming is the only method available to create such severe double curvatures consistently to the required accuracy in this type of complex machined panel. In the process of shot peen forming, there are three methods used to create panel curvature. Firstly chordwise curvature is achieved with peening on one surface only, as in Figure 4(a). In this instance, the compressive stress on the peened surface stretches the metal to cause the change in shape. The limitation of this forming method is that it is only within the elastic range that movement can take place, and therefore only shallow curvatures can be achieved. To obtain greater degrees of curvature, strain peening, Figure 4(b), is used in which the component is held in an unidirectional pre-stressed condition and then it is shot peened on the tensile stressed surface. This means that when the component is released from this stressed condition after peening, the compressive stress is greater in one direction than the other, and it is greater in the direction of curvature formed from the pre-stressing process. The third method of peen forming is by shot peening on the edges on both sides of a piece of material at the same time. This gives elongation to the component because of the stretched material on both faces overcoming the resistance of the core in elastic deformation. This is shown in Figure 4(c). Selectively using two or more of these processes, different shapes with different degrees of curvature can be created using shot peening alone. These are only shallow curvatures, which

4a Curving 4b Stretching

4c Stress Peening Figure 4: Peen forming methods

222 makes them particularly suitable for aircraft components, be they fuselage, wing, or tailplane items. They are however very accurate in their shape and because peen forming is carried out cold the reproducibility of forming is very good.

5

Super-Critical Wing Design

The super-critical wing design requires double curvature of the lower surface of the wing. On these designs, the double curvature is in the form of a saddle back shape, that is, the lower wing panel curvature is concave in the spanwise direction and convex in the chordwise direction. The external surface of the panel is peened at comparatively high intensities to initiate the chordwise shape. The coverage and intensity in specific locations are varied to allow for differences in required curvature and component thickness. Subsequent to this peening, the double curvature is created using methods of pre-stressing the component, see Figure 4b. The double curvature shape begins to appear because the high degree of asymmetric stretching results from the localised pre-stressed peening. This higher intensity shot peening means the edges are actually stretched Figure 4c, and the panel gains length at each edge and thus completes the double curvature shape. This double curvature once achieved is verified on a checking fixture, Figure 3. An accuracy of shape of 1 mm or better is quite usual for this type of forming, even with panels as large as 30 metres long. Invariably the double curvature requirement is not uniform, in that the edges have to be stretched at different amounts on one side of the panel to the other. This means that in plan, the panel can change shape and create a slight banana configuration to the panel. This is usually not desirable, and one way that this can be overcome is that the panel, before shot peen forming, is machined with the banana shape in the opposite direction, so that when the forming operation is completed, the panel will be the correct shape for installation on the wing during production. The process is only really limited by the size of material available. At present panels of 30 metres long by 3 metres wide are being produced to these high levels of accuracy that gives acceptable and consistent assembly.

6

Correction of Distortion

Shot Peening can be used not only to form panels to a given shape from an original piece of material that is flat but also can be used to create flat material from a component that has been distorted during manufacture or heat treatment. The raw material will have stresses of varying and generally unknown magnitude from the casting, rolling, forging, heat treatment and any other manufacturing process. If this stock material is then machined with the majority of the metal removed from one side only there will be a dramatic alteration to the stress pattern in the final machined part, and subsequently distortion could result. Add to this the stresses induced by, perhaps, high speed machining with rapid metal removal and these are all factors that can give rise to distortion, Figure 5, which is usually of such a high magnitude that without remedial action, the component would be unsuitable for assembly. A very successful and metallurgically acceptable method of correcting this distortion is by using the principle of shot peening. This method involves the stretching of material on the concave side of the distorted component.

223

Figure 5: Bulkhead fittings for correction of distortion

Figure 6: Industrial applications

Aircraft components are an extremely good candidate for this correction method as has been indicated previously, but the same technique can be used for industrial applications, Figure 6. Components such as steel drive shafts where they might distort during heat treatment, or even large marine connecting rods or crankshafts distorted through a service problem are both suitable for shot peen correction if their slenderness ratio is approximately 20:1 or greater. The same basic principles as with shot peen forming are followed during correction but it is always a very specialised operation for each distorted part. The very fact that in a batch of components machined on the same NC machine there can be significant differences in the degree of distortion does not make the correction task any easier. With particularly complex machined items even the direction of distortion may vary from part to part produced on the same NC machine. Parts distorted in heat treatment, from the same batch in the same oven can show considerable variations and also service damaged items, like the connecting rods mentioned above will

224 also show indeterminate amounts of distortion. The stresses that cause these distortions are combined with the stresses in the material from original casting, forging, rolling etc. to give variation on a piece by piece basis and the need, therefore is for these distorted parts to be treated individually.

7

Peen Forming Machines

Peen forming is usually performed within a cabinet enclosure by automatic machines. When close tolerances are required, some forming may be performed manually by skilled technicians. Two basic types of machines are used, differing only in how the peen forming media is delivered to the part being formed. In nozzle-type machines, compressed air or gravity is used to propel the steel shot to the workpiece. These machines may have as many as 20 nozzles, and each nozzle is capable of delivering 25 kg of shot per minute to a specific location or area of the workpiece. Pressure gauges, control valves and monitors can independently control each nozzle. The nozzle direction is adjustable so that the optimum angle of impingement can be achieved when the workpiece contains surface areas with unusual geometry. Nozzle-type machines can automatically compensate for varying curvature requirements along the workpiece length or width. Thickness variations, cutouts, and reinforcements, as well as distortion caused by machining stresses or heat treatment, can also be compensated for with these machines. Figure 7a shows a nozzle-type, gantry peen forming machine. In this machine design, the gantry, which houses the nozzles, traverses over the workpiece while the workpiece is stationary. Another machine design, using nozzles or wheels, has the workpiece moving through the stationary machine that houses the nozzles, Figure 7b.

Figure 7a

Figure 7b

Centrifugal wheel peen forming is another method by which the shot media is delivered to the workpiece. These machines use electronic controls to regulate rotating speeds of a paddle wheel, which accelerates the shot at the workpiece. A typical wheel can deliver l36 kg of shot per minute. Production-type centrifugal wheel machines may have 6-8 wheels, providing the machine with a capacity to peen form using more than 900 kg of shot per minute. The ability to deliver shot media at a controlled velocity in such large volumes permits higher production rates on these machines than obtainable on nozzle-type machines. Components formed by centrifugal wheel machines are usually of broad, uniform cross section, with all areas accessible to the shot

225 stream. Indexing the position of the shot delivered to the wheel paddle can make minor changes to the shot stream direction.

8

Lasershotsm Peening and Peen Forming

Conventional Peen Forming is used today on many aircraft, however it has limitations primarily in the thickness of metal that can be altered by the peening effect. Ball sizes up to 6mm can be used but the practicality of throwing, containing and grading this media on a routine basis is difficult. The introduction of Lasershotsm Peening, which has the ability to induce depths of compressive stress 3 to 5 times the depth of conventional peening, with virtually no surface roughening, expands the potential of using this method on a wider range of structures. Laser treatment of metals is not a new phenomenon. Laser cutting, shaping and laser thermal methods are some of the developing techniques, which are finding new applications in many industries. Indeed the principle of peening with lasers has been around for several years. What is relatively new is the speed at which this can be applied, i.e. at last it is a production technique rather than a laboratory tool restricted to exotic applications. Today the speed of processing is achieved using a solid state laser employing Nd:glass slabs and phase conjugation. Prototypes of laser peening machines were developed in the 1970’s and there have been later versions over the last 30 years, but were very much restricted to laboratory processing and were not cost effective techniques as they lacked the high repetition rate required for treating parts in a cost effective manner. A laser appropriate for peening at an industrial level requires an average power in the multi-hundred watt to kilowatt range and an energy of around 100J/pulse and pulse duration of 10’s of nanoseconds. The peening effect generated by the laser is fundamentally introducing a shock wave into the surface and this can, at present, only be achieved by means of tamped plasmas, which are generated at metal surfaces by means of high energy density lasers. The improvements in fatigue performance over conventional shot peening are significant in low cycle/high stress application and where damage tolerant surfaces are necessary. The gain in these situations is primarily due to a greater depth of compressive stress although lack of cold work also plays a part. Figure 9 shows the comparable effect.

Depth mm 00

0. 0.5

1 1.0

1. 1.5

-225 20 -450

Residual 60 Stress -675 10 MPa -900 14

-1125 18

Shot Peened Lasershot Peened by MIC-LLNL

Stress Profile Verified by Peenstress

Figure 9: Lasershotsm peening against shot peening (Inconel 718)

226 When applied to Peenforming applications the results have indicated tremendous potential. However, it is not only the greater section thickness that can be formed, it is also the consistency and “cleanliness” of the technique. Trials to date indicate a repeatability of the shape, perhaps because of the greater depth achieved, not seen to date. In addition the lack of shot flying in all directions within an enclosure enables the forming, measurement and control all to occur in the same area and at the same time. Robotic equipment using iterative techniques are being viewed which should enable formed structures to be developed to precise shapes with no decontamination or surface refinement techniques, which today are time consuming and costly. Figure 10: Lasershotsm peen formed sections Material

Thickness

Radius of curvature attainable by processes Laser peen forming

Shot peen forming

Age/creep forming

Al 2024 - T3

16 mm

1.57 m

12.7 m

Cannot form 2000 series Aluminium

Al 2024 - T3

19 mm

2.85 m

20.3 m

Cannot form 2000 series Aluminium

Al 2024 - T3

25 mm

4m

38.1 m

Cannot form 2000 series Aluminium

Al 2024 - T8

25 mm

6.1 m

30.5

Cannot form 2000 series Aluminium

As in Al2000 series

As in Al2000 As desired in Al7000 series series

Al 7000

9

Summary

The conventional Peen Forming technique has developed significantly over the 50 years since the first application on the Constellation commercial aircraft. The process is today employed on most commercial aircraft and is part of the myriad of manufacturing techniques perfected and tailored to produce the complex products on those aircraft. However, although the process is getting close to its limits of performance, the introduction of Lasershotsm Peen Forming will extend the envelop of performance and possibly one day replace the conventional technique

10 [1] [2] [3]

References Shot Peening Applications, Metal Improvement Co Inc, 2002. Barret, C., “Peen Forming” SME Handbook 1984 Dane, B., et al, “Laser Peening of Metals – Enabling Laser Technology”

225

Optimising the Double-Sided Simultaneous Shot Peen Forming Reiner Kopp, Jörgen Schulz Institute of Metal Forming, University of Technology Aachen, Aachen, Germany

1

Introduction

Shot peen forming is a flexible pressure forming procedure. According to the velocity and mass of the balls striking the components we can induce convex as well as concave curvatures. Hitherto, it has been solely usual to bring both methods of curvature generation to bear one after the other on both sides of a particular component. This study describes the new technology of double-sided shot peen forming. The manufacturing of three-dimensional structural parts may serve as an example of how to identify the mechanisms which come into play during this double-peen technology in order to optimise the process as to both effectiveness and reproducibility. Beside the experimental series of tests, the simulation of shot peen forming, which due to the many difficulties arising from the basic principle is still in its infancy, is being further developed, so that in future we will be able to apply this kind of simulation to rapid advance planning of the process strategy.

2

Principles of curvature Generation

2.1

Single Sided Convex and Concave Shot Peen Forming

From the literature we are familiar with various models of describing the two possible types of curvature produced by single-sided shot peen forming, for instance the plane model of elementary theory [1] or the slip line model [2]. The former is only good for cases of convex forming however, the latter mainly for plain strain and for rigid ideal plastic material behaviour, which represents a considerable limitation. To investigate the mechanisms which bring about a convex or a concave curvature simulations of individual ball hits at various ball velocities were carried out explicitly using the FEMSoftware LS-DYNA3D (Figure 1). This model will be used in the further course of the experiments for the double-sided hitting of balls, one on the top side and one on the bottom side. In what follows we shall first present and discuss the results only. A further detailed discussion of the problems involved in simulating this technology will follow later in this paper. Figure 2 shows the equivalent strain as well as the nodal displacements brought about in the sheet underneath an individual hit at various ball velocities. As expected, these are greater on the upper side of the component. The equivalent strain is reduced the deeper is the hit, and in the region of the bottom side it becomes larger again. This indicates that there is a dynamic bending effect with a short-term overextension of the fibres at the time of contact with the ball.

228 The zones on the bottom side which are affected by plastic strain increase more noticeably at higher ball velocity than are those on the top side, which to a greater extent is also true for the nodal displacements in the sheet (x-direction). At higher ball velocities the dynamic, short-term bending is more pronounced, which produces a greater plastic extension on the bottom side than on the top, and therefore also a concave curvature. Side A: Ball diameter dA = 6.4 mm Mass mA = 1.07 g Side B: Ball diameter dB = 4.0 mm Mass mB = 0.26 g

y z

x Model: 4892 Nodes, 5244 Elements, Element length in forming zone 0.25 mm Material law: Elasto-plastic, Balls rigid Alloy: AlMg3 Flow curve at dj/dt=300 s-1

Software: LS-DYNA3D Geometry: Symmetry in x- and z-direction, Sheet thickness 3 mm, Work piece diameter 25 mm

Figure 1: Simulation of individual hits, single-sided and double-sided

5 m/s

10 m/s

Effective Plastic Strain 15 m/s

A

V

20 m/s 0.010 0.083 0.067 0.050 0.033 0.002 0.000

0.010 0.083 0.067 0.050 0.033 0.002 0.000

Nodal displacements ,x in mm

Figure 2: Effective strains (top) and displacements (bottom) in the sheet (x-direction)

2.2

Double-Sided Shot Peen Forming

The simulations described above were extended for double-sided shot peen forming. A second ball with the properties described in Figure 1 strikes the bottom side of the sheet, doing so however after a short interval of time so that positive countervailing effects caused by both balls striking at the same time may be avoided. A further simplifying assumption concerns the two hit locations, which in reality would hardly ever lie on one axis. A displacement to one side would, however, mean that one would not be able to assume any quarter symmetry. The results are shown in Figure 3.

229 It becomes clear that due to the effects of the balls double hits cause both sides of the component to become sufficiently plastically deformed, so that the high velocity of the balls on the top side will not be necessary to bring about plastification of the bottom side of the component. It accordingly appears from this that if we want to induce a concave curvature in a component, both single-sided and simultaneous double-sided shot peen forming may be used. Effective Plastic Strain

AV

,x in mm

Nodal displacements

0.010 0.083 0.067 0.050 0.033 0.002 0.000

0.010 0.083 0.067 0.050 0.033 0.002 0.000

x

Figure 3: Results of simulation of double hits, vB, upper ball = 13 m/s, vB, lower ball = 10 m/s

3

Applications and Experiments

In the field of the air- and spacecraft industry shot peen forming has been successfully used for many years to form NC-milled components such as aeroplane wings, stringer-strengthened fuselages (Alpha Jet, Airbus) or for segments of the water tank of the ARIANE 4 [3,4,5]. As described in [6]the 1/8 segments of the bottom of the fuel container for the European ARIANE 5 rocket have been shot peened from both sides for several years now, albeit not simultaneously. The numerically controlled shot peening apparatus installed at the Institute of Metal Forming has two independent peening systems, an Injector-Gravitation-Peening system (Side A, ball diameter of 6.4 mm) as well as an airpressure peening system (Side B, ball diameter of 4.0 mm). The conceptual construction accordingly provides for concave curvature of the components in the case of Side A and convex curvature in the case of Side B.

Airpressure System: Ball diameter: 4.0 mm Ball mass: 0.26 g

Injector-Gravitation System: Ball diameter: 6.4 mm Ball mass: 1.07 g

Figure 4: Configuration of the apparatus for double-sided simultaneous shot peen forming

3.1

Pre-Tests

In shot peen forming (given the ball dimensions noted above) it is the direct peening parameters - the mass flow dm/dt and the shot velocity vB as well as the derived, indirect peening parameter

230 represented by the degree of shot coverage A* - that exercise a decisive influence on the result of the peening. Whereas the velocity vB is essential for the direction of the curvature generated, the degree of coverage A* is primarily a measure of the amount of curvature generation. In double-sided shot peen forming these peening parameters have to be co-ordinated for both sides being peened, since they influence one another. Thus, it is not possible to determine the optimal parameters individually by means of pre-tests of the respective separate peening treatments. Figure 4 shows the configuration of the apparatus for double-sided shot peen forming.

3.2

Optimising the Peening Parameters

We thus undertook pre-tests, where the peening pressure and mass flow and thus, too, the peening parameters velocity and degree of coverage were varied on both sides and the curvatures generated subsequently measured. To place these results in a mathematical context, two sine functions were selected in accordance with the considerations outlined above and superimposed in accordance with the form é æ (vK,A + a2 ) × p ö ù æ (vK,B + b2 ) × p ö ÷ ú + AB* × b1 × sin ç ÷ f = - ê A*A × a1 × sin ç ç ÷ú ç ÷ a b3 êë 3 è øû è ø

By means of an optimisation tool [7] the six correction factors a1 to a3 and b1 to b3, three of which respectively influence the form of one sine function per side, were optimised, until a minimum error quotient of calculated and measured values was obtained. Using this mathematical description, the peening parameters can be optimised and co-ordinated with one another.

3.3

Peening of Demonstrator Components

Using the optimised peening parameters we were now able successfully to form a number of demonstrator components. By way of example, a number of contours were produced for a seat, such as a wave and a saddle contour (Figure 5). Simultaneous double-sided shot peen forming allows us to complete-

Figure 5: Various 3D-contours, produced by means of double-sided simultaneous shot peen forming

231 ly plastically deform the areas to be stretched in small local limits, which means that varyingly stretched areas such as we have in the case of a saddle contour may be fairly close to one another without strongly influencing one another.

4

Simulating the Shot Peen Forming Process

4.1

Basic Considerations due to the FEM of Shot Peen Forming

Due to the complex processes involved in shot peening and the great number of parameters affecting the process, it is not possible at present to simulate this process sufficiently precisely by means of the Finite Element Method. The number of balls, the interactions occurring between the balls and the fact that the actual forming process can take several minutes, whereas the individual ball hits occur within a period of 10–4 to10–5 s, cause an extremely high computing effort. In view of the work piece description the large dimensions of the components require a very rough mesh. In order, however, to be able to portray the ball impressions on the peened surface the lengths of the element edges should be very small. To describe the material behaviour simplifying material laws where the material is regarded as a continuum are usually used. Discontinuities present in the material, influenced by the "forming history" of the initial material, instances of elastic stress caused by clamping or preloading and the flow curve description that has to take account of the high strain rates which occur, have a great inluence on the result. Nevertheless, the FEM can still make a contribution to a greater understanding of the mechanisms operating during shot peening.

4.2

Simulation

In simulations the ball hits are first simulated by means of an explicit program module. There then follows an implicit static calculation of springback [8]. For the reasons explained above, the simulations are limited to a fairly small component (Figure 6: Simulation of multiple hits (upper left) and real component (upper right)). Real tests were carried out on sample strips with the dimensions 100 × 20 × 2 mm3 and whose parameters attracted ball velocities and numbers for the simulations. The final curvature achieved after spring-back is greater when simulating than it is in reality. One of the main reasons for this is that the flow curve description applies to a considerably lower strain rate of 300 1/s than that found in reality. Figure 6 shows the equivalent plastic strain and the progression of the residual stress in the sheet. On both surfaces compressive stresses occur, the tensile stresses in the cross section being displaced in the direction of the bottom side of the component.

232 Model: 28101 nodes, 23460 elements, Element length in forming zone 0.5 mm Material law: Elasto-plastic, balls rigid Alloy: AlMg3 Flow curve at dj/dt=300 s–1

Simulation

Real sample part

Equivalent Plastic Strain Av

Residual Stress Ix in N/mm2 50.0 33.3 16.7 0.0 - 16.7 - 33.3 - 50.0

0.30 0.25 0.20 0.15 0.10 0.05 0.00

5

Software: LS-DYNA3D Geometry: 100 × 20 × 2 mm3 Tools: 70 balls diameter 6,4 mm, vB = 17 m/s 140 balls diameter 4,0 mm, vB = 6.1 m/s

Conclusions

With convex curvature only a small layer in the region of the surface is plastically deformed, whereas in cases of concave curvature it is the whole cross section. Simulations of individual hits show that due to a short term, superimposed deflection effect at the time of the hit with greater kinetic energy of the shot brings about greater plastic deformation on the side facing away from the ball. These hit effects can likewise be substituted by means of a peening treatment on both sides. A plastic deformation of the whole cross-section takes place in this case at lower kinetic ball energies than it would be the case with single-sided peening. The degree of shot coverage and the ball velocity have a determining effect on the peening result on both sides and have to be co-ordinated with one another. Knowing the optimised peening parameters, it was possible to manufacture reproducible 3D-contours. The overall plastic deformation brought about by double-sided simultaneous peening makes narrow curvature radii possible, which may be varied as to form and direction in local areas lying close to one another. Simulations are presented, which use several hundred balls to produce deformation energy. Using this concept the first deformation processes can be simulated by using fundamental mechanisms of ball hits, which allow us to reach conclusions about plastic strains and the state of residual stress.

6

Acknowledgements

This study was carried out with the support of Deutsche Forschungsgemeinschaft DFG in Bonn, Germany. We would also like to thank the Kugelstrahlzentrum Aachen GmbH for their co-operation on the shot peening trials.

233

7 [1] [2] [3]

[4] [5] [6] [7] [8]

References R. Kopp, Ein analytischer Beitrag zum Kugelstrahlumformen, Bänder Bleche Rohre 1974, 12, 512-522 R. Hill, The Mathematical Theory of Plasticity, Oxford, 1950 R. Meyer, H. Reccius et al., Shot Peen Forming of NC-machined Parts with integrated Stringers using large Balls, 3rd International Conference on Shot Peening, Garmisch-Partenkirchen, 1987, 327-334 R. Kopp, H.-W. Ball, Recent Developments in Shot Peen Forming, 3rd International Conference on Shot Peening, Garmisch-Partenkirchen, 1987, 297-308 H.-W. Ball, Beitrag zur Theorie und Praxis des Kugelstrahlumformens, 1989, DGM R. Kopp, F. Wüstefeld, W. Linnemann, High Precision Shot Peen Forming, 5th International Conference on Shot Peening, Oxford, 1993, 127-138 R. Kopp, S. Posielek, CAOT - A Computer Aided Optimisation Tool applied on Metal Forming Processes; CIRP Annals, 1999, Vol. 48/1 R. Kopp, J. Schulz, Flexible Sheet Forming Technology by Double-sided Simultaneous Shot Peen Forming; CIRP Annals, 2002, Vol. 51/1

234

Impact Metal Forming Helmut Reccius IHR, Gummersbach. Germany

1

Introduction

2

Tool Design

235 2.1

2.2

Impact Body

Spring

236 2.3

Spring Tension System

237

3

Applications

238

4 [1]

References Hornauer, K.-P., Untersuchungen zur Umformung von Bauteilen durch Kugelstrahlen, Dr.-Ing. Dissertation, RWTH Aachen 1982

239 [2] [3] [4]

H. Reccius, D. Endemann, K.-P. Hornauer, Sheet forming by producing the coverage with a simultaneous working system of balls, ICSP5, Oxford University 1993 R. Kopp, F. Wüstefeld, W. Linnemann, High precision shot peen forming, ICSP5, Oxford University 1993 Guy Levasseuer, Peen forming feature possibility for wing skin fabrication, Metal Finishing News, Vol. 3, May issue 2002

240

233

V

Corrosion and Fretting

234

243

Mechanisms and Modelling of Cracking under Corrosion and Fretting Fatigue Conditions Eduardo R. de los Rios Department of Mechanical Engineering, University of Sheffield, Sheffield, UK

Corrosion Fatigue It is well established that crack growth proceeds at a higher rate in aggressive environments than in air. This complicates even further the prediction of fatigue life, where safety factors are required to account for these uncertainties. However, considering that these correction factors are derived from projected service conditions (which are shortened in time for laboratory testing), make questionable the accuracy of the predictions, although designs are evaluated through service simulation tests. To increase the accuracy of the predictions it is necessary to develop models of corrosion fatigue that incorporate the main aspects of the physical phenomena. These aspects of the environmentally assisted corrosion fatigue of metals in aqueous solutions involve electrochemical processes which include anodic and cathodic reactions. The former is related to an anodic dissolution mechanism and the latter is associated with hydrogen embrittlement. Several possible corrosion fatigue mechanisms related to anodic dissolution have been suggested, including pitting induced crack initiation and short crack growth, film-rupture, dissolution of slip bands at the crack tip and grain boundary oxidation. Of these, pitting corrosion is the most damaging and will be discussed next.

Pitting corrosion Pitting is the result of electrochemical reactions in local cells on the surface of a metal [1], At the site of a pit, corrosion occurs at the local anode, caused by electrochemical differences between one site and its surrounding area at the metal-liquid interface. Pits are usually found at the origin of fracture in industrial machinery parts as well as on test specimens. This implies that pit initiation triggers fatigue crack initiation. Therefore, the quantitative evaluation of pitting is very important in the prediction of corrosion fatigue life. Pitting is a common phenomenon occurring above the pitting potential (Ep) for low carbon steel, medium carbon steel and Al-alloys. However, pitting still occurs below Ep in some Alalloys [2]. In the case of a low stress level and a low loading frequency, pits may play a major role in short fatigue crack initiation and propagation. Akid and Miller [3] observed short fatigue crack initiation at the corner of pits. When pit-depth together with the initiated short crack depth exceeded a critical length, short fatigue cracks propagated at increasing crack growth rates. An example of fatigue cracks initiated at pits is shown in Fig. 1

244

Figure 1. Fatigue crack initiated from a corrosion pit in a high strength steel. Plastic replica after Donohoe [4]

A model of corrosion fatigue was developed by de los Rios et. al [5] which considers the effect of hemispherical pits in the initiation and propagation of fatigue cracks. Experimental evidence has shown that the diameter of the major pit, which eventually created a dominant 1 fatigue crack, increased as a function of time (t), i.e. t 3 .

H c − H o = B( t − t o )

1

(1)

3

H c − H o = B[( N − N o ) / f ]

1

(2)

3

Where to (incubation time), Hc critical pit size, Ho (mean inclusion size) and coefficient B are determined experimentally. Then the number of cycles for pit growth can be calculated as: 3 ⎛ ⎛1⎞ ⎞ N pit = f ⎜ t o + (H c − H o )3 ⎜ ⎟ ⎟ ⎜ ⎝ B ⎠ ⎟⎠ ⎝

(3)

Stress concentration around an inclusion or a pit According to Weiss, Stickler and Blom [6], the field of stress concentration around an isolated micropit was found to extend to approximately twice the radius of the micropit or less. A similar analysis can be used to estimate the effective ranges of stress concentration for a pit. Considering the local pit stress reduces (by a power of the r −2 form) for a semi-infinite body, Equation (4) is proposed to evaluate the stress concentration as a function of distance from the pit centre. 2

⎛H ⎞ K t = (K to − 1)⎜ c ⎟ + 1 ⎝ 2r ⎠

(r ≥ H c / 2)

(4)

Consequently, the local stress surrounding the pit, σ lc , as a function of the distance from the original pit centre will take the following form:

245 2 ⎡ ⎤ ⎛H ⎞ σ lc = σ ⎢(K to − 1)⎜ c ⎟ + 1⎥ , r > H c / 2 ⎝ 2r ⎠ ⎢⎣ ⎥⎦

(5)

This stress is then used to calculate the number of cycles for crack propagation employing a Paris type crack propagation law, i.e.:

da = A(φ)n dN

(6)

Where φ is the crack driving force, e.g. crack tip plastic displacement (CTPD), ΔK, etc. Thus

N crack =

af

da

ao

A(φ)

∫

n

(7)

And total life, for a condition where pit formation is the dominant environmental factor, is: N T = N pit + N crack

(8)

Hydrogen embrittlement For some metals hydrogen embrittlement plays a major role in cathodic reaction-related corrosion fatigue. There are several possible mechanisms put forward to explain hydrogen effects including:(1) a critical hydrogen concentration which induces decohesion, (2) hydrogen environment enhanced crack tip plasticity, (3) corrosion product and roughness induced wake closure, and (4) hydrogen trapping induced intergranular short crack initiation and growth [7,8]. For subcritical fatigue crack propagation, the reversible local plastic flow process may be the dominant mechanism. In the case of hydrogen assisted fatigue crack growth, hydrogen would affect the plastic zone at the crack tip, consequently the effect of hydrogen on the reversible plastic flow has to be considered when applying the crack tip decohesion model. Beachem [9], on the basis of observations made after applying various degrees of deformation to a material degraded by hydrogen (particularly on the lowering of the torsional flow stress in a 1020 steel) proposed a hydrogen assisted cracking theory in which the role of hydrogen is to augment dislocation motion. The 1020 steel case seems to be associated with surface damage [10]. Enhanced dislocation motion by hydrogen is now definitely established, with the extent of the softening being dictated by enhanced screw dislocation mobility, enhanced dislocation injection at surfaces, and the promotion of shear instabilities. Lynch [11] proposed that some environments including hydrogen, promote crack tip dislocation nucleation and high strain localization in aluminium and iron based alloys. Lankford and Davidson [12] demonstrated that higher crack tip opening strains are responsible for the rapid growth kinetics of small fatigue cracks in aluminium alloys stressed in moist air, at least for cracks in single grains.

246 It should be emphasized however that the suggestion of a crack tip slip-softening mechanism is based on the observation of hydrogen lowering the yield strength in ductile materials, while for high strength steel, hydrogen has little effect on yield strength, which is controlled by other microstructural features. For lower strength steels (σy< 700 MPa) varying effects have been found. H.Matsui, S.Moriya and H.Kimura [13] found a decrease in yield strength after electrolytic charging at large hydrogen fugacities. Similar results were also reported by Petch [14] for steels with varying carbon content up to 0.60 pct, by Lee et al [15] for spheroidized 1090 steel, by Ciaraldi et at [16] in an Al-Zn-Mg alloy. However, work hardening as opposed to softening has also been observed in many alloys.

Modelling hydrogen-assisted short crack growth The model of Navarro-de los Rios [17] as extended in [18] was further expanded to incorporate the effect of hydrogen in short fatigue crack growth. The equations which describe the environmental-assisted fatigue process are presented first, and discussed later. The equilibrium equation of all the forces in the crack system is: 1

∫

−1

f (ζ′) P (ζ ) dζ ′+ =0 A ζ −ζ′

(9)

The crack tip open displacement is:

∫

c

CTOD = b f (ζ′) dx

(10)

a

The crack growth rate is: da = A (CTPD ) n dN

(11)

The diffusion equation for hydrogen including hydrogen trapping is: ∂C L ∂C T + = DL∇ 2CL ∂t ∂t

(12)

where CL is the lattice hydrogen concentration, CT the trapped hydrogen concentration, and DL the concentration independent lattice diffusivity. P(ζ) is a function of hydrogen concentration, C(x,t), which is a complicated time-dependent function, and cannot be expressed in a simple expression. Therefore the integral equilibrium equation (9) has to be solved using the diffusion equation 12. Numerical analyses were carried out on a computer using a one-dimensional explicit differential equation with moving boundary. The moving boundary positions are determined by calculating the crack increment cycle by cycle. Further details of the model are given in reference [19]. Theoretical predictions and the experimental data are compared in Table 1. Because subcritical crack

247 growth produces a fresh surface after every cycle, it was assumed that the concentration of hydrogen at the crack tip is always equal to the initial concentration C0 of 0.65 atm ratio. Table 1. The comparison of the experimental fatigue life with the theoretical predictions for the different stress ranges. Al-Li 8090, frequency = 10 Hz, R = 0.1 Number of cycles to failure Nf Experimental data 420,900 442,000 219,000 183,000 109,000 109,000

Stress range 144 MPa

180 MPa 207 MPa

Prediction 407,000

223,000 115,000

Cracks and hydrogen It has been believed for sometime, and also proven experimentally [20], that hydrogen induced cracking (HIC) initiation sites correspond to the location of the highest hydrostatic stress σ max . However, according to a recent investigation using SIMS (Secondary Ion Mass h Spectroscopy) under different ratios of I/II mixed mode loads [21], there are two hydrogen accumulation peaks ahead of the crack tip, i.e. hydrostatic stress induced hydrogen accumulation peak C1H located at the crack tip elastic-plastic boundary and dislocation induced (i.e. plastic strain induced) hydrogen accumulation peak C 2H located at the point of highest equivalent plastic strain, which in turn correspond to the point of highest dislocation density very close to the crack tip, see Fig.2. 4 .5

4 .0

α =0ο α =30ο α =45ο α=60ο

3 .5

M

C /C

O

3 .0

2 .5

2 .0

1 .5

1 .0

0 .5 0

500

1000

1500

D IS T A N C E A H E A D O F N O T C H (μ m )

Figure 2. Distribution of Hydrogen concentration ahead of slit (crack)

248 In terms of these two hydrogen accumulation peaks, the site of hydrogen-induced cracks (HIC) would depend on which peak, in combination with its local maximum normal stress, attains the critical state first. Therefore, for a mode I slit, the hydrostatic stress induced hydrogen accumulation is dominant, and the dislocation induced hydrogen accumulation is rather weak. In this case, the C1H peak plays a more important role than the C 2H peak, and consequently the HIC initiation site is determined by the location of the C1H peak (elasticplastic boundary) Fig. 3(a). For a combined I/II mode loading case, however, the situation is more complicated since: (i) the HIC initiation sites are at an angle θ with the slit direction and (ii) the HIC initiation site may relate to either C1H or C 2H depending on the ratio of KII/KI as well as on the loading process. Indeed, in slow strain rate (SSR) tests, and when KII/KI > 1, cracks did not form ahead of the notch tip but they initiated at the notch surface where the equivalent plastic strain is maximum (see Fig. 3(b)). Conversely, when KII/KI < 1, HIC initiation did first appear some distance ahead of the slit tip. However, in constant load (CL) tests, the initiation site of HIC is always ahead of the notch tip and associated with the position of the C1H peak as shown in Fig. 3(c).

Figure 3. HIC initiation sites at: (a) mode I notch tip; (b) combined I/II mode (KII/KI = 3.1) slit tip (in SSRT tests); (c) combined I/II mode (KII/KI = 3.1) slit tip (in CLT tests).

Fretting Fatigue Fretting occurs when two surfaces are in contact and subjected to oscillatory tangential movement. The repeated shear stresses that are generated by friction during the relative motion causes surface damage known as fretting wear. Moreover, surface cracks are likely to initiate in the fretting wear zone. This will eventually leads to crack growth and can result in a significant decrease in fatigue life of a material. When fretting damage is associated with decreased fatigue performance, the phenomenon is termed as fretting fatigue. Such a fatigue failure mechanism is extremely common in aircraft structural lap joints and turbine/disk contacts.

Characteristic of fretting fatigue The mechanism of surface damage that can cause crack nucleation is complex and difficult to study. Surface damage due to wear occurs when the mating surfaces under normal load are

249 subjected to relative movements. Damage begins with local adhesion between the interface and progresses when adhered particles are removed from the surface. The formation of a fretting scar is typically smaller than a millimetre in depth [22] and they can act as micronotches, raising locally the stress level and providing a site for an emerging fatigue crack. The general characteristic of fretting fatigue damage is shown in Figure 4.

Figure 4. Cross-section of a fretting fatigue crack. Courtesy of R.B. Waterhouse

Fretting fatigue cracks grow initially at a particular angle relative to the surface which depends on the frictional and applied stresses (phase I). As the crack propagates inwards, the contact surface stress decrease. This may lead to crack arrest or if static or alternating stresses exist in the bulk of the material, the crack will change direction and run perpendicular to the surface as a mode I crack. The depth at which this occurs depends on the magnitude of the surface shear stresses, which depend on the coefficient of friction and the normal contact stresses.

Fretting fatigue of aircraft materials: experimental and numerical study Fretting fatigue cracks in mechanical joints are an ongoing problem in airframes. These cracks initiate under conditions which involve three-dimensional contact stresses between the fastener and panels and between panels themselves. Depending on the fastener system there are several locations where fatigue cracks may initiate. For example, in systems with a low or medium clamping force, the crack initiation site may occur in the minimum net section at the edge of the fastener holes (in the form of a part-elliptical corner crack), or at the intersection of the countersink profile and the hole. Such crack initiation sites lead to a shorter fatigue life of the joint. For fastener systems with high clamping force, the fretting mechanism is dominant and cracks invariably initiate at some distance away from the fastener hole. This crack initiation site results in a substantially increased fatigue life compared to cracks initiated at the minimum net section along a row of rivet holes. Therefore, a comprehensive study of fretting fatigue, with associated modelling, is required for the development of a damage tolerance methodology leading to accurate predictions of fatigue life and thence designs optimisation of high-load transfer joints.

250 Fretting fatigue experiments

10

7

10

6

10

5

10

10

4

unpeened shot peened rough shot peened polished 3

0

20

40

60

80

100

120

Fretting Fatigue Life, cycles

Fretting Fatigue Life, cycles

Fretting fatigue tests using Al 2024-T351 specimens and steel and aluminium contact bridge pads, were performed with the axial load amplitude 100 MPa and various values of normal pressure covering the range 10-120 MPa. In all tests fully reversed cyclic axial load was applied with a sinusoidal waveform of 20 Hz frequency. Full details of the experimental set up and test procedure are given in Ref. [23]. The influence of normal load on the fretting fatigue life is shown in Figs. 5 and 6. These figures shows that: (i) there is a considerable reduction in fatigue strength due to fretting; (ii) the fatigue life reduces as the contact pressure increases to a critical value of the normal load; (iii) above this normal load further increase in the normal pressure tends to increase fatigue life. Shot peening significantly increases the fretting fatigue durability, particularly at low contact stresses. It should be noted that the durability of peened but rough specimens in the medium range of contact pressure is higher than that of polished specimens. This correlates with other work reported in the literature, indicating that there are two beneficial effects of shot peening in fretting fatigue. One is the compressive residual stresses, the other being surface roughness [24] Aluminium bridges show a similar trend with normal pressure as for steel bridges. However, the critical normal pressure, (giving the shortest life) is achieved with a normal pressure of approximately 40 MPa using aluminium bridges while it is about 80 MPa for steel bridges. Furthermore, fatigue life at normal pressures above the critical point are higher for aluminium bridges than for steel bridges, while the opposite is true below the critical normal pressure for aluminium bridges (see Fig. 6).

Al 2024-T6 σ xx =100MPa, R=-1 Bridge span=16.5m m 10

6

10

5

Steel Bridge Aluminum Bridge 10

4

0

Contact Stress, MPa

Figures 5 and 6. Fretting fatigue life as a function of contact pressure

20

40

60

80

100

Contact Pressure, (M Pa)

120

251 Stress Analysis To predict the initial growth direction and site of fretting fatigue cracks, the fretting damage parameters need to be quantified. To examine the stress distribution on the specimen during cyclic loading, two-dimensional elastic finite element (FE) analysis was carried out under plane strain condition using the commercial finite element package ANSYS. The stress distribution over the contact interface between the specimen and the bridge is shown in Fig. 7. 60 0

20 0

A l 2 0 2 4 -T 6 σ a ,m a x = 1 0 0 M P a R = -1 σn = 6 0 M P a

600

400

S tre ss, s (M P a )

0 -20 0 -40 0

S tee l A l

-60 0 σx x

-80 0

C o ntact R e gio n

τx y

Range of stress/2 (MPa)

40 0

Al 2024-T6 σ a,max=100MPa R=-1 σ n=60MPa

Steel

Al

Δ σ xx/2

Δ τ xy/2

200

0

-200

-400

Contact Region

-1 00 0 7 .0

7 .5

8 .0

8 .5

9 .0

D ista n ce fro m ce nte r, x (m m )

9 .5

7.0

7.5

8.0

8.5

9.0

9.5

Distance from center, x (mm)

(a) (b) Figure 7. Stress distribution over the contact area obtained from the numerical analysis

Fig. 7(a) presents the tangential and shear stress distributions along the contact area when the axial cyclic stress achieved its maximum value. It shows that these stresses have peak values at only one of the edges of the pad (leading edge). Fig. 7(b) shows the distribution of the stress ranges during cyclic axial loading. It shows that the range of the tangential and shear stress achieve maximum value at the outer edge of the pad. Comparisons with observations made from the test pieces, confirm this site as being the primary crack initiation location. From the FE results, the variation of these stresses with angular locations below the leading edges of the bridge feet at various depths, could be investigated. The result of this investigation is summarized in Fig. 8 where the orientation of peak stresses are compared with the initial direction of crack growth obtained from the experiments. It shows that initial crack directions for both contact materials coincide with the direction the maximum value of the tangential stress range, Δσθθ,max.

252 200

Al

Al 2024-T6 σ a,max =100MPa R=-1

180 160

Steel

Δ σ θ θ ,max

σ θ θ ,max

Δ τ θ θ ,max

Angle, θ (degree)

Exp. 140 120 100 80 60 40 20 0

20

40

60

80

100

120

Contact Pressure (MPa)

Figure 8. Variation of initial crack growth direction with normal pressure.

Fractographic analysis The mechanisms of crack initiation and failure are different between peened and unpeened conditions in fretting fatigue. Fig. 9 summarises these differences.

Unpeened specimens Low contact pressure The initiation of fretting fatigue cracks in unpeened specimens was always in the form of 3D semi-elliptical surface cracks. At low contact pressure only one crack initiated and propagated to failure

High contact pressure At high contact pressure, on the other hand, there was multicrack initiation which leads to coalescence and to the early formation of a through section crack

253

Peened specimens Low contact pressure In peened specimens, failure was associated with a single subsurface crack originating from the corner and propagating as a quarter-elliptical crack

High contact pressure Cracks initiated both at the surface and subsurface. However, all surface cracks were substantially retarded or arrested by the effect of the compressive residual stress left after shot peening 500μm ⎯ Figure 9. Crack initiation points in fretting fatigue

References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14]

R. H. Brown and R. B. Mears, Ind. Eng. Chem., 33, 1941, 1002. R. Akid: Ph. D. Thesis, University of Sheffield. UK. 1986. R. Akid and K. J. Miller, Environment Assisted Fatigue (Ed.: P. Scott, Mechanical Engineering Publications, London, 1990, 415–434. C. J. Donohoe. Ph.D. Thesis, University of Sheffield 1999. E. R. de los Rios, X. D. Wu and K. J. Miller. Fatigue Fract. Engng. Mater. Struct. 19, 1383–1400, 1966. B. Weiss, R. Stickler and A. F. Blom, Short Fatigue Cracks, ESIS 13, (Eds.: K. J. Miller and E. R. de los Rios), Mechanical Engineering Publications, London. 1992, 423–438. J. P. Hirth, Effects of hydrogen on the properties of iron and steel, Met. Trans. A, vol. 11A, 1980, 861–890. H. P. Gangloff, and R. P. Wei, Proceedings of the Second Engineering Foundation International Conference, Small Fatigue Cracks, 1986, 239. C. D. Beachem, Met. Trans., vol. 3, 1972, 437–451. R. A. Oriani and P. H. Josephic, Acta Met., vol. 22, 1974, 1065–1074. S. P. Lynch, Metals Forum, vol. 2, 1979, 189–200. J. Lankford and D. L. Davidson, Fatigue Crack Growth Threshold Concepts, (Eds.: D. Davidson and S. Suresh) Warrendale, PA: TMS-AIME, 1984, 447–463. H. Matsui, S. Moriya and H. Kimura, Mat. Sci. Eng., vol. 40, 1979, 207–216; Mat. Sci. Eng., vol. 40, 1979, 217–225; Mat. Sci. Eng., vol. 40, 1979, 227–234. N. J. Petch, Phil. Mag., Vol. 1, No. 4, 1956, 331–337.

254 [15] T. D. Lee, T. Goldenberg, and J. P. Hirth, Met. Trans. A, vol. 10A, 1979, 439; Met. Trans. A, vol. 10A, 1979, 199. [16] S. W. Ciaraldi, J. L. Nelson , R. A. Yeshe and E. N. Pugh, Effect of Hydrogen on Behavior of Materials, (Eds.: I. M. Beinstein and A. W. Thompson), 1980, 437–447. [17] A. Navarro and E. R. de los Rios (1992), Proc. R. Soc. Lond. A 437, 375–390. [18] Zuyu Sun, E. R. de los Rios and K. J. Miller (1991), Fatigue Fract. Engng. Mater. Struct., 14, 277–291. [19] E. R. de los Rios, Zuyu Sun and K. J. Miller (1994), Fatigue Fract. Engng. Mater. Struct., 17, 1459–1474. [20] R. A. Page and W. W. Gerberich, Metall. Trans. 13A, 1982, 305–311. [21] Gao Hua, Cao Weijie, Fang Changpeng and E. R. de los Rios, Fatigue Fract. Engng. Mater. Struct. 17, 1994, 1213–1220. [22] D. A. Hills and D. Nowell, Mechanic of fretting fatigue, Kluwer Academic Publisher, 1994. [23] U. S Fernando, G. H. Farrahi, and M. W. Brown, Fretting Fatigue, Mechanical Engineering Publications, London, England, 1994, 183–195. [24] R. B. Waterhouse, Fretting Fatigue, ESIS 18 (Eds.: R. B. Waterhouse and T. C. Lindley), Mechanical Engineering Publications, London, 1994, 339–349.

255

Influence of Shot Peening on Stress Corrosion Cracking in Stainless Steel Jürgen Kritzler Metal Improvement Company, Inc.; Unna, Germany

1

Abstract

Shot Peening is usually considered as a sensible method to prevent cracking related to fatigue or stress corrosion. In this paper, I will present the results of static stress corrosion tests of three austenitic materials. The specimens for tensile testing were made from the materials 1.4571 (X6CrNiTi17-12-2), 1.4541 (X6CrNiTi18-10) and 1.4462 (X2CrNiMoN22-5-3). The different shot peening parameters were selected by the Peenstresssm Software. The calculated and measured residual stress profiles are very close.

2

Introduction

Controlled shot peening is an established process that is used to increase the resistance of metal parts to fatigue failure in a great variety of industries, including aircraft, aerospace, automobile, heavy equipment, power generation, petrochemical, etc. The residual compressive stresses introduced by shot peening have a major beneficial effect not only upon metal fatigue but upon all

Material

Medium

Composition Heat Treatment

Composition

Corrosion

Microstructure Surface

Electrical Potential

SCC Fatigue

Temperature

Corr. Fatigue

Stress Operating Stress Compressive Residual Stress

Residual Stress Stress Rate

Figure 1: Influence of critical states on material failures

256 the tensile stress related modes of failures such as stress corrosion cracking, corrosion fatigue, thermal fatigue and fretting fatigue. To these can be added the purely mechanical effects of a peened surface that can reduce galling, improve lubricity, close porosity, increase sealing properties, and even form parts or correct their shape. It is an established fact that stress corrosion cracking (SCC) only occurs in the presence of a critical state in the working material caused by environment conditions and by sufficiently high tensile stress (Fig. 1). The residual compressive stresses induced by the shot peening procedure will have a positive effect on fatigue, on stress corrosion cracking resulting from any oscillation and on tensile crack corrosion in the components. In chemical plants it is not possible to exert influence on the medium and thereby prevent stress corrosion cracking. For this reason the materials and the distribution of residual stresses were varied in the tests conducted.

3

Materials and Experimental Procedures

Three typical materials, all used in chemical plants, were selected for the tests under laboratory conditions. 1.4571 1.4541 1.4462

X6CrNiTi17-12-2 X6CrNiTi18-10 X2CrNiMoN22-5-3

From the chosen materials, tensile specimens (proportional test bars) were made for the stress corrosion cracking tests. An unpeened specimen of each material was solution heated to reach a uniform starting state. The remaining specimens were then peened in line with three different shot peening parameters. The shot peening parameters were selected and optimised with the assistance of the Peenstresssm Software. Figure 2 shows the relationship between the basic shot peening parameters and the distribution of residual stress [Source: B. Scholtes DGM 1990]. The proportional test bars selected were processed in automated equipment and under full production conditions on the Peenamatic shot peening systems. The shot peening medium for all of the specimens was Stainless Steel Cut Wire (SSCW), size 0.8 mm, material 1.4310, hardness 610–670 HV1. The intensities were 5 A, 12 A and 20 A. The degree of coverage was uniform at 100 % for all specimens, and this was controlled by means of the PEENSCAN procedure. The following diagrams indicate the calculated distributions of the residual stress of the various materials. The residual stress values were measured with the hole drilling method and they are denoted in each case by a small rectangle.

Compression

0

v, d, C, HS

HM

v:

Velocity of Shot [m/s]

d:

Shot - size [mm]

C:

Coverage [%]

HS : Hardness of Shot [HRc]

[MPa]

Residual stress

Tension

257

HM : Hardness of Material [HRc] HS HM v, d, HS , H M

0

[mm]

Distance from surface

Figure 2: Shot Peening Parameter vs. Residual Stress Distribution

Res. Stress (MPa)

Residual Stress Distribution

PEENSTRESS ENSAM - MIC

Depth (mm) 5A S 800 Shot V = 9 m/s SS 1.4462 Dir. Z

Figure 3: Residual stress distribution in shot peened 1.4462

Is (MPa)

- 588

Im (MPa)

- 615

PI Im(mm)

0,025

PI I0 (mm)

0,135

258

Res. Stress (MPa)

Residual Stress Distribution

PEENSTRESS ENSAM - MIC

Depth (mm) 12 A S 800 Shot V = 33 m/s SS 1.4462 Dir. Z I5s I5 (MPa)

- 682

Im (MPa)

- 720

PI Im(mm)

0,040

PI I0 (mm)

0,315

Figure 4: Residual stress distribution in shot peened 1.4462

Res. Stress (MPa)

Residual Stress Distribution

PEENSTRESS ENSAM - MIC

Depth (mm) 20 A S 800 Shot V = 69 m/s SS 1.4462 Dir. Z

Figure 5: Residual stress distribution in shot peened 1.4462

Is (MPa)

- 694

Im (MPa)

- 742

PI Im(mm)

0,035

PI I0 (mm)

0,490

259 Res. Stress (MPa)

Residual Stress Distribution

PEENSTRESS ENSAM - MIC

Depth (mm) 5A S 800 Shot V = 9 m/s SS 1.4571 Dir. Z Is (MPa)

- 485

Im (MPa)

- 523

PI Im(mm)

0,025

PI I0 (mm)

0,170

Figure 6: Residual stress distribution in shot peened 1.4571

Res. Stress (MPa)

Residual Stress Distribution

PEENSTRESS ENSAM - MIC

Depth (mm) 12 A S 800 Shot V = 33 m/s SS 1.4571 Dir. Z

Figure 7: Residual stress distribution in shot peened 1.4571

Is (MPa)

- 545

Im (MPa)

- 579

PI Im(mm)

0,035

PI I0 (mm)

0,390

260 Res. Stress (MPa)

Residual Stress Distribution

PEENSTRESS ENSAM - MIC

Depth (mm) 20 A S 800 Shot V = 69 m/s SS 1.4571 Dir. Z Is (MPa)

- 543

Im (MPa)

- 583

PI Im(mm)

0,020

PI I0 (mm)

0,610

Figure 8: Residual stress distribution in shot peened 1.4571

Res. Stress (MPa)

Residual Stress Distribution

PEENSTRESS ENSAM - MIC

Depth (mm) 5A S 800 Shot V = 9 m/s SS 1.4541 Dir. Z Is (MPa)

Figure 9: Residual stress distribution in shot peened 1.4541

- 481

Im (MPa)

- 512

PI Im(mm)

0,025

PI I0 (mm)

0,165

261 Res. Stress (MPa)

Residual Stress Distribution

PEENSTRESS ENSAM - MIC

Depth (mm) 12 A S 800 Shot V = 33 m/s SS 1.4541 Dir. Z Is (MPa)

- 526

Im (MPa)

- 563

PIm(mm)

0,030

PI0 (mm)

0,380

Figure 10: Residual stress distribution in shot peened 1.4541

Res. Stress (MPa)

Residual Stress Distribution

PEENSTRESS ENSAM - MIC

Depth (mm) 20 A S 800 Shot V = 69 m/s SS 1.4541 Dir. Z

Figure 11: Residual stress distribution in shot peened 1.4541

Is (MPa)

- 542

Im (MPa)

- 563

PI Im(mm)

0,020

PI I0 (mm)

0,590

262 The proportional test bars were fixed firmly in a holding device and tested statically for tensile strength. In total, four loading conditions at 50 %, 70 %, 80 % and 90 % of the yield stress (Iy) of the respective materials were applied. The specimens were exposed to a 42 % MgCl2 solution at a constant temperature of 145 oC.

4

Results and Discussion

Table 1 displays the exposure times elapsed until the onset of stress corrosion cracking. Due to the limited time available for the tests, not all permutations were tested. The tests were aborted after an exposure period in excess of 1,000 hours without any findings. Table 1: Exposure times elapsed until the onset of stress corrosion cracking

Material

Shot Peening Para- Tensile Stress meters

1.4541

not peened SSCW 0.8, 5A, 100 %

1.4571

1.4462

0.8 Iy

0.9 Iy

11h

5h

–

3.2 h

–

–

–

–

–

–

–

–

SSCW 0.8, 20A, 100 %

–

>1000 h

–

>1000 h

17h

11.3 h

–

7.5 h

SSCW 0.8, 5A, 100 %

–

–

–

–

SSCW 0.8, 12A, 100 %

–

>1000 h

–

>1000 h

SSCW 0.8, 20A, 100 %

–

>1000 h

–

>1000 h

5.3h

3.3 h

–

1.3 h

–

–

–

–

not peened

not peened

not peened

0.7 Iy

SSCW 0.8, 12A, 100 %

SSCW 0.8, 5A, 100 %

1.4529

0.5 Iy

SSCW 0.8, 12A, 100 %

–

>1000 h

>1000h 2 h

SSCW 0.8, 20A, 100 %

–

>1000 h

>1000h 5.3 h

–

500 h

–

37 h

The results given here demonstrate clearly that controlled shot peening the materials significantly increases their resistance to stress corrosion cracking in all three cases. On inspection of the exposure times of the unpeened specimens it is noted that the material 1.4462 has the lowest resistance to stress corrosion cracking. The application of controlled shot peening at a stress of 0.7 Iy, the exposure time can be increased from 3.3 hours to over 1,000 hours. At a stress of 0.9 Iy, the improvement is not so significant. The results for the material 1.4529 in a peened state are not yet available. The exposure times of 500 hours and of 37 hours permit the conclusion that the shot peened version will generate also no findings. Even under high static loads, the materials 1.4541 and 1.4571 display a marked improvement against stress corrosion cracking. Because of the high testing times required only a few alternatives were investigated. Not even under high static loads did any damage occur, therefore, it may be assumed that no premature damage needs to be anticipated at a reduced load earlier stage.

263 The application of controlled shot peening acquires special importance in the context of the production costs. The production costs for a 5,000 litre storage tank made of the material 1.4541 and including the cost of shot peening would be approximately EUR 48,000. In case of making it from the material 1.4571, the production costs including shot peening would amount to EUR 50,000. The application of the material 1.4462 would increase the production costs to EUR 90,000 without including controlled shot peening.

Cost of production [ EUR ]

90000

80000

70000 60000

50000

40000

30000 20000

10000

0

1.4541

1.4571

1.4539

1.4462

material

Figure 12: Comparison of material costs for a 5000 liter tank

5 • • • •

•

•

Conclusions Up to now tests have only been completed under static load. In plant construction most components are under dynamic stress. In the future tests for corrosion fatigue will be carried out. On the basis of the results obtained the findings will be put into practice. The investigation into the production costs of a 5,000 litre storage tank revealed significant differences. For producing the tank from material 1.4541 or 1.4571 the cost for the controlled local shot peening at the site is already included. Controlled shot peening will significantly increase the resistance to stress corrosion cracking. Earlier research confirms the above findings in respect of ferritic base materials and in NH3 media. On the basis of the present results less expensive materials may be used in plant the construction or in chemical plant without impairing the resistance to corrosion.

264

Investigating the Benefits of Controlled Shot Peening on Corrosion Fatigue of Aluminium Alloy 2024 T351 Sean A. Curtis1, Eduardo R. de los Rios1, Chris A. Rodopoulos1, Jose Solis Romero1, Andrew Levers2. 1

University of Sheffield, Sheffield, UK - 2 Airbus UK, Chester, UK.

1

Introduction

It is widely accepted that the use of controlled shot peening (CSP) can be a beneficial process to improve the performance of metallic components. More precisely, it is recognised that it can increase fatigue life and reduce susceptibility to stress corrosion cracking (SCC), improvements generally accredited to the induced compressive residual stress layer close to the surface of the shot peened material. These two improvements due to CSP have been reported to be the most commonly investigated [1]. On searching the available literature, works on the effects of CSP on corrosion fatigue are far less in abundance. In most cases CSP was shown to increase corrosion fatigue life by introducing an effective compressive residual stress layer, which reduces crack propagation rates during the early stages of crack growth [2-7]. In one case, CSP was said to have had a greater effect at higher stress levels in a spring steel [8], whilst it has also been shown that the largest gain was achieved at lower stress levels in a structural steel [9]. However, contrary to the aforementioned results, evidence has been presented that shows no advantageous effect of CSP, in either a spring or structural steel, under corrosion fatigue conditions [10]. Possibly due to the mechanical nature of the fatigue phenomenon, very little attention appears to be directed towards the electrochemical effects of CSP. In [3], CSP was seen to have no effect on the free corrosion potential, Ecorr, in a 1.0N H2SO4 solution, but the corrosion current density, Icorr, was said to increase with peening. However, in a 3.5% NaCl solution, Ecorr was seen to be nobler and Icorr was reduced. These latter findings are consistent with results observed in [11]. The present investigation has been made as part of an ongoing study into the effects of CSP. Aluminium alloy 2024 T351 is commonly used within the aerospace industry, and is routinely shot peened as part of the manufacturing process. To determine the effects of CSP on the corrosion fatigue behaviour of this alloy, a series of tests were performed. Initially, corrosion fatigue tests were conducted on both unpeened and peened specimens. Electrochemical testing was then undertaken, firstly, to investigate if the shot peening would change the characteristics of the material, and secondly, to help with the interpretation of the corrosion fatigue results.

2

Experimental Procedure

2.1

Corrosion Fatigue

Corrosion fatigue testing was carried out on flat, hourglass profiled specimens of dimensions 200 x 24 x 5mm, with a 10mm wide centre section [Figure 1], in a 3.5% NaCl solution, pH ap-

265 proximately 6.0. The solution was continuously aerated and pumped from a tank to a corrosion cell containing the specimen, and then recycled to the tank throughout the tests. Testing was performed on an Instron 8501 servo-hydraulic test rig at 25 Hz, stress ratio R = 0.1. 10mm Direction of Rolling 24mm

200mm Figure 1: Hourglass specimen for corrosion fatigue testing

Shot peening conditions were 200 % coverage, 45° impingement angle, using steel shot, S110. After peening, the specimens were cleaned using a 70 % nitric acid solution to remove any shot contamination.

2.2

Electrochemical Testing

Electrochemical tests were performed on cylindrical specimens as described in ASTM standard F476-87. Shot peening parameters and solution conditions were identical to those for fatigue testing. Potentiodynamic polarisation tests were conducted to ascertain corrosion characteristics, such as Ecorr and Icorr, by producing Tafel plots (approx ± 600 mV of Ecorr) and linear polarisation plots (approx. ± 20 mV of Ecorr). Observations were made of the peened and unpeened conditions of the alloy’s surface after identical periods in the corrosive solution. This was done, using an optical microscope, to compare some pitting characteristics.

3

Results

3.1

Corrosion Fatigue

The results of the corrosion fatigue tests [Figure 2] show little improvement due to CSP, using the chosen peening conditions. In general, there is a slight increase in life, however, the increase is not consistent, and at some stress levels, the peened tests show a reduction in life. Also shown in Figure 2 are the results of air fatigue tests conducted for comparison, which show only a slight improvement due to CSP.

266

Stress Range (MPa)

400

Air SP Air Corrosion SP Corrosion

300

200

100

f = 25Hz R = 0.1 Corrosion - 3.5% NaCl 104

105

106

107

Cycles to Failure Figure 2: Results of air and corrosion fatigue tests

3.2

Electrochemical Analysis

The free corrosion potential of the material, in both the peened and unpeened conditions, was determined by allowing samples to corrode freely over a number of hours whilst measuring the potential versus a saturated calomel electrode (SCE). The potentiodynamic polarisation tests were then run from approximately –600 mV below this value to +600 mV above it. As seen in Figure 3, in both conditions the material behaves the same. It can be seen that Ecorr values for the peened specimens, were nobler by about 50 mV. To determine Icorr, it was necessary to determine the polarisation resistance Rp. This can be taken as the gradient of the linear portion of the linear polarisation plot, ±20 mV of Ecorr. Icorr can then be calculated using Equation 1.

b A bB DE = Rp = 2.3 I corr bA + bC Di

[1]

267 Where DV/Di is the gradient of the linear portion of the plot, bA and bC are the anodic and cathodic Tafel constants respectively, obtained from the polarisation plot, Figure 3. -200

E (mV) ref SCE

-400

2x Peened -600 Unpeened 2 Unpeened 3 Peened 1 Peened 2 Unpeened 1

-800

3x Unpeened -1000

-8

-6

-4

-2 2

Current Density (log I/A (A/cm )) Figure 3: Potentiodynamic polarisation plot indicating the difference in Ecorr

At least 10 tests were conducted for each condition and the average values of the tests were used for comparison, presented in Table 1. Table 1: Calculated Icorr values for Al 2024 T351

Icorr

3.3

Unpeened

Peened

Ratio

0.328 mA

1.719 mA

1 : 5.2

Optical Microscopy

Surfaces of peened and unpeened specimens were compared after free corrosion for 24 hours, examples are given in Figure 4. The pits in the unpeened material are seen to follow clearly defined lines, or strings of pits, in the direction of rolling. This can be attributed to inclusions, as found in [12], where it was shown that constituent particles, or inclusions, are the primary source for pits, and their locality must be incorporated into probabilistic modelling of pit growth. Such observations are not possible for the peened specimen, and the number of pits is greatly reduced. However, the pits are notably larger, especially perpendicular to the rolling direction, which is the direction of fatigue crack growth.

268

Figure 4a: Unpeened – showing strings of pits in the rolling direction

4

Figure 4b: Peened – showing less, but larger, pits

Discussion

As only a slight benefit was obtained from CSP in air fatigue, it is necessary to investigate the role of the compressive residual stresses obtained. Based on a theoretical analysis [13, 14] it is possible to compare the residual stresses developed in the material with those required to overcome the detrimental effects of the indentations created through CSP. The results of this comparison are shown in Figure 5.

Residual Stress (MPa)

0

Experimental Theoretically Required (based on 350 MPa maximum applied stress R=0.1)

-50

-100

-150

-200 500

1000

1500

2000

Depth (mm) Figure 5: Distribution of residual stresses, measured vs theoretically required

It is demonstrated here, that within approximately 300mm in depth, the residual stresses would not be sufficient to prevent crack growth, however, it might be expected that they would hinder it to some extent. As the early stages of crack growth are known to dominate the fatigue

269 life, it would be expected that having residual stresses higher than those required deeper into he material, would be less efficient in extending the fatigue life. From the electrochemical analysis, it is observed that CSP can change the pitting characteristics of a material. Although less pits were evident in the peened specimen, the increase in Icorr is reflected in the larger size of pit found, thus indicating a higher pit growth rate. An important factor in the larger size of pit is the growth perpendicular to the direction of rolling. The cracks grew in this direction in the tensile specimens. Therefore a crack initiating at a pit of a critical size, and depending on the applied stress level, would initiate earlier due to the increased pit growth rate. As previously stated, the residual stresses would still hinder crack growth, therefore the corrosion fatigue life of the peened material would not necessarily be less than that of the unpeened. It would then be expected that less benefit will be obtained compared to that of CSP for air fatigue, as seen in Figure 2. Further work is required to quantify the observations made. Crack growth measurements would allow any difference in growth rate be identified and the expected earlier crack initiation in peened specimens be realised.

6

Conclusions

For the peening conditions used, CSP is not effective in extending the corrosion fatigue life of Al 2024 T351. To gain maximum benefit from CSP, it is necessary to determine the residual stress levels required to produce a given improvement in fatigue life. If CSP is to be used in corrosion fatigue or SCC conditions, electrochemical analysis should be used to ensure that any changes in pitting characteristics would not be more detrimental than any residual stress benefits gained.

7 [1] [2] [3] [4] [5] [6]

References

K. Tosha, ICSP-7, 1999, 5-10. G. V. Prabhugaunkar, M. S. Rawat, C. R. Prasad, ICSP-7, 1999, 177-183. M. S. Rawat, T. R. Jayaraman, C. R. K. Prasad, Y. Kalpana, ICSP-7, 1999, 184-191. M. S. Baxa, Y. A. Chang, L. H. Burck, Met. Trans. A, 1978, 9A, 1141-1146. M. P. Muller, C. Verpoort, G. H. Gessinger, ICSP-1, 1981, 479-484. T. Nakano, T. Sakakibara, M. Wakita, A. Sugimoto, Soc. Auto. Eng. Japan, 2001, V22, no. 3, 337-342. [7] M. Papakyriacou, H. Mayer, C. Pypen, H. Plenk Jr, S. Stanzl-Tschegg, Int. J. Fatigue, V22, no.10, 873-886. [8] B. Kaiser, ICSP-3, 1987, 667-674. [9] D. Kirk, M. Jarrett, ICSP-2, 1984, 133-142. [10] M. C. Sharma, R. N. V. D. M. Rao, ICSP-4, 1990, 551-560. [11] W. J. Tomlinson, K. P. Smith, Corrosion – NACE, 1983,V39, no. 11, 432-434, (technical note). [12] D. G. Harlow, R. P. Wei, Eng. Frac. Mechs. 1998, V59, no. 3, 305-325.

270 [13] S. Curtis, E. R. de los Rios, C. A. Rodopoulos and A. Levers, Inter. J. of Fatigue, 2002, in press. [14] C. A. Rodopoulos, R. Edwards, S. Curtis, J. Solis Romero, J.-H. Choi, E. R. de los Rios, A. Levers, ICSP-8, 2002.

269

Influence of Shot Peening on the Fatigue and Corrosion Behavior of the Die Cast Magnesium Alloy AZ91 hp Clemens Müller, Roberto Rodríguez Physikalische Metallkunde, Technische Universität Darmstadt, Darmstadt, Germany

1

Abstract

The influence of shot peening and corrosion on the fatigue behaviour of the magnesium die cast alloy AZ91 hp is studied. Investigations on as-cast fatigue specimens show that the endurance limit is significantly reduced by the presence of pores in the near surface region. This effect can be drastically reduced by shot peening, improving the endurance limit of porous material by about 100%. The fatigue behavior of a dense material can also be improved by shot peening, but the effect is rather small (in the order of 20 %). Fatigue experiments under corrosive conditions underline one of the main disadvantages of magnesium alloys, i.e. their pronounced sensitivity to contact corrosion. Iron contamination from steel shot drastically accelerates corrosion. After 3 h in a 5% NaCl solution, corrosion attack is comparable to that observed for polished surfaces after one week. As a consequence of the accelerated corrosion, shot peening with steel shot leads to a strong deterioration of the fatigue behavior, showing clearly the need for nonferrous shot or a careful cleaning of the shot peened surface.

2

Introduction

Increasing demand of light materials in automotive applications focuses interest on magnesium alloys. In load-bearing parts, magnesium alloys are not used to any significant degree due to the lack of information on the fatigue behavior under corrosive conditions. General information on corrosion and fatigue of magnesium and its alloys, is available in the open literature, see e.g. [1,2]. Mayer et al. [3] investigated the influence of salt-water on highfrequency fatigue tests of AZ91. In contrast to tests in ambient air, the high-pressure die cast alloy shows no fatigue limit in salt-water spray tests. Crack initiation takes place at porosity found on the surface. Pitting corrosion also reduces low cycle fatigue strength due to notch effects. The low-pressure die cast alloy shows a large scatter of fatigue strength due to crack initiation at casting defects. Fatigue strength values of pressureless die cast alloy are higher in ambient air than in sprayed saltwater. The same authors [4] investigated fatigue of AZ91 by ultrasonic testing. In salt water, pits are formed from which fatigue cracks initiate. No endurance limit was observed under these conditions. Ferguson et al. [5] used three-point bending tests to investigate fatigue of the alloy AM50 in air, in sprayed salt water and in laboratory water. Endurance fatigue limit values were similar (approx. 100 MPa) in the first two environments while low fatigue strength values and no fatigue limit were observed in water. Schindelbacher and Rösch [6] investigated the mechanical properties of AZ91 as a function of wall thickness of die cast tubes. With increasing wall thickness, a linear decrease of tensile strength, yield strength and elongation to fracture was observed. Stephens et al. [7,8] studied

272 corrosion fatigue of AZ91E-T6 in 3.5% sodium chloride. Fatigue lifetime is significantly decreased in comparison to tests in ambient air. Studying the growth of long cracks, corrosive environment is shown to clearly increase growth velocity. There is only little information available on the influence of shot peening on fatigue properties of magnesium alloys. Wagner [9] showed a considerable improvement of the endurance limit by shot peening of the alloy AZ 80, but the influence of an aggressive environment was not investigated in detail.

3

Experimental

Cylindrical specimens with dimensions shown in Fig.1a of the magnesium alloy AZ91 hp (nominal composition in wt %: 8.8 Al, 0.8 Zn, 0.22 Mn, and balance Mg), where produced by die casting at Norsk Hydro, Norway. After fatigue testing, the grip-section of the specimens was cut off and hourglass shaped specimens were machined with a gauge diameter of 5 mm and a length of 55 mm (Fig. 1b). Shot peening was performed with a ferritic shot SCCW 14 (diameter 0,350,4 mm) with an intensity of 0.41 mmN. For optical microscopy samples were cut perpendicular to the axis of the fatigue specimens and mechanically polished. Fatigue tests were performed on a rotating beam testing machine (R = -1) under laboratory air at a frequency of 50 Hz. In order to investigate the influence of corrosion on fatigue lifetime, some specimens were precorroded in a 5% NaCl solution before fatigue testing.

a. As-cast fatigue specimen

b. Specimen machined from the grip-section of the cast specimen

Figure 1: Dimensions of fatigue specimens

4

Experimental Results and Discussion

Figure 2 shows typical micrographs of cross sections of the as-cast fatigue specimens taken from the gauge length (Æ 6 mm) and the gripping region (Æ 10 mm). In both samples a nearly pore free zone in the surface area is observed, while pores are present in the central area. With

273 increasing diameter, the size of the pores increases too. This is a well-known problem in die cast magnesium alloys limiting good mechanical properties on thin components. In order to quantify the influence of pores on fatigue behavior, i.e. endurance limit, specimens were machined from the gripping area. The diameter of the specimens was chosen such as the surface of the fatigue specimens is in the highly porous zone.

