GEOTECHNICAL MATERIALS : MEASUREMENT & ANALYSIS Raymond J. Krizek Commemorative Symposium August 3, 2002
Preface Raymond J. Krizek Biography Symposium Papers Measurement & Performance Materials & Behavior Analysis & Design Teaching & Management
Northwestern University Memorabilia
The Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium Proceedings CD requires the Adobe Acrobat Reader software to operate. The CD is designed to automatically launch the GeoTec02.pdf file that contains the symposium content when the CD is inserted into a computer. Adobe Acrobat Reader: For those who do not already have Adobe Acrobat Reader installed on their computer, this CD contains an installation file with the Acrobat Reader 5.0.5 software and an Acrobat Reader 5.0.5 software patch for the Windows 95/98/2000/NT operating systems. The vendor provides this software at no charge, and has granted permission for distribution of their software on this CD. To download Acrobat Reader software that runs on other operating systems, or to obtain support, visit the Adobe Systems Inc. web site at http://www.adobe.com. Adobe Acrobat Reader 5.0.5 installation instructions: 1. Open the Adobe Acrobat Reader folder on the CD. 2. Open the file rp505enu.exe. A software installation package will launch and guide you through the installation steps. Note: If you have problems running the software, you may also need to install the software patch, A505AFPI.exe, which is also included in the Adobe Acrobat Reader folder. Open this file after you have completed the installation of the rp505enu.exe file. The installation software will guide you. 3. You should now be able to open the proceedings file. Open Acrobat Reader. 4. Open the proceedings file GeoTec02.pdf that is in the root directory of the CD. Other software requirements: The proceedings file GeoTec02.pdf contains links to several audio files and one video file. Your computer will need audio and video player software, such as Apple Computer Inc.'s QuickTime Player or Microsoft Corporation's Windows Media Player. Each of these programs may be downloaded for free from the vendors' web sites. Links to the vendor sites are listed below and are provided on the CD should you need to download the free player software. QuickTime Player: http://www.apple.com/quicktime/ Windows Media Video: http://www.microsoft.com/downloads/search.asp?
OBTAINING PROCEEDINGS Electronic copies of the proceedings are available for free online at: http://www.iti.northwestern.edu/research/geotechnical_symposium/proc.html Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium Proceedings on CD-ROM The CD-ROM version of the Proceedings is available for $25.00 (this includes shipping and handling costs.) Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium Proceedings - Hard Copy An 8.5”x 11” laser printed hard copy of the Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium Proceedings is available for $100.00 (this includes shipping and handling costs.) Both hard copies and CD-ROMs are available by contacting the following: Infrastructure Technology Institute Knowledge Services Seeley G. Mudd Library for Science & Engineering Northwestern University 2233 North Campus Drive Evanston, Illinois 60208-3530
[email protected] Please make check or purchase order payable to: Northwestern University Prices are subject to change without notice
BIBLIOGRAPHIC INFORMATION Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium, Infrastructure Technology Institute at Northwestern University, Evanston, IL, 2002. Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium Editor: Charles H. Dowding Production editors: Ker Min Chok and Ruth W. Allee Copy editor: Ker Min Chok This publication contains papers prepared for the 2002 Geotechnical Symposium at Northwestern University to honor Raymond J. Krizek for his manifold contributions to teaching and research in geotechnical engineering.