Figure 2: Micrographs of cross sections of the as-cast fatigue specimen: left from the gauge length of the specimen (Ø 6 mm), right from the grip-section (Ø 10 mm)

The die cast alloy AZ91 hp exhibits good fatigue properties (fig. 3), showing a stress amplitude endurance limit of 100 MPa (R = -1). As expected, pores have a detrimental influence on fatigue behavior. In rotating beam testing the presence of pores at the surface reduces the endurance limit by 50 % from a stress amplitude of 100 MPa to approximately 50 MPa (fig. 3). Shot peening can significantly reduce this detrimental effect. After the surface treatment the endurance limit of porous specimens rises from a stress amplitude of 50 to 90 MPa and reaches nearly the value of the pore-free material in the untreated condition (fig. 4). The improvement of fatigue behaviour of pore-free material by shot peening is rather small. The endurance limit rises only by about 20 % from 100 to 120 MPa (fig. 4). Experimental results by Wagner [9] on the magnesium alloy AZ 80 indicate that a more pronounced improvement of the endurance limit can be achieved by very low peening intensities in the order of 0.05 mm N. It remains doubtful, however, whether such low intensities would improve the fatigue behavior in the presence of pores. While shot peening improves fatigue behavior of AZ91 hp under non- (or low) corrosive conditions, it reduces corrosion resistance in a drastic manner. Figure 5 shows photographs of the corrosive attack in a 5% NaCl solution on the polished surface of a fatigue specimen. After 3 hours (fig. 5b), the surface is partially attacked and after one week (fig. 5c) small pits can be observed on the surface. X-ray analysis revealed the presence of aluminium oxide on the surface. After shot peening large pits are present after 3 h exposure to NaCl solution (fig. 6b), and after one week the diameter of the specimen is significantly reduced (fig. 6c). The surface mainly consists of magnesium hydroxide Mg(OH)2 without building up a protective layer. The dramatic acceleration of the corrosive attack is mainly caused by iron contamination of the surface due to the use of a steel shot. Quantitative element analysis by EDX revealed an iron concentration of about 1.6 wt% on the surface, while the concentration in the bulk material of the high purity alloy AZ91 hp is below 0.003 wt %. Disperse iron

274

160

Stress amplitude [MPa]

Stress amplitude [MPa]

160 140 without pores

120 100 80 60 with pores

40 10

4

5

whithout pores 140 SP

120 100 80

SP

60 with pores

40 6

10 10 Cycles to failure Nf

10

Figure 3: Influence of porosity on S-N curves

7

4

10

5

6

10 10 Cycles to failure Nf

10

7

Figure 4: Influence of shot peening on S-N curves

particles are known to accelerate the corrosion process of magnesium alloys by a galvanic coupling [10,11]. In pure magnesium Polmear [12] observed a strong increase of the corrosion rate for iron concentrations above 0.02 wt%, which is two orders of magnitude lower than the iron concentration observed here on the surface of shot peened fatigue specimens.

(Figure 5: Corrosion on the polished surface of a magnesium alloy AZ91 hp specimen a) polished, (b) 3 h 5 % NaCl, (c) 1 week 5 % NaCl

275

Figure 6: Corrosion on the shot–peened surface of a magnesium alloy AZ91 hp specimen (a) shot peened, (b) 3 h 5 % NaCl, (c) 1 week 5 % NaCl,

Pre-testing corrosion for one week reduces the endurance limit (107 cycles) of the as-cast specimen from a stress amplitude of 100 MPa to approximately 50 MPa (fig. 7). After shot peening a pre-testing corrosion of only 3 hours in a NaCl solution causes the endurance limit to drop from 120 MPa to 50 MPa (fig. 8). The fatigue behavior of specimens that were shot peenend and pretesting corroded for one week was not tested, as a load amplitude of 50 MPa already leads to failure after a few cycles and the testing equipment did not allow lower loading. Determination of a 107 endurance limit would not make much sense, as its value will be well below 50 MPa and therefore out of interest for technical applications. The question arises whether mechanical or chemical surface cleaning is an appropriate way to avoid the strong deterioration of the fatigue properties following corrosion. In order to utilize the benefit of shot peening on fatigue properties under corrosive conditions, iron concentration must be completely avoided as iron particles act as local elements for corrosion. Even if the overall surface concentration of iron may be low after cleaning, the local concentration must not exceed a very low value. If non-ferrous shot is used after the use of steel shot or after shot peening of ferrous targets, iron contamination will remain in the shot peening facilities, probably leading to accelerated corrosion. The questions of how much iron contamination is tolerable, and how the contaminated surfaces and shot peening facilities should be cleaned, will be addressed in forthcoming work.

276

160

Stress amplitude [MPa]

Stress amplitude [MPa]

160 140 120

as-cast

100 80 60

1 week NaCl

40

SP

120 100 80 60

SP + 3h NaCl

40

10

4

5

6

10 10 Cycles to failure Nf

10

7

Figure 7: Influence of corrosion on S-N curves of as-cast specimens

5

140

4

10

5

6

10 10 Cycles to failure Nf

10

7

Figure 8: Influence of corrosion on S-N curves of shot-peened specimens

Summary and Conclusions

Shot peening is an appropriate method to improve the fatigue behavior of the magnesium die cast alloy AZ91. Especially the detrimental effect of casting pores on fatigue properties can be reduced. Iron contamination of the surface drastically accelerates the corrosion process in salt water. This leads to a strong deterioration of the fatigue properties after shot peening with ferrous shot. Even local iron-contamination will be sufficient to deteriorate fatigue properties by accelerating local corrosion (pitting). It remains therefore doubtful whether cleaning of the surface (mechanical or chemical) or the use on non-ferrous shot in an iron-contaminated shot peening equipment would be sufficiently effective to avoid the problem.

6

Acknowledgements

Part of this work was supported by Adam Opel AG (Rüsselsheim). The authors would like to thank Prof. L. Wagner (BTU Cottbus) for helpful comments and for conducting shot peening treatments at his institute.

7 [1] [2] [3] [4]

References Y. Kobayashi, T. Sibiusawa, K. Ishikawa, Mater. Sci. Eng. A 1997, 234 A. Eliezer, E. M. Gutman, E. Abramov, E. Aghion, Corr. Rev. 1998, 1 H. Mayer, M. Papakyriacou, S. Stanzl-Tschegg, E. Tschegg, B. Zettl, H. Lipowsky, R. Rösch, A. Stich, Mater. Corr. 1999, 80 H. R. Mayer, H. Lipowsky, M. Papakyriacou, R. Rösch, A. Stich, S. Stanzl-Tschegg, Fatigue Fract. Eng. Mater. Struct. 1999, 591

277 [5]

W. G. Ferguson, W. Liu, J. MacCulloch, in Proceedings of the Second International Conference on Advanced Materials Development and Performance (Eds.: I. Nakabayashi, R. Murakami), University of Tokushima, 1999, 49 [6] G. Schindelbacher, R. Rösch, in Magnesium Alloys and their Application (Eds.: B. L. Mordike, K. U. Kainer), Werkstoff-Informationsgesellschaft Frankfurt, 1998, 247 [7] R. I. Stephens, C. D. Schrader, K. B. Lease, J. Eng. Mater. Techn. 1995, 117, 293 [8] R. I. Stephens, C. D. Schrader, D. L. Goodenberger, K. B. Lease, V.V. Ogarevic, S. N. Perov, SAE Technical Paper Series, SAE International, Warrendale, 1993, 843 [9] L. Wagner, Mater. Sci. Eng. A, 1999, 210 [10] D. Eliezer, E. Aghion, F. H. Froes, Adv. Perform. Mater., 1998, 201 [11] Magnesium Taschenbuch, Aluminium-Verlag Düsseldorf, ISBN 3-87017-264-9, 2000 [12] I. J. Polmear, Mater. Sci. Technol., 1994, 1

277

VI

Fatigue of Fe- and Ni-based Alloys

278

279

Shot Peening and Fatigue Strength of Steels Karl-Heinz Lang, Volker Schulze, Otmar Vöhringer Institut für Werkstoffkunde I, Universität Karlsruhe (TH), Karlsruhe, Germany

1

Introduction

Shot peening is a commonly used production process which changes the material state close to the surface. Depending on the material, the material state and the microstructure nearby the surface of the concerning workpiece, the topography, the residual stress state, the workhardening state and the microstructure may be altered. All these changes may influence the fatigue properties of a component more or less significantly. The lifetime may increase if shot peening is performed with optimized peening parameters. For this effect a smoothing of the surface, a workhardening of surface near areas of the material or the introduction of stable residual stresses can be responsible. The greatest lifetime increase may be reached if all three mechanisms act simultaneous. In [1] the characteristics of surface layers produced by shot peening are described in a systematical overview. In principle, the greatest potential for the lifetime increase is ascribed to residual stresses. Residual stresses may alter the cyclic deformation behavior, promote or retard crack initiation, accelerate or decelerate crack propagation, and may be beneficial or detrimental to finite fatigue life and the endurance limit. The consequences of residual stresses at a concrete application depend strongly on the effects, which are connected with the production of residual stresses like the change of the surface and of the microstructure of surface near areas. Furthermore, the stability of the produced residual stresses under the operating conditions and the mechanical properties of the regarded material are relevant. In particular, the material strength is very important. Low strength materials (for example normalized steels or wrought alloys), medium strength materials (for example quenched and at a medium temperature tempered steels or Ȗǯ-hardened Ni-base alloys) and high strength materials (for example quenched and at a low temperature tempered steel) have to be distinguished in this context. In the present paper a systematic overview of the knowledge about the influence of residual stresses - especially of shot peening induced residual stresses - on the different stages of fatigue process and on the endurance limit of different steels and steels in different states will be given using selected examples. A more detailed survey of this topic is given in [2].

2

Influence of Residual Stresses on the Cyclic Deformation Behavior

The influence of macro and micro residual stresses produced by shot peening on the cyclic deformation behavior is shown in Fig. 1. As characteristic cyclic deformation curves the plastic strain amplitude from stress controlled push-pull fatigue tests (R = ımin / ımax = -1) at different stress amplitudes is plotted versus the number of cycles. The behavior of different

282 heat treated smooth specimens of the steel 42 CrMo 4 (AISI 4140) are compared in unpeened and in shot peened conditions with compressive residual stresses at the surface [3,4]. In the normalized state (Fig. 1a), the onset of cyclic deformation is different in both conditions, since the shot peened specimens with surface compressive residual stresses Vrs = -400 MPa show cyclic softening from the first cycle and higher plastic strain amplitudes during the first cycles for stress amplitudes Va between 250 MPa and 350 MPa. After some numbers of cycles the opposite tendency can be detected and the plastic strain amplitudes of the shot peened conditions are smaller than those of the unpeened material states. However, for the same Vavalues the plastic strain amplitudes of both conditions approach another at relatively high numbers of cycles. Corresponding results for a quenched and at 730°C/2h tempered 42 CrMo 4 are presented in Fig. 1b. In the unpeened condition, the characteristic cyclic deformation behavior of quenched and tempered steels occurs with a quasielastic incubation period which is followed by cyclic softening until crack initiation. After shot peening which generates surface compressive residual stresses Vrs = -400 MPa, the onset of cyclic softening is shifted to smaller numbers of cycles. Furthermore, it is interesting to note that for identical stress amplitudes and comparable numbers of cycles, the higher plastic strain amplitudes are always measured for the shot Figure 1: Cyclic deformation for stress controlled push- peened specimens. Fig. 1c shows a compull loading of (a) normalized, (b) quenched and tempered pilation of cyclic deformation curves for (730°C/2 h), and (c) quenched and tempered (570°C/2h) another quenched and tempered smooth specimens of the steel 42 CrMo 4 in unpeened and (570°C/2h) 42 CrMo 4 steel with a higher shot peened conditions. strength compared with the steel condition in Fig. 1b. In this case, the shot peened condition which has surface compressive residual stresses Vrs = -530 MPa is characterized for all investigated Va-values by small measurable plastic strain amplitudes during the first cycle which diminish or disappear first of all with an increasing number of cycles. After a subsequent regime of quasi-elastic behavior, cyclic sof-

283 tening is dominant which yields to lower plastic strain amplitudes and larger numbers of cycles to failure in comparison with the unpeened conditions. In relatively soft material states, as for example in normalized as well as in quenched and at high temperatures tempered conditions, the consequences of mechanical surface treatments like shot peening or deep rolling on the cyclic deformation behavior are mainly caused by near surface micro residual stresses, i. e. work hardening of the surface layers, because the macro residual stresses are relaxed very soon by cyclic plastic deformation [5-7]. The dislocation structures in the ferrite after the mechanical surface treatment are not stable and change during cyclic loading in energetically more favorable arrangements. The small plastic strain amplitudes of the shot peened or deep rolled conditions and the resulting increase in fatigue life are caused by the restricted mean free path of the mobile dislocations in the work hardened surface layers. In hardened as well as in quenched and at low temperatures tempered conditions, the changes in the cyclic deformation behavior result not only from surface hardening or softening effects but also from the more stable residual stresses.

3

Influence of Residual Stresses on the Crack Initiation

Crack initiation occurs as a consequence of microstructural changes in metallic materials during cyclic loading. Different mechanisms are responsible for their formation [e.g. 8]. If it is accepted that for given materials states at comparable load amplitudes increasing amounts of plastic strain amplitudes lead to decreasing numbers of cycles to crack initiation Ni, it follows that residual stresses may extend, shorten or leave the number of cycles unchanged to crack initiation. However, experimental investigations concerning the influence of macro and micro residual stresses on crack initiation are scarce. This is due to the difficulties connected with the observation of the formation and the propagation of small cracks. A recently published investigation gives a report on the influence of mechanical surface treatments on crack initiation and crack propagation in push-pull loading of steels [9]. In untreated materials, crack initiation normally takes place at positions of high localized slip, e.g. at extrusions and intrusions which are connected with persistent slip bands. However, as shown in Fig. 2 in shot peened and deep rolled conditions of the austenitic steel AISI 304, crack formation occurs later than in the untreated state due to the consequence of numerous obstacles for slip (dislocations, grain and twin boundaries) in the work hardened surface layer which impede localized slip. In these surface work hardened conditions, no persistent slip band is observed at all. Furthermore, crack propagation is slower than in the untreated state due to the effect of Figure 2: Influence of mechanical surface treatments on microstructure and compressive residual the damage evolution of the push-pull loaded austenitic stresses. Similar results are found in [9] steel AISI 304 (Va = 320 MPa, R = -1, a) Nf = 3859, b) Nf = for the normalized steel Ck 45 (SAE 1045), in [10] for the plain carbon steel 4445, c) Nf = 20265) [9]

284 Ck 80 (SAE 1080) and in [11,12] for a high strength spring steel 50 CrV 4 (AISI 6150). However, in some cases, as reported in [13], the crack initiation time of shot peened specimens is sometimes shorter than that of unpeened ones despite increased lifetimes. One example is given in Fig. 3 for the quenched and tempered steel Ck 45 (SAE 1045) in the case of bending fatigue tests in sea water [14]. With the exception of high stress amplitudes, the cracks are formed earlier in shot peened specimens than in ground ones. This finding is attributed to an enhanced crack initiation at micro-notches resulting from shot peening which is obviously supported by corrosion pittings in the case of seawater environment. However, Figure 3: Stress amplitude vs. number of cycles to crack the numbers of cycles to failure of the initiation and to failure of a quenched and tempered steel shot peened conditions are higher than Ck 45 under bending fatigue loading in sea water [14] that of ground conditions.

4

Influence of Residual Stresses on the Crack Propagation

The tip of a propagating crack is surrounded by a typical residual stress field as shown exemplarily in Fig. 4a for a high strength structural steel of the European grade S690QL1 [15,16]. A crack was produced by cyclic loading up to a stress intensity range of 'KI = 47.4 MPam1/2. The distribution of the macro residual stress component Vyrs, which acts perpendicular to the crack flanks, shows maximum compressive residual stresses of -350 MPa at the crack tip. The alteration of this distribution after application of 20 overload cycles with an overload ratio O = 2 ('KI = 94.8 MPam1/2) is given in Fig. 4b. In front of the crack tip, a larger maximum value and a larger area with compressive residual stresses compared to Fig.

a)

b)

Figure 4: Residual stress component Vyrs vs. distance from crack tip of the steel S690QL1 after a mode I-base load of 'K = 47.4 MPam1/2 (a) and an overload of 'K = 94.8 MPam1/2 (b) [15,16]

285 4a are developed. The influence of different overload cycles on the crack propagation rate is shown in Fig. 5 for O = 2 and 3. For both overload ratios a delayed retardation of crack propagation occurs which is more pronounced for O = 3 than for O = 2 due to the effect of overload induced compressive residual stresses. Thus, by sufficient high overloads, crack arrest can be produced [17]. If a crack propagates into a macro residual stress field, the crack propagation behavior can be considerably influenced by magnitude and distribution of the Figure 5: Crack propagation rate da/dN vs. crack length a0l residual stresses. This is demonstrated in for overloads with O = 2 and 3 [15,16] Fig. 6 for differently surface rolled stainless steel AISI 304. Crack velocity is influenced mainly by work hardening and other microstructural changes as well as by residual stresses in the mechanically treated surface layers. The crack propagation rate da/dN, which was determined by analyzing striations on cracked surfaces increases with increasing rolling pressure and is considerably diminished compared with untreated specimens [18]. For practical purposes, it is very important to know that crack propagation through residual stress fields can be described quantitatively by introducing an Figure 6: Fatigue crack propagation rates da/dN in effective stress intensity range 'Keff = annealed and with different rolling pressures deep rolled K max – Kop. Kmax is the stress intensity at steel AISI 304 (stress amplitude Va = 320 MPa, surface residual stresses Vrs = -200 MPa (75 bar), -350 MPA (150 maximum loading and Kop the stress bar), -400 MPa (225bar), -300 MPa (300 bar)) [18] intensity at the loading at which the crack opens. It is important to note that a modeI crack can only grow during that portion of loading cycle where the crack is open. This portion is influenced by the loading conditions itself and the near crack tip residual stress distribution. If the influence of the residual stress contribution on Kop is known, 'Keff can be estimated and used to predict the fatigue crack growth. The validity of this idea could be proved with the crack propagation behavior in the heat-affected zone of welded specimen of SAE 1019 steel specimens in comparison with the one in unwelded specimens [19]. In the welded specimens, the da/dN, 'K-relations are completely changed. Particularly, for small 'K-values the crack propagation behavior is entirely controlled by the welding residual stress state which leads to identical crack propagation rates irrespective of the loading ratio R = Vmin/Vmax. Using 'Keff the crack propagation behavior in macro residual stress fields can be quantitatively described if crack propagation data of macro residual stress free materials are available and the residual stress distribution is known.

286

5

Influence of Residual Stresses on the Lifetime Behavior

S-N curves for alternating bending of normalized Ck 45 steel (SAE 1045) are shown in Fig. 7 [20-23]. The notched specimens had a stress concentration factor kt = 2.5. The stress gradient at the notch root related to the maximum stress (normalized stress gradient Ș) was 5 mm-1. All data are nominal stress amplitudes and are valid for a failure probability of 50 %. The bending fatigue strength was evaluated at an ultimate number of cycles Nu = 107. Downcut milling and up-cut milling, respectively, generated surface residual stresses of 242 and -234 MPa. The corresponding S-N curves are almost identical. A third batch of specimens was annealed 2 h at 700 °C after down cut milling. The annealing results in a reduction of bending fatigue life and bending fatigue strength [20-23]. The alternating bending fatigue strengths of milled smooth and notched specimens with different geometries are plotted in Fig. 8 as a function of the surface residual stresses [20-26]. Again, all data are given for a failure probability of 50 %, and the bending fatigue strengths are nominal stress amplitudes at Nu = 107. With increasing stress concentration factor and decreasing stress gradient, the bending fatigue strength decreases. The influence of the stress concentration factor is clearly visible from a comparison of the specimens with the same value Ș = 2 mm-1, but different values kt = 1.7 and 2.5, respectively. On the other hand, the increase of Ș from 2 to 5 mm-1 at specimens with kt = 2.5 results Figure 7: Alternating bending S-N curves of notched in a significant increase of bending specimens of normalized plain carbon steel Ck 45 after fatigue strength. It is also interesting to note that specimens with kt = 4.4, Ș = annealing, down cut milling and up-cut milling [20-23] 15 mm-1 have a somewhat higher strength than specimens with kt = 2.5, Ș = 2 mm-1. However, with regard to the residual stress state, there is no significant influence on the bending fatigue strength, even though the range of residual stresses covered comes to more than 1000 MPa regarding specimens with kt = 4.4, Ș = 15 mm-1. Careful inspection of the hardness of the specimens tested shows that a positive slope of the lines in Fig 8 is not related to the changing (macro) residual stress state, but to different hardness of the specimen and, hence, differences in Figure 8: Alternating bending fatigue strength of milled the micro residual stress state produced smooth and notched specimens of normalized plain carbon by different machining procedures. After correction of the data points given in Fig.8 steel Ck 45 vs. surface residual stress [20-26]

287 to the same hardness, it turns out that the bending fatigue strength is hardly changed or slightly diminished at most, if the residual stresses change from compressive ones to tensile ones [23]. S-N curves for rotating bending loading of smooth specimens made from normalized Ck 15 steel (SAE 1015) in the as heat treated state and after an additional deep rolling are shown in Fig. 9 [27]. Now, by deep rolling finite fatigue life is increased by one order of magnitude or more, and the rotating bending fatigue Figure 9: S-N curves of specimens made from normalized strength is increased significantly. The reCk15 steel in the as heat treated state and after an sults shown in Fig. 8 and 9 are characadditional deep rolling for rotating bending loading [27] teristic for a low strength state of steels. In Fig. 10, S-N curves which were determined in alternating bending on smooth and notched specimens of quenched and tempered (600 °C/2 h) Ck 45 steel (SAE 1045) as a typical representative of medium strength state of steels are shown [20-23]. Again all data are valid for a failure probability of 50 % and Nu = 107. The S-N curves of smooth specimens in the ground state and after an additional shot peening are compared in Fig. 10a. There is a distinct increase of the bending fatigue strength by shot peening, but a rather small influence on finite fatigue life. In Fig. 10b, S-N curves of notched specimens, which were milled, ground and shot peened after grinding are compared. Again, in the range of finite fatigue life, the influence of the different manufacturing processes is almost negligible. The relative increase of the bending fatigue strength by shot peening is more pronounced compared to smooth specimens. It is interesting to note that milled specimens have a higher bending fatigue strength than ground ones, even though they have lower compressive residual stresses (-159 MPa) at the surface than the latter ones (-221 MPa).

a)

b)

Figure 10: Alternating bending S-N curves of specimens made from quenched and tempered (600°C/2h) Ck 45 steel. a) Smooth specimens after grinding and after additional shot peening. b) Notched specimens after milling, grinding and grinding with additional shot peening [20-23]

288 In Fig. 11, the alternating bending fatigue strengths evaluated from Fig. 10 are plotted as a function of the surface residual stresses. The arrows mark the shift in bending fatigue strengths and surface residual stresses produced by shot peening. Additionally, data points of ground specimens with negligible or tensile residual stresses are included [2123]. In the case of notched specimens, all data points lie on a common line with the slope í0.154 except for ground specimens Figure 11: Alternating bending fatigue strength of with compressive residual stresses at the quenched and tempered (600°C/2h) Ck 45 steel vs. surface surface. Regarding smooth specimens, the influence of tensile residual stresses on residual stress the bending fatigue strength is much more pronounced than the influence of compressive residual stresses. In Fig. 12, S-N curves for alternating bending of smooth specimens (Ș = 1 mm-1) of blankhardened 16 MnCr 5 steel (AISI 5115) determined in the unpeened and various shot peened states including one with electrolytically removed surface layer are shown [28-30]. The corresponding depth distributions of residual stresses are given in Fig. 13. From the comparison of both figures it becomes clear that the surface residual stress is not a suitable parameter for the assessment of the influence of the various treatments on the fatigue behavior.

Figure 12: Alternating bending S-N curves of smooth specimens made from blank-hardened 16 MnCr 5 steel in the as blank-hardened and in additional with different conditions shot peened states including one with an subsequently electro-polished surface (1: as blank-hardened, 2: shot velocity v = 23 m/s, coverage c = 100 %, mean diameter of the shot d = 0.6 mm, 3: v = 53 m/s, c = 100 %, d = 0.3 mm, 4: v = 53 m/s, c = 100 %, d = 0.6 mm, 5: v = 81 m/s, c = 600 %, d = 0.6 mm, 6: v = 53 m/s, c = 100 %, d = 0.6 mm, 100 μm surface layer electrolytically removed) (after [28-30])

Figure 13: Depth distribution of the residual stress in specimens made from blank-hardened 16 MnCr 5 steel in the as blank-hardened (1) and in additional with different conditions shot peened state (3,4 and 5) corresponding to Fig. 12 (after [28-30])

289 The S-N curves of smooth specimens of quenched Ck 45 steel as a typical representative of a high strength state of steels in the ground state and after additional shot peening are compared in Fig. 14a [20-23]. Similar to quenched and tempered specimens (see Fig. 10), shot peening produces a significant increase of the bending fatigue strength. Contrarily to the results of the medium strength steel, however, there is also a very pronounced increase of finite fatigue life, which comes up to one and a half order of magnitude at high stress amplitudes.

a)

b)

Figure 14: Alternating bending S-N curves of specimens made from quenched Ck 45 steel. a) Smooth specimens after grinding and after additional shot peening. b) Notched specimens after grinding, milling and grinding with additional shot peening with shot of the indicated hardness [20-23]

The S-N curves of notched specimens of the same steel state after grinding and after additional shot peening with shot of different hardness are shown in Fig. 14b. Compared to smooth specimens, shot peening produces a much stronger increase of the bending fatigue strength. Again, there is also a remarkable increase of finite fatigue life. Additionally, the S-N curve of milled specimen is included. Finite fatigue life and bending fatigue strength of these specimens are lower compared to shot peened ones, even though they contain very high surface compressive residual stresses of -1200 MPa. In Fig. 15, the bending fatigue strength data already plotted in Fig. 14 are complemented by data evaluated from differently ground specimens [20-23]. The arrows mark the shift in bending fatigue strengths and surface residual stresses produced by shot peening. Similar to the discussion of Fig. 12 and 13, it becomes obvious that the magnitude of surface residual stress is not a suitable parameter for the assessment of the influence of shot peening induced residual stresses on the fatigue strength. One obvious way to account for the influence of (macro-) residual stresses on Figure 15: Alternating bending fatigue strength of smooth the fatigue behavior is to treat them as and notched specimens made from quenched Ck 45 steel local mean stresses. In doing so, one has with different surface conditions vs. surface residual stress to realize that there are several important

290

Figure 16: Haigh-diagram: Bending fatigue strength Rf of smooth and notched specimens made from a medium strength steel vs. residual stress (schematically) (after [13, 31])

differences between (loading-) mean stresses and residual stresses [2]. Fig. 16 schematically shows a High-diagram for smooth and notched specimens made from a medium strength steel [13,31]. The Goodman-approximation is used to account for the influence of residual stresses on the fatigue strength. If the amount of the minimum stress or the maximum stress in smooth specimens does not exceed the critical stress amplitude ıa,crit which is a function of the cyclic yield strength [5], the residual stresses do not relax, and the line AB gives the influence of the residual stress on the fatigue strength. Then, all combinations of residual stress and stress amplitude inside the shaded area do neither result in residual stress relaxation nor in fatigue failure. However, if the amount of the minimum stress or the maximum stress exceeds the critical stress amplitude, it is assumed that the residual stresses relax to the value given by the points A and B, respectively, and the fatigue strength remains constant at the value given by these points. In the case of notched specimens, the cyclic yield strength and the notch fatigue strength (both in terms of nominal stress amplitudes) are less than the respective values of smooth specimens. However, the ultimate tensile strength of notched specimens is larger than that of smooth ones in such a material state because of the triaxial stress state in the interior of the notched specimens. Now, the Goodman-relationship holds between points C and D, and residual stress relaxation occurs outside the bright shaded area. From these relationships, it is expected that the residual stress sensitivity of notched specimens is less than that of smooth specimens. As examples Fig 17 show Haigh-diagrams for smooth and notched specimens made from a high strength steel. In Fig. 17a, the data points give the correlation between the residual stress and the nominal stress amplitude at the surface. In Fig. 17b, the data points give the correlation between the maximum residual stress and the nominal stress amplitude at the locus of the maximum residual stress. In the range of tensile residual stresses, this is always again the surface. The arrows connecting some data points illustrate the residual stress relaxation which occurs during bending fatigue loading. The ultimate tensile strength Rm = 1910 MPa of smooth specimens was taken from [23]. The ultimate tensile strength Rm,no of notched specimens is unknown, but the value 2000 MPa is reasonable for this high strength material state and the stress concentration factor 1.7 [32,33]. The cyclic yield stresses of smooth and notched specimens are also unknown. However, using results from [20-23] a reasonable borderline for residual stress relaxation can be assumed which indicate the loading at which no residual stress relaxation takes place. The corresponding data points are given in Fig. 17b as open square (smooth specimens) and open circle without arrow (notched

291 specimens). Assuming, that residual stress relaxation starts at a limiting amount of the minimum or maximum stress irrespective of the fractions of the mean stress and the stress amplitude, one gets the shaded areas in which the residual stresses are stable. These areas are transferred to Fig. 17a, too. The two open circles connected with an arrow illustrate the change of the amount of the surface residual stress (Fig. 17a) and of the minimum stress (Fig. 17b) during the cyclic loading of notched specimens. In milled specimens, there is also some residual stress relaxation as indicated by the triangles connected with an arrow. Obviously, the amounts of the minimum stresses are reduced to a value corresponding to the borderline, confirming this estimation. Extrapolation of this line to zero residual stress yields the values 1785 MPa (smooth specimens) and 1660 MPa (notched specimens), respectively, which are higher than the corresponding yield strengths [23].

a)

b)

Figure 17: Haigh-diagrams: Bending fatigue strength Rf of smooth and notched specimens made from quenched Ck 45 with different surface conditions, a) vs. surface residual stress, b) vs. maximum residual stress

These findings can be understood, if crack initiation and crack propagation or crack arrest are treated separately. This will be possible with the concept of the locally effective fatigue strength which has its origin in [34,35]. This concept enables quantitative predictions of the effect of the depth distributions of residual stresses on the locus of crack initiation as well as on the fatigue strength. The basic assumption of the concept is that a crack can only be initiated at or below the surface if the local loading stress exceeds the local fatigue strength. Especially in the case of relatively hard materials, e.g. hardened steels, this concept yields to a good estimation of the corresponding properties. For that purpose, it is necessary to have a good knowledge of the depth distributions of the fatigue strength in the residual stress free condition Rf0 ( z ) as well as of the macro residual stress Vrs(z) and the residual stress sensitivity m(z). The calculation of the locally effective fatigue strength Rf(z) as a function of the distance z from the surface is done by the relationship Rf ( z ) Rf0 ( z ) m( z ) V rs ( z ) , where the residual stress sensitivity m(z) approaches the mean stress sensitivity M of the Goodman relationship if the residual stresses are stable. Then, the residual stress sensitivity is determined approximately by m( z ) Rf0 ( z ) / Rm ( z ) [35]. However, if residual stress relaxation occurs, the residual stress sensitivity m is smaller than the mean stress sensitivity M, if the initial residual stress distribution is used [31,35]. However, if the relaxed residual stress distribution is used, the residual stress sensitivity m again approaches the mean stress sensitivity M. The depth distributions of the tensile strength Rm(z) and of Rf0 ( z ) can be estimated from appropriate correlations with measured depth distributions of the hardness (see e.g. [36]). This way by the comparison of the depth distribution of local fatigue strength with the depth distribution of the loading both the locus of crack initiation as well as the specimens or component fatigue strength could be estimated (see e.g. [34,35].

292 It is well-known, that in notched specimens or components which are loaded in the range of the fatigue strength, cracks may initiate in the root of sharp notches and may arrest in a certain depth, where the driving force for crack propagation falls below its threshold value because of the steep drop of the loading stress. In most cases, such sharp notches are not relevant for components [37,38]. However, in the presence of high residual stresses, the interaction of loading stresses and residual stresses may produce strong gradients of the driving force for crack propagation making crack arrest possible even in notches with low stress concentration factor. On the other hand, in smooth specimens with large compressive residual stresses at and below the surface, crack initiation may occur below the surface because the gradient of loading stresses is small. Then, the question arises whether or not the crack can propagate towards the surface where its propagation is hindered by the residual stress field. Both situations are illustrated by the Haigh-diagrams in Fig. 17. In the case of shot-peened smooth specimens, the fatigue strength falls below the values expected by the Haigh-diagrams, because crack initiation occurs below the surface, where no beneficial effect of the compressive residual stresses is effective. Contrarily, the combination of surface loading stress and surface residual stress occurring at the notch root of specimens with kt = 1.7 (Fig. 17a) falls significantly above the corresponding Goodman-line. As already mentioned, this means that cracks initiate at the notch root, but are arrested below the surface. A crack is arrested, when the driving force for its propagation 'Keff (the range of the effective stress intensity factor) falls below the threshold value 'Kth,eff: 'K eff K max K min 'K th,eff (1) In a rough approximation, it is assumed that Kop (the stress intensity factor when the crack opens) equals zero, resulting in

'K eff ( x) k t 'V ( x) S a Y (2) for R = 0.1 and specimens without residual stresses. 'Vn is the range of the nominal stress. Since compressive residual stresses are present in the specimens investigated, the maximum stress is decreased and the minimum stress becomes less than zero, leading to

'K eff ( x) (k t 'V n,max V rs ( x) S a Y K max ( x) (3) As an example notched specimens of case hardened 16 MnCr 5 steel regarded [39]. Therefore, 'Kth,eff also depends on x (the depth below the surface) because of the carbon gradient. Determination of 'Kth,eff with CT-specimens machined out of the core region yield values between 6.0 and 6.5 MPa(m)1/2. For the carburized surface material state a reasonable threshold value of about 4.5 MPa(m)1/2 was assumed [39]. Concerning the shape of the surface cracks which extend along the notch root, a geometry factor Y = 1.12 is assumed. The interaction estimated depth distributions of the residual stresses with the depth distribution of the loading due to a nominal loading stress amplitude Vn,a = 240 MPa and Vn,a = 365 MPa - which corresponds to the nominal fatigue strengths Rn,f of the unpeened and shot peened specimens [40] - results in the depth distributions of 'Keff(x) shown in Fig. 18 together with the depth distribution of 'Kth,eff(x). Regarding unpeened specimens, 'Keff(x) exceeds 'Kth,eff(x) in a depth of only 15 Pm. Hence, cracks once initiated at oxidized grain boundaries will continue propagation until failure and the fatigue limit is determined by the cyclic loading necessary for crack initiation. For shot peened specimens 'Keff(x) approaches the threshold value between 85 and 100 Pm below the surface, and cracks may propagate or not. Hence, the fatigue limit corresponds to the boundary between propagating or non-propagating cracks which initiate at the notch root. It is interesting to note that at depths greater than 150

293 Pm, 'Keff(x) becomes relatively large as a consequence of the high loading amplitude. This means that the significant increase of finite fatigue life by shot peening is entirely based on small crack propagation rates at low distances from surface. Residual stress

15

1/2

stress intensity factor [MPa(m) ]

14

xox

x rs x

13 12 11 10 9

'Kth,eff(x) 'Keff(x) shot peened 'Keff(x) unpeened

oxidized surface layer

Vrs

8 7 6 5 4 3 2 1

crack

0 0.00

a)

0.02

0.04

0.06

0.08

0.10

0.12

0.14

depth below surface [mm]

0.16

b)

Figure 18: Depth distribution of the residual stresses (a) and local effective stress intensity factors at different distances from surface (b)

6

Summary

Regarding residual stresses and fatigue behavior the following conclusions can be drawn: x In a low strength steel, there will be no or only little influence of macro residual stresses on the fatigue strength Rf and on the lifetime in the range of infinite life, because residual stresses relax more or less completely at the latest if the cyclic loading approaches the fatigue strength. A change of the micro residual stress state by work hardening may significantly increase Rf since the resistance against cyclic plastic deformation and hence, crack initiation increases. x In a medium strength steel there is a significant influence of the macro residual stress on Rf since only a small part of Vrs relaxes during cyclic loading in the range of the fatigue limit. However, in the low cycle fatigue range relaxation becomes more complete with increasing amplitude and the influence of the macro residual stress vanishes. Tensile residual stresses are always detrimental to Rf. x In high strength steels there is a pronounced influence of residual stresses on Rf as well as on the lifetime in the range of infinite life because residual stresses are relative stable during cyclic loading. Stress relaxation in the range of Rf only occurs in notched specimens bearing very high compressive residual stresses. Then, the resulting Rf is also high and during corresponding cyclic loading, very high magnitudes of the minimum stress occur which leads to some residual stress relaxation. Contrarily, in the range of high tensile residual stresses and cyclic loadings which lead to infinite life or to technically relevant lifetimes the occurring maximum stresses are much lower and no residual stress relaxation is observed even in the range of low cycle fatigue.

294

7 References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40]

Schulze, V., Characteristics of Surface Layers Produced by Shot Peening. In: Proceedings 8th International Conference on Shot Peening (ICSP 8), 16-20 September 2002, Garmisch-Partenkirchen. Löhe, D.; Lang, K.-H.; Vöhringer, O., in: Totten, G.; Howes, M.; Inoue, T. (eds.), Handbook of Residual Stress and Deformation. ASM International, Materials Park, OH, USA, 2002, p. 27-53. Ebenau, A.; Dr.-Ing. Thesis, Universität Karlsruhe (TH), 1989. Ebenau, A.; Eifler, D.; Vöhringer, O.; Macherauch, E., in: K. Iida (ed.), Proc. ICSP 4, The Japan Soc. of Precision Engineering, Tokyo, 1990, p. 327-336. Löhe, D.; Vöhringer, O., in: Totten, G.; Howes, M.; Inoue, T. (eds.), Handbook of Residual Stress and Deformation. ASM International, Materials Park, OH, USA, 2002, p. 54-69. Kuhn, G., Dr.-Ing. Thesis, Universität Karlsruhe (TH), 1991. Kuhn, G.; Hoffmann, J.E.; Eifler, D.; Scholtes, B.; Macherauch, E., in: Fujiwara, H.; Abe, T.; Tanaka, T. (eds.), ICRS 3, Elsevier Applied Science, London, 1991, p. 1294-1301. Mughrabi, H., in: Ermüdungsverhalten metallischer Werkstoffe. Munz, D. (ed.), DGM-Informationsgesellschaft Verlag, Oberursel, 1985, p. 7-38. Altenberger, I., Dr.-Ing. Thesis, Universität GH Kassel, Forschungsberichte des Instituts für Werkstoffkunde - Metallische Werkstoffe der Universität GH Kassel, Verlag Universitätsbibliothek Kassel, 2000. Almer, J.D.; Cohen, J.B.; Moran, B., Mater. Sci. Eng A 284, 2000, p. 268-279. Berns, H.; Weber, L., in: Residual Stresses in Science and Technology. Macherauch, E.; Hauk, V. (eds.), DGM-Informationsgesellschaft Verlag, Oberursel, 1987, p. 751-758. Berns, H.; Weber, L., in: Shot Peening , ICSP 3. Wohlfahrt, H.; Kopp, R.; Vöhringer, O., (eds.), DGMInformationsgesellschaft Verlag, Oberursel, 1987, p. 647-654. B. Scholtes: Eigenspannungen in mechanisch randschichtverformten Werkstoffzuständen, DGM Informationsgesellschaft, Oberursel, 1990. Herzog, R., Dr.-Ing. Thesis, Universität Braunschweig, 1997. Berichte aus der Werkstofftechnik, Shaker Verlag, Aachen, 1998, ISBN 3-8265-4388-2. Jägg, S., Dr.-Ing. Thesis, University GH Kassel, 1999, Forschungsberichte aus dem Institut für Werkstoffkunde - Metallische Werkstoffe der Universität GH Kassel, Verlag Universitätsbibliothek Kassel. Jägg, S.; Scholtes, B., ICRS 5, Linkoeping, Sweden, 1997, p. 1078-1083. Welsch, E.; Eifler, D.; Scholtes, B.; Macherauch, E., in: Proc. 6th European Conf. on Fracture (ECF 6), Van Elst; Bakker, A. (eds.), Eng. Mater. Adv. Services Ltd (EMAS), Amsterdam, 1986, p. 1303-1320. Altenberger, I.; Scholtes, B.; Martin, U.; Oettel, H., Mater. Sci. Eng. A 264, 1999, p. 1-16. Kang, K. J.; Song, J. H. ; Earmme, Y. Y, Fatigue Fract. Engng. Mater. Struct. 12, 1989, p. 363-376. Hoffmann, J.E.; Löhe, D.; Macherauch, E., in: Wohlfahrt, H.; Kopp, R.; Vöhringer, O. (eds.), Proc. ICSP3, Garmisch-Partenkirchen, DGM Verlag, Oberursel, Germany, 1987, p. 631-638. Hoffmann, J.E.; Löhe, D.; Macherauch, E., in: Macherauch, E.; Hauck, V. (eds.), Proc. ICRS 1, Vol. 1, DGM Verlag, Oberursel, Germany, 1987, p. 801-808 Hoffmann, J.E.; Eifler, D.; Macherauch, E., in: Macherauch, E.; Hauck, V. (eds.), “Eigenspannungen, Entstehung-Messung-Bewertung”, Vol. 2, DGM Verlag, Oberursel, Germany, 1983, p. 287-300. Hoffmann, J.E., Dr.-Ing. Thesis, Universität Karlsruhe (TH), 1984. Syren, B.; Wohlfahrt, H.; Macherauch, E., Arch. Eisenhüttenwesen, Vol. 46, 1975, p. 735-739. Syren, B.; Wohlfahrt, H.; Macherauch, E., Proc. 2nd Int. Conf. Mech. Beh. Mat. (ICM2), Boston, 1976, p. 807-811. Syren, B., Dr.-Ing. Thesis, Universität Karlsruhe (TH), 1975. Traiser, H.; K.-H. Kloos, K.-H., Z. Werkstofftechnik, Vol. 16, 1985, p. 135-143. Schreiber, R.; Wohlfahrt, H.; Macherauch, E., Archiv Eisenhüttenwesen, Vol. 49, 1978, p. 207-210. Schreiber, R.; Wohlfahrt, H.; Macherauch, E., Archiv Eisenhüttenwesen, Vol. 48, 1977, p. 653-657. Wohlfahrt, H. in: Mechanische Oberflächenbehandlungen, Wohlfahrt, H.; Krull, P. (eds.), Wiley-VCH, Weinheim, Germany, 2000, p. 55-84. Scholtes, B., in: Structural and Residual Stress Analysis by Nondestructive Methods : Evaluation, Application, Assessment, Hauk, V. (ed.), Elsevier Science B.V., Amsterdam, 1997. Backfisch, W.; Macherauch, E., Archiv Eisenhuettenwesen Vol. 50,1979, p. 167-171. Wellinger, K,; Dietmann, D., Festigkeitsberechnung, A. Kröner Verlag, Stuttgart, 1969. Macherauch, E.; Wohlfahrt, H., in: Ermüdungsverhalten metallischer Werkstoffe, Munz, D. (ed.), DGMInformationsgesellschaft Verlag, Oberursel, 1985, p. 237-283. Starker, P.; Macherauch, E.; Wohlfahrt, H., Fatigue of Engng. Mater. Struct., 1, 1987, p. 319-327. Winderlich, B., Mat. wiss. u. Werkstofftechnik, Vol. 21, 1990, p. 378-389. Kloos, K.-H.; Fuchsbauer, B.; Adelmann, J., J. Fatigue Vol. 9, 1987, p. 35-42. Kloos, K.-H., in: Kerben und Betriebsfestigkeit, Nowack, H. (ed.), Deutscher Verband für Materialforschung und -prüfung e. V., Berlin, 1989, p. 7-40. Krug, T.; Lang, K.-H.; Löhe, D., in: Blom, A.F., (ed.) Proc. Fatigue 2002, EMAS Ldt.West Midlands, U.K., Volume 2/5, 2002, 955-962. Krug, T.; Laue, S.; Lang, K.-H.; Bomas, H.; Löhe, D.; Mayr, P.; FVA-Forschungsheft Nr. 669, Forschungsvereinigung Antriebstechnik e.V., Frankfurt, Germany, 2001.

293

The Effect of Shot Peening Coverage on Residual Stress, Cold Work and Fatigue in a Ni-Cr-Mo Low Alloy Steel Paul S. Prevey1) and John T. Cammett2) 1) 2)

Lambda Research Inc., Cincinnati, OH, USA U.S. Naval Aviation Depot, Cherry Point, NC, USA (Formerly with Lambda Research)

1

Introduction

The underlying motivation for this work was to test the conventional wisdom that 100% coverage by shot peening is required to achieve full benefit in terms of compressive residual stress magnitude and depth as well as fatigue strength. Fatigue performance of many shot peened alloys is widely reported to increase with coverage up to 100%, by many investigators and even in shot peening manuals.(1) The fatigue strength of some alloys is reported to be reduced by excessive coverage(2). Aerospace(3,4), automotive(5), and military(6) shot peening specifications require at least 100% coverage. Internal shot peening procedures of aerospace manufacturers may require 125% to 200% coverage. Most of the published fatigue data supporting the 100% minimum coverage recommendation was developed in fully reversed axial loading(2,7) or bending(8,9) with a stress ratio, R = Smin / Smax, of –1. The residual stress field arising from an individual shot impact is much greater in extent than the physical size of the impact crater and the resulting surrounding ridge of raised material.(10) Hence, at least some degree of undimpled surface area, less than 100% coverage, should be tolerable in terms of residual stress and fatigue strength achieved by peening. Accordingly, residual stress-depth distributions were determined for specimens peened to various coverage levels. Fatigue performance was tested at R > 0, so that the shot peened surface was loaded only in tension. Additionally, cold work-depth distributions and the effects of thermal relaxation on both residual stresses and cold work were determined.

2

Material

Aircraft quality 4340 steel plate (0.5 in. thick) per AMS 6359F(11) was employed in this work. The material composition is provided in Table 1 below. Table 1: Steel Composition C

Mn

P

S

Si

Cr

Ni

Mo

0.40 0.68 0.015 0.015 0.23 0.79 1.70 0.23

For peening trials, specimens about 33 × 38 mm (1.3 × 1.5 in.) were cut from the plate with the longer dimension oriented along the rolling direction. After hardening and tempering to 38 HRC hardness, specimens were reduced to 9.5 mm (0.375-in.) thickness by low stress grinding.

296 Tensile properties resulting from heat treatment were 1164 MPa (169 ksi) ultimate tensile strength and 1089 MPa (158 ksi) 0.2% offset yield strength.

3

Experimental Procedures

3.1

Shot Peening

Peening was performed via direct air pressure at 482 kPa (70 psi.) through a single 4.7 mm (3/ 16-in.) diameter nozzle aligned to give an 80-degree incidence angle from horizontal. Specimens were mounted on a rotary table running at 6 RPM at a vertical distance of 305 mm (12 in.) from the nozzle outlet. Carbon steel CCW14 conditioned cut wire shot was used at a controlled flow rate of 1.36 kg/min (3 lb./min). The intensity achieved was 0.22 mmA (0.009 in. A). Coverage was determined by optical observation at 20X magnification. The time to achieve 100% coverage was taken as the peening exposure time at which essentially no undimpled areas remained in an approximately 2.5 cm (1.0 in.) square area in the center of specimens. Undimpled areas were easily observed via surface texture contrast between the original ground surface and shot impacted areas. Fractional and multiple coverages were then taken appropriately as ratios of the time for 100% coverage. Coverage is defined in the shot peening literature both in terms of the fraction of area impacted, as used here, and as multiples of the time required to achieve saturation of the Almen strip. The saturation-based definition does not include the effects of the work piece properties, such as hardness and yield strength, which influence dimple diameter and the total area impacted. Assessing coverage as the fraction of the area impacted using optical examination is inherently subjective, but does include the effect of the work piece mechanical properties, and is the method adopted by most shot peening standards(3,4,5,6). In this study, 100% area coverage was achieved in 5.0 minutes (intermittent peening on the turn table) while only 2.0 minutes was required for saturation of the Almen strip under the same peening conditions; a factor of 2.5 difference between the two coverage definitions. To avoid ambiguity, the number of shot impacting the sample per mm2 at 100% coverage was quantified by direct measurement of total collected shot as 336 shot/mm2. The coverage calculated from the dimple diameter and total impacts(12) was 99.8%.

3.2

Residual Stress and Cold Work Determinations

Residual stress measurements were made via x-ray diffraction in the conventional manner from the shift in (211) diffraction peak position using Cr K= radiation.(13,14,15) Subsurface data were obtained using automated residual stress profiling apparatus to alternately measure the residual stress and then electropolish to remove layers.(16) Residual stress measurements made as a function of depth from the peened surface were corrected for relief resulting from layer removal and for penetration of the x-ray beam into the subsurface stress gradient. An irradiate area of nominally 5 × 5 mm (0.2 × 0.2 in.) was used for residual stress measurement, providing the arithmetic average residual stress over the area of an estimated 8400 shot impacts at 100% coverage. Determinations of cold work resulting from peening were made by relating (211) diffraction peak breadths to the equivalent true plastic strains.(17) The distribution of cold work as a

297 function of depth was thus obtained from diffraction peak breadth measurements made simultaneously with residual stress measurements.

3.3

Thermal Relaxation

Following residual stress and cold work determinations, specimens were thermally exposed at 519K (475F) for 24 hours. Residual stress and cold work determinations were then repeated to determine if thermally induced relaxation had occurred.

3.3

High Cycle Fatigue Testing

Fatigue testing in four-point bending mode was conducted at room temperature under constant load amplitude sinusoidal loading at 30 Hz and stress ratio, R, of 0.1. The R-ratio was chosen to avoid compressive overload and the immediate reduction of the compression introduced by shot peening. Bending fatigue specimens were machined with a trapezoidal cross section to ensure fatigue failure from the peened surfaces. The specimen geometry and test fixturing provided a nominally 0.5-in. wide by 1-in. long surface area under uniform applied stress. The central gage sections of fatigue specimen test surfaces were finished by low stress grinding and peening using the same techniques as for specimens in peening coverage trials.

4

Results and Discussion

4.1

Coverage

Figure 1 provides representative photographs of surfaces peened to various coverages. Not included in the scope of this study was determination, by area measurement, of the actual percentage of dimpled surface area for various peening times. In the context of this work, it sufficed to define coverage based upon the time ratio to achieve 100% dimpling of the surface area. As one may infer from the photographs, the percent of area covered at 0.8T approached that at T. The arrow in the photograph for 0.8T indicates a very small undimpled area easily visible when viewed optically at 20X magnification. Undimpled areas in specimens peened for times less than 0.8T are obvious in appearance as they appear in Figure 1. The overall appearance of surfaces peened for times, 2T and 4T, did not change relative to that peened for time, T.

4.2

Residual Stress Distributions

Figure 2 shows residual stress-depth distributions for various coverage levels, including the distribution for the as-ground surface before peening. Except at the lowest coverage level, 3 % (0.03 T), classical shot peening distributions resulted, whereby residual compressive stress magnitudes reached a subsurface maximum and decreased gradually until small tensile stresses occurred at greater depths. For 3 % coverage, the maximum compression occurred either at the surface or at a very slight depth not resolved in the series of measurements taken. The form of

298

3% (0.03T)

10% (0.1T)

100% (T)

20% (0.2T)

200% (2T)

80% (0.8T)

400% (4T)

Figure 1: Surfaces peened to various coverage levels (ratio of time for 100% coverage). The small white bar in each photograph represents 0.25 mm (0.01 in.)

-3

Depth (x 10 m m ) 0

100

200

300

400

30

200

0

-30

-200

-60

-400

Residual Stress (M Pa)

Residual Stress (ksi)

0

-600

-90

Lam bda R esearch 0R -9852

-120 0

2

4

6

8

10

12

14

-800

16

-3

Depth (x 10 in.) 3% 100%

10% 200%

20% 400%

80% As-Ground

Figure 2: Residual stress-depth distributions for various coverage levels; coverage is defined as the ratio of time to produce 100% surface impacts

the subsurface residual stress distribution for 3 % coverage conforms to finite element models of the stress developed in regions between dimples when impact areas are widely separated by

299 twice the dimple radius.(21) Given that the x-ray diffraction results provide an average stress over mostly un-impacted material at the 3 % coverage, the data appear to confirm the FE prediction that even the regions between impacts are in compression. The distributions for coverage levels less than 20 % (0.2 T) exhibited systematic changes with coverage, whereby increasing coverage in this range resulted in increasing compressive stress magnitude at given subsurface depths and an increase in the total depth of compression. Beyond 20 % coverage, there were no further significant changes in stress magnitude at a given depth, other than at the surface, or in total depth of compression. Compression at the surface tended to decrease with increasing coverage above 20 %.

4.3

Cold Work Distributions

Figure 3 shows cold work-depth distributions produced at various coverage levels. Consistent with residual stress-depth distributions, systematic changes in cold work-depth distributions occurred with increasing coverage level up to 20 % (0.02 T). Beyond that level, no systematic changes occurred with increasing coverage. Cold work values for the lower coverage levels were lower than at higher coverages only to a depth of about 0.05 mm (0.002 in.). 3% 100%

10% 200%

20% 400%

80% As-Ground

Cold Work (%)

PERCENT COLD WORK DISTRIBUTION 100 90 80 70 60 50 40 30 20 10 0 0

2

4

6 8 10 -3 Depth (x10 in.)

12

14

16

Figure 3: Cold work-depth distributions for various coverage levels; coverage is defined as ratio of time to produce 100% surface impacts

4.4

Thermal Relaxation

Figures 4 and 5 show the residual stress and cold work depth distributions obtained after thermal exposure at 519K (475F) for 24 hours. This exposure temperature was chosen based upon specification, AMS 13165(4), regarding maximum recommended exposure temperature to avoid residual stress relaxation in shot peened steels. Comparison with pre-exposure results (Figures 2 and 3) revealed changes in both residual stress magnitudes and cold work. Relaxation of both residual stress and cold work occurred at depths less than 0.05 mm (0.002 in.) with the greatest

300 percent changes occurring in surface values. Reduction of surface residual stress magnitudes ranged from 20-30%, and percent reduction of surface cold work ranged from 40-70%. There was no systematic trend with coverage in these reductions although the reductions decreased with depth from the surface, and initial cold work level, to about 0.05 mm (0.002 in.) for all coverage levels. Beyond 0.05 mm depth, where the initial cold work level was less than nominally 5%, there were no significant changes in residual stress or cold work. -3

Depth (x 10 m m ) 0

100

200

300

400

30

200

0

-30

-200

-60

-400

Residual Stress (M Pa)

Residual Stress (ksi)

0

-600

-90

Lam bda R esearch 0R -9852

-120 0

2

4

6

8

10

12

14

-800

16

-3

Depth (x 10 in.) 3% 100%

10% 200%

20% 400%

80%

Figure 4: Residual stress-depth distributions after thermal exposure (475F/24 hr.)

These observations highlight the importance of cold work in residual stress relaxation as has been observed in previous studies of IN718(19) and Ti-6Al-4V.(20) Where cold work values were less than 5%, no relaxation of residual stresses occurred. The implication from these results is that cold work from shot peening, even at less than 100% coverage, is sufficient to induce significant residual stress relaxation in surface and near surface layers at modest temperatures. Where such reduction cannot be tolerated, surface enhancement techniques such as low plasticity burnishing and laser shock(20) which induce low cold work should be considered, or shot peening coverage controlled to provide adequate compression with minimum or controlled levels of cold working.

301 3% 100%

10% 200%

20% 400%

80%

Cold Work (%)

PERCENT COLD WORK DISTRIBUTION 100 90 80 70 60 50 40 30 20 10 0 0

2

4

6 8 10 -3 Depth (x10 in.)

12

14

16

Figure 5: Cold work-depth distributions after thermal exposure (475F/24 hr.)

4.5

Fatigue Performance

Results of limited initial fatigue testing are presented in Figure 6, below. Significant surface and near surface compressive residual stresses were associated with the low stress ground condition. Hence, fatigue life for this condition was intermediate between lives for peened specimens and the electro-polished specimen, which had no residual stresses. Optical fractography revealed that subsurface fatigue origins occurred in all peened specimens and in the low stress ground specimen. No crack initiation sites in peened specimens were associated with undimpled surface areas irrespective of coverage. Therefore, the undimpled surface areas appear to be in compression. These results indicate the beneficial effect of peening relative to unpeened conditions. The results suggest further that, for R > 0 loading, the full benefit from peening can be realized at less than 100% coverage, although the limited number of initial tests did not permit assessment of an optimum coverage level, if any. This finding is in contrast to those of other investigators who have reported that fatigue life decreases dramatically with coverage less than 100%.(8,9) Full S-N curves for a range of coverage were prepared to test the unexpected finding of uniform fatigue strength, independent of coverage. Because the residual stress depth and magnitude was found to be comparable for any coverage greater than 20%, samples were prepared with 20%, 100% and 300% coverage. The fatigue results, presented in Figure 7, indicate no loss of fatigue life or strength for coverage as low as 20%. The fatigue performances for 20% and 100% coverage are essentially equal, given the experimental uncertainty for the limited number of samples tested. Coverage of 300% produced consistently shorter lives and a slightly lower endurance limit than either 100% or 20% coverage. When fatigue testing of shot peened surfaces is conducted in fully reversed loading, (R = – 1.0), the compressive half-cycle superimposes a compressive applied stress on the already highly compressive shot peened surface. The compressive surface then yields in the first few cycles of testing resulting in rapid relaxation of the compressive surface layer. Surface residual stress measurement during fatigue testing reveals that even at alternating stress levels below the residual stress-free material endurance limit, the surface compressive stress can be reduced to 70% of the original level in the first half-cycle.(7) Residual stress measurements on failed sam-

302 ples in the current work showed no significant change in surface compression after 130 and 220 · 103 cycles at R = 0.1 and Smax of 1240 MPa (180 ksi) for either the 100% or 20% coverage samples, respectively. 350

20%

Shot Peened

103 Cycles to Fracture

300

80%

250

200

100%

Low Stress Ground

150

100

50

ELP

0

Figure 6: Bending fatigue lives at 1240 MPa (180 ksi), R = 0.1, for electropolished (ELP), low stress ground (LSG) and shot peened 4340 steel peened to the coverage indicated

The apparent conflict between the lack of dependence of fatigue performance on shot peening coverage reported herein and work previously published is attributed to the stress ratio used in fatigue performance evaluation. Most of the prior studies of the effect of coverage on fatigue have employed fully reversed bending (R = -1) fatigue tests. It is well known that fully reversed bending of the highly compressive shot peened surface can drive the surface beyond yield in compression, causing rapid loss of compression in the initial cycles of the test.(19) The compressive overload relaxation process has been accurately modeled(22) and verified experimentally.(7,22) The benefits of shot peening are then reduced or lost entirely early in the test, depending upon the stress amplitude. The observed fatigue improvement with increased coverage may be due to increasing yield strength with work hardening of the surface with higher coverage. Confirmation of this hypothesis was beyond the scope of the present study, and will be addressed in the future. In tension-tension fatigue testing (R > 0), compressive overload is avoided, and the compressive residual stress survives without significant loss for the duration of the test at alternating stress levels appropriate for high-cycle fatigue failure.

5

Summary

Results from this investigation have clearly demonstrated that complete coverage is not required to produce full benefits of shot peening in 4340 steel, 38 HRC, peened to 0.22 mmA (0.009 in.A) intensity when fatigue tested in tension-tension loading (R = 0.1). Indeed, a coverage level of as little as 20% (0.2T) provided fatigue performance equivalent to full coverage un-

303 HIGH CYCLE FATIGUE DATA 4-point Bending, R=0.1, 30 Hz, RT 1500

OR9852

220 210 200 190

1300

180 1200

170

4340 Steel, 38 HRC Shot Peened, 9A

1100

160 150

MAXIMUM STRESS (ksi)

MAXIMUM STRESS (MPa)

1400

1000 300% Coverage 20% Coverage 100% Coverage

140

900

130 10

3

10

4

10

5

10

6

10

7

CYCLES TO FAILURE

Figure 7: High-cycle fatigue results for shot peened 4340 steel, 38 HRC, at 20%, 100% and 300% coverage

der conditions employed in this study. The principal objective of this work, however, has not been to establish an optimum coverage level although, by extension, such could be established for a given loading spectrum. Rather, it has been to show that full coverage is not required to achieve peening benefits. In a practical sense, this affords potential for significant improvements in current shot peening practice(23) in applications where compressive overload will not occur. Many practical applications of shot peening, from automotive leaf springs to compressor and turbine blades and disks, involve service loading at positive R-ratios. Reductions in peening processing times appear to be possible with obvious attendant economic benefit. Shot peening may be performed to reduced coverage with larger shot than is practical when at least 100% coverage is required, providing deeper compression and reduced cold work without loss of fatigue performance. Reduced cold work shot peening should provide improved thermal stability of the compressive layer.

6

Acknowledgements

The authors gratefully acknowledge the contributions of the staff of Lambda Research and especially Mr. Douglas Hornbach for residual stress measurement and Mr. Perry Mason for peening and fatigue testing.

304

7 [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18]

[19] [20]

[21] [22]

[23]

References Manual of Shot Peening Technology, Wheelabrator-Frye, Mishawaka, IN, 9th ED, (1977) Wagner, L., Lütjering, G., ”Influence of Shot Peening Parameters on the Surface Layer Properties and the Fatigue Life of Ti-6Al-4V”, Proc. ICSPII (1994) pp. 194-200 Aerospace Material Specification, AMS 2430L, Society of Automotive Engineers, United States (1993) Aerospace Material Specification, AMS-S-13165, Society of Automotive Engineers, United States (1997) Surface Vehicle Recommended Practice, SAE J443, Society of Automotive Engineers, United States (1984) Military Specifications, Shot Peening of Metal Parts, MIL-S-113165C, United States (1989) Eigenmann, B., Schulze, V., Vöhringer, O., ”Surface Residual Stress Relaxation in Steels by Thermal or Mechanical Treatment, Proc. ICRS IV, pp. 598-607 (1994) Person, N., ”Effect of Shot Peening Variables on Fatigue of Aluminum Forgings”, Metal Progress, pp. 33-35, July (1981) Meguid, S. A., ”Effect of Partial-Coverage Upon the Fatigue Fracture Behaviour of Peened Components”, Fatigue Fract Eng Mater Struct, 14, pp. 515-530 (1991) Al-Hassani, S. T. S., First International Conference on Shot Peening, 1981, pp. 583-602. Aerospace Material Specification, AMS 6359F, Society of Automotive Engineers, United States (1993). Abyaneh, M., Kirk, D., ”Fundamental Aspects of Shot Peening Coverage Control, Part Three: Coverage Control Versus Fatigue”, ICSP6, pp. 456-463, 1996. Prevéy, P. S., Metals Handbook, ASM International, United States, 1986, v. 10, pp. 380392. Hilley, M. E. ed., SAE J784, 1971. Noyen, I. C. and Cohen, J. B., Springer-Verlag, United States, NY, 1987. US patent No. 5,737,385, 1998. Prevéy, P. S., ”The Measurement of Subsurface Residual Stress and Cold Work Distributions in Nickel Base Alloys”, ASM International, 1987, pp. 11-19. Meguid, S. A., Shagal, G. and Stranart, J. C., ”3D FE Analysis of Peening of Strain-Rate Sensitive Materials using Multiple Impingement Model”, Int, J. of Impact Eng., 27 (2002) 119-134 Prevéy, P. S., ”The Effect of Cold Work on the Thermal Stability of Residual Compression in Surface Enhanced IN718”, Proc. 20th ASM Materials Solution Conference, 2000. Prevéy, P. S., ”The Effect of Low Plasticity Burnishing (LPB) on the HCF Performance and FOD Resistance of Ti-6Al-4V”, Proc. 6th National Turbine Engine HCF Conference, 2001. Smith, P. R. et al., Proc. 5th National Turbine Engine HCF Conference, 2000. Lu, J., Flavenot, F., Turbat, A., ”Prediction of Residual Stress Relaxation During Fatigue”, Mechanical Relaxation of Residual Stress, ASTM STP 993, L. Mordfin, Ed., ASTM, Philadelphia, pp. 75-90, 1988 Patents pending.

305

Effect of Ultrasonic Shot Peening on Fatigue Strength of High Strength Steel Yoshihiro Watanabe, Kaneshi Hattori, Mitsuru Handa Toyo Seiko, Japan

Norihiko Hasegawa, Keiro Tokaji, Masahide Ikeda Gifu University, Japan

Jean Michel Duchazeaubeneix Sonats, France

Abstract A study had been made to evaluate the effect of the Ultrasonic shot peening treatment, developed by SONATS, on fatigue strength. This new shot treatment could give the smooth surface after shot peening because of using the polished ball bearings. To investigate the effect and the merit of new treatment, fatigue test were conducted on cantilever type rotating bending fatigue testing machine. It was confirmed that fatigue strength peened by ultrasonic shot treatment is almost the same as that peened by conventional air peening device. Further, mechanism of the fracture was confirmed between roughed specimens by conventional peening and the smooth one by ultrasonic peening. Key words : Ultrasonic Shot Peening, twofold S-N curve, fish-eye, ultra-high-cycle fatigue

inclusion,

1 Introduction In general, there are two different mechanism of the fracture at the twofold S-N curve in high strength steel, namely surface related initiation in high applied stress region and subsurface crack initiation with a fish eye in low stress region. Further, the fracture mechanism is affected by the surface situation, such as roughness and hardness, etc. On the other hand, shot peening treatment is widely recognized as the one of the method to enhance the fatigue durability by compressive residual stress. However, the fatigue behavior of shot peened high strength steel, which has roughened surface after treatment, is expected to indicate the same twofold S-N curves. Recently, new Ultrasonic Shot Peening, named Stressonic® technology(1), was developed by SONATS, France. The mainly characteristic of Stressonic® could give the smooth surface after operating compared to the conventional shot peening because of using polished bearing ball. So, in this study, fatigue test were conducted on cantilever type rotating bending fatigue testing machine to confirm the effect of Stressonic® on fatigue strength and the merit of operating the Stressonic® compared to the conventional one. Further, this paper describes the influence of shot peening in ultra high cycle fatigue regime. 2 Experimental procedures The chemical composition of test specimens made of JIS SNCM439 is shown in Table 1.

306

Table 1. Chemical composition (wt%) C 0.4

Si 0.22

Mn 0.78

P 0.02

S 0.013

Cu 0.18

Ni 1.78

Cr 0.83

Mo 0.2

Fe Bal.

Table 2. Mechanical properties Tensile strength VB (MPa ) 2274

Elongation V(%) 4

Reduction of area I(%) 39

0.2% proof stress V 0. 2 ( MPa) 1515

φ3

φ1 0

R7

φ8

Figure 1. Geometry of specimen

5. 5

50

2. 0 100

Roughly Machines specimens were austenitized at 850qC for 60 min and oil quenched. Tempering treatment was performed at 620qC for 120 min. Table 2 shows the mechanical properties of specimens after heat treatment. After heat treatment, specimens were polished by grinder to improve the surface roughened by machining. The geometry of specimen is shown in Figure 1. Fatigue tests were conducted by cantilever type rotating bend ing fatigue testing machine of 3,150 r.p.m. in cyclic speed in laboratory atmosphere. 109 cycles were taken as limiting number for evaluation of fatigue properties. The characteristic of the very high cycle fatigue property should be twofold S-N curves: the one for surface fatigue and the other for internal fatigue (2). There are two fatigue limits: the surface fatigue limit, Vws and the internal fatigue limit, Vwi although the latter has not yet observed clearly. The surface fatigue life is shorter than the internal fatigue life. In this study, we evaluate the both fatigue limit to evaluate the effect of Ultrasonic treatment. Shot peening was carried out by Stressonic® and conventional air peening device using three kinds of shot media, bearing ball, tungsten carbide and conditioned cut wire. Figure 2 shows the schematic illustration of Stressonic® process and the appearance of treatment to specimens. A piezo-electric transducer emits the ultrasonic wave at 20 kHz. The waves are amplified when they travel through an acoustic booster, in a housing which contains the parts to be treated and the shot. The shot strike the vibrating walls and are reflected off the surface. Then they collide with one another. The balls are scattered randomly throughout their encasing. A homogeneous treatment is then obtained on the surface of casing. At this study, the amplitude of sonotrode was 90 micrometer. Table 3 (a) and (b) shows the shot peening conditions. Here, shot peening time were decided to obtain 100% coverage at ultrasonic shot peening and 300% coverage for air shot peening. The hardness, density and diameter of each shot media are also shown in Table 3.

307 Table 3 (a). Shot peening conditions on Stressonic® Type of shot media Density (g/mm3 ) Diameter (mm) Hardness (HV) Distance between spec. and Sonotrode Amplitude of Sonotrode Frequency Rotating speed of specimen Quantity of media (g/sonotrode) Peening time(100% coverage) Arc height (mmA)

The ball bearing (SUJ2) 7.86 0.82 HV1000 20mm 90 Pm 20kHz 12r.p.m. 4.2g 90sec 0.21mmA

Tungsten carbide ball (WC) 17.5 0.86 HV1500 20mm 90 Pm 20kHz 12r.p.m. 7.3g 90sec 0.26mmA

Table 3 (b). Shot peening conditions on air type peening device Type of shot media Density (g/mm3 ) Diameter (mm) Hardness (HV) Air pressure (MPa) Distance between nozzle and spec. Rotating speed of specimen Quantity of shot media Peening time (300% coverage) Arc height (mmA)

Rounded cut wire shot (RCW) 7.86 0.88mm HV700 0.25 160mm 12r.p.m. 4,750g/30sec 30sec 0.483mmA

Figure 2. Appearance and illustration of Stressonic® treatment ・Shot media ・Specimen ・Housing ・Sonotrode ・Pre-booster ・Booster

・Piezo transmitter

・Generator

308 Table 4. Surface roughness ( Pm ) Treatment Stressonic Air

Media SUJ2 WC RCW

Ra 0.45 0.75 0.93

Ry 3.46 5.30 7.55

Rz 2.41 4.23 5.74

Figure 3. Hardness distribution

Vickers hardness HV

800

SNCM439 NP SUJ2 WC RCW(Air)

700

600

500 0

200 400 600 800 Distance from surface d (μm)

1000

3 Results 3.1 Surface roughness and hardness distribution Table 4 shows surface roughness after shot peening treatment. Surface roughness of specimens peened by Stressonic® was smaller than that by conventional air peening. The reason of this results was considered that the shot velocity on Stressonic® is lower compared to air type and the surface conditions of SUJ2 and WC, which used for ultrasonic, are much better than RCW for air type. SUJ2 and WC ball were polished to get smooth surface because of the purpose of usage. On the other hand, roughness of WC was bigger than that of SUJ2 because of different density and hardness. Figure 3 shows hardness distribution for the specimens after each shot peening treatment. In case of using WC ball at Stressonic®, the increase of 100HV was observed compared to non-peened specimen. At the comparison between the both treatment, increase by conventional was smaller than ultrasonic. It is explained that the difference of the results is due to the higher hardness of WC and SUJ2 than that of Rounded cut wire. 3.2 Residual stress distribution Figure 4 shows residual stress distribution for specimens peened by both shot. The residual stresses were determined by X-ray diffractometer with sin2ψ - method. The residual stress distribution was obtained by repeating the X-ray measurement and electrochemical polishing

309 Figure 4. Residual stress distribution

(MPa)

0 SNCM439 NP SUJ2 WC RCW(Air)

–200

Residual stress VR

–400 –600 –800

–1000 –1200 –1400 0

20 40 60 80 100 Distance from surface d (μm)

Figure 5. Influence of shot peening treatment on fatigue properties of SNCM439

Stress amplitude Va (MPa)

1600

SNCM439 Rotating bending NP SUJ2 WC RCW(Air) Internal–fracture

1400 1200 1000 800 600 3 10

4

10

5

6

7

8

9

10 10 10 10 10 10 Number of cycles to failure Nf

10

successively. No difference was observed between WC on Stressonic® and RCW on conventional type in the depth direction. However, surface value of conventional type was lower than that of WC. Further, at Stressonic® process, compressive residual stress of WC was much higher than SUJ2. As the same results of hardness, this is derived from the

310 difference in hardness and density. 3.2 Fatigue test Figure 5 shows the S-N curve for the each shot peened specimens. The results of non-peened specimens are also shown. In this figure, (u ) means that specimens failed from the internal crack. The fatigue properties of every shot peened specimens were higher than that of non-peened specimens at the surface fatigue limit, Vws . However, there were no difference at the internal fatigue, Vwi . The reason of this result was considered that the depth of internal crack initiation site is deeper than compressive residual stress layer. (1) On the other hand, the surface fatigue limit of WC shot by Stressonic® was raised up to 1250MPa and highest in all of another treatment. The reason of this result is considered that the surface roughness of WC was better than that of specimen peened by RCW and residual stress profile at both treatments was almost the same, as mentioned before. As the results, new technology, Stressonic®, is useful and is expected to be alternate treatment as the method to enhance the fatigue durability without surface roughened. In addition, this technology should contribute to improvement of environment because of quite small quantity of ball usage as shown in Table 3. 4 Conclusions In the present study, to confirm the effect of ultrasonic shot peening treatment, Stressonic®, on fatigue properties, fatigue test were conducted by cantilever rotating bending fatigue machine by using several shot peening conditions. Following is a summary of the results obtained; (2) Surface roughness of specimens peened by Stressonic ® was smaller than that peened by conventional air type treatment. And almost the same residual stress profile was observed between Tungsten carbide ball (WC) on Stressonic® and Rounded cut wire shot (RCW) on conventional peening. (3) The surface fatigue limit of WC shot by Stressonic® was raised up to 1250MPa and highest in all of another treatment. New technology, Stressonic®, is useful and is expected to be alternate treatment as the method to enhance the fatigue durability witho ut surface roughened. In addition, this technology should contribute to improvement of environment because of quite small quantity of ball usage. 5 References [1] Jean-Michel Duchazeaubeneix, SONATS, France, Proceedings of ICSP 7th , 1999, 444–452. [2] Keisuke TANAKA, Yoshiaki AKINIWA, Proc. of the International Conference on Fatigue in the Very High Cycle Regime, 61–71.