2002 by Infrastructure Technology Institute Evanston, IL 60201 All rights reserved. Unattributed use of text and/or figures is strictly prohibited. Although all information published in these proceedings has been reviewed by the 2002 Geotechnical Symposium Organizing Committee, no warranty, expressed or implied, is made by the Infrastructure Technology Institute or the Department of Civil and Environmental Engineering at Northwestern University as to the accuracy of the data and related materials. The act of distribution shall not constitute any such warranty, and no responsibility is assumed by the Infrastructure Technology Institute or the Department of Civil and Environmental Engineering at Northwestern University in the use of this information or related materials. Printed in the United States of America Printed, August 2002
ISBN 0-9712631-1-6
Preface
This symposium on Geotechnical Materials: Measurement and Analysis was held at Northwestern University in Evanston, Illinois on August 3 2002 to honor Professor Raymond J Krizek. Papers from a number of his students and associates are published in the proceedings. In addition the CD version contains Northwestern memorabilia that includes a video initiating reconstruction of Tech, a project that he co-chaired. This year marks the confluence of a number of decennial anniversaries for him, Geotechnical Engineering, and Civil and Environmental Engineering in general at Northwestern: it is some 10 months since his induction to the National Academy of Engineering, 10 years since the reconstruction of the Civil & Environmental Engineering wing, 10 years since his inauguration of the Masters in Project Management Program, 40 years since he came to Northwestern as a graduate student, 60 years since Geotechnical Engineering was established at Northwestern, and 70 years since he was born. The proceedings begins with a biography of Professor Krizek and a list of his more than 60 doctoral students from 40 countries and their thesis titles. An appendix contains a complete listing of his over 300 publications. Proceedings papers have been organized into four topical areas: Measurement and Performance Materials and Behavior Analysis and Design, and Teaching/Research and Management. These topics represent the diversity of the field of Geotechnical Engineering. The vast range of properties of earth materials has required from the outset an emphasis on the measurement of their interaction with constructed facilities. Thus earth materials and their properties both in situ as well as modified become an important consideration for analysis and design that must take into account the probabilistic variability of the in situ properties. The last topic covers the infrastructure and enterprise of teaching these concepts as well as employing them productively in the process of construction.
Geotechnical Engineering has changed significantly in the 40 years since Ray Krizek began his research, teaching and engineering career. These papers demonstrate these advances and illustrate how he and the geotechnical program at Northwestern have pointed graduates in directions of significant impact. Instrumentation has become miniaturized, electronic, and more precise. Developments in fracture mechanics and physico-chemical properties have propelled quantum leaps of understanding of material behavior. Advances in finite element and probabilistic analysis have allowed phenomenalogical behavior to replace empirical approaches. Finally these advances are being transferred to new generations of engineers through significant changes in the teaching enterprise. Professor Krizek’s diligence, one on one interpersonal interaction, and emphasis on precise written communication have been hallmarks of his teaching legacy. Furthermore his ability to combine fundamental engineering mechanics and clever, multivariable experiments has earned him the enduring respect of both his peers and students. All those who have had the good fortune to have worked and studied with Ray have learned much and thank him for sharing his wisdom.
Charles H Dowding On behalf of Ray Krizek’s students and colleagues
Northwestern University Memorabilia Northwestern Alma Mater & Fight Songs Click on an area or link with a "pointing finger" to open an audio or video file, or internet site. When your media player opens, you may need to click the "play" button to start.
University Hymn (vocal) University Hymn (instrumental) Go! U Northwestern
History Technological Frontier: Click for
1990 video for the Tech reconstruction campaign
QuickTime video
or
Windows Media Player video
Technological Institute Construction History - Transformation of the "Tech" Building into the Technological Campus - Construction of the 1940 "Tech" Building
Geotechnical Engineering at Northwestern University Visit the following sites to download free audio and video player software QuickTime Player: http://www.apple.com/quicktime/
Windows Media Video: http://www.microsoft.com/downloads/search.asp?
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
TRANSFORMATION OF THE "TECH" BUILDING INTO THE TECHNOLOGICAL CAMPUS Charles Dowding Department of Civil and Environmental Engineering, Northwestern University 2145 Sheridan Road, A236, Evanston IL, 60208-3109
[email protected]
CONTINUAL EXPANSION OF TECH Additions to the 1940 Walter P Murphy "Tech" have occurred roughly once each decade since the 1960's. The resulting "technology campus" is summarized photographically in Figure 1. This October 1989 aerial view during the construction of the Materials and Life Sciences Building (MLSB in the foreground) shows all of the additions. The original, 1940, back to back capital “E”s (linked with the two N-S connecting corridors) are at the top of the photograph. The 1961 east addition, provided two eastern "feet" to the E's extended Tech to the edge of the lake at that time. The cost of this addition (in non deflated dollars) was greater than the original building. Figure 2 shows the foundations of the east wings being constructed in September of 1961 with only a sheet pile wall and the beach between the additions and Lake Michigan.