311

Residual Stress Relaxation and Fatigue Strength of AISI 4140 under Torsional Loading after Conventional Shot Peening, Stress Peening and Warm Peening Rainer Menig, Volker Schulze and Otmar Vöhringer Institut für Werkstoffkunde I, Universität Karlsruhe (TH), Karlsruhe, Germany

1

Abstract

Cylindrical rods of 450°C quenched and tempered AISI 4140 were conventionally shot peened, stress peened and warm peened while rotating in the peening device. Warm peening at Tpeen = 310°C was conducted using a modified air blast shot peening machine with an electric air flow heater system. To perform stress peening using a torsional pre-stress, a device was conceived which allowed rotating pre-stressed samples without having material of the pre-loading gadget between the shot and the samples. Thus, same peening conditions for all peening procedures were ensured. The residual stress distributions present after the different peening procedures were evaluated and compared with results obtained after peening of flat material of the same steel. The differently peened samples were subjected to torsional pulsating stresses (R = 0) at different loadings to investigate their residual stress relaxation behavior. Additionally, the pulsating torsional strengths for the differently peened samples were determined.

2

Introduction

Shot peening is widely used to improve fatigue limit and fatigue life of parts and components, e.g. springs, which are subjected to cyclic loading. This is done by well directed induction of compressive residual stresses and work hardening of the near surface regions. Besides amount and penetration depth of the compressive residual stresses their stability is of high importance. In case of conventional shot peening there are many investigations dealing with optimization of the peening parameters to improve the surface characteristics and their impact on fatigue strength. However, there are far less systematic investigations of modified shot peening treatments like stress peening and warm peening. Stress peening is used to induce higher compressive residual stresses into the surface regions. This is done by applying a tensile, bending or torsional load during the shot peening process. At this, the pre-stressing has to act into the same direction as the following operating loads. Warm peening, though, can be described as conventional shot peening conducted at elevated temperatures, usually between 170–350 °C [1–5]. This enables strain aging effects, which are known to increase work hardening rates and UTS and reduce the ductility in this temperature regime [6]. This can lead to increases of fatigue life and limit under stress controlled loading. The present paper deals with surface characteristics, residual stress stability and fatigue strength of cylindrical rod shaped samples of AISI 4140 after conventional shot peening, stress peening and warm peening. Comparisons will be drawn to flat bending samples of the same steel and the same material state.

312

3

Material, Specimen Geometry and Experimental Approach

The investigations were carried out using samples of the steel AISI 4140 with the chemical composition 0.44 C, 1.21 Cr, 0.22 Mo, 0.28 Si, 0.81 Mn, 0.07 Ni, 0.02 P, 0.03 Al, 0.02 S and balanced by Fe (all in wt.-%). The cylindrical rods with a diameter of 5 mm were delivered in 3 m long rods and cut down to pieces of 120 mm length. Then they were austenitized for 20 min at 850 °C, martensitically hardened in oil (25 °C), tempered at 450 °C for 2 hours and cooled down in a vacuum furnace. The shot peening treatments were performed using a modified air blast machine, which allowed peening at 20 °C £ Tpeen £ 410 °C. Details of the warm peening procedure can be found in [1, 5]. In this presentation warm peening was conducted at 310 °C, which was found to be an optimized peening temperature for quenched and tempered AISI 4140 subjected to alternating bending [5]. Cast iron shot S 170 with a hardness of 56 HRC was used at a peening pressure of 1.2 bar with a media flow rate of 1.0 kg/min. The Almen intensity was 0.25–0.27 mmA leading to a full coverage of the sample surface. For peening the device in Fig. 1 was used.

Figure 1: Device for shot peening, warm peening and stress peening of cylindrical rod shaped samples

Conventional and warm peening was conducted by mounting a sample (A) with collet chucks, closing the mechanical clutches (B) and (C) and using the geared motor (D) to rotate the shafts (1–4), which were connected by pairs of toothed wheels (E, F). Applying a torsional pre-load was done by mounting a sample, closing clutch (B) and locking shaft (1). Using a special device it was now possible to twist shaft (2), which applied a torsional pre-stress on sample (A). The pre-stress was controlled using a load cell (G). After reaching the appropriate torsional load the clutch (C) was closed and the pre-stressed system could be rotated with the gear motor (D). Peening was done by using one nozzle with a pendulum motion along the sample axis. The revolution speed of the samples was 125 rpm. The shot peened rods were cyclically tested under pulsating torsional stresses with a stress ratio R = 0 and a frequency of 20 Hz. This was done using torsional testing stands driven by servomotors [7]. The so-called step procedure [8] was used to determine the fatigue limit using 20 samples for each condition. Residual stresses in the 0°, 45° and 90°-direction of the specimen axis were determined using the X-ray technique. Measuring in those three directions allowed to calculate the surface parallel residual stress tensor, whose principle axis after stress peening and torsional loading are at 45° and 135°. The {211}-interference lines of the ferritic phase were analyzed according to

313 the sin2y method [9]. The depth distributions of the residual stresses were determined by iterative electrolytic removal of thin surface layers and subsequent X-ray measurements. Residual stress values measured at the surface after material removal were corrected according to the method of [10]. The half width values were determined as an average of those measured at O = –15°, 0° and +15°.

4

Results and Discussion

Residual stress and half width depth distributions of cylindrical rod samples after conventional peening, stress peening (Jpre /Jpo,2 = 0.51) and warm peening can be seen in Fig. 2. 4.0

200

(a)

3.5

0

(b)

3.0

HW [°2G]

-400

rs

Is [MPa]

-200

-600

conventional peening stress peening warm peening

-800 -1000 0.0

0.1

0.2

x [mm]

0.3

2.5 2.0 1.5 1.0 0.5 0.0 0.0

conventional peening stress peening warm peening 0.1

0.2

0.3

x [mm]

Figure 2: Residual stress (a) and half width (b) depth distributions after different peening procedures

The residual stress distribution after conventional shot peening is similar to the distribution found for flat bending samples [1]. The surface value is about –600 MPa. During stress peening principle stresses are evoked under 45° to the rod axis. This is the direction in which after shot peening clearly increased compressive residual stress values with a maximum of about –860 MPa are found. The depth where the residual stresses change their sign is slightly increased by stress peening. Warm peening yields compressive residual stresses at the surface similar to conventional shot peening. However, at the plateau under the surface the residual stresses are slightly increased. All in all the residual stress depth distributions are comparable with the ones found for flat bending samples [1]. The half width values at the surface region are increased by shot peening. However, warm peening does not lead to further increases, like it was found for flat samples [1, 5]. Fig. 3 shows the pre-stress influence on characteristic values of the residual stress state. The residual stresses found at 45° and 135° as well as the penetration depth of the compressive residual stresses are spread over the ratio of torsional pre-stress and torsional yield strength. In 45° direction the residual stresses are increasing slightly with increasing stress ratio. The residual stress relaxation found in 135°-direction, however, is more pronounced. The depth where the residual stresses change their sign is increasing with increasing pre-stress.

314 I45 I135

0.24

x0

0.23

600

0.22

400

0.21

x0

800

0.20 200

0.19

0 0.0

0.2

0.4

0.6

0.8

0.18 1.2

1.0

Jpre / Jp0,2 Figure 3: Residual stresses in 45° and 135°-direction after stress peening as well as the penetration depth of the compressive residual stresses spread over the ratio of pre-stress and torsional yield strength

Fig. 4 shows the residual stress relaxation of shot peened, stress peened (Jpre /Jpo,2 = 0.51) and warm peened samples which were subjected to pulsating fictitious torsional stresses of * a,s = 400 MPa. The 45°-direction is the direction in which tensile pre-stresses act during stress peening and tensile loading stresses during pulsating loading. In 135°-direction, there are compressive pre-stresses during stress peening and compressive loading stresses during pulsating loading. Fig. 4a shows that there is an obvious quasi static residual stress relaxation (N = 1) for the conventionally peened sample. This relaxation continues with increasing number of cycles and leads to remaining compressive residual stresses of less than 100 MPa. The stress peened sample, though, which yields smaller compressive residual stresses in this direction shows hard-400

0

(b)

(a) -100

-500

-600

-300

Is [MPa]

-400

rs

rs

Is [MPa]

-200

*

135 ° - direction (J a,s= 400 MPa) conventional peening stress peening warm peening

-500 -600

-700 *

45° - direction (J a,s= 400 MPa) conventional peening stress peening warm peening

-800

-900

-700

0

10

0

1

10

10

2

N

10

3

10

4

10

5

0

10

0

10

1

10

2

10

3

10

4

10

5

N

Figure 4: Residual stress relaxation in 135°-direction (a) and 45°-direction (b) due to pulsating loading stresses of * = 400 MPa after shot peening, stress peening and warm peening a, s

315 ly any residual stress relaxation during the first loading cycles, but strong relaxation between 10 £ N £ 104 leading to complete residual stress relaxation. The compressive residual stresses of the warm peened sample are clearly reduced during the first loading cycle like seen for the conventionally peened state, but further residual stress relaxation during cyclic loading is diminished, so that highest compressive residual stresses remain. In 45°-direction, almost no changes of the residual stresses are found for the conventionally and the warm peened sample up to N = 103. However, the stress peened sample shows a strong residual stress reduction within the first 10 cycles. This leads to smaller remaining compressive residual stresses, even though the initial stress state in this direction was clearly higher than for the other peening procedures. This investigation shows that residual stresses obtained through stress peening are strongly diminished if sufficiently high loading stresses are applied. Therefore, the residual stress values after high numbers of cycles are in the same region (135°) or even below (45°) the values found after conventional shot peening. However, warm peening leads to increases in residual stress stability. This is caused by dynamic and static strain aging effect acting during and after the warm peening process. Those effects result in a diffuse dislocation structure, which is stabilized by carbon atoms through the formation of so-called Cottrell-clouds and very small carbides. Similar results were found for differently shot peened flat bending samples [1], where stress peening caused high residual stress values but small resistance against residual stress relaxation. However, warm peening caused clearly improved residual stress stability. Optimized peening temperatures of about 310 °C were found for the bending samples [5]. Cyclic torsional tests with pulsating loads were conducted to determine the pulsating torsional strength of the differently peened variants. The torsional strength of the only quenched and tempered material state of 285 MPa is increased by conventional shot peening up to 400 MPa. Stress peening, however, using a pre-stress of 408 MPa (Jpre /Jpo,2 = 0,51) does not lead to further improvements of the pulsating torsional strength. One reason for this is the high amount of residual stress relaxation seen in Fig. 4. However, it is presumed that for harder material states, where residual stress relaxation does not play such an important role with respect to improvements of the fatigue limit, stress peening should be appropriate to increase the torsional strength. Another reason why no improvements were found in this case could be the increased thickness of the rod shaped samples (5 mm) compared with the thickness of the bending samples (2 mm) which are pulled up for comparison. This leads to a reduction of the stress gradient and could stimulate subsurface cracks. Surprisingly, also warm peening did not improve the pulsating torsional strength, despite improvements of the residual stress stability. It is possible that again the reduced stress gradient in case of the cylindrical rod shaped samples caused subsurface cracks. This, however, must not be generalized as e.g. results of [4] show, where warm peening led to increases of the torsional strength. In case of flat AISI 4140 bending samples, optimized warm peening even led to increases of the alternating bending strength of 59 % compared to the ground state [5].

5

Conclusion

Residual stress and work hardening state, residual stress stability and fatigue limit of quenched and tempered cylindrical rods of AISI 4140 were investigated after conventional shot peening, stress peening and warm peening. The results were compared with results of flat alternating bending samples of the same steel after the same peening procedures were conducted. Conven-

316 tional shot peening and warm peening yield same surface compressive residual stresses of about –600 MPa. Warm peening slightly shifted the subsurface plateau of the conventionally peened sample to higher compressive residual stresses, whereas the penetration depth of the compressive residual stresses did not change for those two peening procedures. After stress peening increased compressive residual stresses (about –800 MPa) were found in the previously prestressed (Jpre /Jpo,2 = 0.51) direction and the penetration depth was slightly increased. The same amount of work hardening at the surface was found for the different shot peening procedures. Stress peening caused higher residual stresses in the direction which was tensile pre-stresses during the peening procedure and lower residual stresses in the direction which was compressed during the peening. This effect was more pronounced for higher pre-stresses. The penetration depth of the compressive residual stresses was increased with increasing pre-stress. Pulsating torsional loading cycles led to pronounced residual stress relaxation for stress peened samples. The overall residual stress decrease after N = 104 led to same or even smaller values compared with conventional shot peening. However, warm peening increased the residual stress stability. This is caused by dynamic and static strain aging effects leading to a diffuse dislocation structure which is stabilized by carbon atoms and very small carbids. The torsional pulsating strength of the only quenched and tempered condition (285 MPa) was improved by shot peening (400 MPa). Stress peening and warm peening did not lead to further improvements for this steel. In case of stress peening this is caused by the high amount of residual stress relaxation. Furthermore, the 5 mm diameter of the rod specimens could stimulate subsurface cracks. It can be stated that surface characteristics and residual stress relaxation behavior is similar to previous results found for flat bending samples. However, in case of bending samples in the same quenched and tempered condition increases of the alternating bending strength were 11 % and 33 % for stress peening and optimized warm peening, respectively, compared to conventionally peening conditions.

6 [1] [2]

References

A. Wick, V. Schulze, O. Vöhringer, Mat. Sc. and Eng. 2000, A293, 191-197. M. Schilling-Praetzel, F. Hegemann, G. Gottstein, Proc. of the 5th Int. Conf. on Shot Peening, (Ed.: D. Kirk), Oxford 1993, 227-238. [3] A. Tange, H. Koyama, H. Tsuji, J. Schaad, Technology for Product and Process Integration (SP-1449), International Congress and Exposition, Detroit 1999. [4] A. Rössler, J. K. Gregory, Ermüdung hochharter Stähle (Ed.: H. Bomas), Berichtsband AWT-Tagung am 21. und 22. Juni 2001 in Weimar, IWT Stiftung Institut für Werkstofftechnik, Bremen 2001, 89-103. [5] R. Menig, V. Schulze, O. Vöhringer, Mat. Sc. Eng. 2002, in print. [6] J. D. Baird, Iron & Steel, 1963, 186-192 and 326-334. [7] T. Beck, B. Denne, K.-H. Lang, D. Löhe: Tagungsband "DVM Werkstoffprüfung 1999", Bad Nauheim, DVM, Berlin 1999, 291-300. [8] M. Hück, Z. Werkstofftechnik 1983, 14, 406-417. [9] E. Macherauch, P. Müller, Z. f. angewandte Physik 1961, 13, 340-345. [10] M. J. Moore, W. P. Evans, Trans. SAE 1958, 66, 340-345.

317

Influence of Optimized Warm Peening on Residual Stress Stability and Fatigue Strength of AISI 4140 in Different Material States Rainer Menig, Volker Schulze, Otmar Vöhringer Institut für Werkstoffkunde I, University of Karlsruhe (TH), Karlsruhe, Germany

1

Abstract

Using a modified air blasting machine warm peening at 20 °C £ T £ 410 °C was feasible. An optimized peening temperature of about 310 °C was identified for a 450 °C quenched and tempered steel AISI 4140. Warm peening was also investigated for a normalized, a 650 °C quenched and tempered, and a martensitically hardened material state. The quasi static surface compressive yield strengths as well as the cyclic surface yield strengths were determined from residual stress relaxation tests conducted at different stress amplitudes and numbers of loading cycles. Dynamic and static strain aging effects acting during and after warm peening clearly increased the residual stress stability and the alternating bending strength for all material states.

2

Introduction

Shot peening is widely used as a mechanical surface treatment for many components such as crankshafts, gears, springs etc. Many studies deal with optimizing peening parameters, but the increases of fatigue life and strength obtained by conventional shot peening are limited [1]. This led to modifying the shot peening process. A possibility is stress peening with tensile pre-loads applied during shot peening [2,3]. However, the superposition of high values of residual stresses with loading stresses can lead to an early residual stress relaxation as soon as the respective yield strength is reached during quasi static and/or cyclic loading [4]. Therefore, in recent years, shot peening at elevated temperatures was investigated [5-10]. It was found that dynamic and static strain aging effects during and after shot peening can stabilize the dislocation structure and therefore contribute to a higher stability of the induced residual stresses. A procedure to evaluate the stability of residual stresses under quasi static loading and to estimate the surface compressive yield strength Re(c),s after shot peening is given in [6] using

Re( c ), s = (I srs )2 + I srsI s*, crit + (I s*,crit )2

(1)

with the initial residual stress value at the surface I srs and the critical stress I s*,crit , which initiates the onset of residual stress relaxation in compression. A measure of the cyclic yield strength at the surface Recycl , s after shot peening can be calculated using the modified Eq. (1) rs 2 rs * * 2 Recycl , s = (s s, N =1 ) + s s , N =1 s a , crit + (s a , crit )

(2)

318 with the critical loading stress amplitude I s*, crit and the corresponding surface residual stress values remaining after the first cycle at the same quasi static load, Irss, N =1 [4,6].

3

Material and Specimen Geometry

Investigations were carried out on steel samples of AISI 4140 (German grade 42 CrMo 4) with the chemical composition 0.42 C, 1.04 Cr, 0.14 Mo, 0.21 Si, 0.71 Mn, 0.01 P, 0.02 Al and balance Fe (all in wt. %). The bending samples were machined from flat material by sawing and milling, and ground to a thickness of 2.2 mm. Details about size and manufacturing can be found in [6]. Afterwards, they were heat treated into normalized (930 °C, 3h), quenched and tempered as well as martensitically hardened material states. Martensitic hardening in oil was done after austenitization at 850 °C for 20 min. Subsequent tempering was conducted at 450 °C (T450) and 650 °C (T650) for 2 h, respectively. In order to eliminate distortions, the specimens were finally ground to a thickness of 2.0 mm. Table 1 shows the yield strength, the UTS and the hardness for the different material states. Table 1: Yield strength and hardness of the different material states material state

yield strength [MPa]

UTS [MPa]

hardness [HV0.3]

normalized

400

740

230

T650

700

940

300

T450

1200

1375

430

hardened

1500

2400

660

4

Experimental Approach

An air blast machine was used to perform the shot peening treatments. Using an upgraded air flow heater system warm peening at 20 °C £ T £ 410 °C was feasible [11]. Cast iron shot S 170 with hardness 56 HRC was used at a peening pressure of 1.2 bar and a media flow rate of 1.0 kg/min. Further details are given in [11]. Residual stresses in the longitudinal direction of the specimens were determined using the X-ray technique. The {211}-interference lines of the ferritic phase were analyzed according to the sin2O - method [12]. The depth distributions of the residual stresses were determined by iterative electrolytic removal of thin surface layers and subsequent X-ray measurements. Residual stress values measured at the surface after material removal were corrected according to the method of [13]. The half width values as a measure of microstructural work hardening were determined as an average of those measured at O = –15°, 0° and +15°. For each S-N curve 25 to 30 specimens were used to determine the alternating bending strength Rab. The failure probabilities were evaluated according to the arcsin P -method [14]. Tests to determine the stability of the surface residual stresses by alternating bending at a fixed initial stress amplitude were performed for different stress amplitudes. For each stress amplitude exactly one specimen was used. At predefined numbers of cycles the tests were interrupted, the surface residual stress values were measured at both sides, averaged and then the tests were restarted.

319

5

Results and Discussion

The influence of shot peening temperatures 20 °C £ Tpeen £ 410 °C on the formation of half width and residual stress values at the surface of AISI 4140 quenched and tempered at 450 °C is shown in Fig 1. The initial half width value of the conventionally peened sample (Tpeen = 20 °C) of about 3.3° 2G in Fig. 1a is slightly increased to about 3.45 °2_ by peening at Tpeen = 290 °C. Further increases of the peening temperature lead to a continuous decrease of this value. The variant peened at 20 °C shows a strong reduction of the half width with increasing numbers of cycles after applying an alternating bending load of I*a, s = 1000 MPa. The half widths of the variants peened at elevated temperatures (T £ 290 °C) remain stable if the fluctuations typical for half width determination are considered. The results of the shot peening induced surface compressive residual stresses for 20 °C £ Tpeen £ 410 °C are shown in Fig. 1b. The initial values (N = 0), each averaged from at least 5 samples measured on both sides, show a slight maximum at peening temperatures of 310 °C and 330 °C. It is conceivable that at those temperatures the speed of diffusing carbon atoms equals the velocity of moving dislocations, which is the basis for profound dynamic strain aging. Through this a very diffuse dislocation structure with a high dislocation density is created. At even higher temperatures thermally induced dislocation movement leads to rearrangement and annihilation of dislocations, resulting in a micro stress relaxation and again decreased compressive residual stresses. For Tpeen = 310 °C and 330 °C a strongly reduced decrease of residual stress relaxation was found when loaded at alternating bending with I*a, s = 1000 MPa. Residual stress and half width depth distributions for variants quenched and tempered at 450 °C and shot peened at different temperatures, including Tpeen = 310 °C and 330 °C, can be seen in [11]. No difference in the half width depth distributions of the warm peened samples can be found. There is a strong work hardening in the surface zone for all elevated peening temperatures. For Tpeen ³ 300 °C it was necessary to increase the hot air flow phot which was used to heat up the samples and to hold the appropriate peening temperature from 1.2 bar to 2.0 bar [11]. This was done without changing the actual peening pressure. This did not affect the surface residual stresses but slightly increased the penetration depth of the compressive residual stresses [11]. 0

3.8

a)

b)

T450

T450

-200

Is [MPa]

3.4 3.2 3.0

rs

HW [°2G]

N=0 N=1 N = 10 N = 100 N = 1000

I*a,s = 1000 MPa

-100

3.6

N=0 N=1 N = 10 N = 100 N = 1000

2.8 2.6

-300 -400 -500 -600

I*a, s = 1000 MPa

-700 -800

2.4 0 20

280

300

320

340

Tpeen [°C]

360

380

400

420

0 20

280

300

320

340

360

380

400

420

Tpeen [°C]

Figure 1: Half widths (a) and residual stresses (b) vs. peening temperature after different numbers of cycles

To investigate the residual stress stability after Tpeen = 310 °C under quasi static and dynamic loading, alternating bending tests were carried out under variation of the applied stress amplitu-

320 de and of the numbers of cycles after which the remaining residual stresses were determined. This is shown in Fig. 2, additionally, results of [6] for Tpeen = 290 °C are given. The critical load for quasi static residual stress relaxation I s*, crit is increased from 500 MPa to 560 MPa. The surface compressive yield strengths Re(c),s calculated using Eq. 1 and the method of [4] are summarized in Table 2. Furthermore results of [6] for samples peened at Tpeen = 20 °C and 290 °C are given. The ratio of the compressive yield strengths Re(c),s after shot peening and the value of the core region Re [4] is also given. It can be seen that Re(c),s /Re is < 1 for all variants. 0 -100

Tpeen = 290 °C Tpeen = 310°C

-200

N=1

N=1 N = 10

4

N = 10

4

-400

*

Ia‚crit = 714 MPa

-500

*

Is‚crit = 500 MPa

rs

Is [MPa]

-300

-600 -700 -800 -900 200

*

Ia‚crit = 750 MPa

*

Is‚crit = 560 MPa 300

400

500

600 *

700

800

900

1000

1100

*

|Is|, Ia,s [MPa] Figure 2: Residual stresses vs. loading stresses or stress amplitudes for samples peened at Tpeen = 290 °C and 310°C

Table 2: Quasi static surface compressive yield strength for 450 °C quenched and tempered AISI 4140 after conventional peening and warm peening material state

Tpeen [°C]

I s*, crit [MPa]

I srs [MPa]

|Re(c),s| [MPa]

Re(c),s /Re

T450

20

-310

-600

801

0.60

290

-500

-660

1008

0.78

310

-560

-700

1186

0.91

Table 3: Cyclic surface yield strength for 450 °C quenched and tempered AISI 4140 after conventional peening and warm peening material state

Tpeen [°C]

I s*, crit [MPa]

Irss, N =1 [MPa]

T450

20

514

-520

895

0.82

290

714

-620

1156

1.07

310

750

-675

1235

1.14

Recycl , s [MPa]

cycl Recycl , s / Re

321 This work softening of the shot peened surface is a consequence of the Bauschinger effect, which is caused by the reversed deformation (compression) compared with the deformation during the peening process (tensile deformation). This effect can be relatively pronounced in quenched and tempered steels and is obviously reduced due to warm peening, especially due to optimized warm peening at Tpeen = 310 °C. The dislocations are locked through diffusion of carbon atoms preferred to edge dislocations and the formation of very small carbides. Dislocation movement is retarded and the residual stresses are stabilized. The loading stress amplitude initiating stress relaxation I s*, crit at cyclic loading is also indicated in Fig. 2. Table 3 summarizes all values necessary to calculate the cyclic yield strengths at the surface after shot peening. Additionally, the ratios of those cyclic yield strengths and the values of the core region are listed. Latter were taken from [15]. The conventionally peened variant shows a work softening at the cycl surface compared to the core region (Tab. 2, Recycl = 0.82). Warm peening even causes a , s / Re work hardening at the surface with ratios of 1.07 and 1.14 for Tpeen = 290 °C and 310 °C, respectively. This is due to the stabilized dislocation structure caused by dynamic and static strain aging during warm peening and during cooling, respectively. In Fig. 3 S-N curves for variants shot peened at different temperatures are compared with the one of the ground state [6]. The values given are valid for a failure probability of P = 50 %. The alternating bending strength of the ground state Rab = 443 MPa [6] is increased by optimized shot peening at Tpeen = 310 °C and Tpeen = 330 °C up to 704 MPa (+59 %). 1000 Tpeen = 310 °C (phot= 2 bar) Tpeen = 330 °C (phot= 2 bar) Tpeen = 290 °C (phot= 2 bar) Tpeen = 290 °C (phot= 1.2bar) Tpeen = 20 °C (phot= 1.2bar) Rab = ground state

900

*

Ia,s [MPa]

800

704 MPa

700

644 MPa 640 MPa

600

530 MPa

500

443 MPa

50 % 400 10

4

10

5

10

6

10

7

10

8

Figure 3: S-N curves for differently shot peened variants in comparison with the ground material state

According to the procedure described before, the com-pressive yield strengths and the cyclic yield strengths were also calculated for different material states which were all peened at Tpeen = 290 °C. Detailed results as well as depth distributions can be found in [16]. Fig 4 shows a summariy of the surface yield strengths found in quasi static and cyclic loading (Fig. 4a) and of the alternating bending strengths found for the ground, shot peened and warm peened condition (Fig. 4b) spread over the Vickers hardness. Note that warm peening of the martensitically hardened material state causes a reduction in hardness from 660 HV to 560 HV. The yield strengths of the shot peened states (Fig. 4a) point out firstly that there is an increase with increasing hardness, and secondly that warm peening generally increases the quasi static as well as

322 the cyclic yield strengths of the different material states, compared with conventional peening. Only the T450 shows clear differences between its quasi static and cyclic yield strength found for the same peening temperature. The alternating bending strengths at a failure probability of P = 50 % (Fig. 4b) show that warm peening leads to increases for all material states, compared with conventional shot peening. This is caused by the higher stability of the dislocation structure and residual stresses caused by dynamic and static strain aging.

a)

D HV

quenched

b)

800

DHV

1200

Rab (P = 50%) [MPa]

cycl

|Re(c),s|, Res [MPa]

1400

1000 800 600

Tpeen = 20°C Tpeen = 290°C

|Re(c),s|

400

Res 300

400

600

T650 normalized

400

warm peened (Tpeen = 290°C) shot peened (Tpeen = 20°C) ground

200

cycl

200 200

T450

500

600

HV 10

700

200

300

400

500

600

700

HV10

Figure 4: Compressive yield strengths, cyclic yield strengths (a) and the alternating bending strengths (b) after different peening procedures vs. the material state

5

Summary and Conclusion

A modified air blast shot peening machine with an upgraded air flow heater system was used to conduct warm peening at 20 °C £ Tpeen £ 410 °C. Optimized peening temperatures for a material state quenched and tempered at 450 °C were identified to be between Tpeen = 310 °C and 330 °C. The influence of different shot peening temperatures on the stability of residual stresses at alternating bending was investigated for AISI 4140 in different material states. Compressive yield strengths |Re(c),s| as well as cyclic yield strengths Recycl , s found after shot peening and warm peening were determined and compared. Warm peening led to increases of |Re(c),s| and Recycl ,s . Alternating bending strengths found after warm peening were increased for all material states compared with those found after conventional peening. This is caused by pinning of dislocations by carbon atom clouds and the creation of very small carbides leading to a highly stabilized dislocation structure with a beneficial effect to the residual stress stability.

6 [1] [2] [3]

References H. Holzapfel, A. Wick, V. Schulze, O. Vöhringer, Härterei-Technische Mitteilungen 1998, 53, 155–163. F. Engelmohr, B. Fiedler, Materialwissenschaft und Werkstofftechnik 1991, 22, 211–216. E. Müller, Hoesch Berichte aus Forschung und Entwicklung 1992, 23–29.

323 [4] [5] [6] [7] [8] [9] [10]

[11] [12] [13] [14] [15] [16]

H. Holzapfel, V. Schulze, O. Vöhringer, E. Macherauch, Mater. Sci. and Eng. 1998, A248, 9–18. A. Wick, V. Schulze, O. Vöhringer, Materialwiss. und Werkstofftechnik 1999, 30, 269–273. A. Wick , V. Schulze. O. Vöhringer, Mater. Sci. and Eng. 2000, A293, 191–197. M. Schilling-Praetzel, F. Hegemann, G. Gottstein, Proc. of the 5th Int. Conf. on Shot Peening, (Ed.: D. Kirk), Oxford, 1993, 227–238. A. Tange, H. Koyama, H. Tsuji, Spring Conf. of the Japan Society for Spring Research (Ed.: M. Allen), Springs 58, 2000. A. Tange, H. Koyama, H. Tsuji, J. Schaad, Technology for Product and Process Integration (SP-1449), International Congress and Exposition, Detroit 1999. A. Rössler, J. K. Gregory, Berichtsband AWT-Tagung “Ermüdung hochharter Stähle” am 21. und 22. Juni 2001 in Weimar (Ed.: H. Bomas), IWT Stiftung Institut für Werkstofftechnik, Bremen, 2001, 89–103. R. Menig, V. Schulze, O. Vöhringer, Mat. Sci. and Eng. 2002, in print. E. Macherauch, P. Müller, Z. f. angewandte Physik 1961, 13, 340–345. M. J. Moore, W.P. Evans, Trans. SAE 1958, 66, 340–345. D. Dengel, Zeitschrift für Werkstofftechnik 1975, 8, 253–261. A. Eifler, Dr.-Ing. thesis, University Karlsruhe (TH) 1989. R. Menig, Dr.-Ing thesis, University Karlsruhe (TH) , 2002, in preparation

324

Thermal Fatigue of Shot Peened or Hard Turned Hot-Work Steel AISI H11 Martin Krauß, Berthold Scholtes Institute of Materials Technology, University of Kassel, Kassel, Germany

1

Introduction

The lifetime of hot-work tools is often limited by the development of thermal fatigue cracks, which occur as a consequence of alternating temperatures connected with thermal stresses [1, 2]. It is well known, that in the case of mechanically loaded components, strength or lifetime can be enhanced by appropriate mechanical surface treatments. This is mostly attributed to near surface residual stress distributions and strain hardening effects [3 - 5]. Although in the case of combined thermo-mechanical loading, near surface microstructures are expected to be less stable compared to isothermal conditions at moderate temperatures, the question arises, whether such beneficial effects can also be realized in the case of thermal fatigue. In what follows, characteristic results of such investigations are presented.

2

Material Investigated and Experimental Details

The specimens used in the present investigation have been machined from a forged ingot, with the chemical composition (wt-%) 0.37 C, 1.2 Si, 0.23 Mn, 4.96 Cr, 1.25 Mo, 0.45 V, 0.003 P, 0.002 S, Fe bal. (AISI H11, german grade X38CrMoV5-1). The steel was heat treated (austenitisated for 20 min at 1015 °C, quenched in oil, two times tempered for 2 h at 625 °C) to achieve a hardness of 43 HRC. In Figure 1 the quasi-homogeneous microstructure of the tempered martensite, with some fine carbides, is shown. Tensile properties at room temperature were: Rp0.2 = 1153 MPa, UTS = 1384 MPa. From heat treated blanks cylindrical specimens with a diameter of 7 mm and a gauge length of 10 mm were machined. To achieve identical starting

Figure 1: Optical micrograph of the annealed hot-work tool steel AISI H11

325 conditions, at first all samples were hard turned with the same machining parameters like the reference state (vc = 150 m/min, f = 0.06 mm, a = 0.25 mm, diluted soluble oil: Plasocut 2000 (7 %), carbide tip: Sandvik DCMT II T3 08-PM P15). This pre-treatment was followed, according to the variation of the surface states, by two different shot peening treatements or by deep rolling. The combined shot peening treatments were carried out on a pneumatically operated industrial blast jet system. In the first processing step for the coarse peened variation blast grain of larger diameter and a several times higher intensity was used, than for the weak peened variation. The second processing step was identical for both variations. For deep rolling a spherical rolling element (Æ6.6 mm) was used with a feed of 0.1125 mm per revolution and a rolling pressure of 150 bar. Thermal fatigue experiments simulating complete suppression of the thermal expansion were carried out under axial loading conditions [6]. For this, a servohydraulic fatigue testing device with a maximum load capacity of 160 kN was used. Heating of the specimens was performed by a 5 kW induction heating system, cooling was realized by forced air, blown by three nozzles around the heating coil. For temperature measurement, a Ni-CrNi thermocouple was spot welded in the middle of the gauge length. The axial strain was controlled and measured with a high temperature capacitative extensometer. All tests were carried out with a triangular temperaturetime-course, a heating or cooling rate of 10 °C/s and a minimum temperature Tmin of 200 °C. Maximum temperatures Tmax of 550, 600, 625 or 650 °C were chosen. At the beginning of each test, specimens were heated up under free expansion to the mean temperature Tm = (Tmin + Tmax)/2. After reaching Tm, the total strain was kept constant by the mechanical loading system. Consequently during the following thermal cycling, out of phase loading conditions with maximum compressive thermal stresses at the maximum temperature result. From the registred stress-temperature hysteresis loops stress amplitudes, mean stresses and plastic strain amplitudes were calculated. Residual stresses were determined by X-ray diffraction technique [7], using the interference of CrKa-radiation at the {211}-planes of the annealed martensite. For stress evaluation, the sin2y-method was applied and the elastic constant ½s2 = 6,09 * 10–6 Mpa–1 was used. Residual stress depth profiles were determined without correction of stress relieve by successive electrochemical materials removal. To estimate micro residual stresses, caused by work hardening, full width at half maximum (FWHM) values of X-ray interference lines were determined.

3

Results

3.1

Near Surface Materials States

Compared with the hard turned reference state (Rmax = 4.7 μm), surface roughness was diminished by deep rolling (Rmax = 1.9μm) and increased by shot peening (Rmax = 9.9 and 11.6 μm resp.). Near surface depth distributions of residual stresses and half-width-values of the different materials states investigated are shown in Figure 2. In all cases, compressive residual stresses were observed immediately at the surface, which were lowest for the hard turned state and highest for the deep rolled one. Considerable differences exist with respect to the thicknesses of the affected surface layers. Whereas in the case of hard turned specimens, only a thin layer with compressive residual stresses was produced, coarse shot peening and deep rolling produced much thicker layers with compressive residual stresses. Similar observations can be made regar-

326 ding depth distributions of interference line FWHM-values. Deep rolled and shot peened specimens exhibit maximum values immediately at the surface, which decrease drastically within a small distance from the surface. Lowest values are observed for the hard turned reference state.

residual stress [MPa]

0

-400

-800

-1200

hard turned weak shot peened coarse shot peened deep rolled

4,5

FW HM [°]

4,0 3,5 3,0 2,5 0,0

0,2

0,4

0,6

distance from surface [mm] Figure 2: Residual stress and full with at half maximum values after different mechanical surface treatments

3.2

Results of Thermal Fatigue Tests

For all materials states investigated, in the main, the same behavior was observed during thermal fatigue loading. Characteristic results are summarized in Figure 3, using specimens of the coarse peening condition. Here the development of stress amplitude and mean stress for the cyclic temperature ranges indicated are given. In the case of an upper temperature of 550 °C, up to 104 cycles a constant stress amplitude, but an increasing tensile mean stress is observed. For lower numbers of cycles, increasing cyclic temperature ranges lead to increasing stress amplitudes as well as mean stresses. The formation of mean stresses is due to the fact, that maximum compressive thermal loading stresses occur at highest temperatures, whereas lowest thermal stresses occur at lower temperatures (out of phase loading conditions) [8]. Due to the temperature dependence of yield strength, compressive plastic deformation at higher temperatures is prevailing, leading to a contraction of the specimen. However, it can be observed, that for higher numbers of cycles, stress amplitudes as well as mean stresses tend to relax, which can be attributed to strain softening processes as well as crack formation and propagation [9, 10]. In the same way, a characteristic influence of the cyclic temperature range on the plastic strain amplitude can be detected (see Figure 3, bottom). It increases with the cyclic temperature range and, except for the lowest temperature range of 350 °C investigated, with increasing numbers of tem-

327 perature cycles. Again, the temperature dependence of the yield strength is decisive, but also damage and strain softening processes at higher numbers of cycles. As expected, fatigue lifetime also decreases with increasing temperature amplitude. 750

mean stress / stress amplitude [MPa]

stress amplitude

500

mean stress 250

temperature: 200 - 650°C 200 - 625°C 200 - 600°C 200 - 550°C

plastic strain amplitude Aa,p [‰]

0 0,4

0,2

0,0 1

10

100

1000

10000

number of cycles [-]

Figure 3: Stress amplitude, mean stress (top) and plastic strain amplitude (bottom) as a function of cyclic temperature ranges during thermal fatigue

Only insignificant consequences of the different near surface materials states on characteristic properties during thermal fatigue experiments have been observed. One example is presented in Figure 4. Here, mean stresses are plotted as a function of the number of thermal cycles for a cyclic temperature range of 425 °C. Clearly different plots can be seen with lowest values for the hard turned reference state and highest values for deep rolled specimens. Results of shot peened specimens can be found between these distributions. This observation can be explained by the amount and distribution of near surface compressive residual stresses, which control the amount of plastic compression during the first loading cycles. Onset of plastic compression is determined by the superposition of residual and thermal loading stresses [11]. This effect is the more pronounced, the thicker the surface layer with compressive residual stresses is. An important consequence of such pronounced plastic deformations already during the first loading cycles is, that near surface materials states are drastically altered. Especially residual stresses relax considerably, as can be seen in Figure 5 for a coarse peening condition and a cyclic temperature range of 400 °C. It is clearly to be seen, that residual stresses relax preferably

328

mean stress [MPa]

300

200

temperature: 200 - 625°C hard turned weak shot peened coarse shot peened deep rolled

100

0 1

10

100

1000

10000

number of cycles [-] Figure 4: Mean stress evolution during thermal fatigue for specimens with different surface conditions

200

residual stress [MPa]

0 -200

-400 -600

temperature range: 200 ... 600°C number of cycles: 0 35 1 100 10 1000

-800

FWHM [°]

4,0 3,5 3,0 2,5 2,0 0,0

0,2

0,4

0,6

distance from surface [mm] Figure 5: Residual stress and FWHM values of coarse shot peened surfaces after thermal fatigue

already during the first loading cycle and decrease further with increasing numbers of thermal cycles. This stress relaxation is much more pronounced compared with the case of pure thermal stress relaxation [12]. FWHM-values decrease also considerably as well as hardness values (not

329 shown here), indicating microstructural alterations in near surface layers during the thermal fatigue process. As a consequence of residual stress relaxation and microstructural alteration during the first stage of thermal fatigue, for the conditions investigated here, no significant influence of the near surface materials state on lifetime was detected. Obviously, such effects can only be expected in the case of moderate temperatures and cyclic temperature ranges connected with small plastic strain amplitudes.

4

Conclusion

To analyze the influence of near-surface materials states on the damage process during thermal fatigue loading, isothermal annealing experiments and thermal fatigue tests with hot-work tool steel AISI H11 (German grade X38CrMoV 5-1) in either shot-peened or hard turned surface condition were performed. Tension-compression thermal fatigue tests were carried out using a servohydraulic testing system, keeping the macroscopic strain of the gauge length constant, while triangular time-temperature cycles were applied. The temperature range was chosen in such a way that numbers of cycles to fracture between 1,000 and 10,000 were achieved. The development of plastic strain amplitudes as well as of mean stresses were analyzed during the tests. All surface types examined revealed similar lifetimes for identical thermal loadings. While the residual stress state during simple annealing was still relatively stable, almost all residual stresses relaxed during a few thermal fatigue cycles. In addition, also the work hardening effects produced by shot-peening in near-surface areas were not stable under thermal fatigue loading conditions. Universal hardness measurements revealed a decrease of work hardening in nearsurface areas, even below the level of the bulk material. Similar results were obtained by performing X-ray diffraction experiments and analyzing FWHM-values.

5

Acknowledgements

The authors are grateful to Böhler-Uddeholm Deutschland GmbH for providing the material, Metal Improvement Company, Unna, for shot peening treatments and AiF Otto von Guericke e. V. as well as VDEh-Gesellschaft zur Förderung der Eisenforschung for financial support.

6 [1] [2]

[3]

References L. Kindbohm, Archiv Eisenhüttenwesen 35 (1964), p. 773–780 C. Rosbrook, R. Shivpuri: A Computer-Aided Investigation of Heat Checking and Die Life Prediction in Die Casting Dies, paper T93-071, NADCA, Rosemont, 1993, p. 181–190 B. Scholtes: Structural and Residual Stress Analysis by Nondestructive Methods (ed. V. Hauk), Elsevier, Amsterdam (1997), p. 590

330 [4]

Mechanische Oberflächenbehandlungen (ed. H. Wohlfahrt, P. Krull), Willey-VCH, Weinheim (2000) [5] I. Altenberger, J. Gibmeier, R. Herzog, U. Noster, B. Scholtes: Analysis and Assessment of Residual Stress States in Mechanically Surface Treated Materials, Materials Science Research Int. - Special Technical Publication, Vol. 1 (2001), p. 275 [6] Y. Pan, K.H. Lang, D. Löhe, E. Macherauch: Cyclic Deformation and Precipiation Behaviour of NiCr22Co12Mo9 during Thermal Fatigue, phys. stat. sol. (a) 138, (1993) p. 133–145 [7] B. Scholtes: Eigenspannungen in mechanisch randschichtverformten Werkstoffzuständen, DGM Informationsgesellschaft Verlag, Oberursel, 1991 [8] B. Kleinpaß, K.-H. Lang, D. Löhe, E. Macherauch: Influence of the minimum cycle temperature on the thermal-mechanical fatigue behaviour of NiCr22Co12Mo9, in: Schriften des Forschungszentrums Jülich, Reihe Energietechnik, vol. 5, 3 (1998), p. 1396 - 1377 [9] R. Hallstein: Das Verhalten von Gußeisenwerkstoffen unter isothermer, thermischer und thermisch-mechanischer Wechselbeanspruchung, Dr.-Ing. Thesis Univ. Karlsruhe (1991) [10] A. Oudin, F. Rézai-Aria: Temperature dependence of thermo-mechanical fatigue behaviour of a martensitic 5% chromium steel, proceedings Euromat 2000, vol. 2, Elsevier, Amsterdam (2000), p. 1053–1058 [11] H. Holzapfel: Das Abbauverhalten kugelstrahlbedingter Eigenspannungen bei 42CrMo4 in verschiedenen Wärmebehandlungszuständen, Dr.-Ing. Thesis Universität Karlsruhe (1991) [12] H. Hanagarth: Auswirkungen von Oberflächenbehandlungen auf das Ermüdungsverhalten von TiAl6V4 u. 42CrMo4 bei erhöhter Temperatur, Dr.-Ing. Thesis Univ. Karlsruhe (1989)

331

Effect of Short-Time Annealing on Fatigue Strength of Shot Peened AISI 4140 in a Quenched and Tempered Material State Rainer Menig, Volker Schulze and Otmar Vöhringer Institut für Werkstoffkunde I, Universität Karlsruhe (TH), Karlsruhe, Germany

1

Abstract

Increases of residual stress stability and alternating bending strength of shot peened AISI 4140 are obtained by subsequent annealing treatments. This is caused by static strain aging effects, which lead to pinning of dislocations by carbon atoms and finest carbides. It will be shown that by well directed annealing of a quenched and tempered AISI 4140 it is possible to maximize the positive effects of static strain aging without causing extended thermal residual stress relaxation. The amount of flow stress increases caused by static strain aging is quantified and correlated with the gain in alternating bending strength.

2

Introduction

It is known that shot peening at elevated temperatures, so-called warm peening, can significantly increase fatigue strength and life of quenched and tempered steels [1-2]. This is caused by dynamic and static strain aging effects [3]. For a quenched and tempered AISI 4140 an optimized peening temperature of about 310 °C was found [2,4]. When steel is thermally treated at elevated temperatures after being plastically deformed, an increase in yield stress due to static strain aging is observed. This could also be obtained by shot peening as a plastic deformation process with subsequent annealing, though this is hardly investigated yet [5]. Such static strain aging effects may be used to increase the fatigue limit and life of peened steels without the sumptuous experimental set-up of warm peening. However, at the same time thermal activated residual stress relaxation must be considered [6]. This paper explains the benefits of static strain aging on residual stress stability and on alternating bending strength of samples conventionally shot peened and subsequently annealed for relatively short times (1 – 60 min) at different temperatures. It will be shown that it is possible to separate the positive influence of static strain aging from the detrimental effects of thermal residual stress relaxation.

3

Material and Experimental Approach

The investigations were carried out using samples of the steel AISI 4140 (German grade 42 CrMo 4) with the chemical composition 0.42 C, 1.04 Cr, 0.14 Mo, 0.21 Si, 0.71 Mn, 0.01 P, 0.02 Al and balance Fe (all in wt. - %). The 110 mm long hourglass shaped samples with a minimal width of 18 mm were machined from flat material and ground to a thickness of 2.2 mm. Then they were austenitized for 20 min at 850 °C, martensitically hardened in oil (25 °C), tempered at 450 °C for 2 hours and cooled down in a vacuum furnace. After the heat treatment, the

332 samples were ground to a thickness of 2.0 mm in order to eliminate distortions in their flatness. The final geometry of the samples can be seen in [1,2]. The shot peening treatments were performed using an air blast machine. The samples were peened from both sides simultaneously in order to avoid distortions. Cast iron shot S 170 with a hardness of 56 HRC was used at a peening pressure of 1.2 bar with a media flow rate of 1.0 kg/min. The Almen intensity was 0.24 mmA. The subsequent annealing was conducted in a salt bath furnace at defined temperature/ time combinations. Residual stresses of the specimens were determined using X-ray technique. The {211}-interference lines of the ferritic phase were determined at 9 O-angles between –60° and +60° using CrKa-radiation and analyzed according to the sin2O-method [7]. Neglecting the elastic anisotropy, a Young’s modulus E = 210 GPa and a Poisson’s ratio n = 0.28 were used. Mechanical testing was carried out on alternating bending machines at a frequency of 25 Hz. Between 25 and 30 specimens were used for each S-N-curve. The results were evaluated using the arcsin P -method [8].

4

Results and Discussion

After shot peening compressive residual stresses I 0FS at the surface of about –660 MPa were measured. These samples were then annealed for 1, 5 or 60 min at different temperatures. The annealing caused corresponding residual stress relaxation, which can be described using a Zener-Wert-Avrami function

I rs (Ta , ta ) m = exp{-[C × exp(-DH / kTa ] × ta } I 0rs

(1)

where I 0FS is the initial residual stress state and IFS (Ta, ta) the remaining amount of residual stresses after annealing at temperature Ta and time ta. DH is the activation enthalpy, k the Boltzmann constant, and C and m are material related constants [6]. The thermal residual stress relaxation, which is increasing with increasing annealing temperature Ta can be seen in Fig. 1a-c (N = 0). For longer annealing times ta the residual stress decline is stronger. With Eq.1 it is possible to determine the activation enthalpy DH as well as the constants C and m by using an iterative mathematical procedure based on a least squares algorithm [9]. The results for C, m and DH found are 4.96·1013 1/min, 0.138 and 2.23 eV, respectively. The calculated activation enthalpy of 2.23 eV is close to the value of the activation enthalpy for self-diffusion of a-iron (.2.6 eV). This implies that the main microstructural process responsible for the residual stress relaxation is controlled by volume diffusion controlled creep, which is determined mainly by climb of edge dislocations. Using Eq. 1 and the values found for C, m and DH the thermal residual stress relaxation is calculated and is spread in Fig. 1a - c (N = 0). It is in good accordance with the values measured. The ratios of the surface compressive residual stresses remaining after annealing plus defined loading amplitudes at alternating bending are also shown in Fig. 1. The applied alternating bending stress amplitude was I*a, s = 1000 MPa. The samples aged only at ambient temperature face a strong compressive residual stress relaxation of almost 50 % after the first bending cycle. Subsequent bending cycles (N = 10, 102, 103) provoke further stress relaxation down to a value of about 20 % of the initial compressive residual stress value. Increasing annealing temperatures lead to an increasing amount of remaining residual stresses, after quasi

333 static (N = 1), and cyclic (N = 10, 102, 103) loading. Above a certain annealing temperature, a maximum of remaining residual stresses is found, depending on the annealing time. This is caused by static strain aging with the formation of carbon atom clouds mainly around edge dislocations and the associated stabilization of the dislocation structure. At further elevated temperatures the remaining residual stresses decrease again. The optimum annealing temperatures for AISI 4140 decrease with increasing annealing time from about 300 °C–350 °C for ta = 1 min to 230 °C–260 °C for ta = 60 min. 1.0

1.0

ta = 1 min

annealing + N cycles *

*

at Ia,s = 1000 MPa rs

N=0 N=1 N = 10 N = 100 N = 1000

0.6

rs

I (Ta,ta,N) / I 0

rs

0.6

rs

I (Ta,ta,N) / I 0

0.8

N=0 N=1 N = 10 N = 100 N = 1000

0.8

ta = 5 min

annealing + N cycles

at Ia,s = 1000 MPa

0.4

0.4

0.2

0.2 0

100

200

300

0

400

b)

Ta [°C]

a)

100

200

300

400

Ta [°C]

1.0

annealing + N cycles

ta = 60 min

*

at Ia,s = 1000 MPa N=0 N=1 N = 10 N = 100 N = 1000

rs

I (Ta,ta,N) / I 0

0.8

rs

0.6

0.4

0.2

c)

0

100

200

300

400

Ta [°C]

Figure 1: Ratios of the surface compressive residual stresses after annealing and defined alternating bending cycles

Static strain aging tests were conducted over a wide range of annealing temperatures and times to determine the yield stress increases DIssa caused by static strain aging. After tensile loading to a total strain of 2 % the tensile samples specimen in a salt bath in the same way as the bending specimen. Afterwards they were reloaded until failure occurred. The yield stress increase caused by static strain aging was measured and is spread in Fig. 2 versus the annealing temperature Ta. There is a maximum in yield stress increase in the temperature range 250 °C £ Ta £ 325 °C for the times investigated. This temperature range corresponds with the range in Fig. 1 for which the combined thermal and mechanical residual stress re-laxation decreases. For higher annealing temperatures over-aging with coarsening of carbides which is combined with dislocation rearrangements occurs and the values of DIssa decrease again. This is more pronounced the higher the annealing times and temperatures are. The kinetics for coarsening of carbides are governed by the slowest participating partner, which is self-diffusion of a-iron instead of diffusion of carbon atoms during the formation of carbon clouds.

334 175

150

,Issa [MPa]

125

100

ta = 1 min ta = 5 min ta = 10 min ta = 20 min ta = 60 min ta = 100 min

75

50

25 200

250

300

350

400

Ta [°C] Figure 2: Yield stress increases in static strain aging tests vs. annealing temperature

It is unknown to which scale the alternating bending strength Rab of the as shot peened state is improved by different annealing treatments, because not only the aging effects have to be taken into account, but also the detrimental effects of thermal residual stress relaxation. Therefore, the alternating bending strengths of shot peened samples were determined after annealing at different temperature/time pairs. For the annealing time ta = 1 min temperatures of 235 °C, 280 °C and 300 °C were chosen to determine the respective alternating bending strengths. According to Fig. 1a those temperatures are equivalent to increasing residual stress stabilization, which was found after mechanical loading of specimens annealed for 1 min. The yield strength increases DIssa shown in Fig. 2 are also increasing at ta = 1 min with increasing annealing temperature up to a maximum of DIssa = 150 MPa at 300 °C. Additionally, alternating bending strengths were determined for ta = 60 min at 220 °C and 255 °C. According to Fig. 1c both variants are taken from the range of highest residual stress stabilization when annealed at ta = 60 min. However, in Fig. 2 at ta = 60 min the maximum of ,Issa is not yet reached at these temperatures. In [1] no increase of Rab was found after the same shot peened steel was annealed at 300 °C / 20 min. However, Fig. 2 shows that for this annealing treatment there is an obvious increase of DIssa , which should improve the alternating bending strength. For clarification, Rab was determined additionally for 300 °C / 20 min. All results of the alternating bending strengths are shown in an extended residual stress Haigh - diagram in Fig. 3 where the alternating bending strengths Rab are spread versus the residual stresses before cyclic loading. The solid Goodman line connects the Rab of the nearly residual stress free ground state [1] (on the ordinate) with the UTS (1375 MPa) at the tensile side of the residual stress axis (not plotted in Fig. 3). The slope yields a residual stress sensitivity of 0.32, which is a value expected for quenched and tempered AISI 4140. The Goodman line is extended to the compressive side of the Haigh – diagram. This describes the expected increase of Rab caused by the induced or remaining compressive residual stresses. However, as soon as the

335 sum of loading stresses and assumed uniaxial residual stresses trespasses the cyclic yield strength Recycl , s residual stress relaxation occurs and no further increase of Rab is possible. This is FS indicated by the dotted lines (I = Recycl , s - s ). The one for the as shot peened state is originating cycl at Re, s = 895 MPa [1] and the one for the shot peened plus annealed state originates at Recycl ,s = 1052 MPa [10] (note, that in Fig. 3 only the range 300 MPa £ Rab £ 900 MPa is shown). The increases of Rab compared with the as shot peened state are between about 60 MPa and 150 MPa despite thermal residual stress relaxation caused by the annealing treatments. This indicates, that the increases of Rab are caused by static strain aging. The extended Goodman line found for 300 °C / 1 min does not describe the respective alternating bending strength satisfactorily, though the tendency to increasing alternating bending strengths is clearly given. It is conceivable, that Rab of a shot peened and annealed state, which would be assumed to be free of residual stresses could also be raised by short time annealing and that therefore it is not appropriate to use the regular Goodman line to describe the alternating bending strengths found after shot peening plus annealing. 900

800

255°C/60min 280°C/1min

700

300°C/20min

Rab [MPa]

220°C/60min

500

300°C/1min

235°C/1min

600

as shot peened [1] ground [1]

400

Tpeen = 20°C + 300°C/1min

300 -1000

-800

Tpeen = 20°C -600 rs Is

-400

-200

0

[MPa]

Figure 3: Extended residual stress Haigh – diagram for shot peened as well as shot peened plus annealed samples of AISI 4140

In Fig. 4 the increases of the alternating bending strength DRab compared with the as shot peened state are spread versus the flow stress increases DIssa introduced in Fig. 2. The increases of the alternating bending strength correlate roughly linearly with the yield stress increases DIssa . After annealing for 1 min the strong increases of DIssa with increasing annealing temperature cause corresponding DRab-increases. This correlates well with the increased stability of the residual stresses seen in Fig. 1a. While annealing for 60 min at 220 °C and 255 °C leads to similar remaining residual stresses after cyclic loading (Fig. 1c), strong changes in the alternating bending strength can be seen from Fig. 4. This shows that not only the residual stress stability but also the static strain aging itself lead to increases of the alternating bending strength. Finally, it can be stated that in contrast to [1] annealing at 300 °C / 20 min also increases Rab of the as shot peened state.

336 180 160

300 °C / 20 min

140

,Rab [MPa]

120

255°C / 60 min

100

235 °C / 1min 80

280 °C / 1min 300 °C / 1min

60 40

220 °C / 60 min 20 0 80

90

100

110

120

130

140

150

160

170

,Issa [MPa] Figure 4: Effect of yield stress increase on the alternating bending strength caused by different annealing treatments

5

Conclusion

Shot peening with subsequent annealing was conducted. The residual stress relaxation behavior caused by the annealing process was modeled using a Zener-Wert-Avrami function. The activation enthalpy for residual stress relaxation was determined to 2.23 eV, which is in the range of the activation enthalpy for self diffusion of =-iron. Therefore, volume diffusion controlled creep is considered to be the main microstructural process that leads to residual stress relaxation. Well directed annealing treatments performed after shot peening clearly decrease the amount of mechanical residual stress relaxation in alternating bending compared to the as shot peened state. This is caused by the formation of carbon atom clouds around dislocations and very small carbides leading to highly stabilized dislocation structures. Tensile tests were conducted to quantify the yield stress increases caused by static strain aging. The results confirmed the temperature/ time ranges for which highest reductions of mechanical residual stress relaxation were found. The alternating bending strengths for selected shot peened and annealed variants were determined and compared with the alternating bending strength of the as shot peened state. Pronounced increases of the alternating bending strength after short-time annealing could be identified. Excessive annealing temperatures lead to over-aging, which is caused by coarsening of carbides and the rearrangement of dislocations. As long as over-aging is prevented it is possible to correlate the alternating bending strength increases with the flow stress increases found in tensile tests.

337

6 [1] [2]

References

A. Wick, V. Schulze, O. Vöhringer, Mat. Sci. and Eng. 2000, A293, 191–197. R. Menig, V. Schulze, O. Vöhringer, A. Wick, 20th ASM Heat Treating Society Conference Proceedings (Ed.: K. Funatani, G.E. Totten), ASM International, 9–12 October 2000, St. Louis, MO, 2000, 257–264. [3] J. D. Baird, Iron & Steel, 1963, 186-192 and 326–334. [4] R. Menig, V. Schulze, O. Vöhringer, Mat. Sci. and Eng., 2002, in print. [5] I. Altenberger, B. Scholtes, Scripta Mater. 1999, Vol.41, No. 8, 873–881. [6] O. Vöhringer,Advances in surface treatment (Ed.: A. Niku-Lari) International Guidebook on residual stresses, Vol. 4; Pergamon Press, Oxford, New York, Paris, 1987, 367–396. [7] E. Macherauch, P. Müller, Z. f. angewandte Physik 1961, 13, 340–345. [8] D. Dengel, Zeitschrift für Werkstofftechnik 1975, 8, 253–261. [9] V. Schulze, F. Burgahn, O. Vöhringer, E. Macherauch, Mat. wiss. u. Werkst. 1993, 24, 258–267. [10] R. Menig, V. Schulze, O. Vöhringer, Z. Metallkunde, 93 (2002), in press.

338

Effect of Shot Peening on Improvement of Fatigue Strength for Metal Bellows H. Okada* , A. Tange* , Kotoji Ando** *NHK SPRING CO.,LTD. ,Kanagawa-Pref, Japan **Yokohama National University, Kanagawa-Pref, Japan

1

Introduction

Generally metal bellows (it is hereafter described as a bellows) are known for a seal to defend a leak being elastic1).Recently small size bore bellows are getting to expect to a seal of pump which was small and can use under the circumstances of high pressure.However, until now, it would be difficult to achieve the fatigue strength for required specifications in most cases. Meanwhile it was known for that shot peening had an effect on improving fatigue strength2).It can be expected that it is difficult to obtain a large compressive residual stress on the inner surface of small bore size bellows by shot peening process.In order to obtain the effect of shot peening on the inner surface of bellows, a new shot peening processes by using a air peening machine and a reflective plate were developed.It was found that this method can apply the shots to the inner surface of bellows effectively, to improve the fatigue life of bellows, comparing with nonshot peening one. It can be also noted that the measurements of residual stress by X-ray diffraction is not applicable to the material of bellow, SUS304 stainless steel. The measurement of residual stress is essential to determine shot peening conditions. Since the structure of SUS304 bellows is mixed with austenite and stress-induced martensite by cold forming the signal to noise ratio is too small to show the peak of X-ray diffraction. In order to solve these problems, the SUS631 stainless steel bellows was selected because it was a good workability and had a lot of martensite. The optimum shot peening conditions were decided by using residual stress of SUS631 bellows.

Figure 1: Cross section of bellows

339

2

Materials and Experimental Procedures

2.1

Materials

The specifications of U-shaped bellows are shown in Table 1. The radius of bent is about 0.5mm due to the small pitch as 1.5mm. The material is austenitic stainless steel SUS304. Fig.1 shows sectional shape cut in axis. Each section of bellows can be said as follows, outer diameter’s area is the top, inner diameter’ area is the bottom, outer diameter’s direction is outside, inner diameter and centerline’s direction is inside. The manufacturing processes of bellows are to set the tube of which diameter is the same as inner diameter of bellows, to the mold divided into two. The interval of mold is contracted with giving internal pressure to the inside of the tube. It must be cautious that an explosion and buckling don't occur during the processes. This process of manufacturing can be well known as hydraulic bulge forming process. Fig. 2 shows the hardness distribution of bellows. The top is hard in comparison with bottom because of work hardening effect. Fig. 3 shows optical microstructure of bellows. It can be seen from the Fig. 3 that the top has a lot of martensite structure in comparison with the bottom.

Figure 2: Hardness distribution of bellows

2.2

Figure 3: Optical microstructure of bellow

Shot Peening Processes

Bellows are usually used under compression. It occurs tension stress in outside of outer diameter and inside of inner diameter, where can be the origin of fatigue fracture in most cases. Therefore, the shot peening should be done in those area. The shot peening machine used here is suction peening machine in air peening machine. The reflection plate shown in Fig4 installed to the nozzle, in order to apply the shots effectively to the inside surface of bellows3). Table 2 shows shot peening conditions. The pressure is from 0.1 MPa to 0.7 MPa. The nozzle diameter is 8 mm in the case of inner, 9 mm (normal size) in the case of outer. The shot materials used is glass beads (550 HV). The glass beads diameter used are from 38 μm to 215 μm. The shot peening processes were carried out under the condition that nozzle was fixed. The turning bellows can be moved from the top to the bottom four times within two minutes. After that, the direction of bellows is made up-down reverse, and it goes in the same action for two minutes. The total shot peening time is, therefore, four minutes. The outside of outer diameter was also shot peened for two minutes by using the standard nozzle without reflection plate.

340

Figure 4: Reflection shot peening system

2.3

Residual Stress Measurement

It would be essential to measure residual stress, to decide shot peening conditions. Therefore, the residual stress measurement by X-ray stress measurement device, was made under the conditions shown in Table 3. The colimeter diameter is 0.5mm since bellows is small pitch. The measurements were made on the outside surface of outer diameter in the circumferential direction. The measurements were done at both (311)C and (211)= because SUS304 steel was mixed with austenite and martensite by cold forming. However, in both cases it was not found the peak value to decide the residual stress. In order to solve these problems, the SUS631 stainless steel bellows was chozen because it was a good workability and had a lot of martensite. The residual stress measurement under the conditions shown in Table 3, became possible by the SUS631 bellows. This means that the SUS631 bellows used for the purpose of deciding the optimum conditions of shot peening, and the SUS304 bellows applied by the optimum shot peening conditions were carried out in the fatigue tests.

2.4

Fatigue Tests

The fatigue test machine of the crank type where the displacement can be adjustable, was used. The frequency is 4 Hz. The fatigue test condition is pulsating stress.

341

3

Experimental Results

3.1

Relationship Between Residual Stress and Pressure

The relationship between residual stress and pressure in glass beads size 58 μm was studied. Fig. 5 shows the results. While the residual stress is –80 MPa in non-shot peened bellows, it can be seen from the Fig. 5 that the residual stress in shot peened bellows is over –700 MPa. The compressive residual stress tends to increase in proportion with increasing the pressure. It was measured –957 MPa in 0.7 MPa pressure. -700

0 N o n - s h o t pe e n in g R e s idu a l s t r e s s ( M P a )

R e s idu a l stress

cM Pay

- 200 - 400 - 600 - 800 -1000 -1200

-800

-900

- 1000

- 1100

0

0 .2

0 .4

0 .6

0 .8

0

50

100

150

200

250

P re s s u re c M Pay G la s s be a ds s iz e c my z

Figure 5: Relationship between residual stress and pressure, glass beads size, 58 μm

3.2

Figure 6: Relationship between residual stress and glass beads size, pressure 0.7 MPa

Relationship Between Residual Stress and Glass Beads Size

Fig. 6 shows the relationship between residual stresses and glass beads size in pressure 0.7 MPa. It can be seen from the Fig. 6 that the compressive residual stress tends to increase in proportion with the smaller glass beads size.

3.3

Relationship Among Pressure, Glass Beads Size and Fatigue Strength

Fig. 7 shows the relationship between the pressure, glass beads size and fatigue strength. It can be seen from the Fig. 7 that the fatigue strength tends to increase in proportion with increasing the pressure. This tendency is more remarkable in the case of a larger glass beads. The optimum conditions of shot peening can be said to be the pressure 0.7 MPa, glass beads size 97 μm in these results. Although it is a large compressive residual stress in glass beads size 38 μm, the fatigue life is low. This reason can be thought to be due to the depth of residual stress that can be discussed later.

3.4

S-N Diagram

Under the optimum conditions of shot peening experimentally decided (the pressure 0.7 MPa and glass beads size 97 μm), the S-N diagram of SUS304 bellows about shot peening and non-

342

( c y c le s )

N u m be r o f c y c le s t o f a ilu r e

1 .E +0 6

1 .E +0 5

38μm 58μm 97μm 1 .E +0 4 0 .2

0 .4

0 .6

0 .8

P re s s u re ( M P a )

Figure 7: Relationship between number of cycles and pressure (Stress amplitude: 485 MPa)

Figure 8: SUS304 S-N diagram Pressure 0.7 MPa, Glass beads size 97 μm

shot peening can be shown in Fig. 8. While fatigue limit of 107 cycles is 240 MPa in non-shot peened bellows that is 420 MPa in shot peened bellows. It can be realized that fatigue strength be improved about 75 % by shot peening in the optimum conditions.

4

Discussion

4.1

The Effect of Residual Stress

It has been normally discussed that the residual stress, work hardening and surface roughness can be the major factors on improvement of the fatigue strength by shot peening. The fatigue test condition is pulsating stress in this time. In order to realize any change of residual stress before and after fatigue tests, fatigue tests and residual stress measurements by using SUS631 bellows, were carried out4). The fatigue test condition is the same as SUS304 bellows. The compressive residual stress before and after fatigue tests was measured. It was 957 MPa before fatigue tests and was –833 MPa after fatigue tests. This result can say that while the small amount of compressive residual stress can be released by fatigue tests, most can be remained after fatigue tests. Therefore, the reason why the fatigue strength was improved by shot peening is, obtaining a larger compressive residual stress – being remained the compressive residual stress during fatigue tests.

4.2

The Effect of Work Hardening

The surface hardness before and after fatigue tests were measured, in order to know whether there has the effect of work hardening for the reason why the fatigue strength was improved by shot peeping. The measuring point was the inside of inner diameter where can be the origin of fatigue fracture in this case. The measurements were carried out from the surface to inside at intervals of 5 μm, total 10 point by micro vickers hardness (load 49mN). Fig. 9 shows the results. Remarking the results of 0.7 MPa/97 μm condition in the Fig.9, which is the optimum shot peening condition, it can be seen that although the bellows before fatigue tests is work-hardened to

343 show higher hardness, it becomes soft after fatigue tests. This can say that the bellows has already been hardened by cold forming and be hardened further by shot peening. Therefore, the work hardening can be one reason why the fatigue strength be improved by shot peening.

460

Non-shot peening 38 μm, 0.3MPa 97μm, ^ 0.7 MPa

Non-shot peening 38μm, ^ 0.3MPa 97μm, 0 ^.7MPa

460 440

Hardness iHV,Load 4

440 420

420

400

400

380

380

360

360

340

340

320

320

300

300

280

280 0 20 40 60 Distance from inside surface (μm) Before fatigue test

0

10 20 30 40 50 Distance from inside surface (μm) After fatigue test

Figure 9: Comparison of hardness distributions

Figure 10: SEM observation of bellows surface before shot peening

4.3

Surface Roughness

The surface of bellows cold formed is usually getting rough by the hydraulic bulge forming processes. It can be seen from the Fig. 10 that there are micro cracks of about 2 μm round the grain boundary when observed with scanning electron microscope (SEM). In order to know the influence of the surface roughness to the fatigue strength, the surface roughness before fatigue tests was measured. The measuring point is the outside of outer diameter. Fig. 11 shows the relationship between the pressure, glass beads size and surface roughness. It can be seen that the surface roughness is improved by shot peening, comparing with non-shot peened bellows. It can be also realized that the surface roughness has no correlation with glass beads size, and tends to be improved in proportion with increasing pressure. It can be considered that the surface of bellows becomes smooth by being struck repeatedly with the particles of glass bead. Therefore, the reduced depth of micro crack and improved surface roughness by shot peening can be one reason why the fatigue strength was improved.

344 4.4

The Effect of Residual Stress Depth

The residual stress mentioned above is the residual stress value on the surface of bellows. The distributions of residual stress against the thickness of bellows were obtained, as shown in Fig. 12. In the Fig. 12, the residual stress measurements were made at the outside of outer diameter before fatigue tests. It can be seen from the Fig. 12, that the residual stress distributions of which the fatigue strength is low, is shallow and those of which the fatigue strength is high, is deep. It can be concluded that the deeper and larger compressive residual stress can delay the development of micro-crack caused by the bulge forming processes5). S u rf a c e ro u gh n e s s ( μ m )

18

38 μm 97 μm

16

14

12

10

38μm, 0.3MPa 97μm,0.7MPa

8 0

0 .2

0 .4

0 .6

0 .8

P re s s u re ( M P a )

Figure 11: Relationship between surface roughness and pressure

5 1. 2.

3.

4.

μm

Figure 12: Distributions of residual stress of bellows before fatigue test

Conclusions It becomes possible to make shot peening to the inside of small bore by developing the special nozzle which is set up a reflection plate to the inside of small bore size bellows. It can be noted that the greater compressive residual stress (-957 MPa) on the inner surface of bellows can be obtained by shot peening. Since the residual stress is not released much during fatigue tests, the fatigue limit of 107 cycles shows 420 MPa in shot peened bellows, comparing with 240 MPa in non-shot peened bellows. It is realized that fatigue strength is improved about 75 % by shot peening. The fatigue strength is improved in proportion with increasing the pressure as for the relationship between pressure, glass beads size and fatigue strength. Specifically, the tendency is recognized remarkably as much as a glass beads size is larger. It can be said that the pressure 0.7 MPa and glass beads size 97 μm be the optimum shot peening condition in these experimental result. It can be thought that deeper and larger compressive residual stress, higher work-hardening(hardness), and smaller surface roughness can be the reason why the fatigue strength of shot peening condition of (3) is the highest.