Figure 1. 1989 aerial view of the Technological Institute at the top of the photo and construction of the Materials and Life Sciences Building in the center of the photo.
Department of Civil and Environmental Engineering Infrastructure Technology Institute Northwestern University, Evanston, Illinois, 60208
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Figure 2. 1961 photo of the foundations for the east addition of the technological institute In the 1970's The Mudd Science and Engineering Library (SEL) and the north east in-fill were added. SEL is the "L" shaped addition east of the 1961 addition and just to the right of MLSB in figure 1. This site became available after the creation of the Miller Lake-fill campus. Contrary to rumor, the sand was not mined from the Sand Dunes National Sea shore, but was from the harbor dredged for the Burns Harbor Bethlehem Steel complex. The October 1976 eastward looking aerial view in Figure 3 shows the finishing touches on SEL and the space for MLSB. The in-fill between the northern first and second "east most" lab wings was built in the 70s and obscures the Second wing in Figure 1. Despite the space provided by these piece-meal additions, Tech resources could not keep up with research demands of the cold war and the looming digital and genetic revolutions, and a drastic solution was needed. In 1980 President Bob Stroz called the Engineering Faculty to a special meeting on a Sunday morning to announce an unusual step for Northwestern, a direct request to Congress for a new building to provide in incremental step in research capacity. While this initiative resulted in the Basic Engineering Research Lab (BIRL) building in the Evanston Research Park rather than adjacent to Tech, it did mark a significant change in NU's approach to the funding of technological building, which would eventually lead to the MLSB addition some nine years later.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Figure 3. 1976 aerial view of the Seely G Mudd Building under construction The 1980's saw the addition of the Catalysis building (attached to the extreme south east Tech wing in Figure 1) and MLSB. Figure 4 shows the December 1984 construction of the Catalysis building sandwiched in between Tech on the north and Dearborn Observatory on the south. And the 1989 addition of the Materials and Life Sciences Building completes the chronological sequence of the construction of buildings shown in Figure 1. RECONSTRUCTION OF "AN OUTPOST ON THE TECHNOLOGICAL FRONTIER" During the 80's it became apparent that not only wasn't Tech big enough, it had hardening of the arteries: the mechanical, plumbing, and electrical systems were over capacity and on the verge of collapse. The building was not air conditioned, and the laboratories were not separated for security and conditioning purposes. Again drastic steps were needed: only a 100 million dollar capital campaign could do the job. The renovation needs were so extensive that the building would have to be gutted and reconstructed from the inside out.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Figure 4. 1984 photograph taken during the construction of the catalysis building Even if the hundred plus million could be raised, there were two seemingly insurmountable obstacles: 1) departments would have to be temporarily relocated while space was renovated, 2) there was no open adjacent space to which to relocate. This project would be like playing the 15 piece sliding number game with no open space. Fortunately the federal grant to construct the MLSB provided suitable adjacent open space. Thus challenge two was solved. The new Masters in Manufacturing and Management (MMM) program with Kellogg provided the momentum to push the first department over the edge into moving. The MMM program was important to the growth of the Engineering School, but much of Mechanical Engineering space that would house the computerized manufacturing work cells had to be extensively renovated to provide a suitable environment. There was no option for ME but to relocate, and Civil Engineering followed suit in the next phase. A video was produced to serve as an inspirational kick off for the campaign that would eventually raise the 120 million dollars to complete the job, and is included in this CD. This video cleverly combined historical and contemporary photographs and computer graphics to visually tell the story of why Tech needed major surgery to survive. Sander Van Oker, an NU alum provided the video voice that proclaimed Tech to be an important "outpost on the technological frontier". The same Computer Aided Design (CAD) graphics employed to design the project were employed in the video, a technique which has now become an industry standard.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
DECISION, PHASING, AND COMPLEXITY OF RECONSTRUCTION The decision to reconstruct the existing building was based largely on economics. Apart from its historical significance and beautiful exterior walls (the Lannon stone with Bedford trim used on the exterior of the building relates it to other university buildings in which this combination of materials has been used effectively), the cost of demolishing the existing building and constructing a new building with equal square footage would greatly exceed the cost of reconstruction. Alternatively, construction of a new building on a different site would still leave the challenge of modernizing the existing building and making it serviceable for another purpose. However, even if the demolition alternative were economically feasible, it would have been impossible to implement because neither space nor funds were available to relocate the extensive teaching and research functions in the building during the. demolition and construction period. These considerations dictated the decision to undertake a multiphase reconstruction of the existing building. The basic plan was to reconstruct the existing building in phases over a period of about eight years. [The actual project lasted some 10 years.] The goal was to create a state of the art environment for research and teaching in such a way that: (1) The overall project cost (including the cost of relocation and temporary service) would be minimized; and (2) each phase would be ready for the beginning of an academic year. Another part of the plan was to undertake any required demolition during the summer so as to minimize noise and disturbance while classes were in session. Accordingly, cost and time were the major criteria to be satisfied. An appreciation for the complexity of the Tech reconstruction project can be obtained by considering the diversity of academic units involved. Eleven different departments and several research centers (which are, in effect, combinations of various department resources) from two different colleges [McCormick - engineering and Weinberg - chemistry, physics] are housed in the building, and each would be intimately affected by the reconstruction effort. Many departments would ultimately occupy space different from that currently occupied, and in some cases two or more moves will be necessary to make the transition. Faculty input was needed to develop customized floor plans for space utilization, while maintaining some semblance of overall uniformity in the finished building, and this task had to be accomplished by a faculty, which, albeit motivated and sincere, is generally inexperienced with this type of activity. Interaction and communication among faculty, administration, architect, designer, manager, and contractor was expected to present innumerable challenges to the patience of all, while approximately one wing per year undergoes reconstruction over the next eight years or so. And all of this had to be accomplished with as little disruption as possible to ongoing research and teaching activity, while simultaneously undertaking a fundraising campaign to acquire the monies to proceed. A major grant [from the McCormick Foundation] provided the means to initiate the project and complete the first few phases, but the timely completion of the project depended on the continued raising of adequate funds. Overview of Reconstruction Master Plan Skidmore Owings and Merrill (SOM) was chosen to design the first phase and create the overall master plan to ensure commonality of all succeeding phases. This master plan consisted of four basic components: exterior, interior, identity, HVAC, and utilities. The main exterior architecture challenge was the addition of penthouse spaces above all laboratory wings for the new heating, ventilation and air conditioning equipment. While the cladding of the penthouses had to be steel to reduce the load on the original structure, the limestone color scheme and fenestration design were maintained as shown in Figure 5. Ventilation louvers were located to continue the window lines and the exhaust was concentrated in one set of stacks to reduce visible clutter. Reconstruction meant complete demolition of the interior of the 1940 building. Only the columns, floors, and outside walls were left as shown by the photograph (Figure 6) of the Mechanical Engineering laboratory wing after demolition. Interior corridors in the lab wings were then rebuilt as outlined by the row of bottom block in the photograph. In addition to replacement of the original utilities, and addition of air conditioning, an Internet infrastructure had to be added that included computer trays hanging from the ceiling in the Figure 6 photograph.