345

6 [1] [2] [3] [4] [5]

References T. Mitsushiba, a vacuum, 26-10 (1983), 757. Shot peening technical association, The method and an effect of shot peening, (1997), p. 13, Nikkan Kogyo Shimbun Ltd.. H. Ishigami, K. Matsui, Y. Jin and K. Ando Fatigue and Fracture of Engineering Materials & Structures (Accept) S. Takahashi, M. Hashimoto, H. Hirose and T. Sasaki, The collection of the Japan Society of Mechanical Engineers papers(A piecies),66-646(2000),101. K. Yamada, M. Ishida, K. Uzumaki, H. Suzuki and K. Teradoko, The collection of the spring technical research meeting spring lecture meeting lecture papers,(1999), 3

346

1

VII Fatigue of Light Weight Alloys

2

339

Property Improvement in Light Metals Using Shot Peening Jean K. Gregory1) and Lothar Wagner2) 1)

Springfield Metallurgical Services, Inc., Springfield, VT, USA Chair of Physical Metallurgy and Materials Technology, Technical University of Brandenburg at Cottbus, Cottbus, Germany

2)

1

Abstract

The goal of this overview is to outline potential improvements in fatigue performance of light metals by shot peening. The overview is aimed at correlating the process-induced changes in surface layer properties (e.g., roughness, near-surface depth profiles of dislocation densities and residual stresses) of titanium, aluminum and magnesium based alloys with the changes in resistances to fatigue crack nucleation and microcrack propagation. Characteristic examples are presented for each alloy system including =, (=+>), metastable > titanium alloys and titanium aluminides, aluminum alloys with various age-hardening conditions as well as magnesium alloys. Depending on the alloy system, shot peening-affected surface layer properties such as process-induced damage, work-hardening and residual stresses and their cyclic stability were found to be the main parameters in affecting fatigue performance of light-weight alloys.

2

Introduction

The reduction of vehicle weight by using light-weight structural alloys such as those based on magnesium, aluminum and titanium for body as well as suspension parts is one of the most promising ways to reduce fuel consumption. Good fatigue performance is an important requirement for these applications. Therefore, the effect of shot peening on the fatigue performance of light-weight alloys is of particular importance. Among the various surface layer properties which are changed through shot peening, e.g., roughness, hardness, residual stresses and crystallographic textures [1], residual compressive stresses play a dominant role regarding fatigue performance since they can substantially hinder and even stop growth of surface cracks [2]. To what degree residual stresses can be effective mainly depends on their depth profile in the surface layer and its cyclic stability. The cyclic stability of the shot peening-induced residual stress-depth profile of the well known (=+>) titanium alloy Ti-6Al-4V is illustrated in Figure 1 [3]. The higher the stress amplitude (for a given number of cycles) in fatigue loading, the higher is the reduction in magnitude of the residual stress-depth profile. In addition to residual stresses, surface damage and work-hardening can be important parameters depending on the alloy system [4, 5]. This overview will present and discuss results which are characteristic for each alloy system. Results on titanium alloys will be contrasted to aluminum and magnesium alloys.

350

Residual stress [MPa]

200 0 Ia

-200

850 MPa

-400

650 MPa

-600

450 MPa

-800 as shot peened

-1000 0

100

200

300

400

500

Distance from surface [mm] Figure 1: Effect of cyclic loading on shot peening-induced residual stress profile in Ti-6Al-4V [1]

3

Experimental

Shot peening was performed on a variety of light-weight alloys based on titanium, aluminum and magnesium and the effects on surface layer properties as well as on fatigue response were studied. In the case of titanium, alloys studied were Ti-2.5Cu, Ti-6Al-7Nb and Ti-3Al-8V-6Cr4Mo-4Zr (Beta C) belonging to the =, (=+>) and metastable > alloy systems, respectively. In addition, a titanium aluminide Ti-47Al-3.7 (Nb, Cr, Mn, Si)-0.5B (C-TiAl) was investigated. The well known AlCuMg based aircraft alloy 2024Al was taken in the tempers T3 and T6 to represent an age-hardenable aluminum alloy. Furthermore, the response of the high strength magnesium alloy AZ80 (Mg-8Al-0.5Si) to shot peening was investigated. Tensile tests on the various alloys were performed on threaded cylindrical specimens having gage lengths and diameters of 20 and 4 mm, respectively. The initial strain rate was 8.3 x 10-4 s1 . Tensile test results are listed in Table 1. To determine the cyclic deformation of the various alloys which is known to affect the cyclic stability of residual stresses, stress controlled low cycle fatigue (LCF) tests were performed on threaded cylindrical specimens having gage lengths and diameters of 20 and 4 mm, respectively. These tests were done in stress control under axial loading at a stress ratio of R = -1 by means of a servohydraulic testing machine. The test frequency was 0.1 Hz. During testing, the axial strain was recorded by strain gages. From the hysteresis loops, the plastic strain was measured at zero load and plotted versus number of cycles. For high cycle fatigue (HCF) testing, hour glass shaped specimens with a gage diameter of 3.6 mm were machined. Part of the specimens was electropolished (EP) to serve as reference. 100 μm were removed from the as-machined surface to ensure that any machining effect that could mask the results was absent. The other part was shot peened (SP) by means of an injector type machine using various shot materials including spherically conditioned cut wire SCCW14 (0.36 mm average shot size), cast steel S 230 and S 330 with 0.6 and 0.8 mm average shot sizes, respectively as well as spherical zirconia based ceramic shot with an average diameter of 0.5 mm. During the peening treatment, the specimens rotated at 1 s-1. The distance between nozzle tip and specimen surface was 45 mm. Peening was done at full coverage to various Almen in-

351 tensities ranging from 0.05 mmN to 0.28 mmA. The change in surface layer properties caused by shot peening was characterized by measurements of surface roughness through profilometry, microhardness-depth profiles, half width breadth- and residual stress-depth-profiles measured by X-ray diffraction techniques. In addition, residual stresses were measured by means of the incremental hole drilling method as described elsewhere [6]. High cycle fatigue (HCF) tests were performed mainly in rotating beam loading (R = –1) at frequencies of about 60 Hz. Some tests were done in axial loading using a servohydraulic testing maching at R = 0.1 and frequencies of about 60 Hz. Fatigue fracture surfaces were studied by SEM. Table 1: Tensile properties of the various light alloys Alloy

Microstructure

E [GPa]

Ti-2.5Cu

equiaxed

107

Beta C

as-solutionized

I0.2 [MPa]

UTS [MPa]

El [%]

AF = ln A0/ AF

610

745

20.0

0.55

80

840

850

25.0

1.12

122

920

995

13.5

0.50

duplex / WQ

114

1030

1120

14.8

0.54

C-TiAl

fully lamellar

170

440

440

1.0

0.01

Al 2024

T3

72

360

550

13.2

0.21

T6

72

420

510

10.9

0.33

as-extruded (L)

44

245

340

12.0

0.20

Ti-6Al-7Nb duplex / AC

600*

AZ80

* in compression

4

Results and Discussion

In addition to the stress amplitude in fatigue loading (Fig. 1), the cyclic stability of shot peening-induced residual stresses will depend on cyclic yield stress of the material. Presumably, materials which cyclically soften will exhibit less stable residual stresses compared to materials which cyclically harden.

4.1

and Metastable

Titanium Alloys

The cyclic deformation behavior of the various alloys used in this study is illustrated in Figure 2 comparing results on the = titanium alloy Ti-2.5Cu (Fig. 2a) with results on the metastable >alloy Beta C (Fig. 2b). Cyclic hardening was observed in Ti-2.5Cu (Fig. 2a) while marked cyclic softening was found in Beta C (Fig. 2b). This difference in cyclic deformation behavior corresponds to the different work-hardening capacity (UTS – I0.2) of the two alloys observed in tensile loading (Table 1) amounting to 135 MPa and only 10 MPa for Ti-2.5Cu and Beta C, respectively. As expected, shot peening

352 10.0

Plast. strain ),pl/2 [%0]

Plast. strain, ),pl/2 [%0]

10.0

Ia = 730 MPa 650 MPa 1.0

630 MPa

Ia = 850 MPa 1.0

820 MPa 700 MPa 0.1

0.1 1

10

100

1000

1

10

Cycles, N

100

1000

Cycles, N

a) Ti-2.5Cu

b) Beta C

Figure 2: Cyclic deformation behavior (R = –1)

700 SP

600 EP

500 400

Stress amplitude, Ia [MPa]

Stress amplitude, Ia [MPa]

leads to marked increases in the fatigue life and HCF strength of Ti-2.5Cu (Fig. 3a) whereas only slight increases in fatigue life and no HCF improvement were observed on Beta C

300

700 600

SP

500

EP

400

300 103

104

105

106

107

103

Cycles to failure, NF

a) Ti-2.5Cu

104

105

106

107

Cycles to failure, NF

b) Beta C

Figure 3: S-N curves in rotating beam loading (R = –1)

(Fig. 3b). Similarly, this comparatively poor response of metastable > titanium alloys to shot peening was observed in work on Ti-10V-2Fe-3Al [7] and LCB [8] which exhibit similar cyclic softening indicating again that residual compressive stresses in this alloy group are not very effective presumably, due to cyclic decay. However, mechanical surface treatments such as shot peening or roller-burnishing of the metastable > titanium alloy Beta C can be used to preferentially age the affected surface layer as demonstrated in Figure 4. The marked hardening in the surface layer which is the result of the effect of cold work prior to aging can stabilize the residual compressive stress field thus increasing the HCF strength as illustrated in Figure 5 [9,10]. Stabilizing of shot peening-induced residual stresses was also observed in selectively surface hardened TIMETAL 21s [11] and in bake-hardening steels [12].

353

Stress amplitude, Ia [MPa]

800 Microhardness, HV 0.025

700

SP + A

600 500

SP 400 300 200 100

700

600 RB +A

500 RB

400

300

0 0

100

200

300

103

400

104

Distance from Surface [μm]

Figure 4: Microhardness-depth profile in Beta C, after shot peening (SP) and preferential aging (SP+A)

4.2

(

105

106

107

Cycles to failure, NF

Figure 5: S-N curves (R = -1) in Beta C, after roller-burnishing (RB) and preferential aging (RB+A)

+ ) Titanium Alloys

(=+>) titanium alloys such as Ti-6Al-4V or Ti-6Al-7Nb tend to exhibit slight cyclic softening. Thus, their response to shot peening is not as good as that of the = alloys but also not as poor as that of the metastable > alloys. However, this alloy group tends to exhibit a so called anomalous mean stress sensitivity [13], i.e., tensile mean stresses can dramatically decrease the HCF strength. This can be important also in fully reversed (R = –1) loading of shot peened conditions since residual compressive stresses are balanced by residual tensile stresses in deeper regions which can lead to subsurface fatigue crack nucleation. S-N curves of the (=+>) titanium alloy Ti-6Al-7Nb with a duplex (primary =in transformed >matrix) microstructure exhibiting the anomalous mean stress sensitivity are shown in Figure 6. 900 Stress amplitude, Ia [MPa]

Maximum stress, ımax [MPa]

900 800 R = -1

R = 0.1

700

600 500

400

800 SP

700 EP

600

500 400

104

105

106

107

Cycles to failure, NF

a) Effect of mean stress

104

105

106

107

Cycles to failure, NF

b) Effect of shot peening (R = -1)

Figure 6: S-N curves in Ti-6Al-7Nb (Duplex/AC)

As seen in Figure 6a, the 107 cycles fatigue strength in terms of maximum stresses at R = 0.1 is not higher than at R = –1. Evidently, the material has a low resistance to fatigue crack nucleation at tensile mean stresses. This behavior directly shows up in the HCF performance of shot

354 peened specimens (Fig. 6b) since fatigue cracks were found to nucleate in subsurface regions (Fig. 7).

Figure 7: Subsurface fatigue crack nucleation in shot peened specimens of Ti-6Al-7Nb

This anomalous mean stress sensitivity in duplex microstructures of (=+>) titanium alloys can be eliminated by increasing the strength of the lamellar (transformed >component through an increased rate of cooling from the duplex anneal, e.g., by using water-quenching instead of air cooling [14]. As seen in Figure 8, the faster cooling rate not only increases the fatigue strength particularly at tensile mean stresses (Fig. 8a), but also drastically increases the fatigue performance of shot peened specimens tested at R = –1 (Fig. 8b). 900 Stress amplitude, Ia [MPa]

Stress amplitude, Ia [MPa]

900 R = 0.1

R = -1

800

700

600

500

400

800 SP

700 EP

600

500

104

105

106

Cycles to failure, NF

a) Effect of mean stress

107

104

105

106

107

Cycles to failure, NF

b) Effect of shot peening

Figure 8: S-N curves in Ti-6Al-7Nb

In both cases, the increased strength of the lamellar portion of the duplex microstructure leads to higher resistances to dislocation motion and crack nucleation at tensile mean stresses thus, to improved fatigue performance.

355 -TiAl

4.3

True stress, I [MPa]

Gamma titanium aluminides are highly sensitive to tensile loading with typical elongations to fracture (El) of only 1 to 3%. However, in compressive loading, marked plastic strains can be achieved without premature failure [6]. This is demonstrated in Figure 9.

-0.15

-0.10

1000

500 TENSION

-0.05

0.05

0.10

True strain, A

COMPRESSION -500

-1000

Figure 9: Stress-strain curves in tension and compression of C-TiAl

Obviously, these plastic strains lead to pronounced work-hardening that can not be observed in a tensile test. By using Considère’s construction where the true stress is plotted vs. conventional strain (in compression), the ultimate tensile strength of the material was estimated to UTS = 965 MPa which indicates marked work-hardening if compared to the yield stress of 600 MPa measured in compression (Table 1). The much lower apparent yield stress in tension (440 MPa) might result from early pore nucleation in tension leading to premature failure without macroscopic plasticity. After shot peening, the microhardness at the surface increases from about 320 to values above 700 HV 0.1 (Figure 10) which is much higher than measured on titanium alloys. Residual stresses as determined by the hole drilling method are shown in Figure 11. Although cyclic deformation of this C-TiAl was not studied, from the marked hardening in monotonic loading (UTS – I0.2 = 360 MPa), it is argued that the material will also cyclically harden which prevents cyclic decay of the residual stresses. Accordingly, the fatigue response of this C-TiAl to shot peening is excellent (Figure 12). The HCF strength increases from 550 MPa of the electropolished reference to 675 MPa after shot peening.

356 200 Residual stress, IR [MPa]

Microhardness HV 0.1

800 700 SP (0.40 mmN)

600

500 400

300

EP

200

0

-200

-400 SP (0.40 mmN)

-600

-800 -1000

0

200

400

600

0

Distance from surface, z [mm]

200

400

600

Distance from surface, z [μm]

Figure 10: Microhardness-depth profile after shot peening of C-TiAl

Figure 11: Residual stress-depth profile after shot peening of C-TiAl

Stress amplitude, Ia [MPa]

900 800 700

SP EP

600 500 kt = 1.0 400 103

104

105

106

107

Cycles to failure, NF Figure 12: S-N curves in rotating beam loading (R = -1) of C-TiAl, effect of shot peening

4.4

Aluminum Alloys

The fatigue response of age-hardenable aluminum alloys to shot peening significantly depends on the induced residual stress profile and its cyclic stability. Both properties highly depend on the aging condition utilized, e.g. higher residual stresses and greater cyclic stability are usually measured in underaged (T3, T4) tempers which cyclically harden compared to peak-aged (T6) or overaged conditions which tend to cyclically soften [14-16]. While the resistance to fatigue crack propagation of the electropolished reference in Al 2024 is already somewhat greater in T3 than in T6 (Figure 13a), after shot peening, crack growth in T3 is much more hindered than in T6 owing to higher and more stable residual stresses in T3 as compared to T6 (Figure 13b). Accordingly, the HCF strength in Al 2024 (Figure 14) is much more beneficially affected in the T3 temper (Figure 14a) than in the T6 temper (Figure 14b).

357 10-7 T6

10-8

10-9

da/dN [m/cycle]

da/dN [m/cycle]

10-7

T3

10-10

SP, Ia = 300 MPa

10-8

T6

10-9

10-10

T3

EP, Ia = 250 MPa

10-11

1

2

10-11

3 4 5 6 7 8 910 ,K [MPa m1/2]

a) Condition EP

1

2

3 4 5 6 7 8 910 ,K [MPa m1/2]

b) Condition SP

350

EP

300 T3

250 T6

200

150 104

105 106 Cycles to failure, NF

a) Condition EP

107

Stress amplitude, sa [MPa]

Stress amplitude, sa [MPa]

Figure 13: da/dN-,K curves of small surface cracks in 2024 Al, rotating beam loading (R = –1)

350

SP T3

300 T6

250

200

150 104

105 106 Cycles to failure, NF

107

b) Condition SP

Figure 14: S-N curves in 2024 Al, rotating beam loading (R = –1)

4.5

Magnesium Alloys

Mechanical surface treatments on magnesium alloys are known to modify the cyclic deformation behavior [18, 19]. Results regarding HCF performance of the wrought magnesium alloys AZ80 and AZ31 have shown that this alloy group responds quite critically to shot peening [2022]. For example, marked life improvements were found on the high strength alloy AZ80 after low intensity peening while increasing the Almen intensity led to a pronounced drop in life (Figure 15). Specimens with this optimum in fatigue performance showed subsurface fatigue crack nucleation while at higher intensities multiple surface crack nucleation sites were observed [20]. This sensitivity of the magnesium alloys to shot peening can also be seen in Figure 16. Low intensity peening is clearly superior to high intensity peening regarding HCF strength. From parallel work [23], it is known that heavier peening not only drastically increases roughness and induces microcracks in magnesium alloys, but also leads to lower near surface residual compressive stresses. If this shot peening-induced damage is removed, e.g. by additional polishing, the HCF strength of heavily peened specimens can markedly be improved (Fig. 17).

107

250 0.05 mmN

200

Cycles to failure, NF

Stress amplitude, Ia [MPa]

358

0.40 mmN

150

SP

EP

100

50

0

104

105 106 Cycles to failure, NF

Figure 15: Effect of Almen intensity on fatigue life in AZ80, rotating beam loading

Stress amplitude, Ia [MPa]

106 105 104 Air

103 EP

107

AZ 80

SCCW 14 glass beads S 330 SCCWS 23

0.2 0.4 0.6 0.8 Almen intensity [mmN]

1.0

Figure 16: S-N curves in AZ80, rotating beam loading (R = –1)

0.65 mmN + surface removal

300

200 0.40 mmN

100 Air

0

104

steel shot 105 106 Cycles to failure, NF

107

Figure 17: S-N curves in AZ80, rotating beam loading (R = –1)

Interestingly, a concomitant shift in fatigue crack nucleation site from the surface to the interior could be observed.

5

Summary

Depending on the alloy system of the various light-weight metals, the fatigue response to shot peening can be quite different. Excellent fatigue response was found in materials which exhibit cyclic hardening as observed in = titanium alloys, C-TiAl and naturally aged aluminum alloys. Comparable poor fatigue response was observed in metastable >alloys and peak-aged aluminum alloys which both exhibit cyclic softening. In (=+>) titanium alloys, additional effects can result from the presence or absence of an anomalous mean stress sensitivity. Magnesium alloys were found to react very sensitively to shot peening presumably caused by the limited room temperature deformability of its hexagonal crystal structure.

359

6

Acknowledgements

The authors would like to thank the Deutsche Forschungsgemeinschaft for financial support.

7 [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23]

References O. Vöhringer, Shot Peening (Eds.: H. Wohlfahrt, R. Kopp and O. Vöhringer) DGM, 1987, 185. J. K. Gregory and L. Wagner, Low Cycle Fatigue and Elasto-Plastic Behavior of Materials-3 (Ed.: K. T. Rie), Elsevier Applied Science, 1992, 588. T. Hirsch, O. Vöhringer and E. Macherauch, HTM , 1986, 41, 166. M. K. Tufft, PhD dissertation, University of Dayton, OH, USA, 1997 M. K. Tufft, Shot Peening (Ed.: A. Nakonieczny), 1999, 264. J. Lindemann and L. Wagner, Light Materials for Transportation Systems (Eds.: N. J. Kim, C. S. Lee and D. Eylon) Center for Advanced Aerospace Materials, 2001, 793. A. Drechsler, T. Dörr and L. Wagner, Materials Science and Engineering A, 1998, 146. J. Kiese, J. Zhang, O. Schauerte and L. Wagner, Shot Peening (Ed.: L. Wagner) WileyVCH, 2002, in press A. Berg, J. Kiese and L. Wagner, Materials Science and Engineering A 243, 1998, 146. L. Wagner, A. Berg, T. Dörr and M. Hilpert, Mechanische Oberflächenbehandlung (Eds.: H. Wohlfahrt and P. Krull) Wiley-VCH, 2000, 179. M. C. Berger and J. K. Gregory, Surface Treatment IV (Eds.: C. A. Brebbia and J. M. Kenny) WIT Press, 1999, 341. A. Rössler and J. K. Gregory, Surface Treatment IV (Eds.: C. A. Brebbia and J. M. Kenny) WIT Press, 1999, 311. J. Lindemann and L. Wagner, Materials Science and Engineering A, 1997, 1118. U. Holzwarth, J. Kiese and L. Wagner, Mechanische Eigenschaften von Implantatwerkstoffen, DVM, 1998, 319. J. K. Gregory and L. Wagner, Fatigue 93, EMAS, 1993, 177. J. K. Gregory, C. Müller und L. Wagner, Metall 47, 1993, 915. T. Hirsch, O. Vöhringer and E. Macherauch, Shot Peening (Ed.: H. O. Fuchs) American Shot Peening Society, 1984, 90. D. Deiseroth, W. Zinn, B. Scholtes, Magnesium Alloys and Their Applications (Eds.: B. L. Mordike and K. U. Kainer) MAT INFO, 1998, 409. I. Altenberger, S. Jägg and B. Scholtes, Werkstoffprüfung 98, DVM, 1998, 55. M. Hilpert and L. Wagner, Magnesium Alloys and Their Applications (Ed.: K. U. Kainer), Wiley-VCH, 2000, 525. M. Hilpert and L. Wagner, Magnesium Alloys and Their Applications (Ed.: K. U. Kainer), Wiley-VCH, 2000, 463. J. Wendt, M. Hilpert, J. Kiese and L. Wagner, Magnesium Technology 2001 (Ed.: J. N. Hryn) TMS, 2001, 281. J. Wendt, A. Ketzmer and L. Wagner, Shot Peening (Ed.: L. Wagner) Wiley-VCH, 2002, in press

350

Fatigue Strength Improvement of Welded Aluminium Alloys by Different Post Weld Treatment Methods Thomas Nitschke-Pagel, Helmut Wohlfahrt Institut für Schweißtechnik, Technische Universität Braunschweig, Braunschweig, Germany

1

Introduction

Special fields where the application of aluminium alloys offer many advantages can be found in the whole field of transportation. Due to strongly increased energy costs the weight factor of transport vehicles like cars, trains and aircrafts has become one of the most important factors because the costs for each vehicle depend directly on its weight. The possibility of substitution of steels by aluminium alloys in fatigue loaded constructions requires the easy applicability of manufacturing techniques where joining processes are a very important part. From steels it is well known that the fatigue strength of a welded construction usually will be very low in comparison with that of the base material. A further problem is, that the notches which are an effect accompanying the welding process will be the more effective the higher the ultimate strength of the base material is. This is the reason because the use of modern high strength steels is not helpful with regard on a weight reduction because the higher potential fatigue strength cannot be found in the welded constructions, if the fabrication procedures are the same.

2

Possibilities of fatigue strength improvement of welded joints

Most of the results of investigations with the aim of a remarkable fatigue strength improvement have been carried out on relatively low strength steels and in some cases also in high strength steels [1,7]. As an example fig.1 shows two different improvement strategies can be used. One way is to use welding procedures which will generate a flat weld seam without sharp notches like TIG-welding or manual-arcwelding (MAW) with special electrodes which may produce a flat weld seam due to a low slag viscosity. The second way is the application of different post weld treatment techniques, where mechanical and thermal treatments can be distinguished. Very effective thermal methods will generate a flat weld toe by remelting this zone with a TIG- or plasma arc without additional use of filler material. Another post weld heat treatment is stress relief annealing Figure 1: S-N-Curves of welded T-joints after which shall reduce high tensile residual stresses. different post weld treatments [2]. However the results are strongly varying because the initial tensile residual stresses mostly are much lower than assumed and the influence of the residual stresses is probably much lower than the influence of weld defects and notches.

361 Mechanical surface treatment methods which can be applied after welding are methods like grinding ( the complete weld seam or only the weld toe), hammer- or needle-peening, shot peening and cold rolling. These methods will reduce the notch effects (grinding) or will increase the resistance against crack initiation by cold working induced surface hardening respectively of the crack propagation by the generation of high compressive residual stresses [6]. Newer methods like the Ultrasonic-Impact-Treatment (UIT) [8] also enable fatigue strength improvements, but the changes of the microstructural properties have Figure 2: Overview of mechanical surface treatment methods for still to be examined. As the the fatigue strength improvement and affected surface features. results of a mild steel in fig.1 reveal each improvement method may increase the fatigue strength but it has to be considered, that different features of the surface are affected by the methods summarized in fig.2 in a different way. For example the increase of the surface roughness due to shot peening is a factor, which compensates a part of the fatigue strength improvement achieved by the generation of compressive residuals stresses, an effect which is very important in high strength steels or in high strength aluminium alloys.

3

Comparison of weld seam improvement and shot peening

A very popular welding method is the Metal-Inert-Gas- (MIG-) welding process which is available in different variations. The process is easy to handle and combines a high productivity with good process stability if a modern welding equipment is used. Unfortunately the weld bead cannot be controlled carefully because high welding speed, current and the speed of the wire are parameters which depend on each other. Therefore the geometry of the weld seam is connected with relatively sharp notches and undercuts at the weld toe which lead to a low fatigue strength o in comparison to the base material. The Tungsten-Inert-Gas(TIG-) welding process enables very flat weld seams with a strongly reduced notch effect but the productivity of the process is low due to a Figure 3: Comparison of the weld seam profile and the hardness low welding speed. Investiga- distributions of MIG- and TIG-welded joints of a cold worked tions on steels have shown AlMg4.5Mn G35-alloy (Rm=350 N/mm2) [3].

362 that a combination of TIG-and MIG- (MAG) processes also enables an acceptable productivity in combination with an improved fatigue strength. However the applicability of different improvement techniques which are very effective in steels cannot be simply assigned on cold formed or age hardened aluminium alloys, which are very sensitive against an oversized local heat input, because the welding process in combination with a broadened heat affected zone (HAZ) will generate a weld zone with a hardness significantly below the hardness of the base material. Wide softening zones are typical for TIG-welded aluminium alloys. As fig. 3 shows TIG-process enables the generation of a very flat weld seam with a smooth notch geometry at the weld toe in comparison to a conventional MIG-welded joint but with a broadened softened zone in relation to a MIG-welded plate [3]. After TIG-dressing the fatigue cracks frequently will start in the softened transition zone from the HAZ to the base material. The notch effect of the weld toe is reduced but without a strong benefit for the fatigue strength. On the other hand a shot peening process well adjusted on the base material has the same beneficial effect as the combination of weld seam improvement and shot peening in steels. Fig. 4 summarizes S-N-curves obtained on a cold formed AlMg 4.5 Mn-alloy [3]. The combination of a peening process with high enough intensity to produce high compressive residual stresses and surface hardening and an additional surface finishing by peening with glass pearls in order to reduce the surface roughness results in a fatigue strength which is as high as that of the base material. The use of a steel shot whith a size well adjusted on the notch radius at the weld toe surface hardening and compressive residual stresses are combined with an improved notch Figure 4: Comparison of the S-N-curves of geometry due to the plastic deformations. AlMg4.5Mn G35-alloy welded joints in the Methods like hammer- or needle-peening also as-welded state and after optimised shot induce compressive residual stresses in combipeening [3]. nation with cold hardening of surface layers. Howeever these processes lead to a higher surface roughness. Therefore newer methods like laser-shock treatment and high-pressure-water peening have been developed with the aim to induce compressive residual stresses without changes of the surface topography. Results of investigations on light-weight alloys reveal that significant compressive residual stresses may be generated however not many results of fatigue test are available.

4

Fatigue strength improvement of TIG-welded aluminium joints

In [4] 5 mm TIG-welded plates of an AlMg4.5Mn0.7-alloy (AA5083) were examined under reversed bending after application of different surface treatment methods. The primary subject of this investigation was to maximize the welding speed during one-layer TIG-welding with the aim

363 current [A]

arc length [mm]

frequency [Hz]

wire speed [m/min]

welding speed [m/min]

shielding gas [l/min]

TIG-AC Ar

450

2

100

1.98

0.90

12 (Ar)

TIG-AC He

290

2

60

2.52

0.83

24 (He)

TIG-AC He/Ar

450

1.5

160

3.42

1.52

29 (He/Ar 90/10)

Table 1: Welding parameters for butt welds of 5 mm AlMg4.5Mn0.7-alloy (AA5083) [4]. to combine a good weld seam Shot peening 1) High pressure water peening 2) quality with an acceptable productivity. In table 1 the parameShot material: S230 Jet pressure: 300 bar ters are summarized which Hardness: 45…55 HRC Distance from nozzle: 45 mm were used for TIG-welding Intensity: 0.008” A Peening time: 2 sec with alternating current. Coverage: 200 % Nozzle diameter: 1.5 mm / 20o The macrographs of the Specification: MI 230 R Overlap: 1 mm TIG-AC- welded plates (fig. 5) reveal the main problem with Table 2: Peening parameters; 1) Metal Improvement Company regard to a fatigue strength 2) IWT, Universität Hannover improvement which is connected with one-layer weldments. On the top side of the weldments the weld seam geometry shows the typical features well known from TIG-weldments. A flat weld seam with a smooth macroscopic notch geometry. The joint TIG-AC welded under Helium seems to have the smoothest notch geometry and the specimen welded with a high welding speed of 1.52 m/min under Ar/He shows a significant undercut of the weld seam. The notch factors of the different welds calculated with the measured geometry parameters are between 1.28 and 1.53. However the fatigue strength under reversed bending was uniformly between 55 and 57 N/mm2 and that is to say significantly below the fatigue strength of the base material (118 N/mm2). The TIG-AC-Ar- and the TIGAC-Ar/He-welded joints were additionally examined after different peening procedures (tab. 2). Fig.6 shows the residual stresses measured at the surface by means of X-rays. In the aswelded state slight tensile residual Figure 5: Comparison of the weld seam profile and the hardwelded joints of an AlMg 4.5 stresses with maximum values of ness distributions of TIG-AC2) [4]. =253 N/mm Mn 0.7-alloy (R m 70 N/mm2 are found in the weld

364 seam. Both surface treatment methods induce compressive residual stresses in the weld seam and in the base material which are approximately on the same level, the greatest compressive residual stresses (-180 are N/mm2) observed in the weld seam after shot peening. A significant difference between both methods can be characterized by the half width distributions of the Figure 6: Distributions of the transverse residual stresses and of the half diffraction lines. widths after welding, high pressure water peening and after shot peening. The half widths across the weld after high pressure water peening are almost the same than after welding. On the other side the shot peening process leads to a significant increase of the half widths in the weld seam and in the base material (fig. 6). The depth profiles of the residual stresses and of the half widths (fig. 7) obtained after incremental electrochemical polishing evidently show, that the effect of the water peening process is limited on the generation of compressive residual stresses in the surface without any changes of the local strength properties which can be characterized by the half widths. On the other side the penetration of the plastic deformations connected with the peening Figure 7: Residual stress depth profiles and half width distributions after shot process is much high pressure water peening and after shot peening. higher. The maximum of the compressive residual stresses can be found in a depth between 0.05 and 0.15 mm, the magnitude of the compressive residual stresses is higher and the depth distribution of the half widths indicates a remarkable cold hardening of the surface layers. The S-N-curves of the different treated specimens determined under reversed bending (fracture probability 50%) are given in fig. 8. In comparison to the base material the fatigue strength decreases significantly as expected to values which are in both cases 50% of the fatigue strength of the base material. The different geometry at the weld toes of the top side of the weld seam which is a consequence of the different welding speeds obviously does not result in a different fatigue behaviour. The fatigue strength is nearly as high as the fatigue strength after conventional MIG-welding. This is caused by the dominating notch effect of the weld toes at the bottom side, where the notch geometry is the same as it can be observed in MIG-welded joints. Because the

365 plates have been welded with one layer a certain control of the weld seam profiles on both sides of the welded plates which is usually possible during TIG-welding due to the separate input of the heat and of the filler material could not be achieved. As fig. 8 reveals both peening procedures which were applied in this investigation lead to an improvement of the fatigue strength. After high pressure water peening at the weld toes the fatigue strength is increased from 57 N/mm2 to 75 N/mm2. This is equivalent to a percentage of more than 30%, however the fatigue strength value is significantly below the base material After shot peening the fatigue strength is increased from 57 N/mm2 to 121 N/mm2 and that is to say to the level of the base material. The significant improvement of the fatigue strength due to shot peening is limited on stress amplitudes, which lead to high number of load cycles. With increasing stress amplitude the benefit disappears. This is in agreement with well known results from literature on steels and aluminium alloys. As summarized in fig. 2 the mechanical surface treatment methods affect three Figure 8 S-N-curves (50% fracture probability) parameters which are important for the of TIG-AC-Ar- and TIG-AC-He/Ar-welded fatigue behaviour, the residual stresses, AlMg 4.5 Mn 0.7-joints in the as-welded state hardness of surface layers and surface and after a mechanical surface treatment [4]. roughness. Due to shot peening beneficial compressive residual stresses are induced and the surface hardness increases. On the other side the surface roughness increases as well which may have a negative effect for the fatigue strength. The magnitude and the distribution of the induced compressive residual stresses will be the more important the higher the ultimate strength of the material is. This is caused by the effect, that in low strength materials the residual stresses will be reduced strongly during the first cycles because the fatigue strength is very close to the yield strength respectively to the cyclic yield strength. With increasing ultimate strength the cyclic yield strength increases stronger than the fatigue strength with consequence of more stable residual stresses. On the other hand the surface hardening effect becomes less important but the surface roughness has to be considered. Therefore in weldments, which are mostly built by relatively low strength metals, the more important factor for the fatigue strength improvement due to mechanical surface treatments is the cold working induced surface hardening. The compressive residual stresses are also effective but they will be reduced strongly at higher load amplitudes. Thus the fatigue strength improvement due to high pressure water peening is much lower than due to shot peening because the surface hardness is not changed significantly. This results frequently in a flat slope of the S-N-curve with the consequence that the fatigue limits at high load amplitudes are as

366 high as in untreated joints. In aluminium alloys the surface roughness is an additional very important parameter because the effectiveness of small notches – and the higher surface roughness can be interpreted as a notch factor – can be much higher than in low strength steels.

5

Conclusions

Different methods for the fatigue strength improvements are available which can be applied on welded joints of steels and light weight alloys. However the recommended way first to improve the weld seam profile and then to apply a mechanical surface treatment only in steels will result in a fatigue strength which is equivalent to the base material. In high strength aluminium alloys the improvement of the weld seam profile will only result in a significantly better fatigue strength, if the weld seam improvement is not connected with a strong softening in the weld seam and in the HAZ due to high local heat input. In welded joints with one-layer welds or in constructions which cannot be welded from both sides the improvement of the weld seam geometry for example with help of TIG-dressing will not result necessarily in an improved fatigue strength. A mechanical surface treatment as demonstrated with shot peening can result in a fatigue strength improvement, where the fatigue strength of the weldments may reach the value of thebase material. This can achieved without an especially optimised welding process. However the avoidance of a broad softening zone is required because results of experiments after a combined treatment (e.g. TIG-dressing and shot peening [4]) have shown, that the softening effects cannot be compensated by the mechanical surface treatments. Several other improvement methods are available, but their principles have to be clarified respectively it has to be examined, in which cases these methods my be applicable under consideration of economical boundary conditions. However the success of each fatigue strength improvement method depends on the possibilities to affect all points of a weldment, where the probability of a crack initiation high. This is also valid for methods like shot peening. If places with sharp notches as the root of fillet welds or the bottom side of one-layer butt welds cannot be affected, an application of a post weld treatment method will result only in a shift of the crack initiation sites without a significant improvement of the fatigue strength.

6 [1] [2]

[3] [4] [5] [6] [7]

References Heeschen, J.: Dissertation Universität Kassel (1986). Haagensen, F.; Dragen, A.; Slind, T.; Orjasaeter, 0.: In: ,,Steel in Marine Structures“, Proc. 3nd Int. ECSC Offshore Conf. (SIMS ‘87), Delft, The Netherlands, June 15-18, 1987, S. 689 - 698, Elsevier Science Publishers B.V., Ed. C. Nordhoek and J. de Back Zinn, W.: Dissertation Universität Kassel (1989) Krull, P.: Dissertation TU Braunschweig (2000). Haagensen, P. J.: IIW-Doc.XIII-WG2-30. August 1994 Wohlfahrt, H., Krull, P. (Hrsg.): Mechanische Oberflächenbehandlungen: Grundlagen – Bauteileigenschaften – Anwendungen. WILEY-VCH-Verlag Weinheim 2000. Haagensen, P. J.: IIW-Doc.XIII-1748-98.

357

Influence of Mechanical Surface Treatments on Notched Fatigue Strength of Magnesium Alloys Bodo Küster1), Matthias Hilpert2), Armin Kiefer3) and Lothar Wagner1) 1) Chair of Physical Metallurgy and Materials Technology, Technical University of Brandenburg at Cottbus, Cottbus, Germany 2) Otto Fuchs Metallwerke, Meinerzhagen, Germany 3) OSK Kiefer, Oppurg, Germany

1

Introduction

High-strength wrought magnesium alloys are considered as potential candidates for application as suspension parts in future automobiles due to their high strength to weight ratio [1-3]. For this application, good HCF performance of notched components is one of the most demanding requirements. While previous work on magnesium alloys has demonstrated to what extent mechanical surface treatments such as shot peening and roller-burnishing can increase the HCF strength of smooth specimens compared to an electropolished reference [4-7], corresponding results on notched specimens are not available yet. From earlier work on titanium and aluminum alloys [8-10], it is known that particularly notched components beneficially react to mechanical surface treatments such as shot peening or deep rolling. This result was explained by the interaction of the notch root stress field with the process-induced residual compressive stresses affecting mainly microcrack growth. In addition, the triaxial stress state at the notch root was thought to result in residual stresses being more stable during cyclic loading compared to those in the surface layer of smooth specimens.

2

Experimental

The wrought magnesium alloy AZ80 (nominal composition in weight percent: 8Al, 0.5Zn, 0.2Mn, balance: Mg) was received as an extrusion from Otto Fuchs Metallwerke, Meinerzhagen, Germany. The rectangular bar had a cross section of 110 × 70 mm (extrusion ratio: 9). The alloy was tested in the as-fabricated condition. Specimens were machined with the load axis parallel to the extrusion direction (L). Tensile tests were performed on threaded cylindrical specimens having gage lengths of 20 mm. The initial strain rate was 8.3 × 10–4s–1. Tensile test results are listed in Table 1. Table 1: Tensile test results on AZ80 (L-direction) E (GPa)

σ0.2 (MPa)

UTS (MPa)

El (%)

RA (%)

44

245

340

12

14

For fatigue testing, circumferentially notched cylindical specimens having a 60° V-notch were machined. After machining, about 50 μm were removed from the notch root surface by

368 electropolishing (EP) to ensure that any machining effect that could mask the results was absent. This electropolished condition was taken to serve as reference. Specimens were tested in rotating beam loading (R = –1) at frequencies of about 60 Hz. Shot peening (SP) was performed with an injector type machine using spherically conditioned cut wire SCCW 14 (0.36 mm average shot size). During the peening treatment, the specimens rotated at 1s-1. The distance between nozzle tip and specimen surface was 70 mm. To determine optimum process parameters with regard to HCF life, the Almen intensity was varied in a wide range. Deep rolling (DR) was performed by a hydraulically driven three-roll device by means of a conventional lathe. The diameter of the hardmetal rolls was 30 mm. The rolls had a 55° cross section and a tip radius of 0.3 mm. During deep rolling, the number of revolutions was kept constant at 45. To determine optimum process parameters in deep rolling, the rolling force was varied in a wide range. In order to keep the geometrical notch factor (i.e., the notch root radius) constant for the various surface treated conditions, the specimens to serve as reference as well as those to be shot peened were machined with a notch root radius of 0.3 mm while specimens to be deep rolled had a starting notch root radius of 0.43 mm. During deep rolling, this radius was also reduced to 0.3 mm. As a result, all specimens had a notch factor of about 2.7. The exact value for each individual specimen was calculated by measuring the net diameter (dn), notch depth (t) and notch root radius (ρ) using the formula (1) for bending [11]:

kt = 1 +

1 é d ê 1+ n 0,12 2 ρ + 4ê 0,45 ê d dn æt ö ê n ç ÷ ρ ρ 2 2 ê ρ ë è ø

ù ú ú ú ú úû

2,66

+ 0,1

dn 2ρ

(1) 1,2

æ dn t öæ t ö + ÷ç ÷ ç è 2 ρ ρ øè ρ ø

After shot peening and deep rolling, the change in surface layer properties was evaluated by roughness measurements taken circumferentially at the notch root and by measurements of the depth profiles of microhardness.

3

Results and Discussion

The S-N curves of the electropolished reference conditions of AZ80 are plotted in Fig.1 comparing the performance of smooth (kt = 1.0) specimens [12] and notched (kt = 2.7) specimens. As expected, introducing a notch clearly deteriorates the fatigue performance in terms of nominal stresses (Fig.1a). To what extent this notch factor of 2.7 reduces the HCF strength can be seen in Fig. 1b where the data of Fig. 1a are replotted in terms of maximum (axial) notch root stresses (σa ·kt). Since notched and smooth specimens have the same run-out stress amplitude of 100 MPa, it can be derived that AZ80 is 100 % notch sensitive in fatigue, a result also found on titanium alloys [13]. Obviously, fatigue crack nucleation determines the HCF strength. Within the finite life regime, notched specimens have an apparent higher life since microcrack growth now determines life. The marked stress gradient below the notch root reduces the driving force for crack growth compared to the smooth specimens thus, increasing fatigue life (Fig 1b).

369

400

EP

Stress amplitude, Ia . kt [MPa]

Stress amplitude, Ia [MPa]

250 200

kt = 1.0

150 100

kt = 2.7

50

EP

350 300 250

kt = 2.7

200 150 100

kt = 1.0

50 0

0

104

105

106

104

107

105

106

107

Cycles to failure, NF

Cycles to failure, NF

a) nominal stress

b) maximum notch root stress

Figure 1: S-N curves of AZ80 (EP) in rotating beam loading (R = -1), comparison of smooth (kt = 1.0) and notched (k t = 2.7) specimens

Typical changes in surface layer properties due to shot peening are shown in Figs. 2 and 3. Compared to the electropolished reference, the shot peening-induced roughness linearly increases with Almen intensity (Fig. 2). An example of a microhardness-depth profile is plotted for a given Almen intensity of 0.48 mmN in Fig. 3. The marked increase of near-surface microhardness from about 82 to 130 HV 0.04 is caused by pronounced work-hardening of the material also seen in tensile testing (Table 1). 160

25

kt = 2.7

140

SCCW 14

120 Rz

15

HV 0.04

Roughness [μm]

20

Ry

10

100 80 60

SP

40

5

Ra

20

kt = 2.7

0

0

EP

0,1

0,2

0,3

0,4

0,5

Almen intensity [mmN]

Figure 2: Roughness values vs. Almen intensity

0,6

SCCW 14 0.48 mmN

0

200

400

600

800

Distance from surface [μm]

Figure 3: Microhardness-depth profile after shot peening (0.48 mmN)

The effect of Almen intensity on fatigue life at constant stress amplitudes in rotating beam loading is plotted in Fig. 4. At the high stress amplitude of σa · kt = 300 MPa, the fatigue life starting with the electropolished condition continuously increases with Almen intensity within the range of intensities utilized (Fig. 4a). At the low stress amplitude of σa · kt = 225 MPa, the fatigue life similarly increases. However, run-outs (107 cycles) were found already at intermediate Almen intensities of 0.38 mmN. The change in surface layer properties due to deep rolling is illustrated in Figs. 5 and 6. Compared to shot peening, the increase in surface roughness caused by deep rolling is much less pronounced (compare Fig. 5 with Fig. 2). One should keep in mind, that deep rolling might even decrease surface roughness if comparison is done with an as-turned as opposed to an elec-

370

Ia.kt = 300 MPa kt = 2.7

SCCW 14 107

Cycles to failure, NF

Cycles to failure, NF

106

105

SCCW 14

104 EP

0,2

0,4

106

105

Ia.kt = 225 MPa kt = 2.7

104 EP

0,6

0,2

0,4

0,6

Almen intensity [mmN]

Almen intensity [mmN]

a) Ia kt = 300 MPa

b) Ia kt = 225 MPa

Figure 4: Fatigue life vs. Almen intensity

14

kt = 2.7

12

DR

160 140

Ry

10

120

HV 0.04

Roughness [μm]

tropolished surface. The penetration depth of plastic deformation caused by deep rolling is typically much greater as compared to that after shot peening (Fig. 6, compare Fig. 6 with Fig. 3).

8 6

Rz

4

80 60 40

2 EP

50

100

150

DR

20

Ra

0 200

F = 100 N

kt = 2.7

0

0

Rolling force [N]

200

400

600

800

Distance from surface [μm]

Figure 5: Roughness values vs. rolling force

Figure 6: Microhardness-depth profile after deep rolling (F = 100 N)

Ia.kt = 350 MPa kt = 2.7

107 Cycles to failure, NF

100

106

105

104 EP

100

200

300

Rolling force [N]

Figure 7: Fatigue life (σa · kt = 350 MPa) vs. rolling force

Interestingly, best fatigue performance was found after deep rolling with low or intermediate rolling forces whereas higher forces led to marked losses in fatigue life (Fig. 7). This deteriora-

371 tion of the fatigue performance after rolling with higher forces is presumably caused by rollinginduced defects such as overlaps and microcracks. Typical metallographic cross sections of the notch root regions are shown in Fig. 8 comparing the various surface treated conditions electrolytically polished (Fig. 8a), shot peened, 0.48 mmN (Fig. 8b), deep rolled, F = 100 N (Fig. 8c) and deep rolled, F = 300 N (Fig. 8 d). Clearly, the high rolling force of F = 300 N leads to material separation in the notch root (Fig. 8d) thus, explaining the observed marked overrolling effect (Fig. 7). In contrast, no process-induced overlaps were found after shot peening (Fig. 8b) or deep rolling using low rolling forces (Fig. 8c) although some roughening is clearly present if compared with the electropolished reference (compare Figs. 8b and 8d with Fig. 8a).

a) Electropolished

b) Shot peened (0.48 mmN)

c) Deep rolled (F = 100 N)

d) Deep rolled (F = 300 N)

Figure 8: Notch root after various surface treatments

From Figs. 4 and 7, the process parameters resulting in the most marked fatigue life improvement after shot peening (0.48 mmN) and deep rolling (F = 100 N) were used to determine the S-N curves in Fig. 9. The 107 cycles notch fatigue strength of the electropolished reference increases in terms of σa · kt from 100 MPa (EP) to 270 MPa (SP) and 350 MPa (DR). Thus, optimum shot peening fully counterbalances the geometrical notch factor of kt = 2.7, while deep rolling even significantly overcompensates this stress concentration (Fig. 9). After sectioning the deep rolled run-out specimen (σa · kt = 350 MPa) and metallographic investigating the notch root region, so called non-propagating microcracks, i.e., cracks which had nucleated at the notch root and had slowly propagated into the interior, were clearly visible (Fig. 10).

372

Stress amplitude, Ia . kt [MPa]

600

kt = 2.7 500

DR (F = 100 N)

400 300

SP (0.48 mmN)

200

EP

100 0 104

105

106

107

Cycles to failure, NF

Figure 9: S-N curves of AZ80 in rotating beam loading (R = –1), comparison of optimum deep rolled (DR) with optimum shot peened (SP) and reference (EP) conditions

Figure 10: Notch root of run-out (σa · kt = 350 MPa) deep rolled specimen

These cracks may have become arrested when the amplitude of the local stress intensity factor at the crack tip ΔK, which is a function of the applied stresses, process-induced residual stresses and crack length reaches the threshold value ΔKth for microcrack growth [10].

4

Summary

From fatigue tests of the electropolished reference, it is concluded that the HCF strength of notched (kt = 2.7) specimens of the high-strength wrought magnesium alloy AZ80 is fully affected by the geometrical notch factor, i.e., the material is 100 % notch sensitive. However, this notch sensitivity in fatigue can be easily overcome by suitable mechanical surface treatments applied before the component is put into service. For example, shot peening is able to fully counterbalance this geometrical notch factor if optimum process parameters are utilized. While deep rolling AZ80 results in the most marked improvement of the notch fatigue strength, it should be taken into account that the component to be treated by deep rolling needs some rotational symmetry. On the other hand, shot peening is much more versatile since it can be applied to almost any component’s geometry.

5

Acknowledgements

The authors gratefully acknowledge the support of this work by the Bundesministerium für Wirtschaft (BMWi) under contract 138/99.

6 [1]

References T. K. Aune and H. Westengen, Magnesium Alloys and their Applications (Eds.: B. L. Mordike and F. Hehmann) DGM, 1992, 221.

373 [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13]

H. Friedrich and S. Schumann, Magnesium 2000 (Eds.: E. Aghion and D. Eliezer) MRI, 2000, 9. G. L. Song and A. Atrens, Advanced Engineering Materials, Wiley-VCH (1999), 11. M. Hilpert and L. Wagner, Magnesium Alloys and their Applications (Eds.: B. L. Mordike and K. U. Kainer) MATINFO, 1998, 271. L. Wagner, Materials Science and Engineering A 263, 1999, 210. J. Wendt, M. Hilpert, J. Kiese and L. Wagner, Magnesium Technology 2001 (Ed.: J. N. Hryn) TMS, 2001, 281. L. Wagner, M. Hilpert and J. Wendt, Light Materials for Transportation Systems (Eds.: N. J. Kim, C. S. Lee and D. Eylon) CAAM, 2001, 205. J. K. Gregory, L. Wagner and C. Müller, Beta Titanium Alloys (Eds.: A. Vassel, D. Eylon and Y. Combres) Editions de la Revue Métallurgie, 1994, 229. A. Drechsler, T. Dörr and L. Wagner, Materials Science and Engineering A (1997), 217. L. Wagner, C. Müller and J. K. Gregory, Fatigue 93 (Eds.: J.-P. Bailon and J. I. Dickson) EMAS, 1993, 471. W. Beitz and K.-H. Grote, Handbook for Mechanical Engineering, Springer, 2001, E 103 (in German). L. Wagner and M. Hilpert, Shot Peening and Blast Cleaning (Ed.: M. C. Sharma) MACT, 2001, 49. C. Gerdes and G. Lütjering, Shot Peening (Ed.: H. O. Fuchs) American Shot Peening Society, Paramus, 1984, 175.

364

Shot Peening of Cast Magnesium Alloys Tomasz Ludian1), Matthias Hilpert2), Armin Kiefer3) and Lothar Wagner1) 1) Chair of Physical Metallurgy and Materials Technology, Technical University of Brandenburg at Cottbus, Cottbus, Germany 2) Otto Fuchs Metallwerke, Meinerzhagen, Germany 3) OSK Kiefer, Oppurg, Germany

1

Abstract

The fatigue behavior of the high-pressure die cast magnesium alloys AM50 and AZ91 was investigated. In both alloys, cast defects led to marked scatter in fatigue performance. To study potential improvements in fatigue behavior, shot peening was performed using various Almen intensities. Both alloys showed marked improvements in fatigue strength compared to an electropolished reference.

2

Introduction

The weight reduction of automobiles is one of the most effective ways for improving fuel consumption since the resistances of a vehicle to rolling, climbing and acceleration are directly dependent on vehicle mass. Therefore, the application of magnesium alloys which density is only roughly 25 % that of steel and 66 % that of aluminum is expected to substantially increase in this decade. While for a limited number of vehicle components, cast magnesium alloys are already in production such as materials for transmission gearbox housings, seat frames and steering wheels, both cast and wrought magnesium alloys are potential candidates for many further applications, e.g., as materials for body and suspension components where they can largely substitute steels and even aluminum alloys [1, 2]. For these automotive applications, the fatigue performance of both cast and wrought magnesium alloys is of particular importance [3, 4]. The aim of this investigation was to outline potential improvements of the fatigue performance of the high-pressure die cast magnesium alloys AM50 and AZ91 through shot peening.

3

Experimental

The cast magnesium alloys AM50 (5Al, 0.5Mn, balance: Mg) and AZ91 (9Al, 1Zn, 0.2Mn, balance: Mg) were received from Audi AG, Ingolstadt, Germany as high-pressure test bar die castings (Fig. 1). From these castings, specimens were taken from the round bars with a diameter of 10 mm. Crystallographic textures were determined by X-ray diffraction and are shown as (0002) pole figures. Tensile tests were performed on cylindrical specimens having gage lengths and gage diameters of 20 and 4 mm, respectively. The initial strain rate was 8.3 × 10–4s–1. For fatigue testing, hourglass shaped specimens (5mm gage diameter) were machined.

375

Figure 1: High-pressure die casting

After, machining, about 200 ìm were removed from the surface of the specimens by electrolytical polishing (EP) to ensure that any machining effect that could mask the results was absent. Shot peening (SP) was performed with an injector type machine using spherically conditioned cut wire SCCW14 (0.36 mm average shot size). After shot peening, the change in surface layer properties was determined by profilometry and microhardness-depth profiles. Fatigue tests were performed in rotating beam loading (R = -1) at frequencies of about 60 Hz in ambient air.

4

Results and Discussion

The microstructures of the high-pressure die cast magnesium alloys are shown in Figure 2. Both alloys AM50 (Fig. 2a) and AZ91 (Fig. 2b) are characterized by massive Mg17Al12 compound at the boundaries of small, cored grains. Presumably, the absence of precipitated discontinuous Mg17Al12 is the result of the rapid cooling during the high-pressure die casting process. Typical pole figures of the cast alloys are illustrated in Figure 3. As expected, the basal planes are randomly oriented in AM50 (Fig. 3a) and AZ91 (Fig. 3b). As opposed to extruded alloys [5, 6] no directionality in properties is likely in cast alloys due to this random basal plane distribution. Tensile properties of the alloys are illustrated in Table 1.

376

a) AM50

b) AZ91

Figure 2: Microstructure of the cast alloys

a) AM50

b) AZ91

Figure 3: (0002) pole figure of the cast alloys

Table 1: Tensile properties of the high-pressure die cast magnesium alloys Material

I0.2 (MPa)

UTS (MPa)

El (%)

RA (%)

AM50

95

165

3

7.0

AZ91

125

170

2

3.5

Examples of typical fracture surfaces of the cast magnesium alloys are illustrated in Figure 4 and Figure 5. At high magnification, some degree of porosity was observed in both alloys AM50 (Fig. 4b) and AZ91 (Fig. 5b). The amount of porosity markedly depended on the location from where the specimens were taken within the test bar [7] (Fig. 6). Close to the feeder, the amount of porosity was clearly higher than in regions far away from the feeder. Since two specimens were taken from each test bar, part of the specimens for mechanical tests had high amounts of porosity while others were fairly free of that (Fig. 6). Obviously, these cast defects are potential sites for fatigue crack nucleation during cyclic loading.

377

a)

overview

b) high magnification

Figure 4: Tensile fracture surfaces of the cast alloy AM50

a) overview

b) high magnification

Figure 5: Tensile fracture surfaces of the cast alloy AZ91

a) AM50

b) AZ91 Figure 6: Longitudinal cross sections of the Ø 10 mm test bars of the high-pressure die castings (compare with Fig. 1)

378 Typical changes in surface layer properties due to shot peening are shown in Figure 7. Shot peening (SP) drastically increases roughness. An example for the dependence of shot peeninginduced roughness in AZ91 on Almen intensity is shown in Figure 7a.

a) Roughness vs. Almen intensity

b) Microhardness-depth profile (SP, 0.38 mmN)

Figure 7: Surface layer properties after shot peening

Shot peening leads to marked increases in near-surface microhardness, particularly for AM50 (Fig. 7b), owing to pronounced work hardening (Table 1). The effect of Almen intensity on fatigue life in rotating beam loading (R = –1) of both cast alloys is shown in Figure 8. For both alloys AM50 (Fig. 8a) and AZ91 (Fig. 8b), the data points can be assigned to two groups, i.e., specimens with low porosity and those with high porosity. On average, there is an order of magnitude difference in lifetime between these two groups (Fig. 8). For both groups, fatigue life steadily increases with an increase in Almen intensity.

a) AM50

b) AZ91

Figure 8: Fatigue life (rotating beam loading, R = –1) vs. Almen intensity

The presented results can be summarized as follows: The change in fatigue performance of the high-pressure die cast magnesium alloys AM50 and AZ91 due to mechanical surface treatments depends on the process-induced surface topography, microhardness and residual stress profiles in near-surface regions. The process-induced residual compressive stresses can overcompensate the detrimental effect of surface roughness [8] since the fatigue life of shot peened specimens is higher than that of the electropolished reference (Fig. 9).

379

a) AM50

b) AZ91

Figure 9: S-N curves in rotating beam loading (R = –1)

Obviously, fatigue life extension by retardation of microcrack growth owing to the residual compressive stress field is greater than the reduction in fatigue life caused by earlier crack nucleation as a consequence of shot peening-induced higher surface roughness.

5

Acknowledgements

The authors would like to thank Audi AG, Ingolstadt, Germany for providing the magnesium castings and Dr. J. Lindemann of BTU Cottbus for texture measurements. Financial support from BMWi through contract 138/99 is gratefully acknowledged.

6 [1] [2] [3] [4] [5] [6] [7] [8]

References T.K. Aune and H. Westenhagen, Magnesium Alloys and their Applications, (Eds.: B. L. Mordike and F. Hehmann), DGM, 1992, 221. H. Friedrich and S. Schumann, Magnesium 2000, (Eds.: E. Aghion and D. Eliezer), MRI, 2000, 9. R. I. Stephens, C. D. Schrader and K. B. Lelase, J. Eng. Mater. Technol, 1995, 293. V. V. Ogarevic and R. I. Stephens, Ann. Rev., Mater. Sci., 1990, 141. M. Hilpert and L. Wagner, Journal of Materials Engineering and Performance, Vol.9, No. 4, 2000, 402. M. Hilpert, Dr.-Ing. thesis, Techn. Univ. of Brandenburg at Cottbus, Germany, 2001 A. Wendt, Dipl.-Ing. thesis, Techn. Univ. of Brandenburg at Cottbus, Germany, 1998 T. Dörr, M. Hilpert, P. Beckmerhagen, A. Kiefer and L. Wagner, Shot Peening, Present & Future, (Ed.: A. Nakonieczny), IMP, 1999, 153.

370

Shot Peening to Enhance Fatigue Strength of TIMETAL LCB for Application as Suspension Springs Jürgen Kiese1), Jiulai Zhang2), Oliver Schauerte1) and Lothar Wagner2) 1)

Volkswagen AG, Wolfsburg, Germany Chair of Physical Metallurgy and Materials Technology, Technical University of Brandenburg at Cottbus, Cottbus, Germany

2)

1

Abstract

The titanium alloy Low Cost Beta (TIMETAL LCB) was specifically developed for automotive applications. The combination of high yield stress, low Young’s modulus, low density and reasonable material costs makes this alloy most suitable for substituting spring steels for suspension springs. Goal of the present investigation was twofold. Firstly, the material’s microstructure was optimized by suitable heat treatments for good yield stress-ductility combinations and best high cycle fatigue (HCF) performance. Secondly, shot peening was applied using various Almen intensities to further improve the fatigue behavior. Fatigue performance of optimum shot peened conditions will be compared with an electropolished reference.

2

Introduction

Besides improvements in engine efficiency, the reduction of vehicle weight by using lightweight structural alloys such as those based on magnesium, aluminum and titanium for body as well as suspension parts is one of the most promising ways to reduce fuel consumption. While both aluminum and magnesium alloys have been introduced already decades ago into industrial scale manufacture of automobiles, there was for a long time no use of titanium alloys in large volume automobile production due to cost arguments. Recently, a low cost beta titanium alloy (LCB) having the composition Ti-6.8Mo-4.5Fe-1.5Al was developed by TIMET specifically for automobile applications [1, 2]. As opposed to other metastable beta titanium alloys, costly alloying elements such as Cr, Nb and V are omitted. In addition, the formulation cost of this alloy is lowered by adding the Mo in the form of a ferro-molybdenum master alloy. For application as suspension coil spring material and by substitution for spring steels, weight savings in excess of 50 % were anticipated [3]. This investigation was performed to optimize for a given volume fraction of primary á phase the final age-hardening treatment for LCB with regard to tensile properties and HCF performance. As is the case with conventional steel springs, shot peening was then applied to further improve fatigue behavior. The shot peening process was widely varied to establish optimum treatments with regard to HCF performance.

381

3

Experimental

The LCB material was received from TIMET, Henderson, NV (USA) as 10 mm bar stock. After working above the beta transus temperature followed by working through the transus, the material had been given a solution anneal shortly below the transus at 760°C. Blanks 50 mm in length were cut from the bar stock and were given various aging treatments for 4h at temperatures ranging from 460 to 540°C. Tensile tests were performed on threaded cylindrical specimens having gage lengths and gage diameters of 20 mm and 4 mm, respectively. The initial strain rate was 8.3 × 10–4s–1. To evaluate cyclic deformation behavior, LCF tests were conducted on threaded cylindrical specimens having gage lengths and gage diameters of 20 and 4 mm, respectively. Axial fatigue tests were performed in fully reversed loading (R = –1) using a servohydraulic tester in stress control at a frequency of 0.1s–1. The axial strain was measured by strain gages. From the hysteresis loops, the plastic strain was measured and plotted vs. number of cycles. HCF tests were done on hour-glass shaped specimens in rotating beam loading (R = –1) at about 60 Hz. An electrolytically polished surface condition was taken as reference. After machining, around 100 mm was removed from the surface by electropolishing to ensure that any machining effect that could mask the results was absent. Shot peening was performed by means of a direct pressure blast system using cast steel shot S 330 (0.8 mm average shot size) and a wide range of Almen intensities. The distance between the tip of the nozzle and the work piece surface was 50 mm. During the shot peening process, the specimens rotated at 1s–1. After shot peening, the change in surface layer properties was determined by roughness measurements through profilometry, microhardness profiles and measurements of macroscopic residual stresses through the incremental hole drilling method. The diameter of the drill was 1.7 mm. The oscillating drill was driven by an air turbine with a rotational speed of about 200.000 rpm. The shot peening-induced strains in the surface layer were measured with strain gage rosettes at drilled depths of about every 20 mm. The measured strains were converted into stresses by using an average Young’s modulus of 115 GPa. The fracture surfaces of the fatigue specimens were studied by SEM.

4

Results and Discussion

The microstructure of the as-received bar stock consists of b grains with about 15 % primary a phase. After aging, the resulting SEM microstructures are illustrated in Figure 1 comparing the conditions aged at 460°C (Fig. 1a) and 540°C (Fig. 1b). As seen from Figure 1, the primary a grains are located at the triple points of the b grains. Aging at 540°C (Fig. 1b) leads to a change in the appearance of the b phase owing to coarser secondary a precipitates (compare Fig. 1b with Fig. 1a). Tensile properties comparing the solution heat treated (as-received condition) with various age-hardened conditions are listed in Table 1.

382

a) 4h 460oC

b) 4h 540 oC

Figure 1: Microstructures of TIMETAL LCB

Table 1: Tensile properties of LCB Heat treatment as-received

E [GPa]

I0.2 [MPa]

UTS [MPa]

El [%]

91

1105

1125

18.7

4h 460°C

112

1605

1685

3.5

4h 500°C

114

1445

1540

7.6

4h 540°C

117

1285

1355

11.9

As expected, the Young’s modulus is lowest in the solutionized condition and increases with age-hardening. This is due to the contribution of fine secondary a being precipitated within the b matrix which also leads to increases in yield stress und tensile strength. With an increase in aging temperature from 460 to 540°C, these stress values drop while a concomitant increase in elongation to fracture is found (Table 1). The cyclic deformation behavior of the various age-hardening conditions is illustrated in Figure 2. For both stress amplitudes of 0.9 s0.2 (Fig. 2a) and 0.8 s0.2 (Fig. 2b), the plastic strain increases with number of cycles, i.e. cyclic softening occurs. This indicates that the microstructure of LCB is cyclically not stable. Such cyclic softening behavior was also observed in other 10.0 4h 460ºC

4h 460ºC 4h 500ºC 4h 540ºC

,A pl /2 [o/oo]

,A pl /2 [o/oo]

10.0

1.0

4h 500ºC 4h 540ºC

1.0

R = -1

R = -1

0.1

0.1

1

10

100

1

Cycles, N

a) Ia = 0.9 I0.2

10

Cycles, N

b) Ia = 0.8 I0.2

Figure 2: Cyclic deformation (R = –1) behavior of TIMETAL LCB

100

383 metastable b titanium alloys, namely Ti-3Al-8V-6Cr-4Mo-4Zr (Beta C) and Ti-10V-2Fe-3Al [4]. The S-N curves of the as-received and various age-hardened conditions are plotted in Figure 3. The lowest HCF strength of about 600 MPa was found for the as-received condition. Although the yield stress or tensile strength values of the age-hardened conditions differ by as much as 330 MPa (Table 1), no significant differences were found in HCF performance. Similar results were reported in work on Beta C, where an upper limit of fatigue strengths (R = -1) was found that apparently could not be exceeded by further increases in yield stress. Presumably, microstructural instabilities during cyclic deformation lead to this limit in fatigue strength of metastable b titanium alloys [5].

Stress amplitude, Ia [MPa]

1200 4h 460°C

1100

4h 500°C

1000

4h 540°C as-received

900 800 700 600

EP, R = -1

500 103

104

105

106

107

Cycles to failure, NF Figure 3: S-N curves (R = -1) of TIMETAL LCB (rotating beam loading in air), electropolished condition (EP)

Shot peening-induced changes of the surface layer properties in LCB are shown in Figure 4. A typical depth profile of microhardness after shot peening is given in Figure 4a. There is only a slight increase in near-surface microhardness, presumably owing to the low work-hardening capacity in LCB (Table 1). The residual macrostresses for the various aging conditions after peening to 0.55 mmA are illustrated in Figure 4b. For the various conditions, there are marked 0

SP (0.55mmA)

520

500

480 460

440

420

400

380

4h 540 oC

360

Residual stress, IR [MPa]

Microhardness, [HV0.1]

540

4h 460 oC 4h 500 oC 4h 540 oC

-200 -400

-600

-800

-1.000

SP (0.55mmA) -1.200

0

200

400

600

800

1.000

0

a) Microhardness-depth profile

100

200

300

Distance from surface, z [μm]

Distance from surface, z [μm]

b) Residual stress profile

Figure 4: Surface layer properties in TIMETAL LCB after shot peening

400

384 maxima in residual compressive stresses at depths of roughly 100 to 150 μm below the surface (Fig. 4b). The effect of Almen intensity on fatigue life at a stress amplitude of sa = 900 MPa is illustrated in Figure 5 indicating a saturation in life improvement at intermediate Almen intensities. From these data, an Almen intensity of 0.55 mmA was utilized for further testing.

Cycles to failure, NF

106

105

EP 104 0

Ia = 900 MPa R = -1

4h 500 oC 0.2

0.4

0.6

0.8

1

Almen intensity [mmA] Figure 5: Fatigue life (Ia = 900 MPa) vs. Almen intensity

S-N curves are shown in Figure 6 comparing optimum shot peened with electrolytically polished conditions for the various aging treatments. As opposed to the electropolished condition (Fig. 3), there is a ranking in fatigue performance after shot peening (Fig. 6) indicating superior HCF life for aging treatments at 500 and 540°C whereas in the finite life regime, aging at 460°C

Stress amplitude, Ia [MPa]

1100 4h 460°C, SP 4h 500°C, SP

1000

4h 540°C, SP 900 EP

800 700 R = -1 600 103

104

105

106

107

Cycles to failure, NF Figure 6: S-N curves (R = –1) of TIMETAL LCB (rotating beam loading in air) comparison of optimum shot peened (SP) with electropolished (EP) conditions

385 seems to be superior. Since suspension springs in automobiles will see fatigue loading in excess of 106 cycles, the overall best performance is given by the condition aged at 540°C. A typical fatigue fracture surface of shot peened HCF specimens of LCB is shown in Figure 7. Fatigue crack nucleation was found below the surface as often observed in titanium alloys. In case of fatigue crack nucleation below the surface, the fatigue strength in rotating beam loading (bending) is affected by the applied stress gradient, the residual stress profile, the mean stress sensitivity of the material and the material’s fatigue strength in vacuum. Further work is needed to understand the comparatively poor fatigue response of LCB to shot peening.

Figure 7: Fatigue fracture surface of a shot peened HCF specimen of TIMETAL LCB

5

Acknowledgements

The authors would like to thank Ms. G. Rodenbeck for performing the LCF tests and for carrying out the residual stress measurements.

6 [1] [2] [3] [4] [5]

References P. J. Bania, Beta Titanium Alloys in the 1990’s, (Eds.: D. Eylon, R. R. Boyer, D. A. Koss), TMS, 1993, 3. P. J. Bania, Metallurgy and Technology of Practical Titanium Alloys, (Eds.: S. Fujishiro, D. Eylon and T. Kishi), TMS, 1994, 9. C. Sommer, Titanium Science and Technology, (Ed.: P. A. Blankinsop), The University Press, Cambridge, 1996, 1836. T. Dörr, Dr.-Ing. thesis, BTU Cottbus, 2000. J. K. Gregory, L. Wagner and C. Müller, Beta Titanium Alloys, (Eds.: A. Vassel, D. Eylon and Y. Combres), Editions de la Revue Métallurgie, 1994, 229.

Effect of Test Temperature on Fatigue of Shot Peened Magnesium Alloys Jens Wendt, André Ketzmer and Lothar Wagner Chair of Physical Metallurgy and Materials Technology, Technical University of Brandenburg at Cottbus, Cottbus, Germany

1

Abstract

The effect of test temperature ranging from –25 to +50ºC on mechanical properties of the highstrength wrought magnesium alloy AZ80 was evaluated. While yield stress and tensile strength within this temperature range continuously increased with decreasing temperature, the 107 cycles notch fatigue strength in fully reversed loading exhibited a marked minimum at T = 0 ºC. Shot peening was found to improve the fatigue strength at all tested temperatures.

2

Introduction

High-strength wrought magnesium alloys are considered for substituting steels and even aluminum alloys as suspension components in future automobiles due to their high strength to weight ratio [1-3]. For this application, good HCF performance of notched components is one of the most important requirements. Since notched components respond particularly well to a shot peening treatment owing to the interaction of the notch root stress field with the process-induced residual compressive stresses [4, 5], the effect of shot peening on the fatigue performance using a wide variation in Almen intensities was studied. In contrast to applications in transmission gear housings and engine blocks, where mechanical properties at elevated (120–150 ºC) temperatures are important, typical temperatures for suspension parts in automobiles are in the range –25 to +50 ºC.

3

Experimental

The wrought magnesium alloy AZ80 (nominal composition in weight percent: 8Al, 0.5Zn, 0.2 Mn, balance: Mg) was received as extrusion from Otto Fuchs Metallwerke, Mei-nerzhagen, Germany. The rectangular bar had a cross section of 110 × 70 mm (extrusion ratio ER: 9). Specimens were machined with the load axis parallel to the extrusion direction (L). The microstructure of AZ80 is shown in Figure 1. The average =-grain size is about 30 μm. A discontinuous precipitation of Mg17Al12 is clearly seen by optical microscopy. Tensile tests were performed at various temperatures on threaded cylindrical specimens having gage lengths and diameters of 20 and 4 mm, respectively. The initial strain rate was 8.3·10–4 s–1. For fatigue testing, threaded circumferentially notched specimens (Fig. 2) with a geometrical notch factor of about kt = 3.4 were machined. After machining, roughly 200 μm were removed

387 from the surface of the specimens by electrolytical polishing to ensure that any machining effect that could mask the results was absent. This electropolished condition (EP) is taken as reference to which shot peened conditions (SP) will be compared.

Figure 1: Microstructure of AZ80

Figure 2: Geometry of fatigue specimens

Shot peening (SP) was performed with an injector type machine using spherically conditioned cut wire SCCW 14 (0.36 mm average shot size). Specimens were shot peened to full coverage using Almen intensities from 0.18 to 0.55 mmN. During shot peening, the surface of the specimens close to the notch was masked by an adhesive tape to ensure that the notch factor was not affected by removal of material from these regions. The exact value of the geometrical notch factor was calculated for each individual specimen by measuring the net diameter (dn), notch depth (t) and notch root radius (H) using the formula (1) for axial loading [6]:

kt = 1 +

1 é d ê 1+ n 0.10 2r + 1.6 ê 0.55 ê dn dn æt ö ê ç ÷ ëê 2 H 2 H èHø

ù ú ú ú ú ûú

2.5

+ 0.11

dn 2r

(1) 1.5

æ dn t öæ t ö + ÷ç ÷ ç è 2 H H øè H ø

After shot peening, the change in surface layer properties was determined by roughness measurements through profilometry, microhardness-depth profiles and measurements of residual stresses by means of the incremental hole drilling method as described in detail elsewhere [7]. Axial fatigue tests were performed in fully reserved loading (R = –1) using a resonance tester and frequencies of about 60 to 70 Hz. These tests were done in an environmental chamber at temperatures ranging from –25 to +50 ºC.

388

4

Results and Discussion

Tensile properties at the various temperatures are summarized in Table 1. As seen in Figure 3, yield stress I0.2 and tensile strength UTS continuously decrease as the temperature increases while both uniform strain Au and fracture strain El increase. Table 1: Tensile properties of AZ80 at various temperatures Temperature

E

I0.2

UTS

eU

El

[°C]

[GPa]

[MPa]

[MPa]

[%]

[%]

100

35

220

320

12.3

17.0

50

46

240

330

7.8

9.3

25

43

245

335

8.9

10.1

0

45

255

340

8.8

9.1

-10

45

260

340

6.9

7.0

-25

49

280

360

7.2

7.3

20

400 375 Stress [MPa]

UTS

325 300

El

275

eU

250

10

5

I0.2

225

Strain [%]

15

350

200

0 -50

-25

0

25

50

75

100

125

Temperature [ºC]

Figure 3: Tensile properties of AZ80 vs. test temperature

Shot peening markedly changes the surface topography as seen in Figure 4. With an increase in Almen intensity from 0.18 to 0.55 mmN, the surface roughness steadily increases. Owing to marked work-hardening in AZ80, the near-surface microhardness significantly increases during shot peening (Fig. 5). Increasing the Almen intensity from 0.18 to 0.55 mmN leads to greater depths of plastic deformation. Shot peening-induced residual stresses as determined by the hole drilling method are illustrated in Figure 6. Shot peening to 0.18 mmN Almen intensity leads to maximum residual compressive stresses at the surface, while peening to 0.55 mmN results in a marked drop of near-surface stresses (Fig. 6). For HCF testing, specimens were shot peened only to the low Almen intensity of 0.18 mmN.

389 EP

SP 0.18 mmN

SP 0.33 mmN

SP 0.48 mmN

SP 0.55 mmN

160

0

150

-10

Residual stress [MPa]

Microhardness HV 0.02

Figure 4: Surface topographies and roughness profiles after shot peening of AZ80

140

130

SP, 0.55 mmN SP, 0.18 mmN

120

110

100

90

80

bulk hardness

70

SP, 0.18 mmN

-20

-30

-40

SP, 0.55 mmN

-50

-60 -70 -80

-90

60

-100 0

100

200

300

400

500

600

Distance from surface [μm]

Figure 5: Microhardness-depth profiles after shot peening

0

100

200

300

400

500

600

Distance from surface [mm]

Figure 6: Residual stress-depth profiles after shot peening

The effect of test temperature on the HCF performance in fully reversed (R = –1) axial loading of notched specimens of AZ80 is shown in Figure 7. For the electropolished (EP) reference condition (Fig. 7a), the S-N curves at temperatures of T = 50, 25 and –25 °C are quite similar and can be characterized in the HCF regime by a common scatterband. However, the HCF performance at the temperatures of T = 0 °C and T = –10 °C is clearly inferior to the other temperatures tested. For example, the 107 cycles fatigue strength of notched AZ80 is only 50 MPa in terms of Ia · kt at a test temperature of T = –10 °C whereas it is 175 MPa at T = 50 °C (Fig. 7a). After shot peening, the temperature dependent ranking of the HCF performance in AZ80 is quite the same (Fig. 7b). Again, the S-N curves at temperatures of T = 50 °C and 25 °C are similar while there is a pronounced drop in HCF performance at lower temperatures of T = 0 °C and T = –10 °C. From Figure 7, the 107 cycles fatigue strengths of the electropolished references and the shot peened conditions were taken and replotted in Figure 8 vs. test temperature. Within the range of temperatures studied, the fatigue strength improvement caused by shot peening is independent

390

Stress amplitude, Ia · kt [MPa]

Stress amplitude, Ia · kt [MPa]

of temperature and amounts to roughly 75 MPa in terms of Ia · kt. This indicates that even at the highest tested temperature of T = 50 °C, no significant thermal relaxation of the shot peeninginduced residual compressive stresses is likely to occur.

EP

400

350 300

250

200

150

100

50

50 ºC 25 ºC 0 ºC -10 ºC -25 ºC

103

104

105

106

107

400

SP, 0.18 mmN

350

300

250 200 50 ºC 25 ºC 0 ºC -10 ºC

150

100 103

Cycles to failure, NF

104

105

106

107

Cycles to failure, NF

a) Condition EP

b) Condition SP

Figure 7: S-N curves in axial loading (R = –1) of notched (kt = 3.4) specimens of AZ80, effect of test temperature

Stress amplitude, Ia · kt [MPa]

250 200 150 100 50 0 -40

EP SP, 0.18 mmN -20

0

20

40

60

Temperature [ºC] Figure 8: Fatigue strengths (R = –1) of notched specimens of AZ80 vs. temperature

Presumably, the marked minimum in 107 cycles fatigue strengths of both the electropolished and shot peened conditions at around T = 0 °C is due to environmental rather than mechanical aspects. The humidity of the lab air in the test chamber was found to be at a maximum at around T = 0 °C. Thus, corrosion fatigue may play a significant role in fatigue of AZ80 at temperatures around 0 °C. Further work is needed to understand the comparatively poor fatigue performance of magnesium alloys at low temperatures.

391

5

Acknowledgements

The authors gratefully acknowledge the support of this work by the Bundesministerium für Wirtschaft (BMWi) under contract 138/99. Thanks are also due to Ms. G. Rodenbeck for carrying out the residual stress measurements.

6 [1] [2] [3] [4] [5] [6] [7]

References T. K. Aune and H. Westengen, Magnesium Alloys and their Applications (Eds.: B. L. Mordike and F. Hehmann) DGM, 1992, 221. H. Friedrich and S. Schumann, Magnesium 2000 (Eds.: E. Aghion and D. Eliezer) MRI, 2000, 9. G. L. Song and A. Atrens, Advanced Engineering Materials, Wiley-VCH (1999), 11. C. Gerdes and G. Lütjering, Shot Peening (Ed.: H. O. Fuchs), American Shot Peening Society, Paramus, 1984, 175. L. Wagner, C. Gerdes and G. Lütjering, Titanium Science and Technology, DGM, 1985, 2147. W. Beitz and K.-H. Grote, Handbook for Mechanical Engineering, Springer, 2001, E 103 (in German). J. Lindemann, D. Roth-Fagaraseanu and L. Wagner, Shot Peening (Ed.: L. Wagner), Wiley-VCH, 2002, in press.

392

Effect of Shot Peening on Fatigue Performance of Gamma Titanium Aluminides Janny Lindemann1), Dan Roth-Fagaraseanu2) and Lothar Wagner1) 1)

Chair of Physical Metallurgy and Materials Technology, Technical University of Brandenburg at Cottbus, Cottbus, Germany 2) Rolls-Royce Deutschland, Dahlewitz, Germany

1

Introduction

Gamma titanium aluminides are attractive candidates for application as blade material in the high-pressure part of the compressor of gas turbines. Compared to the yield stress values at both room temperature and the typical service temperature of 650 °C, the 107 cycles fatigue strengths at these temperatures are extraordinarily high, particularly if related to the material’s density being only half that of the commonly used Ni-superalloys [1, 2]. To further improve the fatigue performance of g(TiAl), mechanical surface treatments such as shot peening or roller-burnishing can be utilized. These treatments induce high dislocation densities owing to plastic deformation in near-surface regions, change the surface topography and generate residual compressive stresses. While the effect of process parameters of mechanical surface treatments on fatigue in structural steels, aluminum and conventional titanium alloys was often studied, no such information is available for gamma titanium aluminides. The present work is part of a project which was undertaken to determine potential improvements in room and elevated temperature applications of g(TiAl) by shot peening.

2

Experimental

The g(TiAl) base ingot with the chemical composition Ti-47Al-3.7(Nb, Cr, Mn, Si)-0.5B was received from Duriron (USA). The material was hipped, extruded and forged to turbine blades. After forging, the turbine blades were heat treated to achieve fully lamellar microstructures. Specimens for mechanical tests were machined from the turbine blade forgings with the load axis in longitudinal blade direction (L) as indicated in Figure 1. Tensile tests were performed on electropolished specimens having a gage length of 20 mm and a gage diameter of 4 mm. The initial strain rate was 8.3 × 10–4 s–1. Compression tests were done on cylindrical specimens with a length of 8 mm and a diameter of 4 mm. The initial strain rate was 2.1 × 10–3 s–1. Fatigue tests were performed on hourglass shaped specimens with a gage diameter of 2.5 mm in rotating beam loading (R = –1). Some tests were also done on cylindrical circumferentially notched specimens having a notch factor kt = 1.7. The specimen geometries for the various mechanical tests are illustrated in Figure 2. Shot peening was performed by means of an injector type system using spherical zirconia based ceramic shot with an average diameter of 0.5 mm. The Almen intensity was varied from 0.08 to 0.40 mmN. All peening was done to full coverage.

Æ4

a) tension

M6

393

L

Æ4

20

b) compression

Æ4

R 30

8

Æ4

Æ 2.5

c) smooth fatigue

d) notch fatigue

Figure 1: Turbine blade forging

~ 0.7

60º R 0.43

10 mm

Figure 2: Specimen geometries for mechanical testing

The change in surface layer properties was characterized using profilometry, microhardness and residual stress-depth profile measurements by means of the incremental hole drilling method [3]. The diameter of the drill was 1.7 mm. The oscillating drill was driven by an air turbine with a rotational speed of about 200.000 rpm. The shot peening induced strains in the surface layer were measured with strain gage rosettes at drilled depths of about every 20 mm. The residual stresses at each depth were calculated from the measured strain gage response using a Young’s modulus of 170 GPa. The fatigue results after shot peening were compared with the electrolytically polished reference. At least, 50 mm were removed from the machined surface to make sure that any machining effect that could mask the results was absent. The thermal stability of the shot peening-induced microhardness and residual stress profiles at an anticipated application temperature of 650 °C was determined by comparing measurements before and after an annealing treatment at this temperature for 50 hours which corresponds to the exposure time for run-outs (107 cycles at 60 Hz).

3

Results and Discussion

The typical microstructure within the blade section of the forgings (Fig. 1) is shown in Figure 3. As seen by optical microscopy, the microstructure is fully lamellar with a packet size of about 150–200 mm. The monotonic stress-strain curves of the material are plotted in Figure 4 comparing the differences in mechanical behavior between tensile and compressive loading. The 0.2 % yield stress in compression (580 MPa) is significantly higher than the corresponding value in tension (440 MPa). The apparently lower yield stress in tension might be the result of early crack nucleation and propagation which is also indicated by the low tensile ductility of only about 1 %. On the other hand, the stress-strain relationship in compression illustrates the mar-

394

True stress, I [MPa]

ked work-hardening capacity of the material which obviously can not be seen in tensile loading due to premature failure.

-0.15

-0.10

1000

500 TENSION

0.05

-0.05

0.10

True strain, A

COMPRESSION -500

-1000

Figure 3: Microstructure of g(TiAl)

Figure 4: Stress strain curves

The S-N curves of smooth (kt = 1.0) and notched (kt = 1.7) specimens of the electropolished reference are shown in Figure 5. Note that the geometrical notch factor of 1.7 reduces the smooth fatigue strength (550 MPa) by only about 10 % to 500 MPa. Thus, the maximum notch root stress Ia · kt at 107 cycles (850 MPa) is by far higher than the smooth (kt = 1.0) fatigue strength (550 MPa). This result was also found in the literature [4, 5] and is in contrast to previous results on (a+b) titanium alloys, where the geometrical notch factor kt fully (100 %) affected fatigue strength.

107

kt = 1.7

900

Cycles to failure, NF

Stress amplitude, Ia (Ia · kt) [MP

1100

700 kt = 1.0

500

106 105

104

kt = 1.7

EP

103

300 10

3

104

105

106

Cycles to failure, NF

Figure 5: S-N curves in rotating beam loading (R = –1), EP

107

EP

0.1

0.2

0.3

0.4

0.5

Almen intensity [mmN]

Figure 6: Fatigue life (Ia = 700 MPa) vs. Almen intensity

The effect of Almen intensity on the fatigue life of smooth (kt = 1.0) specimens in rotating beam loading (R = –1) at a stress amplitude of 700 MPa is plotted in Fig. 6. Starting with the electrolytically polished reference, the fatigue life drastically increases by more than two orders

395 of magnitude. No overpeening effect was found within the range of Almen intensities utilized. Further fatigue testing was done only with specimens shot peened to an Almen intensity of 0.40 mmN. Shot peening clearly increases surface roughness compared to the electrolytically polished reference (Fig. 7). Due to the process-induced plastic deformation, the microhardness in nearsurface regions significantly increases to a maximum value at the surface which is higher than twice the value of the unaffected material in the bulk (Fig. 8). This again illustrates the extraordinarily high work-hardening capacity of the g(TiAl) material (compare Fig. 8 with Fig. 4). For the given Almen intensity of 0.40 mmN, the process-induced depth of plastic deformation is roughly 300 mm. The residual stress-depth profile as measured by the incremental hole drilling method is plotted in Figure 9 indicating residual compressive stresses in near-surface regions. The maximum value of about –800 MPa was found at the surface followed by a gradual decrease within the plastically deformed surface layer (compare Fig. 9 with Fig. 8).

Microhardness HV 0.1

800 700 SP (0.40 mmN)

600

500

400 300

EP

200 0

200

400

600

Distance from surface, z [mm]

Figure 7: Surface roughness profiles (EP-electropolished, SP-shot peened, 0.40 mmN)

Figure 8: Microhardness profile after shot peening (0.40 mmN)

900 Stress amplitude, Ia [MPa]

Residual stress, IR [MPa]

200 0

-200 -400 SP (0.40 mmN)

-600 -800

-1000

800

700

SP

EP

600

500

400 kt = 1.0

300 0

200

400

Distance from surface, z [μm]

Figure 9: Residual stress profile after shot peening (0.40 mmN)

600

103

104

105

106

107

Cycles to failure, NF

Figure 9: Residual stress profile after shot peening (0.40 mmN)

The resulting S-N curve for smooth (kt = 1.0) specimens is shown in Figure 10 comparing the fatigue performance between this shot peened condition and the electropolished reference. As seen in Figure 10, the 107 cycles fatigue strength of the electropolished condition (550 MPa)

396 increases after shot peening to 675 MPa. This fatigue strength improvement by roughly 23 % is significantly higher than those previously determined on shot peened (a+b) titanium alloys such as Ti-6Al-4V [6, 7] and near-a titanium alloys such as TIMETAL 1100 [8]. Presumably, the shot peening-induced residual compressive stresses in g(TiAl) are cyclically quite stable and thus, can significantly suppress microcrack growth from the surface to the specimen interior. As seen in Figure 11, the fatigue crack nucleation site shifts from the surface of the specimens to regions below the surface after shot peening. This shift in crack nucleation site in shot peened specimens is frequently found and is the result of an effective suppression of microcrack growth from the surface to the interior by residual compressive stresses. The resistance to subsurface fatigue crack nucleation depends on the amount and cyclic stability of residual tensile stresses which balance the outer compressive stress field, the mean stress sensitivity of the material and the materials fatigue strength in vacuum. The S-N curves of the notched (kt = 1.7) specimens comparing the shot peened with the electropolished condition are shown in Figure 12. The fatigue strength increases from 500 to 630 MPa, i.e., by roughly 26 %, a value only marginally greater than observed on smooth specimens (23 %). Again, this result indicates little notch sensitivity of the HCF strength of g(TiAl). Stress amplitude, Ia [MPa]

900 800

700

SP

600

EP 500

400

300

kt = 1.7 103

104

105

106

107

Cycles to failure, NF

Figure 11: Fatigue crack nucleation site in smooth shot peened specimens

Figure 12: S-N curves (R = –1) of notched specimens

Although no elevated temperature fatigue tests were done in this investigation, a few critical tests were performed which may shed some light on the material’s response to shot peening in fatigue at 650 °C. The effect of an annealing treatment at 650 °C for 50 hours on the shot peening induced microhardness in the surface layer is shown in Fig. 13. Compared to the as-peened condition, there is only a slight decrease in microhardness which indicates that the shot peening-induced high dislocation density is quite stable at 650 °C. On the contrary, the process-induced residual compressive stresses exhibit a marked decay by the same treatment (Fig. 14). Obviously, creep deformation at 650 °C can transform most of the near-surface elastic strains to plastic strains, thus leading to a marked residual stress relief. Not surprisingly, this annealing treatment at 650 °C for 50 hours significantly deteriorates fatigue performance of shot peened specimens as seen in Figs. 15 and 16. Since both conditions SP and SP + A have a rough surface which contains microcracks, the fatigue strengths are crack propagation controlled. Owing to the residual stress decay in the annealed condition, the threshold value for microcrack growth DKth decreases which reduces the fatigue strength particularly, for notched specimens (Fig. 16).

397 200

700

600

SP (0.40 mmN)

500

400

SP + A

Residual stress, IR [MPa]

Microhardness [HV0.1]

800

SP + A

300

200

0

-200 -400 SP (0.40 mmN)

-600

-800

-1000 0

200

400

600

0

200

Distance from surface, z [mm]

Figure 13: Microhardness profile (0.40 mmN), effect of annealing

900 Stress amplitude, Ia [MPa]

Stress amplitude, Ia [MPa]

600

Figure 14: Residual stress profile (0.40 mmN) effect of annealing

900 800

700

SP SP + A

600

500 400

300

400

Distance from surface, z [μm]

kt = 1.0

800 700

500

104

105

106

SP + A

400

300

103

SP

600

107

kt = 1.7

103

104

Cycles to failure, NF

Figure 15: S-N curves (R = -1) of smooth shot peened specimens, effect of annealing

105

106

107

Cycles to failure, NF

Figure 16: S-N curves (R = -1) of notched shot peened specimens, effect of annealing

Stress amplitude, Ia [MPa]

900 800

SP + A + MP

700 SP + A

600

500

400 kt = 1.0

300 103

104

105

106

107

Cycles to failure, N

Figure 17: Roughness of shot peened and annealed conditions, effect of mechanical polishing (15 mm removed from as-peened surface)

Figure 18: S-N curves (R = –1) of shot peened and annealed conditions, effect of mechanical polishing (15mm removed from as-peened surface)

398 However, polishing shot peened and subsequently annealed specimens can not only significantly reduce surface roughness (Fig. 17) but also markedly improve fatigue performance as seen in Figure 18. For a smooth surface, the fatigue strength is crack nucleation controlled. Therefore, the work-hardened surface layer which is still present after the anneal (see Fig. 13) can increase the resistance to fatigue crack nucleation and thus, improve the fatigue strength. Regarding fatigue performance of shot peened g(TiAl) specimens at T = 650 °C, it may be derived from the annealing experiments that the work-hardened surface layer will be beneficial for improving elevated temperature fatigue strength. However, the peened surface will need polishing to take advantage of the beneficial effect of the high strength of the surface layer on fatigue crack nucleation.

4 [1] [2] [3] [4] [5] [6] [7] [8]

References W. E. Dowling et al., Microstructure/Property Relationships in Titanium Aluminides and Alloys (Eds.: Y.-W. Kim and R. R. Boyer) TMS, Warrendale, PA, 1991, 123. J. Kumpfert, Y.-W. Kim, and D. M. Dimiduk, Mat. Sci. and Eng., A192/193, 1995, 465. T. Schwarz, H. Kockelmann, VDI Report 940, 1992, 99. S. J. Trail and P. Bowen, Gamma Titanium Aluminides (Eds.: Y. W. Kim, R. Wagner and M. Yamaguchi) TMS, 1995, 883. W. V. Vaidya, K.-H. Schwalbe and R. Wagner, Gamma Titanium Aluminides (Eds.: Y. W. Kim, R. Wagner and M. Yamaguchi) TMS, 1995, 867. L. Wagner and G. Lütjering, Shot Peening (Ed.: H. O. Fuchs) American Shot Peening Society, 1984, 194. L. Wagner, Fatigue Behavior of Titanium Alloys, (Eds.: R. R. Boyer, D. Eylon and G. Lütjering) TMS, 1999, 253. T. Dörr and L. Wagner, Surface Performance of Titanium Alloys (Eds.: J. K. Gregory, H. J. Rack and D. Eylon) TMS-AIME, 1996, 231.

399

Mechanical Surface Treatments on the High-Strength Alpha-Titanium Alloy KS 120 Jürgen Kiese1), Jiulai Zhang2), Oliver Schauerte1) and Lothar Wagner2) 1)

Volkswagen AG, Wolfsburg, Germany Chair of Physical Metallurgy and Materials Technology, Technical University of Brandenburg at Cottbus, Cottbus, Germany

2)

1

Abstract

The =titanium alloy KS 120 was specifically developed by Kobe Steel, Japan to possess a fine grained microstructure resulting in high yield stress - high tensile ductility combinations and excellent HCF performance. In the present investigation, shot peening and roller-burnishing were performed to determine to what extent the HCF life of KS 120 can be further improved by mechanical surface treatments. While both shot peening and roller-burnishing increased the fatigue life of KS 120 by roughly one order of magnitude, the 107 cycles fatigue strength of the electropolished reference was not improved.

2

Introduction

Titanium alloys which are well known for their aircraft, chemical and biomedical applications are now entering the market as potential candidates in automotive engineering due to its increasing demand for high-strength light-weight alloys in order to reduce vehicle weight and fuel consumption. Recent developments in titanium application in cars are suspension coil springs for the VW Lupo FSI made of TIMETAL LCB and the exhaust system for the Chevrolet Corvette Z06 made of commercially pure titanium. In both cases, the substitution of titanium alloys for steels resulted in substantial weight savings. Previous work has shown that the fatigue life improvement caused by mechanical surface treatments in =titanium alloys is often more pronounced than that observed in (=+>)- or > alloys [1, 2]. The = titanium alloy KS 120 with the composition Ti-0.5Fe-0.6Si and 0.3 % oxygen is already used for adornments, casings of watches and as facing plates of buildings. Owing to its excellent forgeability, KS 120 may be used in future automotive application as forging material for tire rims and suspension parts. The present investigation was undertaken to evaluate the effects of thermal and thermomechanical treatments on fatigue performance in KS 120 and to determine to what extent the HCF strength can be further improved by mechanical surface treatments such as shot peening and roller-burnishing.

400

3

Experimental

The material was delivered by Kobe Steel, Kyoto, Japan as rolled plate with a thickness of 25 mm. From this plate, blanks 50 × 40 × 25 mm were machined. These blanks were solution heat treated above the > transus temperature at 1050 °C for 0.5h followed by air-cooling (AC). The material was unidirectionally rolled (UR) either at 800 or 900 °C in 8 steps from 25 to 9 mm thickness (j = –1.0)/AC with j = ln h0 /h with h0 = original thickness and h = final thickness. From the rolled material, blanks 9 × 9 × 45 mm were taken in rolling (RD) and in transverse direction (TD) and final heat treated at 700 °C for 2h. From these blanks, tensile specimens were machined having a gage length and gage diameter of 20 and 4 mm, respectively. Tensile properties for both thermomechanical treatments are compared with the as-received condition in Table 1. Table 1: Tensile test results of the various conditions of KS 120 Heat treatment RD

TD

I0.2 [MPa] UTS [MPa] El [%]

I0.2 [MPa]

UTS [MPa] El [%]

as-received

675

800

16.9

745

830

16.0

TMT1*

715

800

19.0

785

860

17.0

TMT2**

700

805

19.5

755

810

15.8

*0.5h 1050°C, UR 900 °C (j = –1.0), 2h 700 °C **0.5h 1050 °C, UR 800 °C (j = –1.0), 2h 700 °C

Crystallographic textures were determined by X-ray diffraction and will be illustrated by (0002) pole figures. Fatigue tests were performed only in TD. To study cyclic deformation behavior, stress controlled LCF tests were performed on threaded cylindrical (d0 = 4 mm, l0 = 20 mm) specimens in fully reversed (R = –1) axial loading using a servohydraulic testing machine. Tests were done at 0.1 Hz. Hysteresis loops were recorded by strain gage measurements. From these hysteresis loops, half of the plastic strain range at zero load (,epl/2) was taken and plotted versus number of cycles. For HCF tests, hour-glass shaped specimens having a minimum gage diameter of 3.6 mm were machined. Part of these specimens was shot peened by means of an injector type machine using spherically conditioned cut wire (SCCW 14) having an average shot size of 0.36 mm. Almen intensities were widely varied to determine conditions for best fatigue life improvements. Other specimens were roller-burnished using a one-roll hydraulic system operating in a conventional lathe. A hard metal ball with a diameter of 6 mm was used. The rolling force was varied in a wide range to determine best fatigue response. The change in surface layer properties caused by these mechanical surface treatments was evaluated by surface roughness measurements and microhardness profiles. In addition, residual stresses were measured by the hole drilling method as described elsewhere [3]. The HCF-tests were performed in rotating beam loading (R = –1) at frequencies of about 60 Hz. An electrolytically polished condition was taken as reference to which the mechanically surface treated specimens were compared.

401

4

Results and Discussion

The as-received microstructure of KS 120 is shown in Figure 1 indicating highly deformed grains as a result of previous hot work.

Figure 1: Microstructure of as-received KS120

Compared to this as-received condition, no significant changes in optical microstructure were found after both thermomechanical treatments TMT1 and TMT2. The (0002) pole figures are illustrated in Figure 2 comparing the as-received crystallographic texture (Fig. 2a) with those after unidirectional rolling (UR) at 900 °C (Fig. 2b) and at 800 °C (Fig. 2c).

RD

TD

a) as-received

b) TMT1

c) TMT2

Figure 2: (0002) pole figures of KS 120

Figure 2a indicates a sharp T-type of texture in the as-received plate since most grains are oriented with the basal planes being aligned in rolling direction and perpendicular to the rolling plane. This explains why the yield stresses in TD are markedly higher than in RD (Table 1) because plastic deformation is difficult in c-direction of the hexagonal unit cell. After unidirectional rolling at 900 °C, this T-type of texture is slightly altered by additional basal pole components (B/T-type of texture) indicating that additional grains are mostly oriented with the basal planes aligned almost parallel to the rolling plane (Fig. 2b). With a decrease in rolling temperature from 900 to 800 °C, the T-pole disappears (Fig. 2c). By comparing the various pole

402

1.0 0.8 I0.2

Depl/2 [o/oo]

0.9 I0.2

TMT1, R = -1 0.1 1

10

100

Cycles, N

Figure 3: Cyclic deformation behavior (R = –1) of KS 120

Stress amplitude, Ia [MPa]

figures in Figure 2, it can be argued that the as-received plate presumably had been unidirectionally rolled at temperatures significantly above 900 °C. Furthermore, reducing the rolling temperature to 750 or 700 °C may result in a fully symmetrical B-type of texture with mechanical properties being isotropic in the rolling plane [4]. The cyclic deformation behavior of KS 120 is shown as an example for the condition TMT1 in Figure 3. After a few cycles of cyclic hardening, marked cyclic softening was observed at both stress levels for most of the fatigue life. 1.000

TMT1 TMT2 as-received

900 800

700

600

500

400

EP, R = -1

300 103

104

105

106

107

Cycles to failure, NF

Figure 4: S-N curves of KS 120 (rotating beam loading in air)

The S-N curves of KS 120 are illustrated in Figure 4 comparing results of TMT1 and TMT2 with the as-received condition. According to the observed differences in yield stress (Table 1), highest 107 cycles fatigue strength was observed on TMT1 (500 MPa) followed by TMT2 (480 MPa) and the as-received condition (400 MPa). Despite the cyclic softening behavior of KS 120, the fractions 107 cycles fatigue strength to yield stress sa107 /s0.2 are fairly high amounting to 0.54 and 0.64 for the as-received and both TMT1 and TMT2 conditions, respectively. Further testing was performed on TMT1 only. The changes in surface layer properties as caused by shot peening are shown in Figure 5. Starting with the electropolished reference (EP), surface roughness values clearly increase with an increase in Almen intensity (Fig. 5a). Owing to shot peening-induced plastic deformation, there is an increase in microhardness in near-surface regions from roughly 300 HV (bulk) to 370 HV close to the surface (Fig. 5b). The magnitude of the shot peening-induced residual compressive stresses and their penetration depth clearly increase with Almen intensity as illustrated in Figure 5c. The effect of Almen intensity on fatigue life at a stress amplitude of sa = 600 MPa is shown in Figure 6. Highest lifetime improvements of roughly one order of magnitude were observed after peening with intermediate Almen intensities. From Figure 6, an Almen intensity of 0.10 mmA was taken as optimum. The S-N curve of this optimum shot peened condition is compared with the electropolished reference in Figure 7. While at intermediate and high stress amplitudes, the fatigue life is improved by shot peening by roughly one order of magnitude, no increase of the 107 cycles fatigue strength was observed (Fig. 7).

403 550

16

Roughness, [μm]

Microhardness [HV0.07]

TMT1

14 12

10

8

Ry Rz Ra

6

4 2

EP

TMT1

500 450

400 350

300

250 200

SP (0.10 mmA)

150

0 0,25

0,20

0,15

0,10

0,05

0,00

Almen intensity, [mmA]

0

250

500

750

1.000

1.250

1.500

Distance from surface, z [μm]

a) surface roughness

b) microhardness-depth profile

Residual stress, IR [MPa]

0 -100

0.10 mmA 0.15 mmA 0.24 mmA

-200

-300

-400 -500

-600

TMT1 SP -700 0,0

0,1

0,2

0,3

0,4

0,5

0,6

0,7

Distance from surface, z [mm]

c) residual stress-depth profiles

Figure 5: Surface properties of KS 120 after shot peening 107

Cycles to failure, NF

TMT1 Ia = 600 MPa

106

R=-1 105

EP 104

103 0,00

0,05

0,10

0,15

0,20

0,25

Almen intensity, [mmA]

Figure 6: Fatigue life (sa = 600 MPa) vs. Almen intensity

Roller-burnishing which did hardly change the roughness of the electropolished reference led to similar increases in near-surface microhardness (Fig. 8) as those measured after shot peening (Fig. 5b). However, the penetration depth of plastic deformation is much greater. Interestingly enough, roller-burnishing was not superior to shot peening with regard to fatigue life (Fig. 9, compare Figure 9 with Figure 6). Very similar to optimum shot peening, the S-N curve after optimum roller-burnishing (F = 500 N) gave an improvement in fatigue life by a factor of about 10 at high and intermediate stress amplitudes as compared to the electropolished reference (Fig. 10). Again, no increase of the 107 cycles fatigue strength was observed. This behavior is quite similar to results on meta-

404 1.000

TMT1 Stress amplitude, Ia [MPa]

900 800 700

EP

SP (0.10mmA)

600 500 400

R = -1 300 103

104

105

106

107

Cycles to failure, NF

Figure 7: S-N curves of KS 120 after optimum shot peening (rotating beam loading in air)

stable >-titanium alloys such as Beta C and Ti-10V-2Fe-3Al which also exhibit cyclic softening behavior [5, 6]. 107

TMT1

500

TMT1 Ia = 600 MPa R = -1

106

450

Cycles to failure, NF

Microhardness, [HV0.07]

550

400

350

300

250

200

RB (500 N)

150 0

250

105

EP 104

103

500

750

1.000

1.250

Distance from surface, z[μm]

Figure 8: Microhardness-depth profile of KS 120 after roller-burnishing

1.500

0

200

400

600

800

Force, [N]

Figure 9: Fatigue life (sa = 600 MPa) vs. rolling force

Comparing the fracture surfaces of fatigue specimens (Fig. 11), it is seen that the fatigue crack nucleation site shifted from the surface for the electropolished condition (Fig. 11a) to subsurface regions for shot peened (Fig. 11b) as well as roller-burnished specimens (Fig. 11c). Therefore, surface roughness is not involved in the HCF failure of mechanically surface treated KS 120. In case of subsurface fatigue crack nucleation, both magnitude and cyclic stability of the residual tensile stresses balancing the outer compressive stress field, the mean stress sensitivity of the fatigue strength and the fatigue strength value in vacuum need to be taken into account. More work is needed to understand why the 107 cycles fatigue strength of KS 120 is not improved by mechanical surface treatments.

405 1.000

TMT1 R = -1

Stress amplitude, Ia [MPa]

900 800 700

RB (500N)

EP

600 500 400 300 103

104

105

106

107

Cycles to failure, NF

Figure 10: S-N curves of KS 120 after optimum roller-burnishing (rotating beam loading in air)

a) EP

b) SP

c) RB

Figure 11: Fracture surfaces of fatigue failed KS 120 specimens, TMT1 (arrow indicates crack nucleation site)

5

Acknowledgements

The authors would like to thank Dr. J. Lindemann for the texture measurements and Ms. G. Rodenbeck for carrying out the LCF tests and residual stress measurements.

6 [1] [2] [3] [4] [5] [6]

References T. Dörr and L. Wagner, Surface Treatment IV, (Eds.: C. A. Brebbia, I. M. Kelly), 1999, 349. L. Wagner, Surface Performance of Titanium Alloys (Eds.: J. K. Gregory, H. J. Rack, D. Eylon), TMS, 1997, 1997. J. Kiese, J. Zhang, O. Schauerte and L. Wagner, Shot Peening (Ed.: L. Wagner), WileyVCH, 2002. M. Peters and G. Lütjering, Report CS-2933, EPRI, 1983. L. Wagner, A. Berg, T. Dörr and M. Hilpert, Mechanische Oberflächenbehandlung (Eds.: H. Wohlfahrt, P. Krull), Wiley-VCH, 2000, 179. A. Drechsler, J. Kiese and L. Wagner, Shot Peening (Ed.: A. Nakonieczny), IMP, 1999, 145.

406

The Effect of Cold Deformation and Surface Treatment on Fatigue Behaviour of Al2O3-Al6061 Composite Material Gaofeng Quan, Wolfgang Brocks Institute for Materials Research, GKSS Research Center, Geesthacht, Germany

407

408

409

410

411

412

413

Effect of Overloads on Fatigue of Shot Peened 2024 Al Vladimír Šupík and Lothar Wagner Chair of Physical Metallurgy and Materials Technology, Technical University of Brandenburg at Cottbus, Cottbus, Germany

1

Introduction

Numerous investigations in the past have shown that shot peening can improve the fatigue performance of structural materials such as steels and aluminum alloys [1-4]. However, this beneficial effect of shot peening is usually demonstrated in constant amplitude tests although components in service hardly see simple constant load amplitudes [5]. Only a limited number of studies on shot peened components has been performed under variable amplitude loading [6]. Since shot peening-induced residual compressive stresses which are known to be the main reason for the improvement of the fatigue performance may not be cyclically stable if overloads of sufficient magnitude occur in service, it is interesting to know to what extent shot peening is still beneficial if variable amplitudes are applied.

2

Experimental

The investigation was performed on the well known AlCuMg-based aircraft alloy 2024 Al. The material was delivered as Æ 12 mm extruded bar. The material was cut into blanks with a length of 50 mm. After solutionizing at 495 °C for 1 hour and water-quenching, part of the blanks was naturally age-hardened at room temperature for at least 5 days (T4 temper) while the other part was given an artificial aging by annealing at 190 °C for 12 hours (T6 temper). Tensile tests were performed on threaded cylindrical specimens having gage lengths of 20 mm. The initial strain rate was 8.3 × 10–4s–1. Tensile test results are listed in Table 1. Table 1: Tensile tests results on 2024 Al E (GPa)

I0,2 (MPa)

UTS (MPa)

IF (MPa)

El (%)

T4

74

350

500

650

18

T6

75

380

450

550

11

To determine the cyclic deformation characteristics of the two tempers, stress controlled low cycle fatigue (LCF) tests were performed on cylindrical specimens having gage lengths and gage diameters of 20 and 6 mm, respectively [7]. These tests were done in axial loading at a stress ratio of R = –1 by means of a servohydraulic testing machine. The test frequency was 1 Hz. During testing, the axial strain was recorded by strain gages. From the hysteresis loops, the plastic strain was measured and plotted versus number of cycles. For high cycle fatigue (HCF) testing, hour glass shaped specimens with a gage diameter of 3.6 mm were machined. For both tempers, part of the specimens was electropolished (EP) to

414 serve as reference. 100 mm were removed from the as-machined surface to ensure that any machining effect that could mask the results was absent. The other part was shot peened (SP) by means of an injector type machine using spherically conditioned cut wire SCCW14 (0.36 mm average shot size). During the peening treatment, the specimens rotated at 1s–1. The distance between nozzle tip and specimen surface was 45 mm. Peening was done to an Almen intensity of 0.28 mmA at full coverage. The change in surface layer properties was characterized by measurements of surface roughness through profilometry and by microhardness-depth profiles. Fatigue tests were performed in rotating beam loading (R = –1) at frequencies of about 60 Hz. During constant amplitude testing, 1000 overload cycles were applied every 10.000 cycles. The fraction overload/baseline stress amplitude was kept constant at 1.3.

3

Results and Discussion

The cyclic deformation behavior of 2024 Al is illustrated in Fig. 1 comparing the results of the T4 (Fig. 1a) and T6 (Fig. 1b) tempers [6]. For the utilized stress amplitudes, the cyclic plastic strain decreases with number of cycles in T4 indicating cyclic hardening (Fig. 1a) while cyclic softening was observed in T6 (Fig. 1b). 10

10

330MPa

1

,Apl/2 [%o]

,Apl/2 [%o]

Al2024, T4

300MPa Ia = 285MPa

0.1

1

Ia = 400 MPa

Al2024, T6

380 MPa 350 MPa

0.1 320 MPa

0

0

100

101

102

103

100

Cycles, N a) T4 (cyclic hardening)

101

102

103

Cycles, N b) T6 (cyclic softening)

Figure 1: Cyclic deformation characteristics in 2024 Al (R = –1)

The microhardness-depth profiles after shot peening are plotted in Fig. 2 comparing again results of the naturally aged T4 (Fig. 2a) and artificially aged T6 (Fig. 2b) tempers. While the penetration depth of plastic deformation in the two tempers is very similar, the near-surface hardness values in T4 are somewhat higher than in T6 whereas the bulk hardness of T6 is slightly higher than T4 (Fig. 2, compare Figs. 2a and 2b). These results can be correlated with the tensile properties (table 1) indicating lower yield stress but higher work-hardening capacity in T4 as opposed to T6. The fatigue performance of the T4 and T6 tempers is shown in Fig. 3 comparing the results of the electropolished reference (Fig. 3a) and shot peened (Fig. 3b) conditions in constant amplitude rotating beam loading. In the electropolished reference, T6 is somewhat superior to T4 presumably caused by the higher yield stress in T6 as opposed to T4 (compare Fig. 2 with Table 1). After shot peening, the fatigue performance of both tempers is markedly improved. Since

415 250

T4

Hardness HV 0.04

Hardness HV 0.04

250

200

150

100

T6

200

150

100 0

200

400

600

800

0

200

Distance from surface [mm]

400

600

800

Distance from surface [mm]

a) T4

b) T6

Figure 2: Microhardness depth profiles after shot peening (0.28 mmA)

both tempers show roughly the same S-N scatter band after shot peening, the fatigue response of T4 to shot peening is more beneficial than that of T6. 350 T4 T6

300 250 200 150 100

Stress amplitude, Ia [MPa]

Stress amplitude, sa [MPa]

350

T4 T6

300 250 200 150 100

104

105

106

Cycles to failure, NF

a) Electrolytically polished (EP)

107

104

105

106

107

Cycles to failure, NF

b) Shot peened (SP, 0.28 mmA)

Figure 3: S-N curves (constant amplitude loading) in 2024 Al, rotating beam (R = –1), effect of aging

This result confirms earlier work and can be correlated with the shot peening-induced residual compressive stresses in T4 being markedly higher than in T6 owing to the more pronounced work-hardening capacity after natural aging [3, 8]. Furthermore, the residual stresses in T4 are likely to be cyclically more stable than in T6 since T4 exhibits cyclic hardening while T6 cyclically softens as seen in Fig.1. The load-time sequence in the fatigue tests with periodic overload cycles are shown schematically in Fig. 4. After 10000 baseline cycles, 1000 overload cycles were periodically applied. The magnitude of the overload was 1.3 times the baseline load. For example, a baseline stress amplitude of 250 MPa was periodically followed by overload cycles of 325 MPa.

416

baseline

overloading

baseline

+ Ia t - Ia 10 000 N

1000 N

10 000 N

Figure 4: Load-time sequence in fatigue testing with periodic overload blocks (schematic)

350 T4 T6

300 250 200 150

Stress amplitude, Ia [MPa]

Stress amplitude, Ia [MPa]

350

T4 T6

300 250 200 150 100

100

104

105

106

Cycles to failure, NF

a) Electrolytically polished (EP)

107

104

105

106

107

Cycles to failure, NF

b) Shot peened (SP, 0.28 mmA)

Figure 5: S-N curves (variable amplitude loading) in 2024 Al, rotating beam (R = -1), effect of aging

The fatigue response of the T4 and T6 tempers to the periodic overloads is illustrated in Fig. 5 comparing the fatigue performance of the electropolished reference (Fig. 5a) and shot peened (Fig. 5b) conditions. As in the constant amplitude tests, the baseline stress amplitude is plotted. In the electropolished reference, no significant differences in fatigue life between the two tempers T4 and T6 are seen (Fig. 5a). Thus, the constant amplitude fatigue performance of T6 is somewhat more affected by the periodic overloads than T4 (compare Fig. 5a with Fig. 3a). This less inferior response of the fatigue performance of T4 to periodic overloads is somewhat more pronounced in the shot peened condition (Fig. 5b, compare Fig. 5b with Fig. 3b). It is argued that the higher cyclic stability of the shot peening-induced residual compressive stresses in T4 owing to the materials cyclic hardening characteristics is still useful if periodic overloads occur. Cyclically stable residual compressive stresses will lead to reduced growth rates of microcracks thus, improving fatigue life [9].

4 [1] [2]

References H. Wohlfahrt, Shot Peening (Eds.: H. Wohlfahrt, R. Kopp and O. Vöhringer) DGM (1987) 563. R. Schreiber, H. Wohlfahrt and E. Macherauch, Arch. Eisenhüttenwes. 48 (1977) 653.

417 [3] [4] [5] [6] [7] [8] [9]

T. Hirsch, O. Vöhringer and E. Macherauch, Shot Peening (H. O. Fuchs, ed.) American Shot Peening Society (1984) 90. T. Dörr, M. Hilpert, P. Beckmerhagen, A. Kiefer and L. Wagner, Shot Peening (A. Nakonieczny, ed.) IMP (1999) 153. P. SkoUovský, P. PalUek, R. KoneUná, L. Várkoly, KonštrukUné materiály, ŽU v Žiline, EDIS (2000) 167. W. Schütz, Shot Peening (H. O. Fuchs, ed.) American Shot Peening Society (1984) 166. T. Dörr, Dr.-Ing. dissertation, BTU Cottbus (2000) L. Wagner and C. Müller, J. Materials Manufacturing & Processing (1992) 423. J. K. Gregory and L. Wagner, LCF 3, Elsevier Applied Science (1992) 588.

418

409

VIII Alternative Mechanical Surface Treatments

410

421

Alternative Mechanical Surface Treatments: Microstructures, Residual Stresses & Fatigue Behavior Igor Altenberger * Department of Materials Science & Engineering, University of California, Berkeley, CA, USA * Now at: Institute of Materials Technology, University of Kassel, Kassel, Germany

1

Abstract

In comparison to the most widely used mechanical surface treatment shot peening, common alternative methods such as deep rolling and less common methods such as laser shock peening, ultrasonic shot peening, water peening or various burnishing methods have been introduced into practical applications only rarely, or for highly specialized components, or are just on the verge from laboratory research into larger scale applications. However, in the future it is expected that these so called “alternative” mechanical surface treatment methods will be more widespread owing to superior benefits for materials’ behavior, improving process technology and dramatically decreasing costs. The basic principles of all mechanical surface treatments are well known: In all cases a localized elastic-plastic deformation in near-surface regions leads to the formation of compressive residual stresses and severe microstructural alterations (usually associated with intense work hardening), enabling the thus strengthened near-surface regions to withstand higher resistance against fatigue crack initiation and propagation. Moreover, in some cases, additional effects may give rise to further fatigue life/strength enhancement such as surface smoothening or deformation-induced phase transformations. At closer look, near surface properties and thus fatigue behavior might be distinctly different for different surface treatment methods. It is the objective of this contribution to shed some light on these basic effects and to propose some basic guidelines for the utilization of ‘optimized’ treatments from a materials science perspective.

2

Introduction

The prime objective of this paper is to summarize the basic mechanical and metallurgical effects associated with specific mechanical surface treatment methods. It is not the aim of this paper to give an overview on the technological aspects of different mechanical surface treatment methods. Such studies can be found elsewhere [1,2,3]. Also, even when comparing the metallurgical alterations by different surface treatments, one should always keep in mind that different process parameters for a single surface treatment method can lead to a broad range of possible properties, thus rendering such comparisons very difficult. A systematic comparative study on mechanical surface treatments from a technological point of view has been presented, for instance, in [3]. There, surface treatment methods such as shot peening, deep rolling, water peening, laser shock peening are also discussed in terms of residual stresses, hardness increase, case depth and effect on stress-life behaviour. Other comparative studies on different mechanical surface treatment methods can be found in [1,2,4–9] with varying thematic emphasis. In this

422 chapter, the most important results of former and recent studies on various mechanical surface treatments will be summarized. Most importantly, recent results on the nature of stability of near surface microstructures under severe loading conditions will be discussed. The following surface treatment methods can be considered as “mechanical”: Shot peening, deep rolling and roller burnishing, laser shock peening, ultrasonic shot peening, water peening, hammering, needle peening, tumbling. The effects of mechanical surface treatments on near-surface properties can be characterized by a multitude of primary and secondary parameters such as case depth, magnitude of residual stresses and hardness, roughness, nature of near-surface microstructures, phase contents, porosity, texture, corrosive properties, quasistatic and cyclic yield strength, stability against mechanical loading under quasistatic and cyclic conditions, stability against thermal loading and stability against thermomechanical loading. The effects on fatigue behaviour are the sum of all these near-surface alterations and therefore even more complex. Firstly, it shall be discussed how different surface treatment methods characteristically affect materials’ near–surface properties. Secondly, the consequences of different surface treatments on fatigue behaviour will be discussed.

3

Near-Surface Alterations by Various Mechanical Surface Treatments

3.1

Roughness

It is irrefutable that surface topography severely influences fatigue life and strength of hard and notch-sensitive materials: The higher the surface roughness the higher the stress concentrations by notches and the lower the fatigue strength. Therefore it is mandatory to not only optimize residual stresses and near-surface microstructures, but also to optimize the surface topography by mechanical surface treatments. Specific mechanical surface treatments can have quite different effects on surface roughness and of course they give rise to a broad range of surface topographies by varying process parameters. Nevertheless some basic guidelines can be stated: In general, for typical as-turned or as-milled surfaces, deep rolling and roller burnishing are the only treatments which diminish surface roughness of machined components significantly. Surface roughnesses Rz of 0.5 to 1 mm are quite common for these treatments. In contrast, shot peening is typically associated with surface roughnesses between 4 and 8 mm, depending on the exact Almen intensity and shot geometry. The surface roughness by laser shock peening and after water peening is hardly altered as compared to the untreated state. Whereas water peening scarcely influences surface roughness, laser shock peening usually increases the surface roughness of the as-machined part slightly [1]. 3.2

“Case” Thickness

One of the most important parameters in surface treatment is the “case”-thickness which is defined as the thickness of the near-surface layer exhibiting compressive residual stresses and strain hardening or in other words the affected depth of the surface treatment. The “case” thickness depends strongly on the surface treatment method as well as on the parameters within one treatment itself such as rolling force, Almen intensity, coverage etc.. The

423 maximum “case” thickness can range between 2–3 mm for laser shock peening and deep rolling [3,8] and 0.1–0.2 mm for water peening. Usually the typical affected depth is around 0.3–0.5 mm for shot peening and around 1 mm for laser shock peening and deep rolling. The thickness of the affected layer through water peening rarely exceeds 0.1 mm [1,10]. The right choice of an “optimized” case thickness should always consider the material state and the loading conditions. For example, a deep “case” is much more essential in push-pull loaded components than in parts subjected to bending with high stress gradients. It should be noted that the deepest “cases” in surface treatment are not caused by mechanical treatments, but by thermochemical (case hardening) or thermal (induction hardening) methods [3].

3.3

Residual Stresses

The formation of compressive residual stresses in near-surface regions by surface treatment is considered as one of the main causes for fatigue life enhancement. In general, the maximum possible amount of compressive residual stresses is much more influenced by the material properties than by the process parameters, e.g. the maximum possible level of residual stress strongly correlates to the yield strength of the surface treated material [2,11]. Therefore, it can be assumed that all mechanical surface treatments generate very similar levels of surface residual stress for the same material if the process parameters are optimized in such a way that maximum compressive residual stresses are formed. On the other hand, it has been shown that different mechanical surface treatments also give rise to different levels of strain hardening [6], thus altering the yield strength differently. Indeed, it appears that treatments with very high deformation grades such as deep rolling lead to slightly higher residual stresses than “low plasticity” surface treatments such as laser shock peening [12], however, much more work in this field is needed to give a systematic assessment. Finally, different mechanical surface treatments are also associated with different stress states (e.g. different degrees of multiaxiality) during the treatment itsself, depending on the contact geometries of the utilized tools, and consequently different residual stress depth distributions: For example ‘hook’-like residual stress depth distributions are quite common for shot peening and deep rolling, but are not very typical for laser shock peening.

3.4

Work Hardening

Most metallic materials exhibit work/strain hardening through mechanical surface treatments. An exception are severely cold deformed alloys and hardened steels which show near-surface softening as indicated by lower FWHM-values in near-surface layers as compared to the bulk FWHM-values. For the assessment of work hardening states it is recommended to use x-ray peak broadening values (FWHM- or half-width values) as a means of characterization instead of simple hardness values, since the latter ones are not as sensitive and can be significantly influenced by residual stresses [2,13]. A difficult issue is the extent of work hardening for different surface treatment methods. Here, again the exact process parameters significantly influence the work hardening state and render a systematic comparison difficult. One possible method of characterizing work hardening of mechanically surface-treated near surface layers is the depthdependent registration of FWHM-values after successive electrolytical removal of material. The

424 obtained FWHM-values can be compared to FWHM-values after uniaxial deformation and thus deformation grades can be estimated at least roughly [14]. Another indirect method for estimating deformation grades in mechanically surface treated materials is to conclude the deformation state by comparing their cyclic deformation behaviour with that of uniaxially predeformed non-surface treated samples [6]. Of course, in that case residual stress effects have to be eliminated (e.g. by hollow-drilling/eroding or annealing) without affecting the work hardening state. An overview on the induced cold work, microhardness increase, dislocation density as well as other factors by different mechanical surface treatments is given in table 1. The results are taken from references [1-12,14-19]. In spite of being somewhat arbitrary, they give a first hint of what magnitudes of work hardening and microhardness increases can be expected for different treatments. The readers are strongly encouraged to complete and expand these results by own investigations and experiences! Table 1: Consequences of various mechanical surface treatments on near-surface properties of metallic materials Amount of Dislocation Estimated Surface Maximum Surface Cold residual density Strain rate microhard“case” depth Rough- work stress ness increase ness Roller bur- @ sYield nishing (low pressure)

Low medium

< 102 s–1

< 60 %

< 0.1 mm

@ 1 μm

Water peening

@ sYield

Low medium

?

?

@ 0.1 mm

1–2 μm < 10 %

Shot peening

@ sYield

Very high 5–8 x 1011 cm–2

103–104 s–1 150 % AISI 304 60 % SAE 1045

0.3 mm

4–8 μm 5–50 %

Explosive hardening

@ sYield

Very high

?

80 %

0.3–0.8 mm < 5 μm

Ultrasonic @ sYield shot peening

High

?

?

0.8 mm

>> 5 μm ?

Gravity pee- @ sYield ning

High

103–104 s–1 ?

0.8 mm

>> 5 μm 10 %

Laser shock @ sYield peening

Medium 2,6 x 1011 cm–2 6,2 x 1010 cm–2

Deep rolling @ sYield

105 s–1

< 102 s–1

?

?

40 % 2024 Al 2 mm 30 % 7075 Al 92 % AISI 316 L 80 % plain carbon steel 130 % maraging steel

1 – 5 μm 1–2 % 7% (Fe-3Si) 10–20 % (Ti-6Al4V)

60 %

£ 1 μm

3 mm

> 20 %

425 Several authors have observed that shot peening leads to most severe work hardening, whereas laser shock peening and water peening lead to significantly lower dislocation densities [9,12,14,20]. Surface states after deep rolling are characterized by intermediate dislocation densities of 1011 cm–2 or lower [6].

3.5

Microstructure

The possible microstructures created by mechanical surface treatment can be extremely manifold and depend strongly on the chosen process parameters as well as on the material itself. Typical near-surface microstructures may involve high dislocation densities (either homogeneously distributed in tangles [20,21] or in various stages of cell-formation [22]), slip bands [17], nanocrystallites [23,24], twinning [17,25], martensitic transformations [18,23] or stress induced precipitates [26]. In all cases, the observed defect structures are strongly influenced by the strain rate of the surface treatment and by the specific glide behavior of dislocations in the material, especially in bcc-metals. “Wavy slip” materials and low or medium strain rates (as for rolling or burnishing treatments) favor the creation of cell-like dislocation arrangements, whereas “Planar slip” materials and high strain rates (as for laser shock peening, explosive hardening, shot peening or water peening) typically produce more homogeneous, tangled dislocation arrangements. In bcc metals, laser shock peened substructures tend to resemble typical low-temperature substructures generated by conventional low strain rate plasticity [27]. Under deformation at high strain rates or low temperatures, the edge components of dislocations can move at higher rates than the screw components which are unable to cross slip, thus preventing the formation of cell structures [28]. Fig. 1 shows typical near-surface dislocation arrangements as a consequence of a high strain rate (here: conventional shot peening) and a medium strain rate (here: deep rolling) mechanical surface treatment of a ferritic steel SAE 1045. A predominant feature of severely surface deformed metals and alloys can also be the formation of a thin nanocrystalline layer. Interestingly, high coverage and high deformation grades seem to promote near-surface nanocrystallization while high deformation velocities appear to lead to the opposite effect. Fig. 2 shows direct near-surface microstructures of deep rolled, of laser shock peened and of shot peened austenitic stainless steel AISI 304 (as prepared by cross-sectional transmission electron microscopy [21–23]). It can be seen that in deep rolled and in shot peened samples nanocrystalline surface layers were formed. However, after laser shock peening no such structures were found; instead, the near-surface microstructure of laser shock peened AISI 304 is characterized by a dense highly tangled dislocation arrangement similar to near-surface microstructures observed after water peening [20].

3.6

Stability under Mechanical and Thermal Loading

A necessary prerequisite for the effectiveness of mechanical surface treatments to enhance fatigue behaviour is the mechanical and thermal stability of near-surface residual stresses or microstructures. Only stable residual stresses influence fatigue strength/life and are usually regarded as local mean stresses, and only stable near surface microstructures (e.g. increased dislocation density) are able to serve as effective dislocation obstacles impeding localized slip and thus crack formation.

426

Figure 1: TEM cross-sectional micrograph of near-surface microstructures in shot peened (left) and deep rolled (right) normalized steel SAE 1045 (Almen intensity 0.175 mmA, rolling pressure 75 bar)

Figure 2: Cross-sectional TEM micrographs of mechanically surface treated austenitic stainless steel AISI 304. Left figure: shot peened (Almen intensity 0.175 mmA, 100 % coverage), Middle figure: deep rolled (rolling pressure 150 bar), Right figure: laser shock peened (power density 7 GW/cm², 200 % coverage)

The stability of macro- and micro residual stresses in mechanically surface treated materials against mechanical (e.g. quasi-static or cyclic) loading is determined by the amount of plasticity

427 during mechanical loading. It has been shown that residual stress relaxation can be directly correlated to the plastic strain amplitude during fatigue. Consequently, it is important to induce high work hardening and deep “cases” by mechanical surface treatments since both increase the cyclic compound yield strength of the surface treated component [6]. Several investigations have been focused on the thermal stability of macro- and micro residual stresses of differently mechanically surface treated materials [2,29,30]. The most notable finding is the observation, that thermal stress relaxation depends strongly on the material state, especially on the surface treatment induced dislocation density. Surface layers with medium dislocation densities (as created by water peening or laser shock peening, for example) showed enhanced thermal stability of residual stresses, whereas surface layers with extremely high dislocation densities (as induced by shot peening) exhibited poor stability against thermal loading [29,30]. This observations can be explained by taking into account the microstructural mechanisms for (macro) stress relaxation: Since stress relaxation is caused by thermally activated climb of edge dislocations, surface layers with severe work hardening are prone to easier stress relaxation by so called pipe-diffusion (with lower activation energies than in bulk diffusion) [31]. The thermal stability of near-surface microstructures can be studied in a direct manner by In-Situ-heating of cross-sectional TEM foils. Fig. 3 depicts the direct near-surface microstructure of deep rolled Ti-6Al-4V for different heating temperatures (holding time for each temperature: 5 min, heating rate: 100 K/10 min). After deep rolling a nanocrystalline grain structure was observed. Successive heating of this TEM-foil yielded vital information about the thermal stability of this surface treatment induced microstructure. It was found that these near-surface nanostructures were thermally quite stable and that -apart from some minor recovery- no visible recrystallization took place below temperatures of approx. 800–900 °C. This implies, that from a fatigue-point of view, near-surface microstructures in deep rolled Ti-6Al-4V are suitable for high temperature applications and serve to improve the fatigue behaviour of this alloy even at elevated temperatures way above the usual service temperature for this alloy in aircraft turbine applications.

Figure 3: In-situ TEM micrograph of deep rolled (rolling pressure 150 bar) Ti-6Al4V surface regions before (left) and after (right) heating to 900 °C. (holding time 5min). Here, the observed nanocrystalline surface layer is thermally stable.

428

Figure 4: In-Situ TEM micrographs of deep rolled AISI 304 surface regions (rolling pressure 150 bar) heated to 500 °C, 600 °C and 700 °C (holding time 5 min) showing a nanocrystalline layer and recrystallization between 600 °C and 700 °C

Figure 5: In-Situ TEM micrographs of laser shock peened AISI 304 (power density: 7 GW/cm², coverage: 200 %) heated to 700 °C, 800 °C and 900 °C (holding time 5 min). This near-surface microstructure exhibits higher thermal stability than the deep rolled surface condition.

429 In a second case the thermal stability of direct near-surface microstructures of deep rolled and of laser shock peened steel AISI 304 was investigated by the aid of In-Situ-TEM (Fig. 4 and 5). It was found that near-surface nanocrystals in deep rolled AISI 304 are thermally stable until approx. 600–650 °C when recrystallization sets in. As mentionend above, laser shock peening did not not induce any nanocrystals, but resulted in highly tangled and dense dislocation arrangements. These near-surface dislocation tangles were stable until even higher temperatures of about 800 °C. Both findings correlate with the excellent and enhanced fatigue behaviour of mechanically surface treated AISI 304 at test temperatures of 600 °C or lower [32].

4

Effects on Fatigue Behaviour

It is known that mechanical surface treatments affect all fatigue stages – from the first dislocation movements until macro crack propagation and final failure. Although a multitude of factors are known to influence the fatigue strength/life of mechanically surface treated metallic materials, three of these factors have been identified as especially influential: the residual stress state, the microstructure and the roughness [2]. More specifically, compressive residual stresses and work hardening improve the fatigue behaviour of surface treated materials significantly, however their effects are assumed to be different for crack initiation and crack propagation. According to [33], for shot peened and room temperature fatigued Ti-6Al-4V, it appears that residual stresses only influence crack propagation, but have little effect on crack initiation. Secondly, work hardening enhances the resistance against fatigue crack initiation, whereas it seems to facilitate crack propagation (table 2). Other investigations on the high temperature fatigue behaviour of deep rolled Ti-6Al-4V indicate, however, that work hardening retards crack initiation as well as crack propagation ( Fig. 6 and [32]). Table 2: Effects of mechanical surface treatment on crack nucleation and crack propagation according to [33] Crack nucleation

Crack propagation

Surface roughness

Accelerates

No effect

Cold work

Retards

Accelerates

Residual compressive stress

Minor or no effect

Retards

The fatigue mechanism in smooth mechanically surface treated samples (e.g. by deep rolling) is mostly initiation controlled, whereas it can be considered crack propagation controlled (damage-tolerant approach) in components with notches or rough surfaces such as in shot peened parts without subsequent polishing. In addition, it has been shown that residual stresses affect the fatigue behaviour of hard materials much more than of soft materials: Hard materials are much more sensitive to residual stresses than soft materials. Therefore, it can be summarized, that compressive residual stresses are the main influential factor on the fatigue strength/life for notched hard materials, whereas work hardening dominates the fatigue strength of smooth soft materials [34]. The following discussion shall confine itself to smooth-bar fatigue conditions and to materials with low and medium yield strength, therefore the effect of work/strain hardening will be treated more closely.

430

Figure 6: Crack growth rates from striation spacings measurements on fracture surfaces of untreated and deep rolled Ti6Al-4V by different rolling pressures

Figure 7: Cyclic deformation curves of deep rolled steel SAE 1045 for different “case” depths obtained

According to Coffin-Manson´s law the fatigue life in the finite life region is controlled by the extent of cyclic plasticity during fatigue: In a double-logarithmic plot the number of cycles to failure is inversely proportional to the plastic strain amplitude after half the number of cycles to failure. The higher the cyclic compound yield strength of a mechanically surface treated material, the lower the plastic strain amplitude. The cyclic compound yield strength depends on a) the volume fraction of the strain hardened regions as compared to soft core regions b) the cyclic yield strength of the work hardened surface layer (which can be determined by methods described in [35]), the cyclic yield strength of the untreated soft core material. Factors a) and b) can be influenced by the choice of mechanical surface treatment and by variation of process parameters (e.g. rolling force, Almen intensity etc.). Fig. 7 shows cyclic deformation curves of deep rolled SAE 1045 with different “case” depths of work hardened material but identical residual stresses and FWHM-values at the surface. It can be seen that the plastic strain strain amplitude is systematically lowered and lifetime is systematically increased with increasing volume fractions of work hardened material. Fig. 8 shows cyclic deformation curves of deep rolled and shot peened SAE1045 with identical “case” depths, but different FWHM-values and therefore different levels of work hardening. The shot peened material condition exhibited much higher dislocation densities than the deep rolled state [6,22] and shows superior cyclic deformation behaviour (lower plastic strain amplitudes) and a longer fatigue life as compared to the deep rolled condition. The fatigue strength/life can therefore be increased by surface treatments which cause high work hardening (e.g. by high dislocation densities or by martensitic transformation) or by surface treatments which induce deep work hardened “cases”. It could be shown that, for push-pull loading, deep rolled SAE 1045 specimens have to have at least 50 % greater “case” depths than shot peened SAE 1045 specimens in order to compensate for the lower near-surface work hardening. These implications are illustrated schematically in Fig. 9. It should be noted that Fig. 9

431

Figure 8: Cyclic deformation curves (Ia = 400 MPa, R = –1) of shot peened and of deep rolled SAE 1045 push-pull specimens with identical “case”depths and residual stresses, but different levels of cold work

cold work

explosive hardening

shot peening

gravity peening

fatigue strength

ultrasonic shot peening

w ater peening roller burnishing

deep rolling laser shock peening

(low pressure)

"case" depth

Figure 9: Cold work, “case” depths and fatigue strength for different mechanical surface treatments (schematically)

is expected to be only valid for push-pull loading of smooth samples with identical or very similar surface roughnesses. Unfortunately, it does not seem to be possible to create “cases” with both maximum work hardening and maximum “case” depth. A good compromise, however, is deep rolling, since it

432

LSP

LSP

delivers deepest cases and still quite high dislocation densities/work hardening (see also table 1). A practical example can be seen in Fig. 10. The superior high temperature fatigue behaviour of deep rolled Ti-6-4 as compared to laser shock peened Ti-6-4 could be correlated to significantly increased cold work in the deep rolled condition, since the stable residual stresses and the case depth were practically identical for both material states after half the number of cycles to failure [12].

Figure 10: Fatigue lifetime enhancement of Ti-6Al-4V by deep rolling and by laser shock peening for test temperatures of 25 °C and 450 °C and stress amplitudes of 750 MPa and 400 MPa, respectively [32].

5 •

• •

6

Conclusions Alternative mechanical surface treatments offer several advantages as compared to shot peening. Main advantages are better surface topography, deeper affected layers and higher thermal stability of surface layers. Since the strengthening mechanisms are quite similar than for shot peening the same fundamental principles for optimized fatigue life/strength improvement apply. A major difficulty in comparing different mechanical surface treatments is certainly the vast range of process parameters. Therefore, further standardization is needed also for alternative treatments.

Acknowledgements

Sincere thanks is expressed to Deutsche Forschungsgemeinschaft (DFG) and to the U.S. Airforce Office of Scientific Research for financial support, and to Metal Improvement Company, Livermore (CA), for the laser shock peening treatment. The author would also like to thank Prof. B. Scholtes, Prof. R.O. Ritchie, Dr. E. Stach, Mr. U. Noster, Mr. T. Ghwilla and Ms. G. Liu for helpful discussions and experimental support.

433

7 [1] [2] [3] [4] [5]

[6]

[7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25]

References H. Wohlfahrt, P. Krull (Eds.), Mechanische Oberflaechenbehandlungen, Wiley-VCH, Weinheim, 2000. B. Scholtes, Eigenspannungen in mechanisch randschichtverformten Werkstoffzustaenden, DGM-Informationsgesellschaft, Oberursel, 1990. H.-W. Zoch, Haertereitech. Mitt. 50 (1995) 207. L. Wagner, Mater. Sci. Eng. A263 (1999) 210. A. Sollich, Verbesserung des Dauerschwingverhaltens hochfester Staehle durch gezielte eigenspannungserzeugung, Dissertation, University Gh Kassel, VDI-Fortschrittsberichte, Reihe 5, Nr. 376, VDI-Verlag Duesseldorf, 1994. I. Altenberger, Mikrostrukturelle Untersuchungen mechanisch randschichtverfestigter Bereiche schwingend beanspruchter metallischer Werkstoffe, Forschungsberichte aus dem Institut fuer Werkstofftechnik – Metallische Werkstoffe der Universitaet Gh Kassel (Ed. B. Scholtes), Band 3, Universitaetsbibliothek Kassel, 2000. M. Gerland, M. Hallouin, H. N. Presles, Mater. Sci. Eng. A156 (1992) 175. J. J. Daly, J. R. Harrison, L. A. Hackel, In: Proc. 7th Int. Conf. on Shot Peening, The Institute of Precision Mechanics, Warsaw, 1999, p. 379. P. Peyre, R. Fabbro, P. Merrien, H. P. Lieurade, Mater. Sci. Eng. A210 (1996) 102. H. K. Toenshoff, F. Kroos, In: Proc. 4th Int. Conf. on Residual Stresses, Baltimore, 1994, p. 615. B. Scholtes, O. Voehringer, Mat. Wiss. u. Werkstofftech. 24 (1993) 421. I. Altenberger, R. K. Nalla, U. Noster, B. Scholtes, R. O. Ritchie, In: Proc. 7th National High Cycle Fatigue Conference, 2002. G. Zoeltzer, I. Altenberger, B. Scholtes, Haerterei. Tech. Mitt. 56 (2001) 347. M. J. Shepard, P. R. Smith, P. S. Prevey, A. H. Clauer, In: Proc. 5th National High Cycle Fatigue Conference, 2000. J. Fournier, P. Ballard, P. Merrien, J. Barralis, L. Castex, R. Fabbro, Journal of Physics 3, Sept. 1991,p. 1467. A. H. Clauer, In: Surface Performance of Titanium (Eds. J. K. Gregory, H. Rack, D. Eylon), TMS, Warrendale, PA (1996), p. 217. A. H. Clauer, B. P. Fairand, B. A. Wilcox, Metall. Trans 8A (1977) 119. J. P. Chu, J. M. Rigsbee, G. Banas, F. V. Lawrence, H. E. Elsayed-Ali, Metall. Mater. Trans. 26A (1995) 1507. J. P. Chu, J. M. Rigsbee, G. Banas, H. E. Elsayed-Ali, Mater. Sci. Eng. A260 (1999) 260. P. Krull, T. Nitschke-Pagel, H. Wohlfahrt, In: Surface Treatment IV (Eds. C.A. Brebbia, J. M. Kenny), WIT press, Southampton, UK, 1999, p. 291. U. Martin, I. Altenberger, B. Scholtes, K. Kremmer, H. Oettel, Mater. Sci. Eng. A246 (1998) 69. I. Altenberger, B. Scholtes, U. Martin, H. Oettel, Haerterei. Tech. Mitt. 53 (1998) 395. I. Altenberger, B. Scholtes, U. Martin, H. Oettel, Mater. Sci. Eng. A264 (1999) 1. K. Lu, J. Lu, J. Mater. Sci. Technol. 15 (1999) 193. I. Altenberger, B. Scholtes, In: Surface Treatment IV (Eds. C.A. Brebbia, J.M. Kenny), WIT press, Southampton, UK, 1999, p.281.

434 [26] A. Nakonieczny, W. Szyrle, J. Suwalski, In: Proc. 6th Int. Conf. on Shot Peening, (Ed. J. Champaigne), San Francisco, 1996, p. 263. [27] W. C. Leslie, in: R. W. Rohde, B. M. Butcher, J. R. Holland, C. H. Karnes (Eds.), Metallurgical Effects at High Strain Rates, Plenum, New York, 1973, p. 571. [28] F. Burgahn, V. Schulze, O. Voehringer, E. Macherauch, Mat.-wiss. u. Werkstofftech. 27 (1996)521. [29] P. Krull, T. Nitschke-Pagel, H. Wohlfahrt, In: Conf. Proc. EUROMAT 98, Vol.1, Lisbon, 1998, p. 519. [30] P. S. Prevey, D. Hornbach, Thermal Residual Stress Relaxation and Distortion in Surface Enhanced Gas Turbine Components, In: ASM/TMS Materials Week, Indianapolis, USA, 1997. [31] O. Voehringer, in [1], p. 29 ff. [32] I. Altenberger, U. Noster, B. Scholtes, R. O. Ritchie, In: Proc. Fatigue 2002 (Ed. A. Blom), Stockholm, 2002. [33] L. Wagner, In: Fatigue Behaviour of Titanium Alloys (Eds. R. R. Boyer, D. Eylon, G. Luetjering, The Minerals, Metals & Materials Society, Warrendale, 1999, p. 253. [34] P. Strigens, VDI-Z 114 (1972) 1193. [35] R. Menig, Doctorate Thesis, University Karlsruhe (TU), 2002, in print.

435

Cavitation Shotless Peening for Improvement of Fatigue Strength Hitoshi Soyama, Dan Odhiambo and Kenichi Saito Department of Machine Intelligence and Systems Engineering, Tohoku University, Sendai, Japan

436

437

438

439

440

441

Comparison of the Residual Stresses Induced by Shot (Stress) Peening and Rolling in Spring Steel Eckehard Müller Fachhochschule Bochum (University of Applied Science), Bochum, Germany

1

Introduction

Today, reduction of weight of material is an actual subject in wide areas of automotive and mechanical engineering. Several possibilities are available, like better utilization of the material by higher hardness or optimizing the construction by finite elements. Two further possibilities are rolling of components and shot peening or the later developed stress peening. Because it is possible to induce compressive residual stresses in dedicated amounts, it is used very often at tensile pulsating load to get a better utilization of the material. The two possibilities have different advantages, which are on the one hand cheaper realization and so huge amount of induced residual stresses like vice versa.

2

Basics

2.1

Shot Peening and Stress Peening

Shot peening is a technology, which is a standard procedure. Peening (in the technical understanding) is the interaction between a particle (with the necessary hardness) with the surface of a working piece. If the particles have a round shape, you call it shot peening [Mü 93a]. In the surface layer (up to 0.5 mm depth) compressive residual stresses are induced. At lower hardness of the working piece, in addition, hardening is achieved. In order to get better results by the peening process, the so called stress peening is used. Here the working piece or component is stressed in the direction of the later loading. After this step the original peening procedure is done and afterwards the unloading. The compressive residual stress profile, which now is achieved, is significantly higher than that gained by normal peening. The result depends on the (torsional-) preload (Jks) Iks during peening [1-12]. 2.2

Deep Rolling

Deep rolling can be regarded as a continuously made plastic deformation of material near the surface. In this procedure a tool (ball, profile roll) is pressed against the work piece with a special pressing force. Here the important parameters are the pressing force and the overlap. This standard procedure is described many times, for example in [13–16].

442

3

Preparation of the specimens

The specimens (the values for rolled parts are in brackets) were pieces of flat steel, which is normally used for leaf springs, with a length of 360 mm (360 mm), a wide of 80 mm (75 mm), and a thickness of 11 mm (15 mm). The edges were rounded in accordance with DIN 59145. The material was normal spring steel 50CrV4 (55Cr3). The specimens were heated up to 880 °C and quenched in oil to get a martensitic structure. Afterwards, they were heat treated at around 460 °C to get a strength of Rm = 1530 (1420) N/mm2. To avoid spreads of the hardness in the surface layer because of decarburization, the specimens were ground to remove a layer of at least 0.3 mm.

4

Experimental setup

4.1

Treatment of the Specimens

4.1.1 Shot Peening In figure 1, the mounting device is shown, in which the samples are put for the further working steps. The sample could be loaded, which is not shown in the figure. The shot or stress peening was done with the help of this device. As shot, conditional cut wire with a nominal diameter of 0.8 mm was used. The peening was done in a shot peener with four wheels in different directions. To verify the calculated loading stress in every sequence, an additional step of loading/unloading is inserted. After each step, the stresses at the surface were measured along and perpendicular to the axis of the specimens. The following basic sequence was carried out: 1. 2. 3. 4.

initial state, heat treated, ground, unloaded specimen plastified at the calculated stress Ib = 1800 N/mm2 loaded specimen with the calculated stresses Iks = 0; 600; 1200; 1600 N/mm2 specimen unloaded

4.1.2 Deep Rolling The other specimens were rolled with the help of a device typ HG 6-9 from Ecoroll. It has a hardmetal ball of 6 mm diameter, which was used to induce the residual stresses. The pressure was p = 300 bar or p = 380 bar. The distance between the different rolling tracks was ,x = 0.02 mm (only 300 bar); ,x = 0.15 mm and ,x = 0.25 mm. The rolling area was 55 mm × 55 mm and was rolled in the way of a meander. 4.2

Measurement of the Stresses

With the help of an X-ray diffractometer (type Rigaku Strainflex MSF-2M) the stresses were determined. It was used as an 9-goniometer with Cr-K=-radiation. In this case the distance bet-

443

force

specimen

140 320 perpendicular

specimen

along

point of measuring and directions

Figure 1: Mounting device and points of measuring for the shot peened specimens

Residual Stress Profil 0

0,1

0,2

0,3

0,4

0,5

0 -200 (N/mm²)

-400 -600

preload

-800

0 N/mm² 600 N/mm² 1200 N/mm² 1600 N/mm²

-1000 -1200 -1400 Depth (mm) Figure 2: Residual stress profiles of the different shot peened specimens

ween the [h,k,l]-layers [2,1,1] is measured. The diameter of the X-ray spot on the surface was 7–8 mm. The determination was done with the help of the sin2O -23 -method [17-19].

5

Results

5.1

Stress Peening

In Figure 2, the residual stress profiles up to a depth of 0.5 mm are shown in dependence of the (pre)load during peening. You obtain the typical distribution, achieved by Hertz pressure. The

444 maximum of the compressive residual stress is at a depth of around 0.15 mm. Without any preload, the residual compressive stress at the surface has a value of I = 610 N/mm², which increases to a maximum of nearly I = 700 N/mm² below the surface. Up to a depth of 0.2 mm, the stress peened samples show compressive residual stress profiles with stress values around 400 N/mm² higher than that of the not preloaded specimen. At a depth of 0.45 mm, the amount of residual stress is due to the plastification of the samples before peening. 5.2

Deep Rolling

The results of the residual stress profiles are shown in figure 3. There is a great difference between the induced residual stresses along the rolling track and perpendicular to the rolling track, which equalize at a depth of 0.4 mm. The compressive residual stress has disappeared at a depth 200

200

-1400 0

0 0

200

400

600

800

1 000

0

1 200

200

400

600

800

1 000

- 200

Figure 2. Residual stress pr

- 400

- 600

ferent shot peened specimens

- 400

- 600

5.2. Deep Rolling

- 800

- 800

P = 380 bar ,x = 0.25 mm

- 1 000

P = 380 bar ,x = 0.15 mm

- 1 000

- 1 200

- 1 200

- 1 400

- 1 400

200

200

0

0 0

200

400

600

800

1 000

0

1 200

- 200

- 200

- 400

- 400

- 600

- 600

200

400

600

800

1 000

- 800

- 800

P = 300 bar ,x = 0.15 mm

- 1 000

P = 300 bar ,x = 0.25 mm

- 1 000

- 1 200

- 1 200

- 1 400

- 1 400

200

depth (μm)

0 0

200

400

600

800

1 000

1 200

- 200

- 400

perpendicular - 600

- 800

- 1 000

1 200

- 200

P = 300 bar ,x = 0.02 mm

residual stress (N/mm²)

- 1 200

- 1 400

Figure 3: Different residual stress profiles induced by deep rolling

along (the track)

1 200

445 of around 0.75 mm. The higher pressure shows a slightly higher compressive residual stress and the nearer the tracks are the less compressive residual stress is induced along the rolling track. The residual stress along the rolling track has the typical profile of Hertz pressure, whereas perpendicular to the rolling track, the maximum of the compressive residual stress is at the surface. 5.3

Comparison between Shot Peening and Rolling

The residual stress distribution at the surface shows a similar shape between stress peening and deep rolling. Along the main loading stress during peening receptively perpendicular to the rolling track the highest compressive residual stresses are achieved (see figure 4). For deep rolling, the compressive residual stress perpendicular to the track is slightly higher than the compressive residual stress achieved by stress peening along the bending direction. In the two corresponding other directions the compressive residual stress is higher.

residual stress -1000

distribution perpendicular

along (N/mm²)

-500

-500

Stress Peening Deep Rolling

-1000 (N/mm²)

Figure 4: Comparison of the residual stress distributions at the surface due to deep rolling and stress peening

6

Conclusions

Deep rolling and stress peening give similar residual stress distributions at the surface. Deep rolling gives a more asymmetrical distribution. The compressive residual stresses after deep rolling are higher and deeper than after stress peening. However, under economical aspects, deep rolling is more expensive.

446

7 [1

[2]

[3] [4] [5] [6] [7] [8] [9]

[10]

[11]

[12] [13] [14] [15] [16] [17] [18] [19]

References E. Müller, L. Bonus: Kugelstrahlen warmgeformter Federn, Konferenzband der internationalen Konferenz zum Thema Federntechnologie in Düsseldorf 1990, European Spring Federation, Cambridge, 1990. E. Müller: Der Einfluß des Plastizierens und des Kugelstrahlens auf die Ausbildung von Eigenspannungen in Blattfedern, Hoesch Berichte aus Forschung und Entwicklung unserer Gesellschaften, Heft 1/92, S. 23 ff. L. Bonus: Auswirkung des Spannungsstrahlens auf die Eigenspannungen von hochvergüteten Bremsspeicher- und Torsionsfedern, Dissertation, TH Aachen, 1994. E. Müller: Spannungsstrahlen von Schraubenfedern, Draht 44, 1993, 1/2, S. 49 ff. E. Müller: Some aspects of stress peening of coil springs for vehicle suspensions, Proceedings of the 5th Int. Conf. on Shot Peening, Coventry University (UK), 1993, p. 341. E. Müller: Die Ausbildung von Eigenspannungen an Torsionsproben beim Spannungsstrahlen, Mat.-wiss. u. Werkstofftech. 27, 1996, S. 354 ff. E. Müller: Eigenspannungsabbau an spannungsgestrahlten Torsionsproben unter dynamischer Belastung, Mat.-wiss. u. Werkstofftech. 28, 1997, S.549 ff. E. Müller, L. Bonus: Lebensdauer spannungsgestrahlter Schraubenfedern unter Korrosion, Draht 48, 1997, 6, S. 30 ff. E. Müller: Die Ausbildung von Eigenspannungen an Flachproben unter dem Einfluß von Spannungstrahlen und Plastizieren, VDI Berichte 1463, S. 419 ff., VDI-Verlag, Düsseldorf, 1999. E. Müller: The Development of Residual Stress at Bending Specimens under the Influence of Setting and Stress Peening, Proceedings of the 7th International Conference on Shot Peening, Warsaw 1999, p. 88 ff. R. Zeller: Verbesserung der Ermüdungseigenschaften von Bauteilen aus Stahl durch optimales Kugelstrahlen unter Zugvorspannung, in DVM-Band Betriebsfestigkeit “Moderne Fertigungstechnologien”, DVM, Berlin, 1991, S. 93 ff. R. Zeller: Influence of stress peening on residual stresses and fatigue limit, in Residual Stresses, DGM-Verlag, Oberursel, 1993, S 907 ff. G. Bernstein, B. Fuchsbauer: Festwalzen und Schwingfestigkeit, Z. f. Werkstofftechnik, 1982, S. 103 ff. E. v. Funkelstein, U. Preckel: Eigenspannungen beim Oberflächen-Feinwalzen, Stahl und Eisen 104, 1984, Nr. 14, S. 657 ff. B. Kaiser: Schwingfestigkeitsteigerung durch Randschichtverfestigungsverfahren, VDIBerichte “Werkstofftechnik”, Düsseldorf, 1984, S. 26 ff. K. H. Kloos, J. Adelmann: Schwingfestigkeitsteigerung durch Festwalzen, Mat.-wiss. u. Werkstofftechnik 19, 1988, S. 15 ff. SAE (Ed.): Residual Stresses, Measurements by X-Ray Diffraction, SAE J 784a, Warrendale (USA), 1971. I. C. Noyan, J. B. Cohen: Residual Stresses, Springer Verlag, Stuttgart, 1987. H. D. Tietze: Grundlagen der Eigenspannungen, Springer Verlag, Wien, 1982.

447

Isothermal Fatigue Behavior and Residual Stress States of Mechanically Surface Treated Ti-6Al-4V: Laser Shock Peening vs. Deep Rolling Ulf Noster1, Igor Altenberger2, Robert O. Ritchie2, Berthold Scholtes1 1 2

Institute of Materials Technology, University Kassel, Kassel, Germany Department of Materials Science and Engineering, University of California, Berkeley, CA, USA

1

Abstract

In this paper, the high-temperature fatigue behavior and residual stress states of a Ti-6Al-4V alloy are investigated after mechanical surface treatment. In particular, the two surface treatments investigated, laser shock peening and deep rolling, were observed to result in significantly different residual stress states. The consequent isothermal fatigue behavior at elevated temperatures, characterized using cyclic deformation curves, are discussed in terms of the stability of the nearsurface work hardening and compressive residual stresses. Despite pronounced relaxation of the residual stresses, both laser shock peening and deep rolling led to a significant improvement in the cyclic deformation behavior and, hence, increased fatigue lifetimes at elevated temperatures as compared to untreated materials states.

2

Introduction

In addition to the commonly used shot peening, laser shock peening and deep rolling are two mechanical surface treatments that are increasingly being used with the objective of enhancing material strength. These treatments can substantially increase the resistance to wear and stress corrosion, as well as enhance fatigue strength. Near surface compressive residual stresses and cold work are predominantly responsible for these effects, and can be related to a delay in surface crack initiation and also a retardation of the near-surface crack growth (e.g., [1–3]). Deep rolling (DR) is typically used for components that are rotationally symmetric (e.g., shafts) and is especially useful in overcoming detrimental notch effects. Besides the abovementioned nearsurface compressive residuals stresses and cold work, in most cases, DR also reduces surface roughness [4]. Laser shock peening (LSP), however, is not limited by component geometry. The treatment is capable of introducing deep compressive residual stress and cold work cases as a consequence of the elastic-plastic shockwave induced in the surface layers during such processing. Furthermore, surface roughness was not markedly influenced by LSP [5]. It is well known that compressive residual stresses have a tendency to anneal out at elevated temperatures. Consequently, an important question arises regarding the efficiency of such mechanical surface treatments for lifetime improvement at high temperatures. For the present investigation, a a+b titanium alloy Ti-6Al-4V was chosen, which is commonly used as turbine blade and disk material in aircraft engines at service temperatures of approx. 300 °C or even higher [6, 7]. To assess the influence of mechanical surface treatments on fatigue lifetimes at such tem-

448 peratures, the stability of the near-surface properties as well as the cyclic deformation behavior was investigated for deep rolled and laser shock peened specimens of this alloy.

3

Materials and Experimental Procedures

Cylindrical specimens with a diameter of 7 mm and a gage length of 15 mm of the a+b titanium alloy Ti-6Al-4V (tensile properties at room temperature: Rp0.2 = 915 MPa, UTS = 965 MPa) were used. After machining, the specimens were stress relieved in vacuo (700 °C, 120 min.). For deep rolling (DR), a spherical rolling element (6.6 mm dia.) was used with a feed of 0.1125 mm per revolution and a rolling pressure of 150 bar. Laser shock peening (LSP) was carried out with a power density of 7 GW/cm2 and a coverage of 200 %. Tension-compression fatigue tests were performed at room temperature and at 450°C (0.4 of melting point) with a standard servohydraulical testing machine under stress control at a load ratio (minimum load/ maximum load), R of –1 (mean stress = 0) and a cyclic test frequency of 5 Hz. For the high temperature tests, specimens were heated with a controlled light heater system. During fatigue, strain was monitored using a capacitative extensometer. Residual stress distributions were measured applying standard X-ray diffraction techniques. Lattice strain measurements were carried out using Cr-K= radiation at the {201}-planes of the hexagonal a-phase. For residual stress evaluation, the well-known sin2O-method was applied and the elastic constant 1/2 s2 = 12.09·10-6 mm2/N was used. Residual stress depth profiles were determined (without correction for stress relief) by successive electrochemical material removal. Near-surface work hardening was characterized by hardness measurements as well as by the full width at half maximum (FWHM) distributions of the X-ray interference lines.

4

Results and Discussion

While laser shock peening (LSP) did not result in any significant alteration of the surface topography, deep rolling (DR) as expected, led to a marked decrease in the roughness (untreated: Rz = 1.7 mm, deep rolled: Rz = 0.8 mm). Hardness of the untreated specimens was about 330 HV 0.1 and both mechanical surface treatments investigated led to an increase in the near-surface hardness of about 10 %. Furthermore, distinct deformation of the grains directly at the specimen surface was observed in the case of DR, whereas in the case of LSP, no such differences were detected. Figure 1 shows an optical micrograph of this deformed near-surface microstructure after DR. From the X-ray diffraction measurements, it was observed that both treatments introduced compressive residual stresses at the surface and in the near-surface regions. Figure 2 shows typical residual stress and FWHM depth distributions after DR and LSP. After DR, maximum compressive residual stress levels of 930 MPa were obtained immediately below the surface and the residual stresses decreased to zero at a depth of 0.5 mm. For LSP, the average compressive residual stress levels were much lower, but the surface layer with residual stresses was considerably thicker. In fact, no significant reduction of the residual stress was detected up to a depth of 0.5 mm. Similar deep residual stress distributions after LSP (sample thickness dependent), have been reported for the same material in [8]. However, the residual stress levels as well as the increase of FWHM are much less pronounced after LSP than after DR. As a consequence of dif-

449

Figure 1: Typical optical micrograph of the near surface microstructure after deep rolling

ferent deformation velocities between both treatments different microstructures are expected too. It should be noted that the observed optical micrographs as well as the residual stress distribution and FWHM states after DR correspond qualitatively to those found in [9] after shot peening of Ti-6Al-4V. Furthermore, in concurrence with the hardness increase after mechanical surface treatment, FWHM distributions indicate the presence of a work hardened layer at the surface of the material.

Figure 2: Residual stress profiles and full with at half maximum (FWHM) values after different mechanical surface treatments

According to [10,11] plastic strain amplitudes are a measure of damage during fatigue of ductile materials. In Fig. 3, such amplitudes, measured during fatigue cycling at a temperature of 450 °C, are given as a function of the number of cycles. Compared to the untreated material state, mechanical surface treatments lower the plastic strain amplitude through the majority of lifetime, which in turn leads to higher lifetimes. Higher near-surface levels of compressive residual stress and FWHM as well as low surface roughness in the case of DR seem to be responsible for the slightly higher lifetimes after DR than after LSP. All material states investigated exhibited an initial increase (cyclic softening) and a subsequent decrease (secondary cyclic hardening) in the plastic strain amplitude. In the case of the untreated material, prior to cyclic softening, the material showed a quasi-elastic incubation phase. The absence of this incubation phase

450 for the mechanically surface treated materials (i.e., the material cyclically softens from the first cycle), is most likely caused by the higher amount of (dislocation density induced) work hardening [12]. The higher work hardening observed after DR, which depends besides the mechanical surface treatment method also on the chosen process parameters, could be the prime reason for the smaller amount of plastic strain amplitude and the higher lifetime compared to the LSP state.

Figure 3: Cyclic deformation curves for the different surface treatment states investigated (stress amplitude 460 MPa, temperature 450 °C)

Fatigue lifetimes for the untreated and DR specimens fatigued at room temperature (25 °C) and elevated temperature (450 °C) are given in Fig. 4. It is clearly evident that besides the higher lifetimes at room temperature as compared to 450 °C, DR leads to lifetime enhancements at room temperature as well as at the higher temperature. The seen enhancements are mostly pronounced in the high cycle fatigue range. This observed improvement in fatigue lifetime due to mechanical surface treatment depends mainly on the stability of the induced residual stresses and the work hardening in the near-surface regions [13]. Already after annealing (40 min, 450 °C), for both DR and LSP material states, the residual stress were markedly reduced (Fig. 5). In Fig. 6, residual stress and FWHM depth distributions for the surface treated and fatigued (450 °C, half the number of cycles to failure) materials states are given. It is evident that cyclic loading (Fig. 6) led to additional residual stress relaxation as compared to the only heat treated states (Fig. 5). For both the surface treatments in this study, residual stress were reduced to nearly the same level. This corresponds with the conception that residual stresses relax if the amount of residual stress and load stress exceeds the material yield strength [14]. In both cases, increased FWHM values immediately at the surface indicate that small amounts of work hardening still exist which seems to be responsible for the observed alteration of cyclic deformation behavior and lifetime enhancement. Obviously, mechanical surface treatments can be a useful tool to enhance high temperature fatigue behavior and lifetime. But this finding must be assessed more carefully, because other

451

Figure 4: Lifetimes of the untreated and the deep rolled material states at room temperature and at 450 °C

Figure 5: Residual stress profiles and FWHM values for DR and LSP surface treatment after heat treatment (40 min, 450 °C)

452

Figure 6: Residual stress profiles and FWHM values for DR and LSP surface treatments after high temperature fatigue up to half the number of cycles to failure (stress amplitude 460 MPa, temperature 450 °C)

materials, e.g. the magnesium alloy AZ31 [15], show different effects at comparable test to melting temperature ratios.

5

Conclusions

Mechanical surface treatments (deep rolling and laser shock peening) of the titanium alloy Ti6Al-4V led to alterations of the near-surface microstructure and residual stress states which can markedly affect the subsequent fatigue lifetime. For the conditions investigated, laser shock peening induced lower near-surface compressive residual stresses and work hardening than deep rolling, but led to thicker cases of the altered material. Thermal exposure or fatigue at higher temperature (450 °C) caused a marked reduction of such residual stresses. However, despite this relaxation, the fatigue behavior was found to be still superior to that of the untreated material. This is believed to be associated with the more stable near-surface work hardening. This implies that, even at elevated service temperatures, mechanical surface treatments of Ti-6Al-4V can be used to improve fatigue properties.

6

Acknowledgements

This work was supported in part by the U.S. Air Force Office of Scientific Research under Grant No. F49620-96-1-0478 under the auspices of the Multidisciplinary University Research Initiative on High Cycle Fatigue to the University of California at Berkeley. Thanks are also due to Metal Improvement Company, Livermore, CA, for performing the LSP treatment and to DFG (Deutsche Forschungsgemeinschaft) for financial support of Dr. I. Altenberger.

453

7 [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15]

References Advances in Surface Treatments (ed. A. Niku-Lari), Pergamon Press Oxford (1987). Short Fatigue Cracks (ed. K. J. Miller, E. R. de los Rios), Mechanical Engineering Publications Limited, London (1992). B. Scholtes: Structural and Residual Stress Analysis by Nondestructive Methods (ed. V. Hauk), Elsevier, Amsterdam (1997), p. 590. Mechanische Oberflächenbehandlungen (ed. H. Wohlfahrt, P. Krull), Willey-VCH, Weinheim (2000) B. P. Fairand, A.H. Clauer, R.G. Jung, B.A. Wilcox, Appl. Phys. Letters 25 (1974), p. 431. D. Helm: Fatigue Behavior of Titanium Alloys (ed. R.R. Boyer, D. Eylon, G. Lütjering), The Minerals, Metals & Materials Society, Warrendale, PA (1999), p. 291. R. R. Boyer: Mater., Sci. Eng. A213 (1996), p. 103. M. J. Shepard, P.R. Smith, P.S. Prevey, A.H. Clauer: Proc. 4th National HCF Conference, USA (1999). L. Wagner: Fatigue Behavior of Titanium Alloys (ed. R. R. Boyer, D. Eylon, G. Lütjering), Minerals, Metals & Materials Society, Warrendale, PA (1999), p. 253. S. S. Manson: NACA TN-2993 (1953). L. F. Coffin: Trans. ASME 76 (1954), p. 931. H. Hanagarth: Dr.-Ing. Thesis Universität Karlsruhe (1989). I. Altenberger, J. Gibmeier, R. Herzog, U. Noster, B. Scholtes: Materials Science Research Int. - Special Technical Publication, Vol. 1 (2001), p. 275. V. Schulze, K.-H. Lang, O. Vöhringer, E. Macherauch: Proc. 6th Int. Conf. Shot Peening (ed. J. Champaigne), p. 403. U. Noster, I. Altenberger, B. Scholtes: Proc. of Magnesium Alloys and their Applications (ed. K. U. Kainer), Willey-VCH, Weinheim (2000), p. 274.

454

Influence of Shot Peening and Deep Rolling on High Temperature Fatigue of the Ni-Superalloy Udimet 720 LI Janny Lindemann1), Kim Grossmann2), Thomas Raczek1) and Lothar Wagner1) 1) Chair of Physical Metallurgy and Materials Technology, Technical University of Brandenburg at Cottbus, Cottbus, Germany 2) Rolls-Royce Deutschland, Dahlewitz, Germany

1

Abstract

The Ni-superalloy Udimet 720 LI is typically used for application as turbine disc material for gas turbines. The present investigation was undertaken to determine if the elevated temperature fatigue strength can be improved by mechanical surface treatments. Shot peening and deep rolling were performed on circumferentially notched (kt = 2.3 and 3.5) specimens. Axial fatigue tests were performed at T = 650 °C using a stress ratio of R = 0.1. The fatigue performance will be compared with electrolytically polished references.

2

Introduction

Udimet 720 LI is a nickel-based superalloy with reduced chromium, carbon and boron content as compared to the original Udimet 720. Typical property requirements are good LCF and HCF strengths at elevated temperatures, creep strength and damage tolerance. Considerable work has been undertaken to optimize these properties by thermal and thermomechanical treatments resulting in microstructural modifications regarding, e. g., grain size, g¢ size and distribution, carbide- and boride-phase content, and grain boundary morphology [1–3]. The present investigation was performed to determine if the elevated temperature notch fatigue performance of Udimet 720 LI can be improved by mechanical surface treatments such as shot peening and deep rolling.

3

Experimental

Cylindrical blanks (Æ 12 × 60 mm) were prepared by wire-guided electrical discharge machining from a fully heat treated turbine disc having a homogeneous fine grained microstructure as shown in Figure 1. Tensile specimens were machined from the blanks with gage diameters and lengths of 4 and 20 mm, respectively. Tensile tests were performed at ambient and elevated (650 °C) temperatures. The initial strain rate was 8.3 × 10–4 s–1. Tensile results are given in Table 1.

455 Table 1: Tensile properties of Udimet 720 LI test temperature

E [GPa]

I0.2 [MPa]

UTS [MPa]

El [%]

AF = ln A0/ AF

RT

220

1120

1535

14

0.13

650 °C

180

1050

1280

32

0.33

Figure 1: Microstructure of Udimet 720 LI

~ 0.8

R 0.43

60º

Figure 2: Geometry of the notched fatigue specimens

Fatigue specimens were machined having a circumferential 60° V-notch as illustrated in Figure 2. The notch root radius was either 0.43 or 0.30 mm. Part of the specimens was electrolytically polished (EP) to serve as a reference. Others were shot peened. Shot peening was performed using an injector type machine and spherically conditioned cut wire (SCCW 14) with an average shot size of 0.36 mm. Specimens were peened to an Almen intensity of 0.24 mmA. Furthermore, some specimens were deep rolled using a hydraulically driven three-roll device operating in a lathe. Deep rolling (DR) was performed on specimens having a notch root radius of 0.43 mm. 55° rolls with a tip radius of 0.3 mm were used. Thus, during deep rolling, the notch root radius of the specimens was reduced from 0.43 to roughly 0.3 mm. After these mechanical surface treatments, the surface layer properties were characterized by measurements of surface roughness through profilometry and X-ray measurements of half width breadths and residual macrostresses. Measurements were taken from the (311) planes. Surface layers were subsequently removed by electropolishing to enable the determination of depth profiles. Fatigue tests were performed in axial loading at R = 0.1 using a servohydraulic testing machine. Tests were done at 650 °C at a frequency of about 60 Hz. Specimens were heated by means of a 3-zone electric resistance furnace. Fracture surfaces were studied by SEM.

3

Results and Discussion

The microstructure of the fully heat treated Udimet 720 LI material is shown in Figure 1. The average C grain size is about 15 mm. Typical surface layer properties caused by shot peening are illustrated in Figure 3. In addition, measurements are shown for material shot peened and an-

456 2.0

SP SP + 50h 650ºC

700 600 500

SP SP + 50h 650ºC

1.5 1.0 0.5

(311)Ni

0.0

400

0

100

200

300

Distance from surface, z [mm]

Residual stress, sR [MPa]

Half width breadth [º]

Microhardness [HV 0.1]

800

0

0

100

200

300

Distance from surface, z [mm]

(311)Ni

-400

-800

SP SP + 50h 650ºC

-1200

0

100

200

300

Distance from surface, z [mm]

Figure 3: Shot peening induced surface layer properties (SP: SCCW 14, 0.24 mmA)

nealed at 650 °C for 50h. This annealing treatment corresponds to the heat cycle that 10 7 cycles run-out specimens in the elevated temperature (650 °C) tests will see. After shot peening to an Almen intensity of 0.24 mmA, the microhardness (Fig. 3a) drastically increases from around 500 in the bulk to values above 750 HV 0.1 at the surface owing to the marked work-hardening capacity of the material at ambient temperature (Table 1). Interestingly, this shot peening-induced microhardness profile is not changed by the annealing at 650 °C for 50 h (Fig. 3a). Obviously, the high dislocation density being responsible for this strengthening effect is stable at 650 °C. On the other hand, the shot peening-induced half width breadth which depth profile is very similar to that of the microhardness is somewhat more affected by the anneal (Fig. 3b). It is argued that the contribution of residual microstresses to the interference line broadening may be responsible for this difference in thermal stability between microhardness and half width breadth (compare Figs. 3b with 3a). Residual microstresses may partially relax without significant losses in dislocation density. Apparently, residual macrostresses are even more affected by the anneal (Fig. 3c). Particularly, near surface residual compressive stresses are markedly reduced presumably, owing to free surface effects while stresses at depths greater than 100 mm are hardly affected. The S-N curves at 650 °C are shown in Figure 4 comparing shot peened with electrolytically polished conditions. The 107 cycles fatigue strength is increased by shot peening from 690 to 780 MPa while no fatigue life improvement is seen in the finite life regime (Fig. 4). In order to determine if this poor fatigue performance of shot peened specimens tested at high maximum stresses where residual stresses will cyclically relax is related to roughness-induced early microcrack growth, the shot peened notch root was mechanically polished to remove all pee-

457 ning-induced microcracks, dents and overlaps. As seen in Figure 5, the roughness in the as-peened condition (SP) was markedly reduced by polishing (SP + MP) and almost as low as in the electropolished reference (EP). However, only slight life improvements in the high stress regime were found after polishing shot peened specimens (Fig. 6). Presumably, crack nucleation is quite early at these high maximum stresses irrespective of near-surface high dislocation densities at smooth surfaces.

Maximum stress, Imax [MPa]

1100 EP SP

1000 900 800 700 600 kt = 2.3 500 104

105

106

107

Cycles to failure, NF Figure 4: S-N curves of notched specimens (R = 0.1, kt = 2.3, T = 650 °C, f = 60 s–1, air)

Roughness values [mm]

Ra Rz Ry

8

6 4

2

Maximum stress, smax [MPa]

1000

10

SP + MP

900 SP

800

700

0 SP

SP + MP

EP

104

105 Cycles to failure, NF

Figure 5: Surface roughness values of the notch root after various mechanical surface treatments

Figure 6: Effect of polishing on LCF-fatigue performance of shot peened specimens (R = 0.1, kt = 2.3, T = 650 °C, f = 60 s–1, air)

458 As opposed to shot peening, much greater depths of plastic deformation can be induced by deep rolling (Fig. 7). However, microhardness values close to the surface of deep rolled specimens are lower than after shot peening. Presumably, the size of the shot being much smaller than the size of the rolls leads to higher near-surface deformation degrees. As seen in Figure 8, deep rolling can drastically improve the fatigue life of the electropolished reference even at very high nominal maximum stresses of Imax = 800 MPa where the life improvement due to shot peening is rather small.

Microhardness [HV 0.1]

800 250 N 500 N 1000 N 1500 N 2000 N

700

600

500 SP 400 0

500

1000

1500

2000

Distance from surface, z [mm] Figure 7: Microhardness profiles after deep rolling with various rolling forces compared to shot peening (SP: 0.24 mmA)

While the process window is quite wide with regard to suitable rolling forces which lead to life improvements of more than three orders of magnitude (Fig. 8), some loss in fatigue life was observed at the highest rolling force used. Metallographic observation of the deep rolled notch roots showed process-induced microcracks (Fig. 9), which could easily act as crack starters during subsequent fatigue testing. As often observed on fatigue fracture surfaces of deep rolled specimens, the stage of early crack growth in residual compressive stress fields where cracks may significantly be hindered to propagate can easily be identified by a shiny zone [4]. It is argued that the width of this zone corresponds to the crack depth where the local stress intensity range as controlled by applied and residual stresses had overcome the threshold value DKth (Fig. 10).

2000 N

106 105

1500 N

500 N

107 250 N

Cycles to failure, NF

108

1000 N

459

104 103 EP

DR

SP

Figure 8: Fatigue life (R = 0.1, Imax = 800 MPa, kt = 3.5, T = 650 °C, f = 60 s–1, air) of the various conditions

Figure 9: Process-induced microcracks in the notch root after deep rolling with a rolling force of 2000 N

4

Figure 10: Typical HT-fatigue fracture surface after deep rolling with an optimum rolling force (F = 500 N, R = 0.1, kt = 3.5, T = 650 °C, air, f = 60 s–1, Imax = 800 MPa, NF = 6.65 × 106)

Acknowledgements

Financial support by the German Federal Ministry for Education and Research (03N3067B/8) is gratefully acknowledged.

460

5 [1] [2] [3]

[4]

References H. Puschnik, G. Zeiler, J. Fladischer, W. Esser, K. H. Keienburg, “Udimet 720 Turbine Blades – Production and Properties”, La metallurgia italiana 82 (1990) 473. D. Furrer, H. Fecht, “Ni-Based Superalloys for Turbine Discs”, JOM (1999) 14. G. Baumeister, J. Albrecht, G. Lütjering, “Influence of Microstructure on the Mechanical Properties of Udimet 720 LI”, Microstructure and Mechanical Properties of Metallic High-Temperature Materials (Eds.: H. Mughrabi, G. Gottstein, H. Mecking, H. Riedel and J. Tobolski) Wiley-VCH, 454. A. Drechsler, T. Dörr, L. Wagner, “Mechanical Surface Treatments on Ti-10V-2Fe-3Al for Improved Fatigue Resistance”, Mat. Sci. Eng. A (1997), 217.

461

Shot Peening and Roller-Burnishing to Improve Fatigue ) Titanium Alloy Ti-6Al-4V Resistance of the ( Marcin Kocan1, Alfred Ostertag2 and Lothar Wagner1 1 Chair of Physical Metallurgy and Materials Technology, Technical University of Brandenburg at Cottbus, Cottbus, Germany 2 Ecoroll AG, Celle, Germany

1

Introduction

It has long been recognized that mechanical surface treatments such as shot peening or roller-burnishing can significantly increase the fatigue performance of structural components. Regarding the application of light-weight alloys, it is known that titanium and magnesium alloys as opposed to aluminum alloys can respond quite critically to a shot peening treatment. For example, a very marked over-peening effect was observed on the high-strength magnesium alloy AZ80 [1, 2], i.e., the fatigue life as a function of Almen intensity first dramatically increased compared to an electropolished reference followed by a drastic drop as the intensity increased. This sensitivity was attributed to the limited deformability by slip of the hexagonal magnesium crystal structure. The response of titanium alloys to shot peening is reported to strongly depend on many factors, as alloy class (a, (a+b) and metastable b) and its cyclic deformation behavior which in turn determines the cyclic stability of the process-induced residual compressive stresses. For example, metastable b alloys exhibited only slight improvements of the fatigue performance while a alloys responded much more beneficially. Further, previous work [3] has shown that the response of the (a+b) titanium alloy Ti-6Al-7Nb to shot peening and roller-burnishing was clearly related to the mean stress sensitivity of the particular microstructure and crystallographic texture. Conditions with an anomalous mean stress sensitivity [4, 5] showed little improvement in fatigue performance as opposed to conditions with a normal mean stress sensitivity. The present investigation was performed on the well known (a+b) titanium alloy Ti-6Al-4V having a typical commercially available mill annealed microstructure. In order to establish optimum conditions with regard to fatigue performance, shot peening and roller-burnishing were performed using a wide variation in Almen intensity and rolling force, respectively. Additional polishing treatments were performed to reduce process-induced roughnesses and microcracks in order to find out if the fatigue behavior can be further improved.

2

Experimental

The (a+b) titanium alloy Ti-6Al-4V was received as Æ10 mm rod hot rolled below the beta transus temperature. Prior to rolling, the ingot had been b-forged. The crystallographic texture was determined by X-ray diffraction and will be presented as (0002) pole figure. Tensile and fatigue specimens were machined in rolling (RD) direction. Tensile tests were performed on

462 threaded cylindrical specimens with gage lengths and diameters of 20 mm and 4 mm, respectively. The initial strain rate was 8.3 × 10–4s–1. Tensile test results are listed in table 1. Table 1: Tensile properties of Ti-6Al-4V I0.2

UTS

El

RA

890 MPa

960 MPa

13 %

46 %

For fatigue testing, hourglass shaped specimens (3.8 mm gage diameter) were prepared. Specimens were turned (T) under well defined process conditions. The turning parameters are listed in table 2. Table 2: Turning parameters of the final passes in the as-machined (T) condition Tool bit

TiN/Al2O3/TiCN-CVD coated WC

Chisel radius

0.4 mm

Infeed

0.2 mm

Feed rate

0.15 mm/rev.

Spindle speed

2500 min–1

Cooling fluid

oil-water mixture (1:10)

One part of these specimens was electrolytically polished (EP) to serve as further reference. Roughly 100 mm were removed from the surface to ensure that any machining effect that could mask the results was absent. Others were shot peened (SP) by means of an injector type machine and a direct pressure blast system for low and high Almen intensities, respectively. Shot peening was performed using cast steel shot S 330 (0.8 mm average shot size). All peening was done to full coverage. Another part of the turned specimens was roller-burnished (RB) using a one-roll hydraulic system with 6 mm hardmetal ball operating in a conventional lathe. Spindle speeds of 36 min–1 (RB1) and 780 min–1 (RB2) were used. Rolling forces were varied in a wide range. In addition, mechanical polishing was performed on some shot peened (SP+MP) or rollerburnished specimens (RB+MP). This was done to reduce process-induced roughnesses and possible microcracks. For the various surface treated conditions, roughnesses were determined by profilometry and residual stress-depth profiles by the incremental hole drilling method as described in [6]. To study cyclic deformation behavior, stress controlled LCF tests were performed on threaded cylindrical specimens with gage lengths and diameters of 10 mm and 5 mm, respectively. Tests were done in fully reversed (R = –1) loading using a servohydraulic testing machine at a frequency of 0.05 Hz. Hysteresis loops were recorded by strain gage measurements. From the hysteresis loops, half of the plastic strain range at zero load (DApl/2) was taken and plotted versus number of cycles. HCF tests were performed on the various surface treated conditions in rotating beam loading (R = –1) at frequencies of about 60 Hz in ambient air.

463

3

Results and Discussion

The microstructure of the Ti-6Al-4V alloy is illustrated in figure 1. This mill annealed structure consists of fairly equiaxed a grains with an average size of about 10 mm and of roughly 20 % transformed b phase.

Figure 1: Microstructure of Ti-6Al-4V

Figure 2: (0002) pole figure of Ti-6Al-4V

The (0002) pole figure of the Ti-6Al-4V alloy is illustrated in figure 2. The mixed basal/ transversal (B/T) type of texture is typical for unidirectional rolling in the (a+b) phase [7]. The cyclic deformation behavior is illustrated in figure 3. Marked cyclic softening was observed at the various stress levels for most of the fatigue life. 10.0

,Apl /2 [o/oo]

Ia = 0.95 I0.2 Ia = 0.90

1.0

Ia =0.80

0.1 1

10

100

1000

Cycles number Figure 3: Cyclic deformation characteristics in Ti-6Al-4V (R = –1)

The surface roughness profiles of both the as-turned and electropolished references are shown in figure 4. Removing a surface layer of about 100 mm of the as-turned surface by electropolishing reduced the measured surface roughness from Ra = 1.7 to 0.2 mm and Ry = 9.4 to

464

Stress amplitude, sa [MPa]

800

700

T 600

EP 500 104

105

106

107

Cycles to failure, NF

Figure 4: Roughness profiles of reference conditions T and EP

Figure 5: S-N curves in rotating beam loading of reference conditions T and EP

1.4 mm. From earlier investigations [8], it is known that this surface layer removal of 100 mm is sufficient to remove also the turning-induced high dislocation densities and residual stresses. The S-N curves of these two reference conditions T and EP are illustrated in figure 5. Interestingly, the 107 cycles fatigue strength of condition T is roughly 80 MPa higher than that of condition EP indicating that the turning-induced high dislocation densities and residual stresses markedly overcompensate the detrimental influence of high surface roughness. Similar results were reported in earlier work on Ti-6Al-4V [9]. It is obvious that any assessment of possible improvements in fatigue performance caused by shot peening or roller-burnishing will highly depend on the reference condition taken for comparison.

30

Roughness, Ra, Ry [μm]

Roughness, Ra, Ry [μm]

30 25

Ry

20 15 10

Ra

5 0 T

0.1

0.2

0.3

0.4

Almen intensity [mmA]

Figure 6: Surface roughness vs. Almen intensity (SP)

0.5

0.6

25

20

15

10

Ry 5

Ra

0 T

200

400

600

800

1000

Rolling force, F [N]

Figure 7: Surface roughness vs. rolling force (RB1)

The influence of shot peening and roller-burnishing on the resulting surface roughness values are plotted in figures 6 and 7. Starting with condition T (fig. 6), the surface roughness first decreases by shot peening SP if low Almen intensities up to 0.22 mmA were applied followed by an increase in roughness at higher Almen intensities. It should be noted that slight polishing after even heavy peening (SP+MP) again resulted in very low roughnesses. Since only about 20 mm were removed from the as-peened surfaces, residual stress and dislocation density profi-

465 les were hardly affected. Similarly to the effect of Almen intensity (fig. 6), roughness values after roller-burnishing (RB) first decrease with rolling force (fig. 7), but then level off at low values at rolling forces higher than about 300 N. The effect of Almen intensity on the residual stress-depth profile is shown in figure 8. With an increase in Almen intensity from 0.12 to 0.48 mmA, the magnitude of the residual compressive stresses close to the surface and the penetration depth of the residual compressive stress field significantly increase. 107

Cycles to failure, NF

Residual stress, IR [MPa]

0 -200

0.12 mmA -400

0.48 mmA -600

-800

106

SP

105

104

-1000 0

100

200

300

400

T

0.1

0.2

0.3

0.4

0.5

0.6

Almen intensity [mmA]

Distance from surface, z [μm]

Figure 8: Residual stress-depth profiles after shot peening (SP)

Figure 9: Fatigue life (Ia = 700 MPa) of SP condition

107

Cycles to failure, NF

107

106

SP SP+MP

Cycles to failure, NF

The effects of Almen intensity and rolling force on the fatigue life at a constant stress amplitude of Ia = 700 MPa are illustrated in figures 9 and 10, respectively. Starting with the as-turned reference (figure 9), the fatigue life owing to shot peening increases by less than one order of magnitude and then levels off at intensities as low as 0.12 mmA, i.e., no over-peening effect was found. Obviously, the pronounced increases in surface roughness at higher Almen intensities (fig. 6) are counterbalanced by opposing effects of dislocation densities and residual compressive stresses (fig. 8). If slight polishing is done after heavy shot peening (SP+MP), the fatigue life dramatically increases as shown in figure 10. This indicates the importance of surface roughness on fatigue performance after shot peening.

105 0.34

0.45

0.48

Almen intensity [mmA]

Figure 10: Fatigue life (Ia = 700 MPa) of SP condition after mechanical polishing SP+MP

RB1 106

105

104 T

200

400

600

800

1000

Rolling force, F [N]

Figure 11: Fatigue life (Ia = 700 MPa) of RB1 condition

466 Not surprisingly, the fatigue life (Ia = 700 MPa) of roller-burnished specimens continuously increases with rolling force (fig. 10) since low roughnesses (fig. 7) are combined with increasing depths of high dislocation densities and residual compressive stresses (fig. 11). Similar polishing, as done on shot peened specimens did not improve fatigue life of roller-burnished specimens. From Figures 9–11, the optimum process parameters for SP and RB1 with regard to fatigue performance were taken and further testing was performed for establishing S-N curves. These results are summarized in figure 12. Effects of optimum surface treatments on improvement of 107 fatigue strengths are summarized in table 3. 900

800

SP+MP SP

700

600

EP

500 104

105

106

107

Stress amplitude, Ia [MPa]

Stress amplitude, Ia [MPa]

900

800

RB1 RB2 T EP

700

600

500 104

105

106

107

Cycles to failure, NF

Cycles to failure, NF

a) Effect of SP (0.34mmA) and SP+MP

a) Effect of RB1 (670 N) and RB2

Figure 12: S-N curves in rotating beam loading of the various surface treated conditions in Ti-6Al-4V

Table 3: Improvements of 107 fatigue strengths after optimum surface treatments Reference condition

SP

SP+MP

RB1

RB2

EP

» 15%

» 20%

» 30%

» 20%

T

» 5%

» 5%

» 15%

» 5%

Comparison of RB1 and RB2 conditions indicates that for Ti-6Al-4V, a lower deformation rate in roller burnishing (36 min–1 spindle speed) is superior to the higher deformation rate (780 min–1 spindle speed). No such effect of spindle speed was observed in parallel work on 42CrMo4 and 54SiCr6 [10]. More work is needed to understand why the response of Ti-6Al-4V to such a variation in deformation rate in roller-burnishing is different from that of steels.

4 [1] [2]

References M. Hilpert and L. Wagner in: Magnesium Alloys and Their Applications (Ed.: K. U. Kainer), Wiley-VCH, 2000, 463. M. Hilpert and L. Wagner in: Magnesium Alloys and Their Applications (Ed.: K. U. Kainer) Wiley-VCH, 2000, 525.

467 [3]

U. Holzwarth, J. Kiese and L. Wagner, Mechanical Properties of Implant Materials, DVM, 1998, 319. (in German) [4] J. Lindemann and L. Wagner, Materials Science and Engineering A, 1997, 1118. [5] A. Drechsler, T. Dörr and L. Wagner in: Light Materials for Transportation Systems (Eds.: N. J. Kim, C. S. Lee and D. Eylon), Center for Advanced Aerospace Materials, 2001, 793. [6] J. Lindemann, D. Roth-Fagaraseanu and L. Wagner in: Shot Peening (Ed.: L. Wagner), Wiley-VCH, 2002. (in press) [7] M. Peters, Dr.-Ing. dissertation, Ruhr-University of Bochum, 1980. [8] L. Wagner in: Surface Performance of Titanium Alloys (Eds.: J. K. Gregory, H. J. Rack, D. Eylon), TMS, 1997, 199. [9] W. Trojahn, Stud.-thesis, Ruhr-University of Bochum, 1980. [10] D. Wierzchowski, A. Ostertag and L. Wagner in: Shot Peening (Ed.: L. Wagner) Wiley-VCH, 2002. (in press)

468

Fatigue Performance of the Mechanically Surface Treated Steels 42CrMo4 and 54SiCr6: Shot Peening vs. Roller-Burnishing Dariusz Wierzchowski1) Alfred Ostertag2) Lothar Wagner1) 1) Chair of Physical Metallurgy and Materials Technology, Technical University of Brandenburg at Cottbus, Cottbus, Germany 2) Ecoroll AG, Celle, Germany

1

Introduction

Mechanical surface treatments such as shot peening (SP) and roller burnishing (RB) are commonly used in industrial applications to improve fatigue life and fatigue strength of cyclically loaded engineering components. These treatments lead to surface layer properties of the workpiece different from those in the bulk. For example, the yield stress in near-surface regions increases due to cold work and resulting high dislocation densities. Owing to the local plastic deformation, residual stresses are generated. In addition, the surface topography is changed. Depending on the surface treated material, other property changes can result from stress-induced martensitic transformations and/or modifications in near-surface crystallographic textures [1]. Work has shown that the fatigue performance of mechanically surface treated high-strength steels is mainly affected by near-surface residual compressive stresses which can largely suppress crack growth from the surface into the bulk of a component [2]. Increasing the strength level of the steels may increase the cyclic stability of process-induced residual compressive stresses and thus, their contribution to the improvement of the fatigue performance [3]. However, the stronger the steels the smaller is the strength differential between shot material and workpiece. As notch sensitivity of steels typically increases with an increase in tensile strength, greater contributions of surface roughness to fatigue crack nucleation resistance may result [4]. The goal of the present investigation was to determine possible strength effects on the improvement of the high cycle fatigue (HCF) performance of high-strength steels by shot peening. For comparison, the effect of roller-burnishing which leads to low roughness was also investigated.

2

Experimental

The investigation was performed on the structural steel 42CrMo4 and the spring steel 54SiCr6. The steels were delivered as Ø10 mm bar materials. Chemical compositions are given in table 1. While the spring steel was delivered in quenched and tempered condition, the structural steel was given an austenitizing treatment at 850 °C for 30 min followed by oil quenching. The material was tempered at 450 ºC for 2 hours followed by air cooling.

469 Table 1: Chemical composition (wt. %) of the tested steels

42CrMo4 54SiCr6

C

Si

Mn

P

S

Cr

Mo

Fe

0.41 0.55

0.34 1.44

0.79 0.70

0.022 0.007

0.002 0.006

1.18 0.70

0.16 –

balance balance

Tensile tests were performed on threaded cylindrical specimens having gage lengths and diameters of 20 and 4 mm, respectively. The initial strain rate was 8.3 × 10–4 s–1. Tensile properties are listed in table 2. Table 2: Tensile properties of the tested steels

42CrMo4 54SiCr6

I0.2 [MPa]

UTS [MPa]

I0.2/UTS

El [%]

RA [%]

HV10

–

1375 1865

1440 2055

0.96 0.91

10.0 9.9

56 58

455 600

–

For HCF testing, hour-glass shaped specimens with a minimum diameter of 3.8 mm were machined. Turning (T) was done using cubic boron nitride (CBN) tool bits (0.8 mm nose radius) under cutting fluid cooling in a CNC lathe operating at a rotational speed of 2500 rpm, a feed rate of 0.1 mm/rev and a depth of cut of 0.1 mm/pass. After machining, part of the specimens was electrolytically polished (EP) to serve as reference. About 120 ìm were removed from the surface to ensure that any machining effect that could mask the results was absent. Shot peening (SP) as well as roller-burnishing (RB) were performed on as-turned specimens. Shot peening to low Almen intensities (LSP) was carried out using spherically conditioned cut wire (SCCW) in an injector type system while for realizing high Almen intensities (HSP), rounded cut wire (RCW) was used in a pressure blast system. All peening treatments were done to full coverage. Shot properties are listed in table 3. As seen in the SEM pictures (fig. 1), the SCCW shot is almost perfectly spherical while the RCW shot is just rounded with some edges still present. Table 3: Properties of the shot material Shape

Ø mm

HV1

SCCW

Spherical

0.34

610

RCW

Rounded

1.00

640

Shot peening of both steels was done either at low intensity of 0.20 mmA (LSP) or at high intensity of 0.55 mmA (HSP). Some tests were done with specimens being first heavily peened followed by light peening (HLSP) [5]. Part of the HLSP specimens was mechanically polished (HLSP+MP) using fine grained SiC paper to decrease the shot peening-induced surface roughness. Roughly 20 and 25 ìm were removed from the as-peened surfaces of 42CrMo4 and 54SiCr6, respectively.

470

1 mm

a) SCCW

1 mm

b) RCW

Figure 1: Geometry of the shot material

Roller burnishing was performed in a conventional lathe using a one-roll hydrostatic system [6]. The diameter of the hardmetal ball was 6 mm. The rolling parameters were as follows: 0.2 mm/rev feed rate, 1 pass, 36 rpm rotational speed. For optimum roller-burnishing regarding fatigue life response, rolling forces of 220 N and 680 N were chosen for 42CrMo4 and 54SiCr6, respectively. Surface roughness was measured by a profilometer. Residual stresses were determined by the incremental hole drilling method as described elsewhere [7]. Fatigue tests were performed in rotating beam loading (R = –1) at 100 Hz.

3

Results and Discussion

Roughness profiles and values for the various conditions of both steels are given in figure 2. Lowest roughness values were measured for the conditions EP, HLSP+MP and RB (fig. 2). Much higher roughnesses were determined for the various shot peened conditions with comparatively low, intermediate and high values for LSP, HLSP and HSP, respectively. Comparing now roughness values between 42CrMo4 and 54SiCr6 for the same treatments, it is seen that roughness values are identical for EP, while most mechanical surface treatments lead to roughnesses being higher in 42CrMo4 than in 54SiCr6. Presumably, this is caused by the difference in yield stress of the steels resulting in more marked local plastic deformation in the lower strength 42CrMo4 (I 0.2 = 1375 MPa) compared to the higher strength 54SiCr6 (I 0.2 = 1865 MPa). Residual stresses as measured by the hole drilling method for the various surface treatments on 42CrMo4 are illustrated in figure 3. Only very small residual compressive stresses were found after turning (T) while much higher stresses were observed after the various shot peening treatments and after roller-burnishing (RB). Compared to light peening (LSP), heavy peening (HSP) led to somewhat lower residual compressive stresses in regions very close to the surface, whereas in deeper depths much higher stresses were observed [9]. As expected, heavy peening followed by light peening (HLSP) does not significantly change the residual stresses profile of the HSP condition. Residual stresses after roller-burnishing (RB) were similar to these shot peened conditions. However, stresses close to the surface were higher (fig. 3). While the magnitude of the induced residual stresses was so-

471 Surface treatment 42CrMo4 54SiCr6 42CrMo4 54SiCr6

EP

T

42CrMo4 54SiCr6

LSP

42CrMo4

Ra Rz Ry [mm] [mm] [mm] 0.3 1.0 2.0 0.3 1.2 2.0 1.3 7.6 11.0 0.8 3.8 5.8

Roughness profiles

15 mm/cm 45 mm/cm

1.8 10.2 13.4 1.2 7.4 8.2

7.1 34.4 45.2 HSP

54SiCr6

2.3 12.6 17.8 3.3 16.2 19.4

42CrMo4 HLSP 54SiCr6

42CrMo4 54SiCr6 42CrMo4 54SiCr6

HLSP + MP RB

1.5

8.6 10.8

0.3 0.3 0.6 0.5

1.6 1.3 2.6 2.2

3.1 2.3 4.0 3.2

Residual stresses, IR [MPa]

Figure 2: Surface roughness values and typical profiles for the various conditions

T

0

LSP

-200

RB -400 HSP -600 HLSP -800 42CrMo4 -1000 0.0

0.1

0.2

0.3

0.4

0.5

Distance from the surface, z [mm] Figure 3: Residual stresses for the various surface treatments in 42CrMo4

mewhat higher in 54SiCr6, the ranking among the various surface treatments was very similar to the results on 42CrMo4. The HCF results in terms of S-N curves for the various surface treatments are illustrated in figures 4–6 comparing results on 42CrMo4 (a) with those on 54SiCr6 (b).

472 1600 54SiCr6

42CrMo4

1100

1000

900 800 EP

700 T

600

Stress amplitude, Ia [MPa]

Stress amplitude, Ia [MPa]

1200

500

1500

1400

1300 T

1200

1100

EP

1000

900 104

105

106

107

104

Cycles to failure, NF

105

106

107

Cycles to failure, NF

a) 42CrMo4

b)54SiCr6

Figure 4: S-N curves in rotational beam loading, effect of turning (T) as opposed to electropolishing (EP)

As seen in figure 4, turning can decrease (fig. 4a) or increase (fig. 4b) the HCF strength of the electrolytically polished reference. Thus, often found statements in the literature about potential improvements of the HCF strength caused by shot peening or roller-burnishing (e.g., 20 %) are questionable if the reference (often denoted as “not peened”) is not well defined. The effects of the various shot peening treatments on the HCF strengths are shown in figure 5. Since shot peening was performed on as-turned specimens, this condition (T) is also shown for comparison. 1600 54SiCr6

42CrMo4

1100

1000

900 HLSP + MP

800

700

LSP T

600

HSP

500

Stress amplitude, Ia [MPa]

Stress amplitude, Ia [MPa]

1200

1500

1400

1300

HLSP + MP

1200

T

1100

HLSP LSP

1000

HSP

900 104

105

106

107

104

Cycles to failure, NF

a) 42CrMo4

105

106

107

Cycles to failure, NF

b) 54SiCr6

Figure 5: S-N curves in rotating beam loading, effect of various shot peening treatments

Heavy peening (HSP) decreases the HCF strengths of both 42CrMo4 (fig. 5a) as well as 54SiCr6 (fig.5b) while slight peening (LSP) gave improved results. Best fatigue performance was observed for the condition HLSP + MP particularly for 54SiCr6 (fig. 5b). Since the residual stress profiles of HSP and HLSP + MP hardly differ (fig. 3), the marked difference in HCF performance between these two conditions is mainly due to roughness effects. Fatigue cracks were nucleated at the surface for all shot peened specimens indicating that surface roughness was directly involved in the crack nucleation process [8].

473 The effect of roller-burnishing on HCF strengths is illustrated in figure 6. For both 42CrMo4 and 54SiCr6, roller-burnishing (RB) led to marked improvements of the fatigue performance of the as-turned (T) conditions. 1600 54SiCr6

42CrMo4

1100

1000

900 RB

800

700 T

600

Stress amplitude, Ia [MPa]

Stress amplitude, Ia [MPa]

1200

1500 RB

1400

1300

1200 T

1100

1000 900

500 104

105

106

107

104

106

107

Cycles to failure, NF

Cycles to failure, NF

a) 42CrMo4

105

b) 54SiCr6

Figure 6: S-N curves in rotating beam loading, effect of roller-burnishing

Comparing the results after roller-burnishing (fig. 6) with those after conventional shot peening (fig. 5), it is obvious that roller-burnishing is by far superior. Only if the shot peening-induced high surface roughness is reduced as is the case in HLSP + MP, similar increases of the HCF strengths as observed after roller-burnishing can be expected (compare fig. 6 with fig. 5).

4 [1] [2] [3] [4] [5] [6] [7] [8] [9]

References O. Vöhringer, Shot Peening (H. Wohlfahrt, R. Kopp and O. Vöhringer, eds.) DGM (1987) 185. H. Berns and L. Weber, Shot Peening (H. Wohlfahrt, R. Kopp and O. Vöhringer, eds.) DGM (1987) 647. H. Wohlfahrt, Shot Peening (H. Wohlfahrt, R. Kopp and O. Vöhringer, eds.) DGM (1987) 563. R. Schreiber, Dr.-Ing. dissertation, TH Karlsruhe (1976). S. Sato, K. Inoue and A. Ohno, Shot Peening (A. Niku Lari, ed.) Pergamon Press (1981) 303. Hydrostatic radial deep-rolling tools HG3-9E45º , Ecoroll AG (1996) J. Lindemann, D. Roth-Fagaraseanu and L. Wagner, Shot Peening ( L. Wagner, ed.), Wiley-VCH (2002), in press. H. Wieser and H. Zitter, Shot Peening (D. Kirk, ed.) Oxford University (1993) 191. N. Hu, X. Lin, J. Yao and K. Jin, Shot Peening (H. Wohlfahrt, R. Kopp and O. Vöhringer, eds.) DGM (1987) 541.

474

Process Control Techniques for Laser Peening of Metals Rob Specht, Fritz Harris, Laurie Lane Metal Improvement Co., Paramus, NJ, USA

Dean Jones Rolls-Royce plc, Bristol, UK

Lloyd Hackel, Tania Zaleski, John Halpin Lawrence Livermore National Laboratory (LLNL), Livermore, CA, USA

Mike Hill UC Davis, CA, USA

Wilfried Wübbenhorst Metal Improvement Co., Haan-Gruiten, Germany

1

Abstract

Laser Peening, also known as LasershotSM and Laser ShockTM Peening, is a surface treatment, which can induce compressive stresses in metals at depths exceeding 1 mm. This produces a more damage tolerant component, which resists fatigue and Stress Corrosion Cracking (SCC) failures better than components treated with conventional shot peening. A new Laser Peening facility was brought on line in early 2002 by MIC in Livermore, CA. This facility utilizes a Lawrence Livermore National Laboratory (LLNL) designed solid state laser employing Neodymium doped laser glass slabs and phase conjugation technology to enable high energy & high laser repetition rate combined with excellent beam quality. Laser Peening process parameters have been identified which will impact on the final depth of compressive stresses in a metal component. Process control techniques widely utilized in conventional shot peening have been adapted for use in Laser Peening with good success. As a result, Suppliers and Users of Laser Peening have a reliable method of monitoring the critical process parameters in a manner that will lead to consistency of production operation and repeatability of end results.

2

Introduction

Various types of surface treatments have been used by industry for many years to induce beneficial compressive stresses in metals. These include fillet rolling, cold expansion of holes and most commonly, shot peening. The significant improvements in resistance to fatigue, fretting fatigue and stress corrosion that result from imparting residual compressive stresses are well known. Shot peening has been the most widely used process because of its ability to induce these stresses efficiently and inexpensively on components of a complex geometry. The depth of the zone of compressive stresses produced by shot peening will vary depending on the metal, but is typically around 0.010″. Although depths of 0.030″ are possible, the surface finish may be unacceptably roughened at the high shot peening intensity levels required to achieve residual compressive stresses to this depth.

475

3

Laser Peening Science

Laser Peening creates shock waves at the metal surface, which drive compressive stresses into the metal to depths exceeding 0.040″, and with magnitudes at the surface comparable to those produced by conventional shot peening. These deeper residual stresses have proven to significantly enhance the damage tolerance of critical components subject to fatigue or SCC. Technical studies also demonstrate that thermal relaxation of the induced residual stresses from Laser Peening is much less than that compared to Shot or Gravity Peening. Initial studies on laser peening of materials were done at the Battelle Institute in Columbus, OH from about 1968 to 1981 [1, 2]. Figure 1 shows the basic sequence of events during laser peening. Laser energy densities (or ”fluence”) of 50–250 J/cm2 are utilized with a laser spot size of 2–5 mm and the energy is delivered within a time frame of 10–30 ns. (This translates to a total power output of 4–12 GW/cm2). Laser beam (60 to 200 J/cm2) 10 to 25 ns Impact diameter 2 to 5 mm

Water “Tamp” Confines pressure

Plasma

100% UTS

Laser light is rapidly absorbed and forms a 106 psi plasma. The tamping layer confines the plasma & drives the pressure pulse into the metal. Residual stress depth is 1 à 3 mm

Laser Light Absorption layer Maximum Surface Stress ~ 60% UTS

1 mm

Near-planar shockwave profile enhances penetration

Figure 1: Laser peening mechanism

A thin layer of light absorbing material (typically paint or tape) on the metal surface absorbs the initial laser burst and produces a plasma which is inertially confined by a 1-3 mm surface layer of water (known as an inertial “tamp”.) Shock pressures of up to 6.9 · 109 Pa are created on the metal surface; which first sends a planar shock wave into the metal and then ejects the water tamp off of the surface. The intense shock wave produces a strain rate in the metal surface that is well in excess of the spherical pressure pulse produced by shot peening. This enables the compressive stresses to be imparted deeper beneath the surface. Depending on the material properties of each metal (Young’s Modulus, Poisson’s Ratio, density, strength, etc.), the laser’s operating properties will be adjusted to deliver a surface pressure shock wave sufficient to cause the metal to plastically deform. Whereas a material such as 2024 T3 Aluminum might only require a laser fluence of 60 J/cm2 to peen the surface, a high strength steel might require a fluence in excess of 200 J/cm2 for proper laser peening. To achieve the full benefit from laser peening, it is generally necessary to make anywhere from 2–4 separate passes with the layer, including refreshing of the ablative/absorption layer.

476 After the initial laser peening pass, the surface is denser & more receptive to transmitting the compressive stresses to a greater depth. The additional laser peening passes increase the depth of the imparted residual stress, but generally do not increase the residual stress levels at the surface. As an example of the capabilities of the laser peening process, Figure 2 shows the residual stress induced in Inconel 718 by Laser Peening and contrasts it with typical results achieved by shot peening.

Inconel 718

Figure 2: Comparison of depth of residual compressive stresses from Laser Peening vs. Shot Peening

P. Prevey et al of Lambda Research compared the relaxation of compressive residual stress occurring at temperatures ranging from 230 °C to 425 °C for shot peened, gravity peened and laser peened Ti-6Al-4V and Inconel 718 [3]. For shot and gravity peening the repeated dimpling of the surface resulted in a highly cold worked layer. Conventional shot peening produces from 10 % to 50 % cold work. Gravity peening utilizes fewer impacts with larger shot producing a less cold worked surface layer. However, the laser peening process produced remarkably little cold working of the surface (1 % to 2 %), because the shock wave accompanying laser peening more evenly distributes the cold work beneath the surface. The authors found that the initial thermal relaxation of highly cold worked surfaces (either shot or gravity peened) can be far more rapid than previously realized and can result in a 50 % loss of the compressive stress at elevated temperatures. However, the laser process, producing minimal cold working of the surface, has exhibited striking resistance to thermal relaxation. No detectable relaxation was produced in the tests at the lower temperature and at the highest temperature, 425 °C, only a small loss occurred near the surface. In testing of Ti-6Al-4V jet engine fan blades, researchers have shown the laser treatment to be superior to other technologies for strengthening and protecting of new and previously damaged blades from fatigue failure [4].

477

4

The LLNL/MIC Laser

The Laser Programs Directorate at LLNL has been a world leader in developing high energy Nd-glass slab lasers for fusion applications for the past 25 years. At present LLNL is in the process of building the National Ignition Facility (NIF), which will produce over 2 million Joules of energy per pulse. It is out of this technology base that LLNL & MIC starting working together in 1997 to adapt a 1.053 nm (infrared wavelength) Nd-glass laser for use in Laser Peening. The specific LLNL/MIC laser used for laser peening is comprised of a single master oscillator and one or more sets of power amplifiers [5]. It utilizes several unique technologies relative to older technology rod lasers:

4.1

Neodymium Laser Glass Slab Technology

The laser glass is configured in a rectangular slab shape, instead of a cylindrical rod. Whereas total energy storage of a laser glass is dependent on the volume of the glass (roughly 0.5 liters of Nd doped glass volume to store 25 Joules of energy), the ability of the glass to dissipate heat will be dependent on the shortest distance to a heat sink (typically cooling water.) Assuming equivalent length of comparable slab and rod laser amplifiers, the heat is much more easily transferred out of the slab (4× faster) by the surrounding fluid. This enables more rapid laser firing rates as the internal thermal gradients (which would distort the beam) are quickly minimized. In addition, the laser light is directed through the slab so that the beam propagates by zig-zag bouncing off the slab faces, which aids in averaging out wavefront distortions due to thermal gradients. Conventional lasers employing cylindrical rod designs naturally produce a round output beam spatial profile. Treatment of extended areas then requires overlapping spots in an inefficient manner. The naturally rectangular spot profile of the LLNL slab laser technology facilitates full coverage by placing each laser spot adjacent to the next.

4.2

SBS Phase Conjugation

The LLNL/MIC Laser employs a patented LLNL architecture, which includes SBS phase conjugation for correction of residual wavefront distortions. The SBS phase conjugator allows generation of a high power beam with nearly diffraction limited beam quality. By correcting for thermal aberrations, the LLNL design enables extraction of average powers up to the mechanical limit of the laser glass. Without the SBS phase conjugator, the beam quality would rapidly degrade as the laser average output power is increased leading to reduced focus control of the beam and eventually less energy on the target. Reduced beam quality can also lead to intensity ”hot spots” within the laser and consequently to self-damage of the laser. The SBS phase conjugator eliminates these potential problems. In April of 2002, MIC started up a production facility for laser peening in Livermore, CA using the LLNL laser technology & the laser peening techniques that were jointly developed by LLNL & MIC. By combining the benefits of slab laser technology and SBS Phase conjugation, the MIC Production laser was designed with capability for a 6 Hz laser firing rate per laser head

478 and with capability for two heads that can fire in tandem or independently. If required, the firing rate can be increased to 10 Hz by the use of stronger laser glasses. Each laser head is capable of generating up to 25 Joules of energy. The laser spot shape is square/rectangular and is selected to range in size from 2–5 mm square (depending on the fluence [J/cm2] required by the metal being laser peened.) The production MIC laser peening system also incorporates a 6-axis robot with repeatable dimensional accuracy to 0.005″. This enables the component surface being laser peened to be ”presented” to the laser at a consistent focal length and at an angle as close as possible to perpendicular to maintain the square laser spot and maximize the pressure wave.

5

Important Laser Peening Parameters

The goal of laser peening is to impart residual compressive stresses to a metal component. From a macro viewpoint, the two primary factors which determine this are (A) the amount and quality of the energy being emitted by the laser, and (B) the amount of energy received by the component. The major laser output parameters which in combination will determine the effectiveness of a specific laser peening pulse are Fluence (J/cm2) & Power Density (GW/cm2). Since it is difficult to measure these parameters directly for high power lasers, the constituent parameters are monitored instead: • • •

Total laser energy – (J); Size/area of laser ”spot” – (cm2); and Duration of laser pulse – (ns).

Besides these parameters defining the individual laser beam properties, two other parameters that speak to the total energy imparted to the target during laser peening are: •

•

Total % unpeened surface area per laser pass. This will be dependent on the laser peening spot size, shape (circle or square) and pattern. The MIC standard spot pattern strives for 0 % unpeened area per pass. Total number of passes. It is common to have 2 to 5 laser peening passes. As the component surface is compressed/densified, it is theorized that it is easier for subsequent laser peening passes to drive the residual stresses to a deeper level.

A recent laser peening study by J. Rankin et al [6] also highlighted the importance of the beam uniformity to achieving consistent compressive residual stresses near the surface. A 7049 T73 aluminum test coupon was laser peened twice using the LLNL/MIC laser with a 1.5 mm corner section of a 5.0 mm square beam intentionally attenuated. Each individual layer had 100 % coverage as the squares were peened adjacent to each other in a close packed tile pattern. The resulting surface residual stresses on the coupon were only 150 MPa compared to 400 MPa of a coupon which was peened exactly the same, but without the corner attenuation. This effect is reduced at depths greater than 0.3 mm beneath the surface, where the residual compressive stresses were comparable. Thus, one might conjecture that having a uniform laser beam is more

479 important in those instances where inhibiting crack initiation is of primary importance. Other observations of this study: • • •

No difference in residual stress profiles was noted when all variables were held constant except for laser spot size which was varied from 3.2 to 5.0 mm square; A 50 % overlap between the first and second laser passes was found to yield a more consistent residual stress profile beneath the surface than that from a 10 % overlap; and As might be expected, laser peening at a fluence of 60 J/cm2 produced a deeper layer of compressive residual stresses than that produced by laser peening at 45 J/cm2.

Factors which can impact on the amount of laser energy & power which is received by the target are: • •

6

Consistency/integrity of the absorptive layer (paint or tape). Laser peening of an area with an incomplete or damaged absorptive layer could lead to non-uniform beam energy. Also critical to the laser peening process is for the water tamping layer to flow over the laser target area in laminar (non turbulent) flow.

Control Techniques for Laser Peening Parameters

The typical component that an OEM considers for laser peening is a high value ”mission critical” component, where the residual compressive stresses produced by conventional shot peening have proven insufficient. As such, preliminary development work will involve iterative testing of varying laser peening parameters; using X-Ray diffraction measurements of actual residual stress profiles, rig tests and even field evaluations to confirm the expected benefit in life extension. When the desired component life extension evaluation is satisfactory; the laser peening parameters and process are frozen & a Process Control Plan is instituted. This Process Control Plan ensures that the appropriate residual stress profiles identified during the Development program will be consistently reproduced on Production components. MIC is utilizing electronic monitoring of the laser beam properties in combination with Almen strip test coupons to monitor and control the laser peening process:

6.1

Electronic Monitoring of Laser Output

Figure 3 is a schematic of how the laser output is electronically calibrated, monitored and controlled. The absolute calibration of the laser energy output (10–25 J) is accomplished by inserting a NIST calibration traceable calorimeter in the beam path. At the same time an additional calorimeter detector, which is sampling a small constant percentage (~0.2 %) of the beam energy, is calibrated against the total energy output. During production laser peening, this second detector readings are used as a reliable proxy for the full energy output of the laser, which is directed onto the component being laser, peened. At the same time a third detector constantly monitors the time in nanoseconds of the laser pulse. By knowing the laser energy (J) and pulse width (ns), we can compute the Power (GW).

480 Calorimeter for absolute energy calibration Target Part

Nd:glass Laser Beam splitter (~0.2% reflectivity) Laser Power Supply

Laser Control System

Calorimeter moved into beam for calibration & out for operation Detectors for monitoring energy and pulse width for every laser shot

Process Monitoring and Data Logging Computer

Figure 3: Calibration, monitoring and control of laser output

To compute the Fluence (J/cm2) and Power Intensity (GW/cm2), we need an accurate representation of the surface area of the focused laser beam. Although laser spot dimensions can be defined in theoretical terms such as Full Width Half Maximum (FWHM), a more practical method is to physically measure the dimensions of a single laser peened depression at the point where the laser is fully focused on the target. Since the optics of a laser are fixed, as long as future laser peening hits are done at the same focal length as the component is manipulated, the spot size will remain constant. Beam uniformity is monitored by sampling of the near field output of every laser pulse using a video camera. The pixels of the camera are linear in their response, so their output represents a relative measurement of the local intensity within the beam profile. This pixel-by-pixel intensity pattern is acquired by a frame grabber that stores a matrix array of intensity vs. position within the beam. The profile is analyzed and compared to acceptable profiles. Finally, the laser output data from each laser pulse during the processing of a component is captured and logged. At the conclusion of a run, statistics are generated on parameters such as mean output energy vs. time, standard deviation in the energy vs. time, pulse width vs. time and standard deviation in pulse duration vs. time.

6.2

Almen Strips for Process Control in Laser Peening

Almen strips have traditionally been used for the control of conventional shot peening processes, where the level of shot intensity imparting compressive residual stresses into a material is correlated with the deflected arc height in an Almen strip processed to the same parameters. During the development of the laser peening technology, many different control methods have been developed and reviewed. However, it was felt that the continued use of the Almen strip, combined with electronic monitoring, data logging and feedback control of the laser would give optimal process control.

481 It was demonstrated that the amount and depth of compressive residual stress imparted into a component by Laser Peening could be correlated with levels of arc height deflection of an Almen strip. Thus, Almen strips can be used as a simple and effective check on the equipment set up prior to process start up, or even during production to evaluate any levels of process parameter drift. MIC & Rolls Royce conducted trials to compare the results of using conventional 1070 spring steel Almen C strips, as well as strips manufactured out of a Titanium alloy, which was similar to the component being laser peened. It was found that the spring steel Almen C strips produced much more consistent levels of deflection at constant laser fluence and power settings than the Titanium strips. It was determined that this was primarily due to the superior methods of manufacture of the Almen C strips and the tight controls for hardness, flatness & dimension under which they were manufactured. Almen strips can also be used to check how the laser energy and power is being received by the component being laser peened. It is a final check of laser beam integrity issues such as near field image and alignment. At the laser target area two critical items that will impact on how efficiently the energy is received and converted into residual compressive stresses in the component are the integrity of the water tamping (or “inertial confinement”) layer and the receptivity/ integrity of the laser light adsorptive/ablative layer. It is well known that if either of these factors is compromised, then energy transfer to the substrate will be reduced and the desired levels of residual stress will not be produced. In addition to measurement of the amount of deflection of the Almen strips after laser peening, it is also useful to visually assess the peened surface and the post laser peened condition of the ablative layer. An appropriately laser peened surface will be characterized by crisp indentations in the metal surface and absence of damage to the ablative layer in the region outside of the laser peened spot. Maintaining the integrity of the ablative layer during laser peening is important, as the possibility exists for the localized disruption & turbulence of the water tamping layer, which will inhibit the transfer of the full laser energy into the component.

7

Summary

Combined with electronic monitoring and data logging of the laser output, the use of Almen strips as a process control tool for laser peening can give OEM’s an added measure of confidence that a laser peening process on a critical part can be reproduced and monitored in a production environment. By using the same techniques utilized for control of conventional shot peening, broad industrial acceptance of the laser peening technology should be more readily accepted. Thus, laser peening can simply be viewed as an added capability extension of conventional shot peening and surface enhancement technologies widely used in industry today.

8 [1] [2]

References B. P. Fairand and B. A. Wilcox, J. Appl. Phys. 43 (1972) 3893. H. Clauer, B. P. Fairand and J. Holbrook, J. Appl. Phys. 50 (1979) 1497.

482 [3]

[4]

[5]

[6]

P. Prevey, D. Hombach and P. Mason, ”Thermal Residual Stress Relaxation and Distortion in Surface Enhanced Gas Turbine Engine Components,” Proceedings of ASM/TMS Materials Week, Indianapolis, IN, September 15-18, 1997. S. R. Mannava, W. D. Cowie, A. E. McDaniel, ”The Effects of Laser Shock Peening (LSP) on Airfoil FOD and High Cycle Fatigue”, 1996 USAF Structural Integrity Program Conference, December, 1996. B. Dane, J. Wintemute, B. Bhachu and L. Hackel, ”Diffraction limited high average power phase-locking of four 30J beams from discrete Nd:glass zig-zag amplifiers,” postdeadline paper CPD27, CLEO ‘97, May 22, 1997, Baltimore, MD. J. E. Rankin, M. R. Hill, J. Halpin, H-L Chen, L. A. Hackel and F. Harris, ”The Effects of Process Variations on Residual Stress Induced by Laser Peening,” Sixth European Conference on Residual Stress. 2002.

483

High Temperature Fatigue of Mechanically Surface Treated Materials Igor Altenberger1, Ulf Noster2, Berthold Scholtes2, Robert O. Ritchie1 1 2

Department of Materials Science & Engineering, University of California, Berkeley, CA, USA Institute of Materials Technology, University Kassel, Kassel, Germany

1

Introduction

The most well known effect of mechanical surface treatments on metallic materials is the improvement in fatigue properties. It is therefore not surprising that most of the archival literature on mechanical surface treatments, such as shot peening, deep rolling and laser shock peening, deals with the effect of near-surface properties on fatigue behavior. Most of these studies, however, are confined to room temperature fatigue behavior; in comparison, the effect of mechanical surface treatment on fatigue behavior at high temperatures has been rarely investigated [1-5]. The reason for this disparity can be found in the popular belief that fatigue strength improvement by mechanical surface treatments is mainly due to the presence of compressive residual stresses, and since such stresses should anneal out at elevated temperatures, mechanical surface treatments for high temperature applications would appear questionable. However, this view may be over simplistic as there is always a possibility that the residual stresses may be at least partially stable at elevated temperatures [6]; in addition, other factors may be involved, such as the nature of the near-surface microstructure. Accordingly, it is the objective of this study to examine the role of mechanical surface treatments on the high temperature fatigue behavior of several metallic engineering materials. Moreover, it is the aim of this work to clarify what are the critical temperature “thresholds” at which near-surface microstructures and residual stresses become unstable and whether this can explain the observed fatigue behavior.

2

Fatigue Behavior

We begin by reviewing the high temperature stress-controlled fatigue of mechanically surface treated metallic materials. For all materials studied, the range of homologous temperatures was between 0.35 and 0.6 Tm where Tm is the melting temperature. Fig. 1 presents the cyclic deformation behavior of a titanium alloy Ti-6Al-4V (bimodal microstructure) at a temperature of 450 °C (0.4 Tm) in the untreated and mechanically surface treated condition. In this case, laser shock peening (intensity 7 GW/cm2, coverage 200 %) and deep rolling (rolling pressure 150 bar, spherical rolling element Æ 6.6 mm) were selected as surface treatments. Both untreated and treated materials exhibit an initial increase in the plastic strain amplitude (cyclic softening) with number of cycles, followed by a decrease (cyclic hardening) until fracture. The untreated microstructure shows a quasi-elastic incubation period before cyclic softening, in contrast to the mechanically surface treated conditions, where softening starts immediately with cycling. The early onset of softening is most likely caused by the high degree

484

Figure 1: Cyclic deformation behavior of Ti-6Al-4V at behavior of AISI 304 450 °C (Ia= 460 MPa, R = –1, f = 5 Hz)

Figure 2: Cyclic deformation at 450 °C (Ia = 200 MPa, R = –1, f = 5 Hz )

of work hardening (increased dislocation density) in the mechanically treated surface layers. With this alloy, both deep rolling and laser shock peening lead to an improvement of fatigue life compared to the untreated state. The deep rolling gave better life enhancement, consistent with a more pronounced reduction in plastic strain amplitude. The high temperature cyclic deformation behavior of an austenitic stainless steel AISI 304 in different surface conditions is shown in Fig. 2 for a test temperature of 450 °C (0.4 Tm) [7]. Both shot peened and deep rolled conditions show pronounced secondary cyclic hardening, with plastic strain amplitudes decreasing with increasing lifetime, particularly for the deep rolled material. The untreated material, conversely, exhibited the highest plastic strain amplitudes combined with the shortest lifetime. It can be seen that, even at 450 °C, the fatigue lifetime can be considerably enhanced by mechanical surface treatments compared to the untreated material (Fig. 3). For the selected process parameters, at all temperatures, deep rolling enhanced the fatigue life more effectively than shot peening. It

Figure 3: Stress controlled fatigue lifetimes of AISI 304 in different surface treatment states for temperatures from room temperature to 650 °C (Ia = 280 MPa, R = –1, f = 5 Hz)

485 should be noted that a reduction of plastic strain amplitude due to mechanical surface treatment does not always lead to enhanced fatigue lifetimes. In fact, for a wrought magnesium alloy AZ31 [8], it was found that despite a reduction of plastic strain amplitude at test temperatures above 100 °C by surface treatment no lifetime increase was observed. This was attributed to a strong influence of creep to the damage mechanism [5].

2.1

Stability of Near Surface Residual Stress

To examine why mechanical surface treatments influence the high temperature fatigue behavior of various materials so differently, the stability of the near surface properties involving both residual stresses and near surface microstructures is considered during fatigue loading. The depth profiles of near-surface residual stresses and half-width values (FWHM = full width at half maximum) of x-ray diffraction peaks are shown in Fig. 4 for the deep rolled magnesium AZ31, after cycling at different temperatures up to half the number of cycles to failure. It is apparent, that the residual stresses, as well as half-width (FWHM) values, are markedly unstable during isothermal fatigue at temperatures above ~100 °C; indeed they are eventually reduced to the initial values of the virgin material. Here, the relaxation in both macro- and micro-stresses occurs primarily by thermal annealing, although aided by the cyclic deformation [8,9]. The corresponding stability of near-surface macro- and micro-stresses in the titanium alloy Ti-6Al-4V after isothermal fatigue at 450 °C i.e., at a similar homologous temperature is shown in Fig. 5 for the deep rolled and laser shock peened conditions. Here for both surface-treated conditions, the compressive residual stresses at the surface relax significantly after half the number of cycles to failure, from initial values of –700 and –400 MPa down to –100 to –200 MPa. Interestingly, the stress relaxation at this temperature affects mainly the macro-stresses, whereas inhomogeneous micro-stresses, which are responsible for line-broadening (FWHM-values),

Figure 4: Relaxation of near surface residual stresses and FWHMs in deep rolled magnesium alloy AZ31 (rolling pressure 100 bar) after stress-controlled fatigue at temperatures from 100 to 300 °C (Ia = 75 MPa, R = –1, f = 5 Hz, Nf/2 cycles (Nf = cycles to failure)).

486

Figure 5: Near surface residual stress and FWHM-depth profiles of deep rolled (rolling pressure 150 bar) and of laser shock peened (peening intensity 7 GW/cm2, coverage 200 %) Ti-6Al-4V before (left) and after (right) high temperature fatigue (Ia = 460 MPa, R = –1, f = 5 Hz, N = Nf/2)

are more stable at this loading condition (Ia = 460 MPa) and temperature (450 °C). Additionally, pronounced pure thermal stress relaxation was found for this material. A similar trend in the stability of the macro- and micro-stresses can be seen for a mechanically surface-treated austenitic stainless steel AISI 304 (c.f., Fig. 6 and 7). In this alloy, the near surface work hardening seems to be more stable than the macro residual stresses after high temperature fatigue, especially in the temperature range £ 450 °C. Fig. 6 shows residual stress values at the surface of deep rolled and fatigued AISI 304 (after half the number of cycles to failure at different temperatures). After cycling at 450 °C, the surface residual stresses relax by more than 50 %, whereas at 650 °C, they relax by more than 70%. In general, shot peened specimens exhibited more stable FWHM values after high temperature fatigue (Fig. 7), owing to a higher initial dislocation density at the surface; however, the corresponding residual stresses relaxed more severely than in deep rolled samples [10]. An important finding of this work on Ti-6Al-4V, AZ31 and AISI 304 is that work hardening (elevated FWHM-values) is still prevalent at temperatures higher than where the (macro) resi-

N = Nf/2 Ia = 280 MPa

-100

FWHM (°)

Residual stress (MPa)

0 -50

-150 -200 -250 -300 -350 unloaded

-400 0

100

200

300

400

500

600

700

Temperature (°C)

Figure 6: Surface residual stresses of deep rolled AISI 304 (rolling pressure 150 bar) before and after stress-controlled high temperature fatigue at different temperatures (Ia = 280 MPa, R = –1, f = 5 Hz, N = Nf/2)

1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0

unloaded

deep rolled

shot peened

0

100

200

N = Nf/2 Ia = 280 MPa

300

400

500

600

700

Tem perature (°C)

Figure 7: Surface FWHM-values of deep rolled and of shot peened (peening intensity 0.120 mmA) AISI 304 before and after stress-controlled fatigue at different temperatures (Ia = 280 MPa, R = –1, f = 5 Hz, N = Nf/2)

487 dual stresses have already relaxed, which is also known for the case of pure thermal stress relaxation [3].

2.2

Stability of Near Surface Microstructures

Cyclic and thermal stability of near-surface microstructures in mechanically surface-treated metallic materials is crucial for fatigue life improvement, especially if residual stresses are known to be unstable. For room temperature fatigue, the microstructural and loading factors that determine the stability of the near-surface microstructures are well known [11,12,13,14]. For example, in AISI 304 austenitic stainless steel, the formation of a near-surface nanocrystalline layer and a strain-induced martensitic transformation lead to very stable near-surface microstructures for room-temperature fatigue, even at high stress and strain-amplitudes [12]. By examining the variation in microstructure with depth below the surface, using cross-sectional transmission electron microscopy (XTEM), the nanocrystalline surface layers and transformed martensitic regions, together with the near-surface FWHM-values, all appeared to be cyclically stable before and after cycling at room temperature [12]. Preliminary results also exist for high temperature fatigue in AISI 304. Fig. 10 shows XTEM images of the near-surface microstructure for deep rolled AISI 304 before and after fatigue at 450 °C (Ia = 280 MPa, cycling frequency f = 5 Hz) for half the number of cycles to failure. Interestingly, the microstructures initially observed after deep rolling again appear to be quite stable after fatigue at high temperature, although a slight decrease in the dislocation density could be detected microscopically and by FWHM-measurements. In situ heating of TEM foils of the deep rolled condition yielded recrystallization temperatures for the “nanolayer’ of 600–650 °C.

Figure 8: Cross-sectional TEM micrograph of deep rolled AISI 304 (rolling pressure 150 bar) of direct surfaceregions before (left) and after (right) high temperature fatigue (sa = 280 MPa, R = –1, f = 5 Hz, N = Nf/2) revealing nanocrystalline regions.

488 In contrast, the magnesium alloy AZ31 in the deep rolled condition showed no such stability in the near-surface microstructures and the residual stresses after fatigue loading at temperatures as low as 100 °C (Fig. 6). In fact, this alloy shows typical annealing and recrystallization behavior, e.g., induced deformation twins vanished after only 10 sec at 300 °C (Fig. 11). This behavior was also observed after longer exposures at lower temperatures [9].

Figure 9: Near-surface microstructure of deep rolled AZ31 before (above) and after (below) high thermal exposure (300 °C, 10 seconds)

3 1.

2. 3.

4.

Conclusions There are marked differences in the response of materials to mechanical surface treatments, performed in order to enhance high-temperature fatigue resistance. While some mechanically surface treated materials show poor high-temperature fatigue properties, such as the magnesium alloy AZ31 at temperatures of 0.4–0.6 Tm, other materials, such AISI 304 austenitic stainless steel, at 0.4–0.5 Tm, show an improvement in high-temperature fatigue behavior compared to the untreated condition. Inhomogeneous micro residual stresses (as indicated by FWHM-values) generally relax at higher temperatures than macro residual stresses. The critical temperature threshold for lifetime improvement of mechanically surface-treated states was found to occur at the temperature where the near-surface work hardening (inhomogeneous micro residual stresses) “anneals out”, rather than at the relaxation of the macro residual stresses for all the materials investigated. A useful parameter to assess the high-temperature fatigue behavior of mechanically surface treated materials is the plastic strain amplitude during cyclic loading. Improvement in life can be expected if mechanical surface treatments, such as laser shock peening and deep rolling can induce significantly deep work-hardened and cyclically stable surface layers, so as to lead to a finite reduction of the plastic strain amplitude during most of the lifetime.

489

4

Acknowledgements

This work was supported by the U.S. Air Force Office of Scientific Research under grant No F49620-96-1-0418 under the auspices of the Multidisciplinary University Research Initiative on High Cycle Fatigue. In addition, thanks are due to the German Science Foundation (DFG) for financial support (grant-no. AL 558/1-1) and to the Metal Improvement Company, Livermore, CA, for performing the laser shock peening.

5 [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14]

References Luetjering, G., The Minerals, Metals & Materials Society, Warrendale, PA, 1999, p. 291. Wang, R., Zhang, X., Song, D. and Yin, Y., Proc. 1st Int. Conf. on Shot Peening, Edited by A. Niku-Lari, Pergamon Press, Oxford, UK, 1982, p. 395. Gray, H., Wagner, L., Luetjering, G., Fatigue Prevention and Design, Edited by J.T. Barnby, Chamelion Press, 1986, p. 363. Hasegawa, N., Watanabe and Y., Kato, Y., Proc. 5th Int. Conf. on Shot Peening, Edited by D. Kirk, Oxford, UK, 1993, p.157. Altenberger, I., Noster U., Scholtes, B. and Ritchie, R.O., in: Fatigue 2002, EMAS, Stockholm, 2000, in print. Holzapfel, H., Schulze, V. and Voehringer, O., Mater. Sci. Eng., Vol. A248, 1998, p. 9. Altenberger, I., Gibmeier, J., Herzog, R., Noster, U. and Scholtes, B., Materials Science Research Int. – Special Technical Publication, Vol. 1, 2001, pp. 275-284. Noster, U., Altenberger, I. and Scholtes, B., Proc. Magnesium Alloys and their Applications, Edited by K.U. Kainer, Verlag Wiley-VCH, Weinheim, 2000, p.312. Noster, U., Altenberger, I. and Scholtes, B., Surface Treatment V, Edited by C. A. Brebbia, WIT press, Southampton, UK, 2001, p.3. Altenberger, I. and Ritchie, R. O., unpublished research, University of California, Berkeley, 2002. Altenberger, I., Scholtes, B., Martin, U. and Oettel, H., Haerterei Tech. Mitt., Vol. 53, 1998, p. 395 Altenberger, I., Scholtes, B., Martin, U. and Oettel, H., Mater. Sci. Eng., Vol. 264, 1999, p 1. Altenberger, I. and Scholtes, B., Mater. Sci. Forum, Vol. 347-349, 2000, p. 382 Martin, U., Altenberger, I., Scholtes, B., Kremmer, K. and Oettel, H., Mater. Sci. Eng., Vol 246, 1998, p. 69.

490

Roller Pressing or Shot Peening of Fir-Tree Root of LPT Blades of 500 MW Steam Turbine Mahesh C. Sharma Maulana Azad College of Technology, Bhopal, India

1

Abstract

An experimental investigation was carried out to explore the possibility of replacing pressure rolling of Fir-tree root of L.P. Blades by shot peening. For this purpose it was necessary to find out shot peening parameters which could achieve the required depth of compression and residual stress as pressure rolling does. Depth is especially crucial when very poor water condition come into effect. Further more surface roughness is a concern for bearing lands. It was necessary to achieve residual stress not lower than -300 MPa down to a depth of 0.25 mm and roughness value Rz must not be larger than 22 μm after shot peening. In the present investigation surface roughness and residual stress data obtained with different peening parameters were reported. It was observed that pressure peening with 10 mm convergent divergent nozzle, 1 to 1.2 mm cast steel shots, at 0.32A peening intensity, 3 kg/cm2 tank pressure 12" stand off, 125 to 150% coverage gave residual stress value -418 MPa at 0.25 mm depth in 12% chrome steel test pieces. The surface roughness too was with in required range (Ra = 2.80 μm). Residual stress on an actual blade profile cut section was also verified. Thus roll pressing can be replaced by shot peening.

2

Introduction

The fir-tree root of LP Blades are fitted into the corresponding groove of the turbine disc. They undergo very small amplitude cyclic movement at the root in close contact and under heavy pressure. The minute relative displacement of contact surfaces subsequently causes the rupture of asperity, which then oxidizes. Severe surface damage may soon appear due to the influence of the applied service stresses of the moving parts. Fatigue micro cracks will develop and propagate leading to eventual part rupture. It was reported that shot peening improved fretting fatigue and fretting fatigue strength increases with increasing roughness [1] The complete interaction fretting/shot peening may be represented diagramatically as follows [2] (table 1):

491 Table 1: Interaction fretting/shot peening Fretting Is Due To

Shot Peening Imparts

Contact pressure

Bearing capacity

Normal stress

Residual stress

Displacement amplitude

Peak elasticity

Debris

Debris absorption

Lack of lubrication

Lubrication agent retention

Micro cracks

Micro crack blockage

Hardness

Favorable microstructure

Friction

Favorable morphology

If the decision is to be made to shot peen a particular turbine blade – type of shot peening parameters are determined. The desired distribution of residual stresses is chosen depending on the steel shot. This distribution is characteristic of the component material upon shot peening. However, not only the material behavior during shot peening of component is important, but it has to be checked in a second step whether the chosen residual stress distribution is compatible with the geometric conditions. Experiments with notched specimens have shown that shot peening of notches has a beneficial effect only if the notch radius to be shot peened is several times larger than diameter of shot used [3,4]. Table 2: Chemical composition and mechanical properties of blade material C 0.18

S 0.010

Hardness

260 BHN

Impack energy

100 Joules

UTS

880 MPa

P 0.015

Si 0.31

Mn 0.48

Ni 0.38

Cr 13.68

Table 3: Surface finish for the blade Profile

3.2 μm

Root

2.5 μm

Inlet edge, exit edge and fillet

0.8 μm

3

Experimental Work

An actual turbine blade was cut at the end of fir-tree profile and its root were shot peened by pressure peening system as shown in the figure 1. First the peening parameters were decided by shot peening block samples. About twenty block samples of 2" × 3/4" × 1/4" were cut by les-

492 ser cutting to make residual stress free test blocks. These test blocks were peened to achieve different peening intensities using suitable peening parameters. At different peening intensities surface roughness and residual stress distribution in the depth direction was experimentally investigated. The recommended residual stress was –300 N/mm2 at 0.25 mm depth at the fir-tree root and surface roughness lesser than 22μm. Residual stress data for shot peened blade root to an intensity 0.35A using 1 mm cast steel shots were as follows: Table 4: Residual stress data for shot peened blade root Blade Root Spot - 1

Blade Root Spot - 2

Depth (mm)

Residual Stress (MPa)

Depth (mm)

Residual Stress (MPa)

0.0

- 786

0.00

- 780

0.05

- 639

0.10

- 539

0.10

- 627

0.17

- 553

0.20

- 463

0.30

- 376

0.25

- 461

0.32

- 393

0.40

- 319

Above results showed that required residual stress distribution at fir-tree root is possible by shot peening with even 1 mm steel shots. This shot diameter was about one fourth the root radii of the blade. Further investigations were carried out using few sizes of bearing balls and shots. 3.1

Ball peening with 3.1 mm Steel Bearing Balls using pressure peening

In order to induce greater depth of penetration of required magnitude of residual stress value – 300 MPa at 0.5 mm depth or even –500 MPa at 0.5 mm depth. Following peening parameters were used. Table 5: Peening parameters Comp. Air supply tank pressure

5 kg/cm2

Discharge of compressed air

2300 litre/min

Pressure pot pressure

1.5 kg/cm2

Convergent divergent nozzle of throat dia

8 mm

Orifice diameter

12 mm

Half valve opening for uniform flow of balls peening intensity

0.5A

Stand off

1 ft. 2

At higher tank pressure of 2 kg/cm peening intensity was enchanced to 0.6A

Ball peening gave surface roughness valve Ra = 4.4 to 6.1 μm

493

Figure 1: Left : Photographs of turbine blade cut section placed on turn table of pressure peening unit and two positions of nozzles for peening the fir-tree root; right: blade fir-tree root actual profile to full size

Ball peening with slightly higher velocity which could be obtained by increasing tank pressure to 2 kg/cm2 and an intensity was enchanced from 0.5A to 0.6A. It was found that at a depth of 0.5 mm residual stress was even greater than –500 MPa observation for three different ball peening condition were plotted as shown in figure 2. Figure 2 shows that with intensity 0.5A it was possible to induce –500 MPa up to a depth of 0.4 mm, while with 0.6A at 2 kg/cm2

Figure 2: Residual stress distribution by 3.1 mm ball peening using pressure peening

494 pressure –500 MPa could be induced to a depth of 0.5 mm. Primary peening with smaller shots 0.8 mm and then ball peening was not benefitial as compared to peening at higher intensity with 3.1 mm ball.

3.2

3.1 mm Ball Peening with Syphonic System

Block Samples were ball peened on Syphonic System developed at our end [5]. Following peening parameters were used. • •

Bearing Ball 3.1mm dia Syphonic Peening System with compressed air supply from a compressor of 15 hp, 1460 rpm and 1850 litres/minute discharge air tank pressure 5 kg /sq. cm., stand off 2" exposure time 3 minutes. Residual Stress and surface roughness data were as shown in Figure 3

Figure 3: Residual stress distribution by 3.1 mm ball peening using Syphonic peening

With peening intensity 0.28 A it was possible to attain depth of penetration greater than 0.3μm and residual stress at this depth was also greater than –300 MPa. Surface roughness was well within limits Ra = 1.36 μm, Rz = 7.85 μm. Using same peening parameters but with shot size 4.5mm in place of 3.1mm block samples were peened and residual stress distribution was measured by X-ray diffraction. The results were plotted as shown in Figure 4. It was observed that depth of penetration was 0.55 μm and magnitude of residual stress was –300MPa at even lower peening intensity of 0.27 A. Shot Peening with Shot size compatible to Blade root Geometry and its effect on residual stress and surface finish. Block sample were shot peened using shot size: 1.2 mm, stand off: 8 Inch, Air Pressure: 2.0 kgf/sq.cm, Peening Intensity: 0.23 A, nozzle: 6 mm throat diameter. Fig. 5 shows residual stress distribution & surface roughness produced by Shot peening using 1.2 mm steel shots & Pressure peening system.

495

Figure 4: Residual stress distribution and surface roughness produced with 4.5 mm steel bar

Figure 5: Residual stress distribution and surface roughness produced by 1.2 mm steel shots

4

Results and Discussion

Above experimental results showed that ball peening using 4.5 to 3.1 mm steel ball introduced higher magnitude of residual stress distribution compared to 1 mm steel shots with lower surface roughness. However due to geometric constraints we had to go for smaller shots and it was observed that 1.2 and 1mm steel shots at peening intensities 0.23A and 0.35A respectively could again produced residual stress distribution and surface roughness within limits. Therefore

496 Shot peening can replace roller pressing of fur-tree roots of LP Blades of 500 MW Steam Turbine. Residual stress results were as follows: Table 6: Residual stress results Material Removed After polishing (mm)

Stress MPa

Peening Parameters % Coverage

0

–493.9

Shot dia 1.00

0.050

–482.2

Intensity 0.35 A

0.100

–461.6

ts = 15 Sec.

0.150

–358.6

Pressure Peening

0.200

–333.2

Peening Pressure

0.250

–310.6

3 kg/ Sq.cm

150 %

Rz

13.82

Figure 6: Roughness of blade at smaller and middle groove before peening was Ra = 0.70 μm and Rz 4.40 μm. When block sample was shot peened with 1 mm steel shots to 0.35 A as per above data it was possible to attain –310 MPa residual stress at a depth of 0.250 mm and Ra = 2.46 μm, while Rz = 13.82 (with in limits).

5

Acknowledgements

Sponsorship from BHEL, Haridwar and BHEL R & D Hyderabad is duely acknowledged.

497

6 [1] [2] [3] [4]

[5]

References Y. Watanabe, N. Nasegawa, H. Endow: An effect of peening on fretting fatigue, ICSP-7 Proc. p. 127. Yves Le Guernic, France: Shot peening retards “Fretting”, ICSP-3, p. 281. C. M. Verpoort and C. Gerdes: Influence of shot peening on material properties and the control of shot peening of turbine blades, p. 60. M. C. Sharma & R. K. Joshi: Improvement of fretting fatigue performance of large steam turbine blade material using ball peening, Proc. of Int. Conf. ICSP & BC-2, MACT, Bhopal, India, Sept. 2001, p. 132. M. C. Sharma & N. Nath : Development of ball peening nozzle, ICSP-7, p. 432.

498

Comparison of Surface Characteristics and Thermal Residual Stress Relaxation of Laser Peened and Shot Peened AISI 4140 Rainer Menig, Volker Schulze and Otmar Vöhringer Institut für Werkstoffkunde I, Universität Karlsruhe (TH), Karlsruhe, Germany

1

Abstract

Laser peening is a relatively new mechanical surface treatment which causes deep zones bearing compressive residual stresses. This is accomplished by applying shock waves to the material surface using short laser pulses. Quenched and tempered steel AISI 4140 was laser peened using a Nd:glass slab laser with a pulse energy of 25 Joule and a wavelength of 1053 nm. Afterwards the mechanically affected zones were analyzed according to their topography, residual stress state and work hardening state and compared to shot peened samples. It is well known that the effect of mechanical surface treatments strongly depend on the stability of the induced residual stress state. Therefore, annealing treatments at different temperatures and times were performed to analyze the thermal residual stress relaxation behavior. Using an iterative mathematical procedure based on a least squares algorithm the activation enthalpy for thermal residual stress relaxation was determined and the responsible mechanisms were identified. The results were evaluated according to the amount of cold work caused by the different surface treatments.

2

Introduction

Laser shock processing also called laser peening is a newly developped surface hardening procedure, which affects surface zones in the mm-range without causing increases in the surface roughness in the same magnitude as shot peening. Investigations about laser induced shock waves have been conducted for almost fourty years [1, 2]. However, using laser peening to influence the surface characteristics of metallic parts was hardly investigated until the last decade of the 20th century [e.g. 3-5]. Mainly Nd:glass laser, Nd:YAG laser or XeCl-Excimer laser are used to induce pulses with widths in the nanosecond regime and intensities in the GW/cm2 range to modify the surface characteristics. The high intensity of the laser irradiation is absorbed causing ablation at the material. The ablation as well as the repulsion of the expanding plasma results in a high amplitude shock wave. Direct and confined ablation is distinguished [6]. The plasma caused by direct ablation is created directly at the surface. Very high intensities are necessary to induce sufficiently high shock waves, because the plasma can expand into the surrounding atmosphere. Moreover, the direct coupling of the laser irradiation as well as the resulting plasma with the material results in unfavorable thermal stresses with the creation of tensile residual stresses. Confined ablation, however, uses a transparent and a thermo-protective coating. The transparent coating, e.g. water, is transmitted by the laser beam and prevents a free plasma expansion into the atmosphere, which leads to intensified shock waves travelling into the material. Instead of the material, the opaque (black paint, metallic foil) thermo-protective coating is ablated within a very thin layer. Thus, no thermal stresses of the laser peened metal

499 occur. After recombination of the plasma the hot vapor expands and induces a compressive shock wave that creates surface parallel plastic deformations, which result in a compressive residual stress state. The industrial operational area of laser peening is not widely spread yet. It is used e.g. to improve the crack resistance of turbine blades [7] or to prevent stress corrosion cracking of austenite stainless steels in power plants [8]. Furthermore, sporadic applications in the automotive and medical industries are known. In contrast to this shot peening is a very established procedure to improve the fatigue properties of metallic components for many years. Therefore a comparison of the surface characteristics after laser and shot peening is necessary to improve the number of applications of laser peening. This will be given in the present paper for the quenched and tempered steel AISI 4140. Additionally the thermal residual stress relaxation behavior will be compared in order to get informations on the stability of the residual stress states induced.

3

Material, Specimen Geometry and Experimental Approach

Investigations were carried out on steel samples of AISI 4140 steel (German grade 42 CrMo 4) with the chemical composition 0.42 C, 1.04 Cr, 0.14 Mo, 0.21 Si, 0.71 Mn, 0.01 P, 0.02 Al and balance Fe (all in wt. %). The samples were machined from flat material by sawing, milling and grinding. Those used for laser peening had a final geometry of 110 mm x 25 mm x 5 mm. The samples used for shot peening had a thickness of only 2 mm. Afterwards, they were austenitized for 20 min at 850 °C, martensitically hardened in oil (25 °C), tempered at 450 °C for 2 hours and cooled down in a vacuum furnace. Thermal residual stress relaxation after laser peening and shot peening was conducted by annealing in a salt bath furnace at defined temperatures and time. The laser peening was accomplished at the Lawrence Livermore National Laboratories in Livermore CA, USA, using confined ablation. The laser used was a Nd:glass slab 25 Joule 1053 nm wavelength laser providing a very uniform rectangular “top-hat” shaped beam footprint. A laser light absorptive material was spread on the material surface in order to achieve a confined ablation. The absorptive coating was struck by the laser beam during processing and provided for the generation of the plasma discharge while protecting the surface. The plasma continued to absorb the laser light until the pulse was completed. A thin damping layer of water flowed over the process target and was used to restrain the plasma long enough for shock wave generation to occur into the part. The water was then blown clear of the surface as the remaining plasma was absorbed by the atmosphere. The samples were processed using two overlapping layers of laser shots (3.5 mm × 3.5 mm) at 168 J/cm2 energy density and 18 ns pulse width. The second layer of shots overlapped the first layer by 50 %. The shot peening treatments were performed using an air blast machine. Cast iron shot S 170 with a hardness of 56 HRC was used at a peening pressure of 1.2 bar with a media flow rate of 1.0 kg/min. The samples determined for shot peening were peened from both sides simultaneously in order to avoid distortions. The Almen intensity was 0.24 mmA leading to a full coverage of the sample surface [9]. The shot peening treatments of the samples used for thermal residual stress relaxation were carried out using shot S170 with a hardness of 44-48 HRC and a peening pressure of 1.6 bar [10]. However, no differences of the induced residual stresses and the Almen intensity were found for the two different peening procedures.

500 The resulting roughness and surface structure were measured using a confocal white light microscope (Nanofocus). Residual stresses of the specimens were determined using CrKa -Xrays and apertures of 0.3 and 2 mm diameter. The {211}-interference lines of the ferritic phase of the investigated steel were analyzed according to the sin2y method [11]. The depth distributions of the residual stresses were determined by iterative electrolytic removal of thin surface layers and subsequent X-ray measurements. Residual stress values measured at the surface after material removal were corrected according to the method of [12]. The half width values were determined as an average of those measured at O = –15°, 0° and +15°.

4

Results and Discussion

Fig. 1 shows the surface structure after laser peening in a region of 6.5 mm x 6.5 mm. Periodical surface structures are created due to pulse overlap. The surface shows recurrent notches of about 5 mm with a distance of about 1.7 mm, which is half of the laser beam cross section, caused by the 50 % overlap. Whereas the roughness of 2.8 mm is not changed by laser peening it is increased to 8.1 mm due to shot peening [9]. Because laser peening is only conducted on one side, the induced inhomogeneous plastic deformation causes a curvature height of 0.94 mm. (a)

(b)

4 3 2

z [µm]

1 0 -1 -2 -3 -4 0

1

2

3

4

5

6

x [mm]

Figure 1: Surface structure (a) and profile line (b) of AISI 4140 after laser peening with two overlapping (50 %) layers

Not only the surface roughness of the laser peened samples but also the residual stresses are influenced by the pulse overlap. Fig. 2 shows that the compressive residual stresses measured using X-rays and an aperture with a diameter of 0.3 mm are varying between –400 MPa and –750 MPa. The highest compressive residual stresses occur in a distance of about 3.5 mm, which matches with the width of the pulse cross section. This indicates that the resulting residual stress state is mainly influenced by the hit of the final, second pulse. However, the oscillation of the residual stresses does not provoke tensile residual stresses like usually found for separately distributed peening shots and is positive for the fatigue properties, therefore. Residual stress depth distributions measured using a 2 mm aperture of laser peened and shot peened [9] samples are shown in Fig. 3. It can be seen that the mean surface values are at about –600 MPa for both variants. While shot peening causes a residual stress plateau underneath the surface till about 0.1 mm, the laser peened sample shows a continuing decrease from the surface value with increasing depth. The penetration depth of the compressive residual stresses is strikingly increased from 0.17 mm to 0.87 mm by laser peening compared with shot peening. In

501 -200

laser peening -300

-500

rs

Is [MPa]

-400

-600

-700

-800 0

2

4

6

8

10

12

x [mm]

Figure 2: Surface residual stresses after laser peening with two overlapping (50 %) layers

Fig. 4 the half width depth distributions as a measure of the microstructural work hardening state are given. The core value of the laser peened sample is only slightly raised towards the surface, which is typical for laser peening [3]. The shot peened sample, however, shows distinct work hardening close to the surface. The core value of about 2.75 °2q is raised towards the surface to a value of about 3.25 °2q. 1000 800 600

rs

I [MPa]

400 200 0 -200 -400 -600

laser peening shot peening [9]

-800 -1000 0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

2.0

2.2

x [mm]

Figure 3: Residual stress depth distribution of laser peened and shot peened samples of AISI 4140

Laser peened samples were annealed in a salt bath at 300 °C as well as 450 °C for 1 min, 10 min, 100 min and 1000 min, respectively. The corresponding residual stress relaxation at the surface is shown in Fig. 5. It can be seen that there is increasing thermal residual stress relaxation with increasing annealing time and temperature. It can be described using a Zener-Wert-Avrami function

srs (Ta , ta ) = exp{-[C × exp(-DH / kTa ) × ta ]m} srs0

(1)

502 4.0

3.5

HW [°2G]

3.0

2.5

2.0

laser peening shot peening [9]

1.5

1.0 0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

2.0

2.2

x [mm]

Figure 4: Half width depth distribution of laser peened and shot peened samples of AISI 4140

600

500

300

rs

|Is | [MPa]

400

200 Laser peening shot peening [10] data / calculation data / calculation

100 300°C 450°C 0

0

10

0

10

1

10

2

10

3

10

4

t [min]

Figure 5: Measured and calculated thermal residual stress relaxation at the surface of laser peened and shot peened samples

where I 0FS is the initial residual stress state and IFS (Ta, ta) the remaining amount of residual stresses after annealing at temperature Ta and time ta. DH is the activation enthalpy, k the Boltzmann constant, and C and m are material related constants [13]. With Eq. 1 the activation enthalpy DH as well as the constants C and m are determined using an iterative mathematical procedure based on a least squares algorithm [10]. The results for C, m and DH are found to be 2.912 × 1015 1/min, 0.17 and 2.49 eV, respectively. The calculated activation enthalpy of 2.49 eV is close to the value of the activation enthalpy for self-diffusion of =-iron (2.6 eV), which implies that the main microstructural process responsible for the residual stress relaxation is volume diffusion controlled creep, which is determined mainly by climb of edge dislocations. Using Eq. 1 and the values found for C, m and DH the thermal residual stress relaxation is calculated. The obtained progression, also shown in Fig. 5, describes the results of the X-ray measurements well. Additionally, results of thermal residual stress relaxation of the shot peened variant [10] are spread. The thermal residual stress relaxation found for 300 °C and 450 °C is

503 very similar to the results of the laser peened samples. An activation enthalpy DH = 3.29 eV and constants C = 1.22 · 1021 1/min and m = 0.122 for the shot peened samples are found, also using the iterative mathematical procedure [10]. However, those results were obtained using 5 different temperatures between 250 °C and 450 °C with 9 times each. This might be the reason that the activation enthalpy differs somewhat from that found for the laser peened variants although the residual stress relaxation behavior in Fig. 5 is absolutely comparable. The respective relaxation of the surface half widths as a measure of the work hardening state can be seen in Fig. 6. The surface half widths after peening [10] are slightly lower than seen in Fig. 4 [9] due to the usage of a different aperture, thus, they coincide with those found after laser peening. The relaxation behavior of the work hardening state at the surface is similar for both variants if the typical fluctuations occurring at half width determination are considered. 3.4 3.2 3.0

HW [°2G]

2.8 2.6 2.4 2.2 2.0

Laser peening shot peening [10] 300°C 450°C

1.8 1.6

0

0

10

1

10

2

10

3

10

10

4

t [min]

Figure 6: Measured thermal half width relaxation at the surface of laser peened and shot peened samples

At the first sight, the results of the thermal residual stress relaxation are in contrast to results reported previously. In [14] e.g., the thermal residual stress relaxation of Ti and Ni alloys used in compressor and turbine stages was investigated at engine temperatures (Ti-6Al-4V: 325 °C–475 °C, Inconel 718: 525 °C–675 °C). It was found that the relaxation of compressive residual stresses induced by shot peening is far more rapid than it is for laser peened variants. Even though the initial residual stress state at the surface was higher after shot peening, the absolute values were already lower after loading for 10 min in the temperature range given above. [14] correlates the rate and amount of thermal residual stress relaxation with the degree of cold work. They assumed that mechanical surface treatments which cause the least cold work retain compressive residual stresses for the longest time or at the highest temperatures. The work hardening caused by laser peening was about 4.2 % and 6 % for Ti-6Al-40 and Inconel 718, respectively. Shot peening, however, caused cold work of 75 % and 30 % for the two alloys. For AISI 4140 the same amount of cold work was found after laser peening and shot peening (Fig. 6). In consideration of the results of [14] this may explain why also the same rate of compressive residual stress relaxation was obtained. Therefore, the maximum operating temperatures of laser peened AISI 4140 components are similar to those of shot peened parts.

504

5

Conclusions

Specimens of quenched and tempered steel AISI 4140 were laser peened and shot peened and compared regarding surface characteristics and thermal residual stress relaxation. Laser peening was conducted with confined ablation using two overlapping (50 %) layers of shots, which caused a periodical surface structure with recurrent notches of about 3 mm at a distance of 1.7 mm. This distance coincides with half the distance of the laser beam cross section. Residual stress measurements with a small aperture of 0.3 diameter revealed oscillations at the surface between –400 MPa and –750 MPa, which occurred in a periodical scheme, determined by the 50 % overlap. If using a aperture of 2 mm diameter, laser peening and shot peening led to similar average surface residual stresses of –600 MPa. However, the depth of the compressive residual stresses was strongly increased by laser peening. Thermal residual stress relaxation behavior was investigated and modelled using an Avrami-function. Comparable residual stress relaxation was found for both surface treatments. The reason may be the same amount of cold work obtained at the surface, which was found previously [14] to determine the rate of thermal residual stress relaxation.

6 [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14]

References R. M. White, J. Appl. Phys. 1963, 34, 2123. N. C. Anderholm, Bull. Am. Phys. Soc. 1968, 13, 388. P. Peyre, R. Fabbro, et al., Mat. Sci. and Eng. 1996, A210, 102–113 J. Kaspar, A. Luft, Prakt. Metallogr. 2000, 37 (4), 181–193. J. P. Chu, J. M. Rigsbee et al. Met. and Mater. Trans. 1995, 26A, 1507–1517. R. Fabbro, J. Fournier et al., J. Appl. Phy. 1990, 68, 775. S. R. Mannava, US Patents US5591009A, US5584662A, US5584586. Y. Sano, N. Mukai, et al., Nuc. Inst. Met. Phys. Rev. (Japan) 1997, B121, 432. A. Wick, V. Schulze, O. Vöhringer, Mat. Sci. and Eng. 2000, A293, 191–197. V. Schulze, F. Burgahn, O. Vöhringer, E. Macherauch, Mat. wiss. u. Werkst. 1993, 24, 258–267. E. Macherauch, P. Müller, Z. f. angewandte Physik 1961, 13, 340–345. D. Dengel, Zeitschrift für Werkstofftechnik 1975, 8, 253–261. O. Vöhringer, Advances in surface treatment (Ed.: A. Niku-Lari) International Guidebook on residual stresses, Vol. 4; Pergamon Press, Oxford, New York, Paris, 1987, 367–396. P. S. Prevey, D. J. Hornbach, et al., 17th ASM Heat Treating Society Conf. Proc. (Eds.: D. L. Milam et al.), 15–18 Sept. 1997, March 1998, ISBN: 0-87170-610-5.

495

IX Modeling

496

507

Finite Element Simulation of Shot Peening – A Method to Evaluate the Influence of Peening Parameters on Surface Characteristics Jochen Schwarzer, Volker Schulze, Otmar Vöhringer Institut für Werkstoffkunde I, University of Karlsruhe (TH), Karlsruhe, Germany

1

Introduction

Shot peening, known for its potential to improve fatigue strength of metallic parts, can be seen as a multiple and progressively repeated elastic-plastic interaction between the surface and the shots. With each impact the target undergoes local plastic deformation while the shot is moving into the material. After the contact between the target and the shot has ceased, compressive residual stresses remain at the surface and small tensile residual stresses in the inside. Developing a model to analyze the process of shot peening is useful for several reasons; to be able to predict the material state after peening without having to conduct costly experiments and to be able to optimize peening processes. In contrast to prior studies found in the literature [1-4], a more detailed approch by simulating several single impacts on a 3-dimensional surface is chosen to model the shot peening process.

2

Finite Element Modelling

2.1

Model Geometry and Boundary Conditions

The model used for the shot-peening analyses was realized in ABAQUS/Explicit and consists of an infinite steel sheet of thickness 0.85 mm and multiple half-spheres. The target is represented by a three-dimensional mesh of 1.5 × 1.5 × 0.85 mm, surrounded by infinite elements. The use of infinite elements provides “quiet“ boundaries by minimizing the reflection of dilatational and shear wave energy back into the finite element mesh. The boundary conditions on the target’s base fix the model in z-direction. The mesh consists of 372000 8-node linear brick elements with reduced integration and hourglass control. In order to achieve sufficient discretisation it is graded in all three directions so that the smallest element occurs in the middle of the target area with an element size of 0.008 mm. Based on measurements of the cast steel shot used for experimental verification, half spherical rigid surfaces of diameter 0.56 mm are used to model the shot. Each rigid surface is connected to a point mass and a rotary inertia element providing the properties of a full sphere. Figure 1 shows the mesh and one half-sphere after the impact of 19 shots.

508

Figure 1: Discretised model used in shot peening simulation

2.2

Material Properties

Great importance was attached to the description of the target material properties. The quenched and tempered steel AISI 4140 (German grade: 42 CrMo 4, Re = 1263 MPa, Rm = 1373 MPa) was chosen therefore. A constitutive law which describes the influence of temperature and strain-rate on the flow stress on the basis of thermally activated dislocation slip was implemented into the finite element code using a user subroutine VUMAT. Accordingly, the flow stress s0, depending on strain-rate A0 and temperature T, is calculated using [5]:

I 0 = IG +

I*0

é æ kT ln ( e e ) ön ù 0 ê1 - ç ÷ ú DG0 ê è ø úû ë

m

(1)

where sG is the athermal proportion of the flow stress, DG0 is an activation enthalpy and s0*, A0 , n, m are further material dependent parameters describing the thermal flow stress component. k is the Boltzmann constant. The work hardening behavior was modelled using a “generalized voce” constitutive equation according to [6]:

sG = sG0 +

s1

+ q1e éë1 - exp -q0 e s1 ùû

(2)

where sG0 and q0 describe the initial yield stress and hardening rate, while s1 and q1 determine the asymptotic characteristics of the hardening. The material constants were determined by a numerical fit of data obtained from tensile tests at different temperatures and strain-rates to the material law similarly to [5]. By comparing the shape of simulated and experimentally produced

509 shot impacts, the material law could be validated. To describe the contact between shot and target, isotropic Coulomb friction with a coefficient of friction m = 0.4 is used. The cast steel shot is modeled as a rigid body with a mass density of 7.85 g/cm3. Elastic or plastic behavior of the shot is not beeing considered in the modelling.

2.3

Impact Order of the Shot

To achieve a realistic modelling of a shot peening process with full coverage an arrangement of the spheres was chosen that provides a closest packed dimple pattern on the surface (Fig. 2). The gray marked inner area which can be approximated with a circle was used for the calculation of residual stress profiles.

4

6 7

8 5

1

4 6

2

8 7

4 5 8

impact order of the shot

4

3 4

5

=

5

Figure 2: Arrangement and order of impacts as well as area evaluated to calculate the residual stress profile

As will be shown later, the fact that the shots impact one after the other instead of impacting simultaneously has great influence on the developing residual stresses. Simplifying a model by using its symmetry and modelling only a part of it does always imply several shots impacting simultaneously. To prevent this, symmetry wasn’t considered in the modelling. The impacts in the inner circle occur one after the other. For computational costs it was allowed that non adjacent shots in the outer circle impact simultaneously. An other aspect concerning the impact order is the number of predecessors and successors around each dimple. In the chosen arrangement each dimple is surrounded by 6 further dimples. Each of the seven inner dimples, which were used for the calculation of residual stress profiles, has a different number of predecessors and successors.

2.4

Analysis of Residual Stress Profiles

In the scale of the dimple size, shot peened surfaces don’t show a uniform distribution of residual stresses. This is shown by experimental results [7] and will later be shown in the performed analyses of seven impacting spheres. Residual stress profiles, usually measured by X-ray diffraction, give an average of stresses in an area covered by the X-ray. To achieve comparable analysis results a mean of residual stresses in a representative area has to be calculated at each depth. The chosen area in the finite element model consists of a circle enclosing the seven inner dimples in figure 2. Within this area an aver-

510 age of the residual stresses parallel to the surface weightened with the element-size is calculated. To exclude a direction dependence resulting from non-perpendicular shots, the mean of residual stresses in x- and in y- direction is calculated.

Figure 3: Residual stresses in x-direction after 1 impact

3

Results and Discussion

3.1

Effect of Adjacent Impacts

In preliminary studies the effect of several successively impacting shots, set at a certain distance to each other, on the residual stress distribution was studied. Therefore, an impact of a single shot in the middle of the target model was simulated. The initial velocity of the shot was 35 m/s. This first impact was surrounded by 6 further impacts so that the borders of the remaining dimples had contact to their neighbors. The residual stress distribution in x-direction after the first impact is represented in figure 3. The maximum of residual stresses parallel to the surface occurs in a depth of approximately 0.05 mm. In this region compressive residual stresses of 1600 MPa are calculated. The remaining dimple diameter is 146 mm, its calculated depth is 9.4 mm. 400

IRS,X after i impacts

200

IRS,X after 7 simultaneous impacts

0

IRS,11 in N/mm

2

-200 -400

i= 5 4 7 6 3

-600 -800 -1000 -1200

2

-1400

1

-1600 -1800 0.00

A

Figure 4: Residual stresses in x-direction after 7 impacts

0.05

0.10

0.15

0.20

B

Figure 5: Residual stresses along the path AB after the number of impacts given in the respective curve

511 After the impact of the 6 further shots a redistribution of residual stresses parallel to the surface occurs (Fig. 4). The compressive residual stresses under the first impact get reduced by the impact of the following shots. Figure 5 shows the development of residual stresses along the path AB from figure 3 after each impact. The first 4 succeding shots cause a remarkable reduction of the maximum compressive residual stress while the position of the stress maximum is shifted into the target material. The next impact causes an increase in compressive stresses with its maximum at 0.075 mm depth. It is also responsible for large compressive residual stresses at the surface. The last shot doesn’t give a significant change of the stress state along the path. These results show the interference of adjacent shot impacts. Succeding impacts have great influence on residual stresses under precedent impacts. In contrast to these results figure 6 presents the analysis of the 7 simultaneously impacting shots. In this case the stress state develops differently. The maximum of the compressive residual stresses is situated below the central dimple. Its value exceeds –1300 MPa and its position is closer to the surface. The respective residual stress profile is shown in figure 5 using a dashed line. At depths up to about 0.05 mm the workhardening induced during previous shots reduces the plastic strains and therefore the residual stresses induced. At higher depths the residual stresses after succeeding impacts are higher because the region of plastic strains was shifted to higher depths due to workhardening at small depths.

Figure 6: Residual stresses in x-direction after 7 simultaneous impacts

3.2

Analysis and Verification of Surface Layer Characteristics

The above presented shot peening model was used to simulate the shot peening process. The shot-peening parameters were comparable to performed verification experiments conducted with an air-blast machine: A shot velocity of 35 m/s and shot diameter of 0.56 mm was chosen. In the analysis all shots impacted perpendicular to the surface. Figure 7 presents the resulting residual stress profile compared to an experimentally obtained one measured at an AISI 4140 which was shot peened with similar parameters [8]. The calculated profile is of the same shape but it shows larger maximum compressive stresses sRS,max and larger surface stresses sRS,surf. The position of maximum compressive residual stresses xmax and of zero stresses x0 is comparable to the experimental results.

512 500

2

IRS [N/mm ]

0

-500

-1000

simulation exp. results measured with X-ray diffraction [8]

IRS,max

-1500 0.0

xmax

x0

0.2

0.4

0.6

x [mm]

Figure 7: Calculated and experimental residual stress pofile

3.3

Effect of Coverage

In order to determine if the chosen dimple configuration can be compared to 100 percent coverage further 19 spheres were added to the model. Their impact location has been changed by rotating the inner and outer circles so that the center of impact of the last 19 spheres was placed on the point of contact of two former dimples. The resulting residual stress profiles doesn’t show significant change compared to the 19shot-model. Only a slight increase in maximum compressive stresses can be noticed. This shows that the 19-shot-model achieves a saturation state of the residual stress profile which isn’t significantly changed by further impacts.

3.4

Effect of Shot Velocity

The effect of shot velocity on residual stresses was studied by calculation of residual stress profiles resulting from different shot velocities (Fig. 8). With increasing shot velocity there is no significant change in calculated surface and maximum compressive stresses. The position of maximum compressive stresses moves into the material with increasing shot velocity. The same behavior is shown by the zero-crossing of the stress profile. These results are in good agreement with experimental results obtained by variing the pressure in an air blasting machine [9].

3.5

Effect of Shot Diameter

The effect of variing the shot diameter on residual stress profiles was calculated using a constant shot velocity of 35 m/s. The analysis results show a strong dependence of the position of the maximum compressive stress on shot diameter (Fig. 9). The zero-crossing of the stress profiles does also move into the target with increasing shot diameter. In contrast to that the surface

513 0

0.30

IRS,surf IRS,max

-200

xmax x0

0.25

-400

0.15 -800

x [mm]

2

IRS [N/mm ]

0.20 -600

0.10 -1000 0.05

-1200

0.00

-1400 20

30

40

50

60

70

80

shot velocity [m/s]

Figure 8: Influence of shot velocity on the residual stress profile

residual stresses and the maximum compressive residual stresses don’t show any dependence on shot diameter. Experimental results from [9] also show an increase of x0 with growing shot diameter whereas sRS,max is not affected from the shot size.

3.6

Effect of Impact Angle

To study the influence of the impact angle on the residual stress profile the impact angle was variied at a constant shot velocity of 35 m/s. Within an analysis each of the 19 spheres had the same direction. The calculated results show a decrease of surface and maximum residual compressive stresses with increasing impact angle. The position of maximum compressive residual 0.5

0 IRS,surf IRS,max

-200

xmax x0 0.4

0.3

-600 -800

0.2

-1000 0.1 -1200 0.0

-1400 0.5

1.0

1.5

2.0

2.5

multiple of inital shot diameter

Figure 9: Influence of shot diameter on the residual stress profile

x [mm]

2

IRS [N/mm ]

-400

514 0

0.20

IRS,surf IRS,max

-200

xmax x0

2

IRS [N/mm ]

-600 0.10 -800 -1000

x [mm]

0.15

-400

0.05

-1200 -1400

0.00 0

10

20

30

40

50

60

impact angle [°]

Figure 10: Influence of the impact angle on residual stress profiles

stresses and the zero-crossing of the residual stress profile also show a strong dependence on the impact angle. Experiments conducted with an air-blast machine variing the peening angle of a quenched plain carbon steel show the same effect [10].

4

Conclusions

A three-dimensional finite element model was developed to simulate the shot peening process. Multiple dynamic spherical indentations were examined and a method for the calculation of residual stress profiles was presented. The effect of shot velocity, diameter, coverage and impact angle upon residual stress profiles was examined and discussed. The results reveal that the depth of the compressed layer is significantly increased by growing shot velocity or shot diameter. A significant influence on surface and maximum compressive residual stresses could not be shown. An increase in the impact angle reduces surface and maximum residual stresses as well as the depth of the residual stresses. The current work indicates that the proposed finite element model is capable to capture the shot peening process, thus implying its potential as an effective tool for the prediction of residual stress profiles.

5 [1] [2] [3]

References S. A. Meguid, G. Shagal, J. C. Stranart, Finite Elements in Analysis and Design, 1999, 31, 179–191. K. Schiffner, C. Droste gen. Helling, Computers and Structures, 1999, 72, 329–340. K. Han, D. Periæ, A. J. L. Crook, D. R. J. Owen, Engineering Computations, 2000, Vol.17, No.5, 593–619.

515 [4]

S. T. S. Al-Hassani, K. Kormi, K. Webb, D. C. Webb, Proceedings ICSP 7, 1999, 217–226. [5] V. Schulze, O. Vöhringer, Metallurgical and Materials Transactions A, 2000, 31A, 825830. [6] C. Tome, G. R. Canova, U. F. Kocks, N. Christodoulou, J. J. Jonas, Acta metall. 1984, Vol. 32, No.10, 1637-1653. [7] B. Scholtes, Eigenspannungen in mechanisch randschichtverformten Werkstoffzuständen, Ursachen-Ermittlung-Bewertung, DGM-Informationsgesellschaft, Oberursel, 1990. [8] R. Menig, Dissertation, University of Karlsruhe, in preparation. [9] A. Wick, H. Holzapfel, V. Schulze, O. Vöhringer, Proceedings ICSP 7, 1999, 42-53. [10] A. Ebenau, O. Vöhringer, E. Macherauch, Proceedings ICSP 3, 1987, 253-260.

516

Shot Peening and Coverage A. Tange, H. Okada NHK Spring Co., Ltd. ,Yokohama, Japan

1

Introduction

As for the processing conditions of shot peening, there are the amount of shot, shot size, arc height, coverage, etc. However, when remarking to the coverage, no systematical study concerned with the relationship with the fatigue strength, can be seen. It would not be also certain that the over peening where fatigue strength will reduce if processing time is too long, exists1). It can be said that it is not clear about whether the optimum conditions between the coverage and fatigue strength can be decided. However, it is difficult to specify the value of coverage because the measuring method of coverage is not designated. Therefore, it is difficult to discuss the relationship between the coverage and fatigue strength. In this paper, the following studies to specify the value of coverage were made: [1] experimentally obtain the relationship between the diameter of dent formed by shot peening, shot size, the shot movement energy (shot speed and shot weight) and the hardness of peened work, and express the experimental results to the formula by multiple regression analysis. [2] it was shown that the coverage can be simulated by the Monte Carlo method. (3) in order to study over peening phenomenon, the fatigue tests with the coverage of 36 %, 98 % and 1000 %, were carried out. It was realized that there were no fatigue strength reduction with 1000 % coverage.

2

Experimental Procedures

2.1

Formulation of Dent Diameter

2.1.1 Shot Peening Machine The shot peening machine employed here is a type of impeller. The diameter of impeller is ö350mm, and the shot peening machine is for experimental studies. Fig.1 shows point A where the direction of an impeller-tangent be coincident with the base. In order to obtain the maximum arc height at the point A, the control gauge was adjusted. Therefore, it can be expected that the most shots are perpendicularly applied in the point A. To make the sample for measurement of dent diameter and the sample for fatigue tests, the all shot peening processes were carried out under the point A.

517 2.1.2 The Specimen for Dent Diameter Measurement The specimen for dent diameter measurement was taken as a width of 19mm, a length of 71mm, and 5mm of thickness. The material used was SUP 10(JIS G 4801), and the hardness of specimen was 666 HV, 505 HV, and 438 HV, prepared by quenching and tempering. In order to remove decarburized layer of both sides, 0.2 mm was removed by wrapping processing, to have mirror finished surface.

The specimen for dent diam hardness of specim empering. In order to rem

Figure 1: Shot peening position

2.1.3 Steel Shot The shot used was the round cut wire of hardness 680 HV. The shot size is Æ1.2 mm, Æ1.0 mm, Æ0.77 mm, and Æ0.4 mm, and was carefully sieved. Consequently, the average weight of 50 shots, M was 9.92 mg, 5.80 mg, 2.58 mg and 0.37 mg, respectively. 2.1.4 Shot Peening Conditions The shot peening conditions are shown in Table 1. The shot speed, V can be calculated by, V = 1.3 2p (ID/2) IRPM/60, where the circumferential speed of impeller (mm/sec) calculated from the diameter of impeller, ID = 350 mm, and the number of impeller revolutions per minute, IRPM, be multiplied by 1.3 times2). 2.1.5 Measurements of Dent Diameter The measurements of dent diameter were carried out by the 23 times magnifying glass. Since the dent was not a true circle, it was assumed as an ellipse. The dent diameter was decides by the geometric average of the long and short axis.

518 2.2

Measurements of Surface Roughness

The measurements of surface roughness were carried out about the specimens of coverage 100%.

Generation of random numbers X and Y Specification of coordinates that shot hits Dent diameter is specified from a formula. (1)

Repetition

Dent is reproduced on a virtual coordinates plane Addition of the coordinates mark contained in dent Removal of the duplicate coordinates point The sum total of the coordinates mark in dent (all coordinates mark= 100% coverage) Figure 2: Simulation method of coverage

2.3

Simulation Method of Coverage

The simulation of coverage showed Fig. 2 was carried out about the following procedures by using personal computer. It was presupposed that the 4 × 4 mm real plane corresponds to the virtual coordinates plane that consists of 175 × 175 points. A coordinates point has from (1, 1) to (175, 175), and the point has 175 × 175 = 30625 point in the area of 4 × 4 = 16 mm2. In order to specify a random coordinates point as if rain falls on this coordinates plane, it was simultaneously generated two different random number from 0 to 1 in the direction of X and Y, and the generated random number is assigned to the division into equal parts from 1 to 175. Next, the circle of the size equivalent to dent diameter applied by the actual shot, is placed focusing on the coordinates specified with two previous random numbers. The number of all the coordinates

519 points included in a circle here is integrated, and if it compares with all coordinates, the coverage of one dent will be calculated. According to the definition of coverage, the coverage when applied many dents must avoid the duplicated integration of the overlap part of dent.

2.4

Relationship Between Coverage and Fatigue Strength

The fatigue test is a cantilever rotary bending fatigue test. The hardness of specimen is 530 HV. The shot peening conditions are, shot diameter 1.0mm, and impeller revolutions number 3500 rpm. The specimen self-rotates by 60 rpm under the point A of the Fig. 1. The coverage was adjusted by changing shot application time as 3, 36 and 360 sec.

3

Experimental Results and Discussions

3.1

Formulation of Dent Diameter, C

Generally, the dent diameter tends to increase in proportion with increasing shot speed and shot weight. The dent diameter also tends to increase in proportion with decreasing the hardness (HVW) of material applied. Therefore, the dent diameter, C, was formulated like equation (1), where H is density. C = a1 (M V2)1/3 HVW–1/3 + a2 (M/HVW)1/2 + a3 (3M/4p/H)1/3 + a4

(1)

with a1 = 0.074, a2 = 726.728, a3 = 0.049, a4 = 0.011, C: dent diameter (mm), M: weight(kg), V: speed (mm/s), HVW : vickers hardness, p: the circular constant, H: density (g/mm3) The 1st clause is based on the assumption that the volume of dent and deformation resistance (hardness), are proportional to the shot movement energy. The influence of shot weight in the 2nd clause and the influence of shot size in the 3rd clause are taken into consideration. By using the equation (1), partical regression coefficients, a1, a2, a3, and a4 can be obtained from multi-

Figure 3: Relationship between experimental and regression value about dent diameter

Figure 4: Relationship between experimental value and and regression value about roughness

520 ple regression analysis of experiment data. Fig. 3 shows the comparison between the experimental values and regression values. The Fig. 3 shows that equation (1) has sufficient accuracy, as the regression equation of dent diameter. 3.2

Formulation of Maximum Surface Roughness, R

It will be expected that the formulation for the surface roughness is also possible by the similar method of the dent diameter. The same processes can be carried out by using equation (2). Fig. 4 shows the comparison between the regression results and experimental results. It can be said that the regression equation accuracy of roughness is bad compared with dent diameter since coefficient of determination adjusted for the degree of freedom, R*2 is low. It can be assumed that the main reasons be due to relatively high scattering of the surface roughness. R = a1 (M V2)1/3 HVW–1/3+a2 (M/HVW)1/2+a3 (3M/4p/H)1/3 +a4

(2)

with a1 = 0.020, a2 = 2602.19, a3 = 0.726, a4 = 0.104 3.3

Simulation of Coverage

Fig. 5 shows the comparison between the measured coverage and the simulated coverage for application time. The shot peening conditions used are that the shot is 0.77 round cut wire, impeller revolutions number is 3500 rpm, and the amount of shot application is 122,000 pieces/sec. The shot application density under these conditions can be measured about 1.81 pieces /mm2/ sec. The hardness of specimen is 545 HV. The measured coverage can be obtained by measuring paper’s weight of the enlarged photograph shot-peened surface, comparing the shot-peened surface and un-shot-peened surface. When simulated dent diameter by equation(1), the dent diameter can be determined as Æ0.297 mm. The time change of coverage by using the random number series 1 and 2 of MS-DOS BASIC was calculated. The simulations show good relationship with measurements. It can be realized that the actual coverage be simulated by using this method.

Figure 5: The Simulation result of coverage

Figure 6: Relatoinship between coverage fatigue strength

521 3.4

Relationship Between the Coverage and Fatigue Strength

It can be seen from Fig. 5 that the fatigue strength tends to become higher in proportion with increasing application time (coverage rate). The coverage was measured 36 % when shot peened for 3 sec. As for the coverage for 36 sec and 360 sec, it can be presumed as follows. The dent diameter can be calculated as 0.406 mm from the equation (1) and the maximum surface roughness can be calculated as 0.053 mm from the equation (2). Then, under the condition of coverage 36 %, and 3 sec shot application time, the shot application density can be measured as 3.44 pieces/mm2 by experiment. Therefore, the shot application density of 36 sec and 360 sec can be calculated as 41.2 pieces/mm2 and 412 pieces/mm2 respectively. The Coverage of 36 sec and 360 sec can be estimated as 97.9 % and 1000 %, following to the coverage simulation method of Section 2.3. Considering the above, the discussions can be extended to the relationship between the coverage and fatigue strength. Fig. 7 shows the distribution of residual stress. It can be explained from the Fig. 7 that the fatigue strength of coverage 36 % is small because the residual stress is low. On the other hand, since the coverage of 97.9 % and 1000% do not have a difference in the residual stress distributions around the surface which can show great influence on fatigue strength, and the surface roughness also does not show a large difference between the coverage of 97.9 % and 1000 % such as 0.045 mm for the coverage, 97.9 % and 0.048 mm for the coverage, 1000 %, further new parameter can be required to explain the difference of fatigue strength.

Figure 7: Distributon of residual stress

4 1. 2. 3.

Conclusions Applying various shot peening on the plate of hardness 666, 505, and 438 HV, it was found that formed dent diameter and roughness can be expressed by a equation. The method to simulate the coverage by generating a random number, can be proposed. It could be realized that that fatigue strength increases in proportion with increasing coverage.

522

5 [1] [2]

References M. Hirose, Shot peening (1955), P 133, Seibundo Shinko Ltd.. A. Oono, Study on Shot peening processing method of spring material (1985), P5

523

Example of the Computer Simulation of Shot Peening Process Aleksander Nakonieczny1), Pawe » Borkowski2), Pawe» Wymys»owski 2) 1) 2)

1

Institute of Precision Mechanics, Technical University of Warsaw, Warsaw, Poland Institute of Air Technical and Applied Mechanics, Technical University of Warsaw, Warsaw, Poland

Materials for simulation

40 HM steel was taken as the considered material. For the need of simulation of shot impacts at the same position, a cylindrical form of specimen was assumed. In other cases cuboidal specimens were considered. Cast iron shot of the diameter of 0.4 mm thrown with the velocity of 80 m/s was considered for calculations. Material data of specimens and shot assumed for simulation are shown in Table 1. Table 1: Material data of specimens and shot used for simulation of shot peening process

Material data

Shot ST170

Specimen material 40HM

Young’s modulus E [MPa]

2.15 · 105

2.06 · 105

Tensile strength Rm [MPa]

750

1079.0

Yield point Re [MPa]

380

882

Coefficient k

0.09

0.07

Modulus of strain hardening Eu [MPa]

19.35 · 103

14.42 · 103

Poisson’s ratio n

0.295

0.29

Densityr [kg/m3]

7800

7800

Friction coefficient m

0.1

0.1

It was assumed that after the onset of plastic deformation, materials show a decrease of longitudinal modulus of rigidity defined by the coefficient k, and when stresses attain the value of immediate tensile strength Rm material shows an ideal plastic flow.

2

Mathematical models

For determination of the displacements, strains and stresses the Finite Element Method was used for dynamic and static problems in elastic-plastic states with area contact [1]. The search function were the nodal displacements u, from which the individual components of nodal strain and stress vectors were determined. Non-linear equation of motion was solved by NewtonRaphson’s method [1] using at each iterative step the linearized equation of motion [1]:

524

[ M ]{u} + [C ]{u} + [ K ]{u} = {F a }

(1)

where: [M] - mass matrix, [C] – damping matrix [K] – rigidity matrix {u} – nodal displacement vector

{u} – nodal velocity vector {u} – nodal acceleration vector

{F a} – loading vector with boundary conditions (Figs. 1 and 2) and initial conditions for time t = 0 and for individual partial specimen volumes:

{u} = {0}; {u} = {0} for specimen ;

ëê n ûú {u} for shot = -V ;

{u} = {0}

where: |n| – normal vector to the specimen surface V- initial velocity of shot ball Unknown values of nodal velocities un +1 and nodal accelerations un +1 at the end of each iterative step were expressed with the aid of Newmarke’s scheme in function of the known values of un, un and un at the beginning of step and the search value of displacement un+1 at the end of step.

S - symmetry Figure 1: 2-D symmetrical-axis model

Figure 2: 3-D model

525 Two models were used for calculations: •

a two-dimensional, symmetrical-axis model for static and dynamic analyses of shot impacts at the same position on specimen surface (Fig. 1) • a three-dimensional model for dynamic analysis of successive adjacent shot impacts (Fig. 2). In the majority of calculations the damping was neglected because of lack of data. Replacement possibility of dynamic ”transient” type problem by a static one could essentially shorten the calculation time. In the equation of motion (1) it would mean the neglecting of elements related to the velocities and accelerations:

[ K ]{u} = {F a }

(2)

Two possible solutions of the static problem were assumed: • •

forced action as a force acted on the shot ball displacement forced action.

At the beginning, the dynamic two-dimensional, symmetrical-axis model with a single shot ball was calculated. Afterwards from the relevant temporal sub-steps the following values were read off: • •

maximum force value acted on specimen maximum ball displacements from nodes, which were not subjected to plasticizing.

These values were applied in several nodes on the ball surface. Effective plastic strains (6) and stresses according to the Huber-Mises-Hencke’s hypothesis (7) were averaged in nodes. The grow of plastic strains Depl in the equation (5) was determined from the associated flow law (4), where : Q – plasticity potential, l – plasticity coefficient (indexes mean: el – elastic state, pl – plastic state, tot – total, n – step number, ex , exy ,.....- components of strain tensor, n’ – reduced Poison’s ratio)

{de } = l ìíî ¶¶sQ üýþ pl

(3)

{e } = {e } + {De } pl n

pl n -1

pl

(4) 1

2 2 3 2 2 é ù2 2 2 eeq = ê e x - e y + e y - e z + e z - e x + 2 g xy + g yz + g xz ú û 2 1 + v¢ ë

(5)

el Ieq = E eeq

(6)

1

v¢ =

el 1 æ1 ö e eq - ç - v ÷ tot 2 è2 ø eeq

(7)

526

3

Simulation results

The following results concern the dynamic analysis. Values of internal stresses, strains and deformations obtained at static analysis were 1.5 times lower, which proved, that the dynamic problems cannot be replaced by static analysis cases presented above.

3.1

Case of the multiple shot impacts at the same position (2-D model)

For the determination of strains and stresses in case of multiple shot impacts at the same position calculations were made for 13 successive impacts, after which a stabilization of the internal stress state was reached. Each shot ball impinged the specimen surface after the previous ball was rebounded. Damping effect was neglected. Fig. 3 presents the maximum values of effective plastic strain, which grew with the number of successive impacts. Internal stress states in the x direction after first, seventh, ninth and last – thirteenth impact are shown on Fig. 4. It can be noted, that after 13 impacts a practical stabilization of strains was obtained. Internal stress range in the x-direction produced by the successive impacts and its maximum value have grown of about 1.5 times during the simulated impacts from the first to the thirteenth one. 0,055

0,05

0,045

0,04

A eq pl 0,035

0,03

0,025

0,02 0

1

2

3

4

5

6

7

8

9

10

11

12

13

balls num ber Figure 3: Maximum plastic strains A eleq in function of shot balls number

527

1

7

9

13 Figure 4: Residual reduced stresses Ix [MPa] after first, seventh, ninth and thirteenth shot ball impact

3.2

Case of the successive adjacent shot impacts (3-D model)

Calculations showed, that the maximum values of internal stresses occurred at the depth of 0.0258 mm beneath the specimen surface. Initially, a model with balls impinging specimen surface with the space between impacts equal to ½ of ball diameter d was calculated. Then, model with additional impacts at the half of distance between the ball-marks from previous case was calculated. Variations of those stresses are presented on Figs. 5 and 6, which show the half of specimen cross-section in XZ plane. After last impact, there is a displacement of the area where

528 A

Ix

C

B

0.0259 m m d istan ce fro m su rface 0

0,05

0,1

0,15

0,2

x [m m ]

0,25

0,3

0,35

0 -100 -200

[MPa]

-300 -400 -500 -600 -700 -800

Figure 5: Internal stresses Ix in the specimen cross-section with XZ plane after impingements of specimen surface with shot balls A, C and B at points x = 0,1 and x = 0,2 mm

I

C

A

B

y

0 .0 2 5 9 m m d is ta n c e fr o m s u r fa c e x [m m ] 0

0 ,0 5

0,1

0 ,1 5

0 ,2

0,25

0,3

0,35

0,4

100

0

-100

[MPa]

-200

-300

-400

-500

-600

-700

Figure 6: Internal stresses Iy in the specimen cross-section with XZ plane after impingements of specimen surface with shot balls A, C and B at points x = 0,1 and x = 0,2 mm

maximum values of internal compressive stresses occur. It can be seen, that concentration of impacts results in distinct equalization of the stress state.

4

Conclusions and Final Remarks

By using the simulation results it is not possible to completely define the surface state after shot peening. This is caused by insufficient numbers of shot impacts for 3-D model and by shot im-

529 pact concentrated at one position only for model 2-D. The solution can have only a qualitative character. Calculations show, that the maximum values of residual stress occur in subsurface region of the specimen. Plastic zone has the range of ½ of ball radius. Results obtained from the presented example of application of simulation calculations confirm principally the assumptions consolidated on the base of practical experience. It was confirmed that the maximum compressive stresses occur at the depth corresponding with so-called Bielajev point. It was also found that in the examined case a stabilization of internal stress and strain level is obtained after a relatively small number (13) of shot impacts at the same place. That level is approximately 1.5 times higher than level obtained after first shot impact. Successive shot impacts at the proximity to the previous impact points lead to an equalization of the stress state in the sphere of individual material layers. So, the results obtained in the present work, which has only a recognizing character, didn’t bring many new elements to the practical knowledge concerning the shot peening process. However, its practical value consists in fact, that the elaborated calculation method makes possible to take into consideration the data variations relative to the treated material, shot and process parameters. Thus, it is possible to use this method for optimization of process parameters and for forecasting of internal stress states, even with an admission of simplified shot peening models. It is an important information, necessary for many attempts to predict the durability of shot peened machine parts at the assumed work conditions [5].

5 [1] [2] [3] [4]

[5]

References A. Nakonieczny: MOCIP nr 127-129 (1994). S. T. S. Al.-Hassani, K. Kormi, K. Webb and D. C. Webb: Proc. of the 7th ICSP (ed.: A. Nakonieczny), IMP Warsaw (1999), pp. 217 - 227. Analysis Theory Reference 000855 Eight Edition, CAS IP, Inc. P. Borkowski, P. Wymys»owski – Appendix to Report nr 12.4.01.117.1, Division of Materials Strength of the Department of Mechanics, Power Engineering and Aviation, Technical University of Warsaw (2001). P. Mczy½ski, A. Nakonieczny, P. Borkowski, P. Wymys»owski, S. Janowski: Report nr 12.4.01.117.1, Division of Materials Strength of the Department of Mechanics, Power Engineering and Aviation, Technical University of Warsaw (2001).

530

Modeling of Fatigue Behavior due to Shot Peening Conditions Marsha K. Tufft GE Aircraft Engines, Cincinnati, OH, USA

1

Introduction

The beneficial effects of shot peening have long been recognized. One of the major reasons for shot peening is to induce a beneficial surface condition (compressive stress layer and altered microstructure) that acts to retard the development and propagation of surface cracks. If surface crack formation and propagation can be suppressed, longer component operating lives can often be attained. Dörr and Wagner [1] demonstrated that shot peening was effective in retarding crack propagation of existing cracks, even when peening was applied after the development of cracks. Luetjering and Wagner [2], and others have recognized, however, that shot peening can also cause the equivalent of fatigue damage. This effect has received considerably less attention. There is increasing interest in methods to predict life capability of shot peened parts, and in the use of models that enable a designer to select a robust level of shot peening that will optimize the life benefit, minimize manufacturing costs, and avoid potential life degradation from “overpeening“. This paper examines six different approaches to assessing shot peen impact on life. Four approaches focus on fatigue crack initiation life, or rather the life to failure in the absence of preexisting cracks. One method deals with crack propagation life. The final approach attempts to correlate surface residual stress state with residual life remaining at the time of inspection. Of these six approaches, only two offer general predictive tools: one of fatigue initiation life, the other of crack propagation life. The other methods provide alternate ways for analyzing and using specific fatigue and/or residual stress data. This paper examines some of the challenges and limitations in using each of these methods. Where possible, these methods are demonstrated using data from a shot peening Design of Experiment (DOE) conducted on Rene’ 88DT, a nickel-base superalloy, as documented in references [3, 4, 5, 6]. It must be noted that life prediction methods are engineering attempts at modeling complex physical processes, and will therefore always be limited by inadequate understanding and inability to model the significant elements of physical reality. All models are wrong – by definition they are approximations at best – but some are useful. The most useful are substantiated by data covering the relevant conditions of interest. One must be cautious when trying to apply a model to conditions outside the validated set of conditions – physical reality is often complex and nonlinear, and does not always cooperate with attempts at extrapolation.

531

2

Challenges and Approaches to Modeling Life Behavior

Perhaps the greatest challenge with any life prediction method is obtaining the data necessary to generate a useful model. Life behavior can be affected by many factors: • • • •

• •

operating conditions (“mission” – including stress & temperature profile, minimum, maximum and mean stresses experienced, and time or duration at specific conditions) operating environment (air, vacuum, salt water, etc.) geometry (including resulting stress concentrations, stress gradients and stress state) surface condition – topography & microstructure (low stress ground & polished, turned, broached, reamed, shot peened, etc. & various combinations; each of these processes may have additional specific parameters which need to be defined in order to adequately characterize the surface state) residual stress state (tensile or compressive stress at surface, maximum stress magnitude and depth of residual stress layer) material (chemistry, processing method – cast or wrought, heat treatment, microstructure).

A variety of methods are reviewed here. Some are very specific and are able to incorporate the details of operating conditions, geometry, and specific surface condition since they are derived from specific component or test data. Others can be used as more general predictive tools, but will generally result in reduced correlation with specific data since they are not as well adapted to assessing the variety of factors which can affect component life.

2.1

Weibull Analysis of Data at Specific Test Conditions

One life prediction approach is to fit replicate specimen fatigue test data collected at specific test conditions to statistical distribution functions. The Weibull distribution [7] is often used to analyze failure data including analysis of complete systems or components. When used for analysis of actual component failures it is particularly powerful because the specific details of operating conditions, geometry, surface condition & material can be accounted for. The cumulative distribution function is given as: F t = 1 - e

- ëé t - t0 / hûù

>

(1)

Figure 1 shows the results of a Weibull analysis for 5 different shot peen populations compared against a low stress-grind & polish (LSG+P) baseline [3], while Table 2 summarizes the minimum and median lives for each population. Note that the 50 % line gives the median life. A – 3I life would correspond to a CDF value of 0.135 %. Shot peening behavior observed fell into three categories: 1) good – life benefit, 2) low – life degraded, and 3) transition – some low and high life results at the same peening condition. If one ignores the transition group, all other populations exhibit a high slope characteristic of a “rapid wear out” mode. This suggests that shot peening reduces the crack initiation time (by accumulating plastic strain, which is equivalent to fatigue damage); however it also increases the crack propagation life due to the beneficial surface state imparted.

532

Figure 1: Weibull Analysis Results (1000 °F, stress level chosen to give approx. 100,000 cycle nominal life.)

Table 1: Two parameter Weibull analysis results of shot peen DOE data symbol shot

group

SCALE (average life)

SHAPE Interpretation of (failure SHAPE factor mode)

Minimum Median Life Life (~.135 %) (~50 %)

155,845

12.80

rapid wear out

93,000

CCW14 B – transition 102,808

2.13

LCF (mixed modes) 4,600

87,000

CCW14 C – low

25,891

7.71

rapid wear out

25,000

CCW31 D – all

146,672

17.48

rapid wear out

100,500

143,000

CCW14 E – light pn.

220,353

10.01

rapid wear out

114,000

212,000

CCW14 A – good

11,000

151,000

Compared with LSG+P data, shot peening resulted in tighter scatter of low cycle fatigue results. For shot peening conditions that resulted in “robust” life conditions – groups A, D and E, the increase in the minimum life capability is perhaps more dramatic and significant than the improvements to median life capability. For this particular example, a minimum life of approxi-

533 mately 17,000 cycles would be predicted for LSG+P specimens, and a median life of ~100,000 cycles. By comparison, group D (CCW31 shot peened specimens) resulted in a minimum life of ~100,500 cycles and a median life of 143,000 cycles. However, group C (CCW14, overpeening conditions) resulted in a minimum life of 11,000 cycles and median life of 25,000 cycles. To summarize, a Weibull analysis can be a powerful tool for evaluating life capability of homogeneous populations. That is, for data sets with similar peening conditions, geometry, operating conditions, etc. Very useful life predictions can be obtained with this approach. The drawback is that very specific data are needed. This is more useful as a field management tool, rather than as a design tool for selecting component shot peening conditions.

2.2

Fatigue Crack Initiation Analysis

Fatigue behavior is often broken down into two phases: fatigue crack initiation and fatigue crack propagation. Initiation assessment often relies on the use of S-N (stress vs. life) curves, generated by statistical analysis of test data for the material of interest over a range of stresses at specific temperatures. A “minimum” life curve is used to establish safe operating limits for critical components whose failure might present a safety concern. However significant changes in process capability can often be assessed by comparing average life behavior. So, both average and minimum life capability are of interest. Due to the variety of factors affecting fatigue behavior, it is very difficult to come up with a single analytical method to assess life impact due to a variety of shot peening conditions. Fatigue life correlation and comparative assessment can be attempted for specific conditions. 2.2.1

Modified Goodman Formula with Surface Roughness Measurement

A Modified Goodman Formula approach developed by Li, Mei, Duo and Wang [8] provides a fairly general predictive model. This method requires representative surface roughness measurements to formulate an appropriate stress concentration factor (Kt), as well as knowledge of the residual stress state and information about baseline material life behavior. It can be applied only to materials that demonstrate an endurance limit. Extensive 3D analysis was used to relate a geometric stress concentration factor (Kt) as a function of specific surface roughness parameters, Rt (peak dimple depth) and S (dimple spacing). As a result of this work, an appropriate Kt can be generated for any peening condition for which the appropriate surface roughness parameters can be obtained. Surface roughness measurements are non-destructive and relatively easy to obtain. Finally, a modified Goodman formula was used to predict life, incorporating the residual stresses as a mean stress effect and the Kt as a stress multiplier. The Goodman relation can only be used for materials that exhibit a fatigue strength. Not all materials exhibit this behavior. Further, it is not clear whether this method works over a wide range of peening conditions, from damaging to beneficial. Since Rene’ 88DT does not exhibit a fatigue strength, it was not possible to apply this method to the reference data set documented in references [3-6].

534 2.3

Analyzing Fatigue Data to Define Robust Process Windows

In some cases, the goal is not to quantify the life capability due to shot peening, but to ensure that the process window defined for a particular component is robust and does not result in degraded life capability. Designed Experiments and Damage Maps are two methods that have been used as simple models to identify parameter settings which result in acceptable life behavior. 2.3.1 Analysis of Variation (ANOVA) Approach Although less useful as a predictive fatigue life model, designed experiments (also called DOE’s) using ANOVA methods can be used to evaluate life impact due to a variety of process conditions at a selected test condition. In fact, the strength of these methods is in their ability to efficiently identify the few significant factors from the many potential factors. However, resources are generally limited and are therefore focused on assessing the process parameters at an operating condition selected to meet design needs. So, the goal is normally to define and validate a robust process window, not to develop and validate a life prediction tool over the range of operating conditions that a component might see in actual operation. For analysis of life capability as a function of various process parameters, the use of a normalized life parameter, “stdev” can be very useful:

é log N obs - log N avg ù û stdev = ë é log N avg - log N -3s ù / 3 ë û

(2)

Here, Nobs represents the observed life at failure of a specific test bar. Navg represents the average (median) life for the stress and temperature condition for LSG+P data, and N–3I represents the minimum LSG+P life. As a result, |stdev| > 3 indicates test results which are very uncharacteristic of the average LSG+P population. Approximately 68 % of data points should be within |stdev|<1, while 95 % should be within |stdev|<2, and 99.7 % should fall within |stdev|<3. Table 2 summarizes the significant factors identified by the Rene’ 88 DT study [3]. The DOE data provide evidence of significant interactions between peening parameters. A total of nine effects, including all four main effects, 3/6 two-way interactions, 1/4 three-way interactions and the single four-way interaction were found to be significant at the 95 % confidence level. When multiple factor interactions become significant, this indicates that one or more of the factors does not produce the same trend in life behavior over all levels of the other factors. This is illustrated in the two-way interaction plots in Figure 2 (a-c), using the normalized life parameter, ‘stdev’. From these plots, it can quickly be seen that CCW31 shot produced uniformly good life results over the range of peening conditions evaluated by the study. By comparison, intensity, % coverage and incidence angle all exhibited significant impact on life capability of specimens peened with smaller CCW14 shot.

535 Table 2: ANOVA summary of shot peen DOE results #

Factor

Pr > F

1

shot

0.0001

2

shot*coverage

0.0003

3

coverage

0.0005

4

shot*intensity

0.0015

5

intensity

0.0018

6

shot*incidence angle

0.0038

7

incidence angle

0.0161

8

shot*intensity*angle*coverage

0.0446

9

intensity*angle*coverage

0.0498

“stdev”

Main Effects & Interactions which are significant at the 95 % confidence level. (Pr < 0.05 are significant.) Normalized lives analyzed. Arcsine transformation used to reduce scatter in residuals: arcsine(stdev/6). 1

1

1

0

0

0

-1

-1

-1

-2

-2

-2

-3

-3

-3

-4

-4

6

10

“A” Intensity (mils) CCW14

CCW31

a) Intensity x Shot Interaction

-4 100%

800%

% Coverage CCW14

CCW31

b) Coverage x Shot Interaction

45

90

Incidence Angle (degrees) CCW14

CCW31

c) Angle x Shot Interaction

Figure 2: Plots of significant two-way interactions from DOE. ‘stdev’ plotted on Y axis

Although many factors were identified as being significant to life behavior, this analysis does not shed any light on the physics behind the behavior. Results from a separate investigation [4, 5, 6] suggest that shot velocity is a critical physical parameter which can trigger a change in life behavior for Rene’ 88DT. Unfortunately, shot velocity measurements are difficult to make and are not a standard part of the shot peening process controls used today. But this illustrates the potential for lurking variables that cannot be directly controlled or evaluated in some DOE’s. 2.3.2 Damage Map Approach Designed experiments permit powerful and efficient statistical analysis of the data. However, they require planned experiment designs. “Messy” or “happenstance” data can not be analyzed using ANOVA methods. An alternative is to plot the data as a function of the parameters of interest, using symbols to distinguish “good”, “neutral” or “low” life behavior. Figure 3 shows an example.

536

% Coverage

800%

400%

100%

Consistent above average life behavior

4

Consistent below average life behavior

Transitional behavior – additional factors that have not been identified on this map (incidence angle, velocity) may be influencing behavior

stdev > +1 (above average) -1 < stdev < +1 (~average life)

stdev < -1 (below average)

6

8

10

Intensity (mils “A” scale) Figure 3: Damage map illustrating interpretation of CCW14 data from DOE [4]

It is easier to miss significant factors using this approach. However, it can be used to help evaluate messy or happenstance data, and to select test conditions of interest for future designed experiments.

2.4

Predicting Peened Surface Fatigue Crack Propagation Behavior

Fracture mechanics (FM) is a well-established discipline that enables life predictions when initial cracks are known to exist. The effect of shot peening can be incorporated into FM analysis by modeling the residual stress profile and using the principles of linear superposition via Green’s functions. This affects the mean stress intensity factor. It is important to calibrate the predictions against test results to better account for stress relaxation effects due to thermal and cyclic loading. Residual stresses are readily measured using x-ray diffraction techniques. If desired, the raw data can be curve-fit to provide a more continuous definition. A functional form that appears to work well for many materials, is shown in equation (3) and illustrated in Figure 4.

s RS x = A × exp [ - x / l ] × sin B × x + C

(3)

Prior work shows that the depth of the compressive stress layer is nearly a linear function of shot peen intensity [6]. Variations in shot peen process parameters which result in significant changes to residual stress profiles can be accounted for in a FM analysis by modeling the different profiles. FM calculations using a variety of Rene’ 88DT shot peen profiles showed very minor differences for small changes in intensity, or variations of incidence angle and % coverage over a small intensity range. For example, data from 6 to 8A peening conditions could be regressed as one population. Data from 6 to 12N could also be regressed as a single population, having a slightly different benefit than a 6-8A model. FM provides a general predictive tool that is appropriate when pre-existing cracks are known to exist, and for assessments that assume the presence of an initial crack. It is capable of mode-

537 50 0 0

0.005

0.01

0.015

-50 -100 data

-150

curve fit -200 Depth (inches) Figure 4: Sample residual stress profile and corresponding curve fit, peening condition: CCW31 shot, 6A intensity, 45 degree incidence angle, 800 % coverage

ling complex stress & temperature missions, as well as a variety of stress gradients, component and crack geometries.

2.5

Residual Life Prediction as a Function of Surface Residual Stress

Bradley, Berkley, and Fairbank [9], and Berkley [10] describe the use of surface residual stress measurements (taken non-destructively using x-ray diffraction) to correlate with residual life of aircraft engine components. This work is still in the development stages. This approach assumes that the component surface is initially in a state of compression, due to shot peening or other surface treatment, and that the residual stress condition of the component declines toward tension with service cycles and / or hours. One of the concerns with this approach is that x-ray diffraction surface residual stress measurements show a high level of variation. They are also subject to errors caused by the presence of a sub-surface stress gradient as well as difficulties in interpreting surface results [11]. Another question is whether the surface stress relaxation is significant, and correlates with % life consumed. Figure 5 shows some normalized near surface residual stress profiles that have experienced static thermal exposure or cycling at elevated temperature. Figures 5a-c were taken from specimens made from Rene’ 88DT, a nickel-base superalloy, while 5d profiles were taken from Marage 250, a steel. Figure 5a shows that progressive relaxation occurs as strain range is increased at a given temperature on Rene’ 88DT. No clear trend of relaxation as a function of cycles was observed. In fact, compressive residual stresses increased slightly for some cases. This is probably within the normal scatter of residual stress measurements for this peening condition. Figure 5b shows that progressive relaxation of residual stresses also occurs with increasing temperature for Rene’ 88DT, even in the absence of cyclic loading. A greater difference in relaxation behavior was observed as a function of time at 1300 F than was observed at lower temperatures. As it appears that thermal exposure can accomplish the same amount of surface stress

538 relaxation as cyclic loading without any consumption of life, this should complicate the use of surface residual stress as a reliable correlation to residual life. Figure 5c shows residual stress profiles taken from LCF specimens after failure. These specimens were all peened to the same damaging shot peening condition. Two specimens had small amounts of the damaged surface layer removed by electropolishing. All three specimens were then cycled to failure. Residual stress measurements were taken on the remaining gage section a small distance away from the fracture surface. Although surface residual stresses did go tensile for the two electropolished specimens, the third specimen exhibited compression on the surface even after failure. This suggests that surface residual stress may not be a reliable or robust indicator of residual fatigue life. Finally, Figure 5d illustrates the thermal stability of residual stress profiles in Marage 250 as a function of thermal exposure temperature. Unlike Rene’ 88DT, very little variation in residual stress profiles was observed over a range of temperatures and peening conditions. This demonstrates the importance of obtaining data for the specific material, shot peening conditions and operating conditions of interest. Although this methodology would provide a unique approach to damage monitoring, there are several questions about the robustness and general applicability of the approach. It appears that much more validation is needed. 0.2

0.2

0

0 0

0.5

1

1.5

0

2

0.2

0.4

0.6

0.8

-0.4

75F-0.00%-0 cycles 1150F-0.40%-200 cycles

-0.4

-0.6

1150F-0.55%-200 cycles 1150F-0.70%-200 cycles

-0.6

-0.8

1150F-0.85%-200 cycles 1150F-0.55%-4000 cycles 1150F-0.70%-4000 cycles

-0.8

-1

-1

a) Rene’ 88DT: CCW31 – 7A/45/100% - 200 to 4000 cycles at 1150F

b) Rene’ 88DT: CCW31 – 7A/45/100% - 5 to 100 hours thermal exposure 0.2

0.4 0.2

0 0.0

0.0 0.0

0.5

1.0

1.5

2.0

0 mils removed

-0.8

1 mil removed

-1.0

1.5 mils removed

-1.2

0.2

0.4

0.6

0.8

1.0

1.2

-0.2 -0.4

-0.4 -0.6

1.2

75F-5 hrs 1000F-5 hrs 1150F-5 hrs 1300F-5 hrs 75F-100 hrs 1000F-100 hrs 1150F-100 hrs 1300F-100 hrs

1150F-0.85%-4000 cycles

-0.2

1

-0.2

-0.2

-0.6

75F

250F - 16 hours

400F - 16 hours -0.8

600F - 16 hours

-1

c) Rene’ 88DT: CCW14 – 10A/45/800% - cycled to failure

d) Marage 250: S110 – 10A/45/800% - 16 hours thermal exposure

Figure 5: Normalized residual stress profiles illustrating the effects of a variety of thermal exposure & cyclic loading conditions. Stress (Y axis) normalized by reference stress. Depth (X axis) normalized by shot peen intensity.

539

3

Conclusions

There presently appears no single robust process for fatigue life prediction of shot peened components, nor is development likely, given the differences in material behavior and ranges of use. However there are a number of approaches and tools that can be used to assess the influence of peening on life, depending on the specific need. All methods and tools rely heavily on test data to generate the model elements and/or validate their use.

4 [1]

References

T. Dörr and L. Wagner: 1996. “Effect of shot peening on residual life of fatigue pre-damaged 2024 Al.” Sixth International Conference on Shot Peening, Oct. 1996, pp. 174-183. [2] L. Wagner and G. Luetjering: 1981. “Influence of shot-peening on the fatigue behavior of titanium alloys” First Int’l. Conf. on Shot Peening, Sept. 1981, Pergamon Press, pp. 453460. [3] M. K. Tufft: 1999. “Shot peen impact on life, part 1: designed experiment using Rene’ 88DT,” Shot Peening Present & Future, Institute of Precision Mechanics, pp. 244-253. [4] M. K. Tufft: 1999. “Shot peen impact on life, part 2: single particle impact tests using production shot,” Shot Peening Present & Future, Institute of Precision Mechanics, pp. 254263. [5] M. K. Tufft: 1999. “Shot peen impact on life, part 3: development of a fracture mechanics / threshold behavior predictive model,” Shot Peening Present & Future, Institute of Precision Mechanics, pp. 264-273. [6] M. K. Tufft: 1997. “Development of a Fracture Mechanics / Threshold Behavior Model to Assess the Effects of Competing Mechanisms Induced by Shot Peening on Cyclic Life of a Nickel-base Superalloy, Rene’ 88DT.” Ph.D. Dissertation, University of Dayton, Dayton, Ohio. [7] R. B. Abernethy: 1993. The New Weibull Handbook. Published by Robert Abernethy, 536 Oyster Road, North Palm Beach, FL 33408-4328. [8] J. K. Li, Yao Mei, Wang Duo, Wang Renzhi: 1992. "An analysis of stress concentrations caused by shot peening and its application in predicting fatigue strength," Fatigue and Fracture of Engineering Materials & Structures, Vol. 15, No. 12, Dec. 1992, pp. 12711279. [9] E. F. Bradley, S. G. Berkley, R. J. Fairbank: 1996. “The Role of Residual Stress in Low Cycle Fatigue of Gas Turbine Engine Disks.” San Antonio ASIP Conference. [10] S. G. Berkley: 2001. Presentation given at Aeromat 2001, 12th Aeromat Conference & Exposition, Long Beach, CA. [11] Paul Prevey: 1991. “Problems with non-destructive surface x-ray diffraction residual stress measurement.” Practical Applications of Residual Stress Technology, ed. C. Ruud, Materials Park, OH: American Society for Metals, pp. 47-54.

540

Finite Element Impact Modelling for Shot Peen Forming Tao Wang, Jim Platts University of Cambridge, Cambridge, UK

Andrew Levers Airbus UK Ltd, Chester, UK

1

Abstract

A finite element analysis of multiple random impacts involved in shot peen forming is presented. An explicit dynamic algorithm for modelling up to 1000 impacts and a static algorithm for spring-back simulation are combined to achieve a final curved shape after shot peening on a small sized aluminium 2024-T351 sample. The development of the plastic zones, residual stresses, and the final deflection is given by simulating the experimental conditions. The comparison between simulation results and experiments shows that the finite element impact modelling is able to investigate the macroscopic effects (e.g. curvature) of shot peening as well as the microscopic effects (e.g. local plasticity and residual stresses).

2

Introduction

Shot peening forming involves numerous randomly distributed impacts on the surface of metal sheet. The individual plastic zones created by each impact have a tendency to expand sideways and combine to curve the peened metal sheet and to induce residual stresses in it. For the understanding of peening mechanics, previous numerical work has progressed from the finite element simulation of single indentation or impact to multiple in-line or regularly distributed impacts. Follansbee and Sinclair [1], and Sinclair et al [2] investigated the quasi-static indentation of an elastic-plastic half-space by a rigid sphere using an axis-symmetric finite element method. Meguid and Klair [3] modelled a plate co-indented by two smooth, flat, and rigid punches under plane-strain condition to investigate the interaction between two plastic zones. Grasty [4] modelled nine uniformly distributed indentations by 3D finite elements and extrapolated the deflection to be compared with experiments. Al-Obaid [5] developed a three-dimensional dynamic FE program to determine the residual stress and plasticity of the peened material by applying a dynamic patch load distributed on the target surface. Levers and Prior [6] suggested that the extrapolation of single shot impact to a realistic peening process is difficult and the dynamic explicit FEM for multiple impact modelling is attractive in its efficiency and application to practical peening processes. Nevertheless, they found that the explicit method is truly dynamic so that the model can not reach a state of static equilibrium by simply extending the computation time. Therefore, they finally introduced a temperature profile approach based on shell elements to model the residual stress profile through thickness and hence the corresponding macroscopic deflection. None of these investigations, however, has dealt with the problem of randomly distributed impacts, which are physically involved in shot peening processes. It is also worth noting that the

541 investigation of peen forming processes has concentrated on the macroscopic effect of shot peening by neglecting individual impacts based on shell assumption [6~10]. However, if peening mechanics is concerned, a 3D FEA is preferred because the normal impact pressure is actually the most important load to induce the resulting effects. Particular attention was therefore devoted to two steps: (1) Verify the feasibility of combining dynamic and static FE algorithms together for modelling numerous randomly distributed impacts in peen forming to calculate the final curved shape. (2) Obtain the overall trend of the development of peen formed curvatures, plasticity, and residual stresses.

3

The Experiment

Aluminium 2024-T351 sheets are used for this study. Square samples 20 × 20 mm with 4mm nominal thickness are chosen because the distribution of peens within this surface area can be approximated as a uniform distribution for the peening machine being used. However, for larger surface areas a Gaussian distribution was expected by Holdgate [11]. Cast steel S660 peens are used in an air-blast machine. The mass flow rate is kept as a constant 13.67 g/s and the air pressure is adjustable from 20 psi to 50 psi and exposure time from 10 s to 25 s. The nozzle is kept vertical to the worktable at a distance of 300 mm. Unlike Almen tests, specimens are not constrained by a holder during peening. Instead, only adhesive materials (e.g. tapes) around the test pieces are used to prevent the specimen from moving laterally due to air blast. Therefore, the specimen is unconstrained and is free to bend and elongate. After peening, the specimen is cold mounted in an acrylic material to preserve the deformed configuration. It is then cut through thickness near the initial diagonal of the square surface in a precision cutting machine. The section is finally ground down to a grit size of 1200 for the measurement of deflection and micro-hardness in a micro-indentation machine. The curvature is determined by approximating the measured points with a second order polynomial function. A convenient method to measure the curvature is to measure a number of points on the unpeened bottom surface by a coordinate measurement machine and to approximate these points by a sphere. These two methods give similar results. When measuring the deflection by both methods, measurement points are taken at an approximate distance equal to the sample thickness away from the edge.

4

The Finite Element Analysis

The finite element package ABAQUS is used to simulate the procedure corresponding to the experimental operation. Because of its efficiency, the Explicit dynamic algorithm is used to simulate the numerous impacts. However, as stated above, since a static solution is necessary, the Standard static algorithm is combined to provide the resulting deformed shape as a spring-back analysis. Steel S660 shots are assumed as spherical rigid surfaces since they are harder than the aluminium target sheet. In addition, the analytical rigid surface gives a good computational efficiency. Each shot is associated with a mass element and a rotary inertial element, which are calculated according to the density of the shot and its spherical volume. The shot density H is given by 7500 kg/m3 and the shot radius R is 0.968 mm as the average of those measured by a number of

542 randomly selected shots. The simulated velocity v is from 21~35 m/s estimated for the experimental machine [11] under the given air pressure and mass flow rate. The velocity is assumed to be in the vertical direction and applied to each shot as an initial condition. No other translation and rotation velocities are applied to shot. The aluminium 2024-T351 target sheet is modelled by only a quarter to save computation cost. The elements used are 3D continuum elements C3D4. At least 9 elements are arranged through the thickness. The element dimensions in the target plane are uniform 0.3 × 0.3 mm. 0.2 × 0.2 mm dimensions are also used for detailed coverage investigation. The material property of the sheet is defined by its uni-axial tensile test, regardless of the rolling direction and the dependence on strain-rate. Its stress-strain relationship is approximated by a power function. Corresponding to the unconstrained peening in experiments, a rigid surface is arranged under the target sheet to provide support. The friction coefficient between them is assumed as 0.4. A program is written to create 3D uniformly random distributed shots over the top surface of the sheet to model the shot stream. Up to 1000 shots can be simulated in the current study. The friction coefficient between each shot and the sheet is also assumed as 0.4.

5

Results and Discussion

5.1

The Development of Curvature

The impact velocity is determined by measuring the indentation diameter on 2 mm thick aluminium 5251-H22 sheets after peening to a low coverage. The analytical results in Wang[12] and experimental results for the same material in Holdgate [11] can be used to estimate the impact velocity. The estimated velocity is very similar to the one measured for a sister machine at Airbus UK [13]. In addition, in order to compare the macroscopic deflection from finite element model with experiments, the analysis time in FE analysis needs to be calibrated. Because the real time is too long for an explicit analysis, the FE analysis has to be accelerated by arranging one impact immediately after the other. Thus, the calibration of the FE analysis to experiments is needed. The calibration is conducted by comparing the curvature results from both FE analy2.5

3.5

40 psi peening FEM 31m/s 20 psi peening FEM 21m/s

Test (15s peening) FEA

2

2nd Poly of FEM

2.5

Curvature (1/m)

Curvature (1/m)

3

1.5

2 1.5 1

1

0.5

0.5 0

0 10

20 30 Impact velocity (m/s)

40

Figure 1: The comparison of FE analyses with experimental results for 15 s peening

0

10 20 Exposure time (s)

30

Figure 2: The comparison of FE analyses for 21 m/s and 31m/s impacts with experimental results for 20 psi and 40 psi peening

543 ses and experiments assuming the velocity determined by peening pressure is accurate. This gives a calibration factor of the FE analysis time to the real exposure time for the model being used, which is a function of the impact velocity. Based on the calibration, the comparison of FE analysis with experimental results is shown in Figure 1 for 15 s peening under different peening pressure and in Figure 2 for 20psi and 40psi peening under different exposure time.

5.2

The Development of Coverage

With the finite element model, the plastic coverage [14] can be calculated by analyzing the plastic region. However, to determine the plastic coverage is much more difficult than the visual coverage using experimental methods. The development of equivalent plastic strains on the impact surface for 31m/s impacts is given in Figure 3 from the finite element analysis. The minimum equivalent plastic strains are set to 7 · 10–3, which is approximately 1 tenth of the representative strain 0.2 · a/R [15] The black region indicates a high-plasticized zone and the white indicates an approximated non-plasticized zone. The surface plastic coverage was calculated by an image-processing program, which calculated the ratio of the plasticized area to the total area. The specific values are given under their corresponding diagrams. The development of the plastic coverage Cpr (FEA) together with the calculation of the visual coverage ratio [11, 16] is shown in Figure 4.

Figure 3: The development of the plastic coverage

544 100%

60% 40%

Depth (mm)

Coverage ratio

80%

Plastic Coverage Ratior (FEA)

20%

Visual Coverage Ratio

0% 0

100

200 300 Im pact num ber

400

0 -0.2130 -0.4 -0.6 -0.8 -1 -1.2 -1.4 -1.6 -1.8 -2 -2.2

Microhardness HV 150

170

190

15s(Equivalent FEA 600 Impacts) 20s(Equivalent FEA 800 Impacts)

results of microFigure 4: The comparison of the plastic coverage ratio with visual coverage ratio

Figure 5: Experimental results of micro-indentation for 40 psi peening

The experimental microhardness values through the thickness of the peened sample are shown in Figure 5. The depth is measured from a local top of the peened surface under a 40× microscope. The FEA results of the through-thickness plasticity along the diagonal section is shown in Figure 6 for 100~800 impacts (31m/s) in the undeformed configuration for a clear view of the thickness. Both FEA and tests give an estimation of the plastic depth about 1~1.2 mm. It is worth noting that the analysis for a single impact [12] gives an estimation about 0.97 mm, which is not considerably changed by multiple impacts.

5.3

Residual Stresses

The development of the residual stress Ix along the diagonal section is shown in Figure 7. The white colour indicates tensile stress field. The distribution of Iy has a similar contour so it is not shown here. It is interesting to note that for a partial coverage (200 impacts, coverage ratio = 50~60 %) a large percent of the impacted surface is actually under a compression state. A compressed layer is clearly shown in this figure and it can be compared with the plastic layer shown in Figure 6.

6

Conclusions

The applicability of impact modelling to investigate the mechanics of shot peen forming on a small sized sample is demonstrated through the above analysis. The method of combining the dynamic explicit and static algorithm for a peen forming purpose is able to investigate the macroscopic effects of shot peening as well as microscopic effects. However, previous numerical work on impact modelling concentrated on microscopic effects, such as the residual stress and local plasticity. In fact, Kopp and Ball [17] argued that the quantitative calculation of residual stresses by axis symmetric single indentation or impact model is of little practical use and suggested further multiple impact modelling.

545

Edge

Edge 0

Centre

4

4

3

3

2

2

1

1

0

Centre

4

4

3

3

2

2

1

1

0

0

(a) 100 Impacts

(a) 100 Impacts

(b) 200 Impacts

4

4

3

3

2

2

1

1

0

0

(b) 200 Impacts

4

4

3

3

2

2

1

1

0

0

(c) 600 Impacts

(c) 600 Impacts

(d) 800 Impacts

Figure 6: The development of the throughthickness plasticity

(d) 800 Impacts

Figure 7: The development of residual stress Ix

As Meguid [14] envisioned that the plastic coverage develops faster than the visual surface coverage, the finite element simulation gives a firm proof to his suggestion. In addition, a layer with compressive residual stresses develops as the plastic layer is formed. These discoveries can therefore be used to explain that partial-coverage peening to some extent could also enhance the fatigue strength of metals. The layered structures, such as the plastic layer and the compressed layer, are clearly shown by the present FEA. Like Grasty’s squeezed-layer model [4], a more efficient model by applying layered plastic deformation using shell elements for realistic sized components of peen forming is being considered.

546

7 [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17]

References P. S. Follansbee & G. B. Sinclair, Int. J. Solids Structures, 1984, 20, 81–91 G. B. Sinclair, P. S. Follansbee & K. L. Johnson, Int. J. Solids Structures, 1984, 21, 865–888. S. A. Meguid & M. S. Klair, J. Mech. Work. Tech., 1985, 11, 87–104. L. V. Grasty, PhD Thesis, Cambridge University, 1992. Y. F. Al-Obaid, Computers & Structures, 1994, 52, 693–703. A. Levers & A. Prior, J. Mat. Proc. Tech., 1998, 80, 304–308. R. D. VanLuchene & E. J. Cramer, J. Mat. Eng. Perf., 1996, 5, 753–760. X. M. Kang, T. He, Z. E. Ma & Y. M. Kong, J. Aero. Manufacturing Eng., 1997, 6. D. S. Gardiner & M. J. Platts, ICSP7, Institute of Precision Mechanics, Warsaw, 1999, 235–243. D. S. Gardiner, PhD Thesis, Cambridge University, 2001. N. M. D. Holdgate, PhD Thesis, Cambridge University, 1993. T. Wang, Internal Report, Cambridge University Engineering Department, 2001 A. Levers, Personal communication, 2001 S. A. Meguid, Fatigue & Fracture of Engineering Materials & Structures, 1991, 14, 515–530. D. Tabor, The Hardness of Metals, Clarendon Press, Oxford, 1951 D. Lombardo & P. Bailey, ICSP6, San Francisco, USA, 1996, 493–503. R. Kopp & H. W. Ball, ICSP3, Deutsche Gesellschaft fur Metaukunde, Garmisch Partenkirchen, 1987, 298–307.

547

Theoretical Analysis of Beneficial and Detrimental Effects of Controlled Shot Peening in High Strength Aluminium Alloys Chris A. Rodopoulos 1 Rachel E. Edwards 1, Sean Curtis 1, Jose Solis Romero 1, Jung-Hee Choi 1,2, Eduardo de los Rios 1 and Andrew Levers 3 1 Division of Aeronautical Applications, Department of Mechanical Engineering, University of Sheffield, Sheffield, UK 2 Hyundai Motor Company, S. Korea. 3 Airbus UK, Chester, UK

1

Introduction

For many years controlled shot peening (CSP) was considered as a surface treatment of questionable benefits. This impression was fuelled by contradictory results from fatigue experiments [1,2]. It is now clear that the performance of CSP in terms of fatigue depends on the balance between its beneficial (compressive residual stress and work hardening) and detrimental effects (surface roughening) [3,4]. Hence, in order to achieve a favourable fatigue performance, the role of those effects has to be analysed and understood. To achieve such undertaking it is essential to consider their interaction with other parameters such as the nature of the target material and the loading conditions. This work brings together two micromechanical models, (i) for notch sensitivity [5] and (ii) for fatigue life [6]. The former assesses the effect of surface roughening, whilst the latter incorporates the residual stress distribution and work hardening on fatigue life calculations. Combination of the two models allows the determination of the residual stress distribution to meet specific improvements in fatigue life (improvement life factor, ILF). Using the ILF methodology, the effects of CSP can be scrutinised against stress level, surface roughness and ILF value.

2

Modelling the Surface Roughness

On shot peened surfaces, cracks are likely to form at micro-notches (dents). Early studies from Smith [7] and Tanaka [8] indicate that the propagation of cracks from notches depends on the bluntness of the notch, given by (a/r)0.5 (r is the notch radius). Despite the numerous models published in the literature, as illustrated by the work quoted in [9], most models fail to provide a relationship between the geometry of the notch and the microstructure of the material, except by that provided by Vallellano et al [5,10]. According to their work, the nominal stress in a notched member is given by,

sinom =

s app Zi

(1)

548 where Iapp is the applied stress, Iinom is the distribution of the nominal stress ahead of the notch root as a function of the distance from the notch i, mapped as i = 2 a/D (a: crack length) and Zi is the notch factor given by,

i éê b a ùú Zi = + 2 a + b ê li 1 + li úû ë li =

1 é êa a 2 - b2 ë

1

2

a + iD / 2 2

(2) 2

ù - a 2 + b2 - b a + iD / 2 ú û

where i=1,3,5,…

2a 2b and b = represent in a dimensionless form the notch depth = and D D the notch half width >. The parameter D represents the distance between two successive barriers. In the case where grain boundaries are considered being the dominant barrier, D is regarded as the grain diameter. Li et al [11], proposed that the elastic stress concentration Kt introduced by multiple micronotches in CSP, is somehow lower than the one determined in the case of a single notch of similar depth and width. The above finding reflects the uniformity of the micro-notches on the surface. According to Li, the resulting Kt from CSP is given by, The parameters a =

æR ö K t = 1 + 2.1ç t ÷ èSø

(3)

where the parameters Rt and S are respectively the mean of peak-to-valley heights and the mean spacing of adjacent peaks in the surface roughness profile. In the case of a semi-elliptical notch and a high degree of uniformity (CSP coverage percentage of more than 100%), Eq.(3) can be written as,

æaö K t = 1 + 2.1ç ÷ è 2b ø

(4)

At the beginning of this section it was pointed out that the bluntness of the notch could significantly affect the strain generated at the root of the notch and consequently the propagation rate of the crack. In light of that, Smith and Miller [7] proposed that Kt should be determined by,

Kt = 1 + 2

a r

(5)

where r is the notch root radius. In the case of a semi-elliptical notch, the notch root radius can be approximated by r = a 2 / g and thus Eq.(5) can be rewritten as,

549

Kt = 1 + 2

g a

(6)

where g is the notch half width that considers the bluntness of the notch. By equating Eq.(6) with Eq.(4), the stress concentration due to multiple micro-notches can be expressed in terms of a single notch by,

Kt = 1 +

3

a b

(7)

Modelling the Fatigue Life in CSP Components

In [4,6] it was proposed that the fatigue life of polycrystalline materials can be determined by, 1- m2

æ iD ö dn1i 1 ic çè 2 ÷ø N= å A2 i =1 nòi CTOD m2 nci

(8)

s

where A2, m2 are parameters from the Paris law of crack propagation, CTOD is the crack tip opening displacement and nsi , nci are limit values of n1 as defined in Figure 1.

n1=nic

n1=ni+2c

D

Crack, i Crack, i+2

n1=ni+2s

n1=ni+4s

Figure 1: Schematic showing the position of n1 prior (nc) and after (ns), the unblocking of the crack tip plasticity

In Eq.(8) the parameter nsi and nci represent respectively, the position of the crack tip at the beginning and end of each interval i of crack growth. These two parameters are calculated by,

550 æ p s - sip arrest ö nci = cos ç ÷÷ ç2 s2 è ø i 2 nsi = nci - 2 i

(9)

where I2 is the flow resistance of the material and Iip arrest is the Kitagawa-Takahashi formula for a plain specimen,

sip arrest =

mi s FL m1 i

(10)

with sFL denoting the fatigue limit of the plain material. In Equation 10 the parameter mi / m1 represents the effect of the grain orientation factor. More details can be found in [4,12]. From Eq.(8) it is clear that the number of cycles required by the crack to propagate an i number of half grains, depends solely on the parameter nci . For CSP components, the parameter nci has to be modified in a way that it would take into account the roughening of the surface and the crack closure stresses generated by the residual stresses.

nci ,CSP

æ s CSP ö ç p Z - si arrest ÷ i ÷ = cos ç ç 2 s2 - s1i ÷ ç ÷ è ø

(11)

where the parameter Zi is given by Equation 2. The Kitagawa-Takahashi formula for the case of CSP, Iip arrest , is given by [4], æ mi sCSP FL - s1i =1 ö sCSP + s1i ÷ Z i i arrest = ç m i è 1 ø

(12)

where sCSP FL = sFL + s1i =1 . Hence, Equation 12 is rewritten as, æ mi s FL ö sCSP + s1i ÷ Zi i arrest = ç è m1 i ø

4

(13)

Introducing the Improvement Life Factor (ILF)

In order to increase the life consumed at each grain and consequently the overall life of the CSP component, we make use of a predetermined ILF,

551 1- m2

æ id ö dn1i ,CSP ç2÷ 1 è ø ILFxN = A2 ni ,òCSP CTOD m2 nci ,CSP

(14)

s

where the values of ILF are in percentage. Solution of Equation 14 in terms of CTOD yields,

é ni ,CSP - ni ,CSP Di 1- m2 ù c s ú ln ê ê ú A2 ´ ILF ´ N 0.69(m2 - 1) û ln CTOD = + ë m2 m2

(15)

In the case of a plain/unpeened material, Equation 15 is written as,

é ni , p - ni , p Di 1- m2 ù c s ú ln ê ê ú A2 N 0.69(m2 - 1) û ln CTOD = + ë m2 m2

(16)

The fact that the value of CTOD at the position nc, where the crack tip plasticity is able to overcome the microstructural barrier, is identical for both the peened and the unpeened material (for the same loading conditions) allows Equations 15, 16 to be equated,

é ni ,CSP - ni ,CSP Di 1- m2 ù c s ú ln ê ê ú A2 ´ ILF ´ N 0.69(m2 - 1) û= + ë m2 m2

é ni , p - ni , p Di 1- m2 ù c s ú ln ê ê ú A2 N 0.69(m2 - 1) û + ë m2 m2

(17)

Simplification of Equation 17 gives,

nci ,CSP = ILF ´ nci , p - nsi , p + nsi,CSP

(18)

From Equation 18, the closure stress I1i can be determined. It should be noted that due to the complexity of Equation 18, a computational solution is advised. Figure 2 shows the calculated crack closure, I1i , for several conditions of loading and treatment.

552 0

0

0

-50 -50

-200

-300 K=1.4

-100 Closure Stress (MPa)

Closure Stress, I1i (MPa)

Closure Stress, I1i (MPa)

-100

-150

-200

t

-300

-400 1

10

100

1000

-150

-200

ILF=5% ILF=1%

-250

t

K=1.25

-100

I=350 MPa I=280 MPa

-250

1

i=2a/D

(a)

10

100

1000

1

10

100

i=2a/D

i=2a/D

(b)

(c)

1000

Figure 2: The effect of (a) surface roughness (b) ILF and (c) applied stress on the distribution of closure stress in a 2024-T351 CSP component under mode I loading. The parameters used in the calculations are: ILF = 5 %, s = 350 MPa, s2 = 450 MPa, sFL = 220 MPa and D = 52 mm.

5

Discussion and Conclusions

In this work the effects of CSP on fatigue damage are analysed and modelled. Surface roughness is modelled as a local increase in the far-field stress. Hence, the treated surface has a higher propensity to initiating and propagating short fatigue cracks. Compressive residual stresses are considered as crack closure stresses and regarded as one of the beneficial effect of CSP. Residual stresses tend to reduce the intensity of the far-field stress by introducing a closure stress on the crack flanks. Thus, crack propagation rate of peened material is expected to be lower than that of unpeened material. Finally, strain hardening is expected to reduce the propagation of short fatigue cracks by increasing the resistance of the material to the generation of crack tip plasticity. A predetermine improvement in terms of fatigue life can be calculated by introducing the ILF factor into the number of fatigue cycles consumed in every grain. The above approach allows the mathematical modelling of the balancing between the beneficial and detrimental CSP effects. At first the analysis reveals that the magnitude of the closure stresses should always attain a maximum at the surface. Such distribution minimises the premature initiation of a “visible” fatigue crack. Secondly, the depth distribution should be able to counteract the stress gradient generated by the surface roughness. Further analysis allows the assessment of parameters such as the far-field stress level, the ILF and the surface roughness. From Figure 2, the effect of the above parameter is quantified in the following order, starting from the most decisive: a) Surface Roughness. The analysis reveals that a 12% increase, measured in terms of Kt, in the surface roughness requires a 47% increase in the closure stress magnitude to allow a 5% increase in per grain fatigue life. Additionally, a higher Kt would require deeper closure stresses; b) Far-Field Stress Level. In principle, high far-field stress levels require high magnitude and deeper closure stresses. This agrees with findings published extensively in the literature, that CSP will have a minimum effect, or in some cases a detrimental effect, on the low cycle fatigue region; and c) ILF. The analysis reveals that CSP components are not so sensitive to different ILF values. The above conclusion agrees with experimental data showing that short cracks propagate almost irrespective of the crack closure stress levels.

553 It should be noted that since the methodology is expressed in terms of crack length, it can be easily adjusted to incorporate relaxation profiles of residual stress and strain hardening.

6

Acknowledgements

The authors are indebted to numerous industrial organisations and governmental bodies which supported their research for many years. Special citation should be given to the EPSRC, The Royal Academy of Engineering, The British Council, Airbus UK, The Hellenic Aerospace Industry and the Mexican CONACYT.

7 [1]

References

P. O’Hara in Surface Treatment IV, (Ed. . Brebbia and J. M. Kenny) WIT Press, 1999, 321–330. [2] L. Wagner and G. Lütjering in Shot Peening, Pergamon press, 1981, 453–460. [3] P. K. Sharp and G. Clark DSTO-RR-0208, Defence-Science and Technology Organisation, Royal Australian Air Force, Australian Ministry of Defence, 2001, (declassified). [4] S. Curtis, E. R. de los Rios, C. A. Rodopoulos and A. Levers, Inter. J. of Fatigue, 2002, in press. [5] C. Vallellano, A. Navarro and J. Domínguez, Fatigue Fracture Engineering Materials Structures, 2000, 23, 113–121. [6] E. R. de los Rios, M. Trull and A. Levers, Fatigue Fracture Engineering Materials Structures, 2000, 23, 709–716. [7] R. A. Smith and K. J. Miller, Inter. J. Mech. Sci., 1978, 20, 201–206. [8] K. Tanaka, Inter. J. Fract., 1983, 22, R39–R45. [9] S. Suresh, Fatigue of Materials, Cambridge University Press, 1991. [10] C. Vallellano, A. Navarro and J. Domínguez, Fatigue Fracture Engineering Materials Structures, 2000, 23, 123–128. [11] J. K. Li, M. Yao, D. Wang., R. Wang, Fatigue Fracture Engineering Materials Structures, 1999, 15(12), 1271–1279. [12] E. R. de los Rios and A. Navarro, Philosophical Magazine 1990, 61, 435–449.

554

Finite Element Simulation of Shot Peen Forming Yuansong Zeng Beijing Aeronautical Manufacturing Technology Research Institute, Beijing, P. R. China

1

Abstract

In this paper, both the impact model and the forming process simulation of the shot peen forming are introduced. In the impact model, effects of the shot diameter and velocity on the value and depth of the compressive stress and plastic layer are analyzed with dynamic explicit finite element method(FEM). A new equivalent thermal loading method is presented for the simulation of the shot peen forming process. The deformation of the standard Almen Strip and a wing skin with ribs are simulated with the method using static implicit FEM software. Results show that the method is very effective if the relationships between thermal loads and the shot peening strength are suitably set up

2

Introduction

For a number of years, the aircraft industry has used shot peening as a forming process to some metal parts, such as airplane wing skins [1-2]. The basic principle of the shot peen forming is a partial forging process. When the surface of a metal part is repeatedly hit at a high velocity by small steel shot, a thin plastically deformed layer is formed beneath the impacted area, and residual stresses generated in the part result in an equilibrium of forces and moments, therefore, obtains a permanent convex curvature. Figure 1 shows the typical residual stress distribution throughout the part cross-section. It can be seen from the figure that two surface layers are all compressive stress, which can improve the resistance of the metal part to fatigue and stress corrosion . The most disadvantages of the shot peen forming is that the forming process is very difficult to be controlled. The reason is many process variables can affect the curvature generated by shot peening, such as the shot diameter D, the shot velocity V, the shot coverage ratio D, the shooting angle, the material of shots and sheets. And it is a random process for the impact of a large number of tiny shots against the surface of a metallic component. Therefore, for a long time, the shot peen forming is a try-and-error process. The experience of operators play a very important role in the process. With the developing of the computer and finite element method, it is possible to use FEM simulation to predict the deformation of parts, to design peening process and to decrease the times of experiments.

555

Figure 1: The residual stress distribution generated by shot peening

3

Impact Process Simulation

Because the shot peening process is comprised of the impacting of many single shot on the surface of parts, it is also very important to research the impacting process of one shot on a metal part. For the impact of one or several shots, it is easy to simulated with FEM, and it is also significative to examine the effect of varying the shot size and velocity on the local stress field. A finite element model as shown in Figure 2 is used to simulate the impacting process of one steel shot on the surface of a part. On account of the highly dynamic impact characteristic of the shot peening, a dynamic explicit finite element program, ANSYS/LS-DYNA, was used to simulate the process. Dynamic effect, contact and elasto-plastic nonlinear parameters are taken into consideration.

Figure 2: The impacting model of one shot

556 The thickness of the part is 3mm and the material is LY12CZ, which the properties are shown in Table 1. It is supposed that the effect of the strain rate hardening on the deformation is omitted. Table 1: The properties of aluminum alloy YL12CZ Young's modulus [MPa]

Shear modulus [MPa]

Poisson's ratio

Yield stress [MPa]

Strength [MPa]

71000

27000

0.32

300

426

From the Figure 3 it can be seen that with the increasing of D, the maximal compressive stress gradually move from the surface to the core zone, its value is also increased, and the stress at the surface changes from the compressive stress to the tensile stress, which express that shots with small diameter should be used for getting the compressive stress on the peened surface. The same situation take place if increasing V, except for the value of the maximal compressive stress is not changed as shown in Figure 4. That is to say, we can obtained the desired depth and value of the compressive stress through adjusting D and V. 200

Stress, MPa

100

0

-100

D=0. 6mm D=1. 2mm D=2. 0mm D=2. 6mm D=3. 2mm

-200

-300

.

-400 0

0. 5

1

1. 5

2

2. 5

The distance from the peened surface, mm

3

Figure 3: The effect of D on the stress(V= 60m/s)

From the distribution of the effective plastic strain shown in Figure 5 and Figure 6, the depth of plastic layer after peening can be obtained very directly. Results shown that with the increasing of D and V, the depth of plastic layer and the value of the effective plastic strain will be increased. Although, the impact model is able to directly simulate the impact process of finite shots, and the stress distribution and plastic layer also can be obtained, it is impossible to simulate the impact process of many shots, and the final shape of the part is also not easily predicted.

557 100 50

Stress,MPa

0 -50 -100

32m/s 40m/s 48m/s 56m/s 64m/s 72m/s

-150 -200 -250 -300 0

0. 5

1

1. 5

2

2. 5

3

The distance from the peened surface, mm Figure 4: The effect of V on the stress (D= 0.046 inch)

0. 4

Effictive plastic strain

0. 35

D=0. 6mm D=1. 2mm D=2. 0mm D=2. 6mm D=3. 2mm

0. 3 0. 25 0. 2 0. 15 0. 1 0. 05 0 0

0. 5

1

1. 5

2

The distance from the peened surface, mm

2. 5

Figure 5: The effect of D on the hardening layer(V= 60m/s)

Figure 7 is a numerical model which can be used to simulate the effect of D on the stress distribution. In this model, the central shot is evenly surrounded by six shots. So, if the shot coverage ratio can be changed with the changing of the distance L, which will lead to the change of stress distribution under the center shot. From Figure 8 it can be seen that the depth of the compressive stress increased with the increasing of D.

558 0. 35

0. 3

32m /s 40m/s 48m/s 56m/s 64m/s 72m/s

Effective plastic strain

0. 25

0. 2

0. 15

0. 1

0. 05

0

-0. 05 0

0. 2

0. 4

0. 6

0. 8

1

1. 2

1. 4

The distance from the peened surface, mm Figure 6: The effect of V on the hardening layer( D = 0.046 inch)

Figure 7: The numerical model of 7 shots

1. 6

559 100

50

0

Stress,MPa

0

0. 5

1

1. 5

2

2. 5

3

-50

12. 75%

-100

22. 67% -150

51. 01% 90. 69%

-200

-250

The distance from the peened surface, mm

Figure 8: The effect of D on the stress(V = 60 m/s, D = 0.046 inch)

4

Forming Process Simulation

Some method to simulate the shot peen forming process have been presented [3]. One method is to input a set of initial stresses to the finite element mesh, so that when the analysis starts elements are immediately subject to a residual non-equilibrating stress field that causes deformation. However, this instantaneous application of the residual stress can be difficult to analysis successfully. Another method is to apply pressure to the faces of elements near the surface of the component. But the value and lasting time of the pressure is very difficult to determine. The third method is to use the thermal loads to set up the residual stress profile across the thickness of the sheet [3]. In this method, the load application is straightforward because the existing input methods for thermal analysis can be used only by the temperature profile and the coefficient of thermal expansion. But in practical shot peen forming process, shot coverage is often not more than 50 %, the 100 % shot coverage is very seldom used so the residual stress field is very difficult to tested precisely. Therefore, the application range of this method is very limited. In present, a new method is presented, which also use the thermal loads to produce the equivalent deformation not the residual field. That is to say, if the deformation on a part induced by a thermal load is the same as that of a set of peening variables, the thermal load will be regarded as the equivalent load of the set of peening variables. A relationship between thermal loads and the deformation can be set up with Almen Strip, which can be expressed as follow. C = F(T,>)

(1)

Where, C is the arc height of Almen Strip, i.e. the shot peening strength, T temperature, > the coefficient of thermal expansion in the plane of the sheet, the expansion in the thickness direction is omitted in order to agree with the practical situation. And the relationship between the

560 arc height of Almen Strip and the main peening variables also should be set up according to experimental data: C = f(D, V, D, …)

(2)

Using above two equations, it is possible to simulate all kinds of shot peening process with finite element method regardless of complex residual stress field. It is supposed that the thermal loading process is a steady process, then the shot peen forming process can be simulated with static implicit FEM. Figure 9 is the deformation of an Almen Strip simulated by MARC software. Multi-layer shell elements in the simulation are adopted to describe the grads of temperatures from the peened surface to the other surface. It can be seen the deformation is very same as that of peening process. It should be stated that the rigid constraint of the un-peened area to the peened area should be take into account. Otherwise, the simulated shape will not same as the practical shape.

Figure 9: The deformation of an Almen strip after peening

The complex structure are also able to be simulated with the method. Figure 10 is the peening track of a wing-skin panel with strengthening ribs along the longitudinal direction, where shade areas are peening bands. The shape of the panel after peening will be nearly a single curvature outline according to the peening track. Figure 11 is the shape simulated with the thermal loading. The simulated shape of the panel is very in agreement with that of peening process. These results show that the shot peen forming process can be simulated using thermal loads according to the rule of equivalent deformation.

Figure 10: The peening trcak of a wingskin with ribs

561

Figure 11: The deformation of a wingskin with ribs after peening

6

Conclusions

The effects of the shot diameter, the shot velocity and the shot coverage ratio on the stress distribution and plastic layer were analyzed by impacting model with the dynamic explicit finite element method. On the basis of experimental data, the shot peen forming process is able to be modeled with static implicit finite element method through the equivalent thermal loads with the help of the Almen Strip.

7 [1] [2] [3]

References Kishor M. Kulkarni, John A. Schey and Douglas V. Badger. Investigation of shot peening as a forming process for aircraft wing skins. J. Applied Metal Working, 1981, 1(4), 34–44. R. D. Vantuchene and E. J. Cramer. Numerical modeling of a wing skin peen forming process. J. of Materials Engineering and Performance, 1996, 5(6), 753–760. Andrew Levers, Alan Prior. Finite element analysis of shot peening. J. of Materials Processing Technology, 1998, 80-81, 304–308.

563

Author Index1 A

H

Altenberger, I. 421, 447, 483 Ando, K. 23, 191, 338

Hackel, L. 474 Halpin, J. 474 Handa, M. 31, 305 Harada, Y. 200 Harris, F. 474 Hasegawa, N. 305 Hattori, K. 31, 305 Hill, M. 474 Hilpert, M. 367, 374 Hutmann, P. 3

B Bailey, P. 96 Borkowski , P. 523 Bosshard, R. 65, 83 Bozzolo, W. 130 Brocks, W. 406 C Cammett, J. 295 Champaigne, J. 108, 114 Choi , J-H. 547 Citran, G. 167 Clausen, R. 89 Curtis, S. 264, 547

I Iida, K. 173 Ikeda , M. 305 J Jain, A. 137 Jones, D. 474

D de los Rios, E. 101, 243, 264, 547 Deo, H. 83 do Nascimento, M. 37 Duchazeaubeneix, J.-M. 31, 305

K

Frey, T. 185 Friese, A. 53

Karuppanan, S. 101 Kersching, U. 78 Ketzmer , A. 386 Kiefer, A. 367, 374 Kiese, J. 380, 399 Kittel, S. 44 Kocan, M. 461 König, G. 13 Kopp, R. 161, 227 Krauß, M. 324 Kritzler, J. 255 Küster, B. 367

G

L

Gregory, J. 349 Grossmann, K. 454 Guelorget, B. 173

Lane, L. 474 Lang, K.-H. 281 Lauro, M. 130

E Edwards , R. 547 Eto, H. 23, 191 F

1

Page numbers refer to the beginning of the article

564 Levers, A. 101, 264, 540, 547 Lindemann, J. 392, 454 Linnemann, W. 44 Lohmar, J. 53 Lu, J. 173 Ludian, T. 374

Ritchie, R. 447, 483 Rodopoulos, C. 101, 264, 547 Rodríguez, R. 271 Romero, J. 101, 264, 547 Roth, M. 208 Roth-Fagaraseanu, D. 392 Ruhland, S. 73

M Maki, S. 200 Marconi, P. 130, 167 Matsui, K. 23, 191 Menig, R. 311, 317, 331, 498 Modi, S. 83 Mori, K. 200 Müller, C. 271 Müller, E. 441 N Nadkarni, S. V. 137 Nakonieczny , A. 523 Nitschke-Pagel, Th. 360 Noster, U. 447, 483

S Saito, K. 435 Schauerte, O. 380, 399 Scholtes, B. 161, 324, 447, 483 Schulz, J. 161, 227 Schulze, V. 145, 281, 311, 317, 331, 498, 507 Schwarzer, J. 507 Sharma, M. 83, 137, 490 Soules, J. 181 Soyama, H. 435 Specht, R. 474 Stangenberg , J. 89 Šupík , V. 413

O O’Hara, P. 217 Odhiambo , D. 435 Okada, H. 338, 516 Ostertag, A. 461, 468 P Petit-Renaud, F. 119 Pfeiffer, W. 185 Pieper, G. 73 Platts, M. J. 540 Prevey, P. 295 Q Quan, G.

T Tange, A. 338, 516 Tokaji, K. 305 Torres, M. 37 Tosha, K. 173 Tufft, M. 530 V Vöhringer, O. Voorwald, H.

281, 311, 317, 331, 498, 507 37

W 406

R Raczek, T. 454 Reccius, H. 234 Retraint, D. 173

Wagner, L. 349, 367, 374, 380, 386, 392, 399, 413, 454, 461, 468 Wang, T. 540 Watanabe, Y. 31, 305 Wendt, J. 386 Wierzchowski, D. 468

565 Wohlfahrt, H. 360 Wortman, C. 208 Wübbenhorst, W. 474 Wüstefeld, F. 44, 53 Wymyslowski , P. 523

Z Zaleski, T. 474 Zeng, Y. 554 Zhang, J. 380, 399 Zinn, W. 161

566

567

Subject Index1 A Adhesive bonding 208 Advanced peen forming 53 Aero engine components 13 Aeronautical steel 37 Aging aircraft components 181 Alloy – aluminium 208, 264, 311, 317, 331, 360, 406, 413, 498 – magnesium 137, 271, 547, 367, 374, 386 – titanium 349, 380, 392, 399, 447, 461 Almen – gage calibration 108 – intensities 161 – procedure 65 – round 83 – strips 114 Alpha-titanium alloy 399 Alternative mechanical surface treatments 421 Aluminides 392 Aluminium 181, 208, 264 Annealing, short time 331 Austenite 191 Automotive components 23 Automotive engineering 3 B Bearing ball 31 Bellows, metal 338 Bending 137 Bending stress 191 Bonding, adhesive 208 C Calibration, Almen gage 108 Cast magnesium alloys 374 Cavitation shotless peening 435 Ceramics 185 1

Characteristics – geometrical 167 – mechanical 167 – surface 173, 498, 507 CNC control 44 Cold deformation 406 Cold work 295 Comparison 468, 490, 498 Component – aero engine 13 – aging 181 – dynamically stressed 78 – shape 217 – microstructure 145 Composite material 406 Compressive stress 191 Computer simulation of shot peening 523 Conditions, shot peening 530 Conductivity changes 181 Controlled shot peening 44, 264, 547 Conventional shot peening 311 Corrosion 243 – behavior 271 – cracking 255 – fatigue 264 Coverage 96, 295, 516, 540 Cracking, mechanisms 243 Cracking, stress 255 Cut wire shot 78 D Deep rolling 447, 454 Deformation, cold 406 Dent diameter 516 Development of coverage 101 Die cast alloy 271 Double-sided simultaneous shot peen forming 227 Dynamically stressed components 78

Page numbers refer to the beginning of the article

568 E Emission-free shot peening 73 Endurance life 181 Engine components 13 Enhance fatigue strength 380 Erosion 137 F Fatigue – behavior 421, 447, 530 – conditions 243 – limit 23 – performance 392, 468 – properties 145 – resistance 130 Fatigue strength 137, 264, 271, 281, 295, 311, 317, 331, 338, 360, 367, 380, 386, 406, 413, 435, 454, 461, 483, 516 – high strength steel 305 – thermal 324 Finite element simulation 507, 540, 554 Fir-tree root 490 Flourescent tracers 96 Foil, hard-metal 200 Forming tools 234 Fretting fatigue conditions 243 G Gears 23 Geometrical characteristics

167

H Hard-metal foil 200 Hardness 23, 37 Hard turned steel 324 Heat cycles 208 High-strength alloy 399, 547 High strength steel 37, 305 High temperature fatigue 454, 483 Hot-work steel 324

Impact model 540, 554 Impact velocities 89 Improvement of fatigue strength 338, 360, 435, 461 Induced residual stresses 441 Influence 317 – of pre-annealing 173 – of the velocity 161 Isothermal fatigue behavior 447 L Laser – peening 217, 447, 474, 498 – welding 53 Lifetime – enhancement 13 – prediction 181, 530 Light metals 349 Lining 200 Load capacity 185 Low alloy steel 295 LPT blades 490 M Magnesium alloy 137, 271 Martensite, strain-induced 191 Material behavior 185 Material state 331 Mechanical characteristics 167 Mechanical surface treatment 3, 367, 399, 421, 447, 468, 483 Mechanisms of cracking 243 Metal bellows 338 Metal forming, impact 234 Metal surface 200 Methods of control 130 Microstructure 145, 421 Modelling – cracking 243 – impact 540 – fatigue behavior 530 Multiple shot impacts 523

I Impact body 234 Impact metal forming

N 234

Notched fatigue strength

367

569 O Optimising 227 Optimized warm peening Overload 413

317

P Parameter – peening 507 – shot peening 83, 119 Passenger car industry 3 Peen forming – advanced 53 – process automation 44 – simultaneous 227 Peening – coverage 96 – laser 474, 498 – medium 161 – optimized 317 – parameter 507, 516 – shock 447 – shotless 435 – stress 441 – vacuum-suction 73 Pelton wheels 130 Post weld treatment methods 360 Pre-annealing, influence 173 Pressing, roller 490 Process – automation 44 – control 474 – parameters 119 Property improvement 349 Q Quenched material state

331

R Relaxation 208, 311, 498 Residual compressive stress 101, 191 Residual stress 130, 167, 208, 295, 421, 441 – evaluation 65 – relaxation 311, 498 – stability 317

– state 145, 161, 447 Resistance, improvement 461 Retained austenite 191 Roller burnishing 461, 468 Roller pressing 490 Rolling in spring steel 441 Roughness 89, 516 S Shock peening 447 Short-time annealing 331 Shot diameter 554 Shot impacts, multiple 523 Shotless peening 435 Shot peening 13, 130, 137, 145, 167, 185 – conditions 530 – control 547 – conventional 311 – coverage 101, 295 – effect 338, 392 – emission-free 73 – forming 540, 554 – influence 317, 255, 271, 454 – parameter 83, 119 – residual stress 208 – simulation 507, 523 – specimens 161 – steel 324 – treatment 37 – ultrasonic 31, 305, 507, 523 Simultaneous shot peen forming 227 Spring steel 234, 441 Stainless steel 255 Steam turbine 490 Steel 324, 468 – aeronautical 37 – fatigue 281 – high strength 305 – low alloy 295 Straightened Almen strips 114 Strain-induced martensite 191 Strengthening of ceramics 185 Strength, fatigue 281, 305, 311, Stress – bending 191

570 – compressive 191 – concentration 243 – corrosion cracking 255 – evaluation 65 – peening 311, 441 – relaxation 311, 498 – residual 101, 208, 295, 311, 317, 421, 441, 447 – stability 317 Superalloy 454 Surface – characteristics 173, 498, 507 – layers 145, 173 – metal 200 – roughness 516 Surface treatment 399, 406, 421, 447, 468, 483 – mechanical 3, 367 Suspension springs 380

Theoretical investigation 101 Thermal fatigue 324 Thermal residual stress relaxation 498 Torsion 137 Torsional loading 311 Tracers, flourescent 96 Treatment – mechanical 399, 421, 447, 468, 483, 399 – surface 406 Tungsten carbide ball 31

T

W

Tempered material state 331 Tempering 114 Test temperature 386 Theoretical analysis 547

Warm peening 311, 317 Welding 360 Work, cold 295 Workhardening state 145

U Udimet 454 Ultrasonic shot peening

31, 305

V Vacuum-suction peening 73 Velocity, influence 161

Blood Shot

Kill Shot

Kill Shot

Kill Shot

Double Shot

Drop Shot

Drop Shot

Espresso Shot

Blood Shot

Last Shot

Dissolve Shot

Double Shot

Hot Shot

Bank Shot

Shot Girl

One Shot

Hot Shot

Last Shot

Blood Shot

Dead Shot

Shot Girl

Dead Shot

Second Shot

One Shot

One Shot

One Shot

One Shot

Sharp Shot

Double Shot

Head Shot

Shot Peening

Blood Shot

Kill Shot

Kill Shot

Kill Shot

Double Shot

Drop Shot

Drop Shot

Espresso Shot

Blood Shot

Last Shot

Dissolve Shot

Double Shot

Hot Shot

Bank Shot

Shot Girl

One Shot

Hot Shot

Last Shot

Blood Shot

Dead Shot

Shot Girl

Dead Shot

Second Shot

One Shot

One Shot

One Shot

One Shot

Sharp Shot

Double Shot

Head Shot

Recommend Documents