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Figure 5. View of the south side Tech loading Dock Besides reconcentrating departments, the main interior design signature is location of utilities in the ceilings of the lab wing corridors as shown in Figure 7. Ceiling exposure of the utilities allows for ease of maintenance and reconfiguration as laboratory needs evolve. As can be seen the floors are lightened from the original brown to the new white tile and each wing's floor trim is color-coded by department. The old numeric room numbering scheme, wing-floorclockwise 10 ft distance (eg. 2245) was changed to alpha-wing and numeric-street address (eg. A245), and the basement level, B, was renamed garden, G. Spatial identity is defined by addition of glass pained wooden doors at the junctions of the interior and laboratory corridors that name the departments within. Offices are confined to the interior and east-west corridors, which provide a clear distinction between office and laboratory space for air conditioning and security purposes. Classrooms and undergraduate laboratories are restricted to interior corridors of the garden, and first floors, except for the third floor library reading room and special second floor rooms.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Figure 6. This 1992 photo is a typical example of an area gutted before renovation There are two separate HVAC systems to minimize conditioning costs. Air for offices and classrooms spaces are continuously recirculated, while laboratory air is a passed through once as required by code. Before leaving, laboratory air passes through heat exchangers to capture heat energy. The enormous volume of equipment necessary to add air conditioning and heat exchangers is all housed on the new floor added above each wing. From a student-life standpoint there are two large changes, the most popular of which is the construction of a new cafeteria, Tech Express, in former undergraduate chemistry lab space. Pedagogically, most undergraduate laboratories are now concentrated in the central garden level below the cafeteria. This concentration brings freshmen together. In addition, six undergraduate computer laboratories have been established to support the new Engineering First curriculum.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Figure 7. Completed laboratory corridor with exposed utilities and color coded floor
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
CONSTRUCTION OF THE 1940 “TECH” BUILDING Charles Dowding Department of Civil and Environmental Engineering, Northwestern University 2145 Sheridan Road, A236, Evanston IL, 60208-3109
[email protected]
John C. Sanderson, who graduated from Northwestern in 1932 and retired from Northwestern as the University Architect, was the Resident Construction Superintendent and wrote an M.S. thesis, "Construction of the Northwestern Technological Institute Building." This interesting thesis details the construction and its history, and is summarized below. The building contractor, R. C. Wieboldt , whose home later became the current presidential mansion, was in personal charge of all phases of the project and did little else other than to supervise the job. Of course, who wouldn't take advantage of the opportunity of supervising such a large and interesting a project when it was only a three or four block walk away along the Lake Michigan shore. Financing the $5,000,000 job was shared between Northwestern and the Murphy Foundation. Northwestern would advance monthly payments (in the amount of approximately $300,000) to Wieboldt and then apply to the Murphy Foundation for reimbursement, after verification by a New York architectural firm. Initially the process required six weeks by mail, which meant Northwestern was out $600,000 for two weeks. This delay forced the team to invent electronic funds transfer -- albeit a rather crude version by today's standards -- by relying on phone conversation rather than written documents. Speaking of written documents, there were over 6000 letters and shop drawings issued during the two years of the construction. As is still customary today (alas, some things never change) each subcontract or manufacturer prepared special "shop drawings" showing in detail the construction of their component and the manner in which it was fitted into the building. Those who have built a kitchen will not be surprised that the largest number of shop drawings were those necessary for the laboratory furniture. A large number of last minute decisions about cabinetry seem to be part of any project. The second largest category is for the infrastructure components of Tech: electrical, plumbing, ventilation and heating. Copying of the drawings cost some $10,000. That was an enormous sum in those days. For comparison two freight elevators were purchased and installed for this $10,000. To make room for Tech, the Phi Kappa Psi house was moved and rotated from its former position, parallel and immediately south of the Sigma Chi house. Also the first Patton Gym and Dearborn observatory were demolished and moved respectively. The new Patton gym was immediately rebuilt. Completion of this enormous building in two years required that the concrete columns and floors be poured during the winter. Pouring of concrete requires temperatures of at least 50 degrees F. This environment was provided by enclosing Tech in a cocoon of over 400,000 square feet of canvas and warming the interior with coke fired salamanders. On December 2, 1940 an ominous sign of the coming world conflagration marred an otherwise excellent project as a fire broke out at 8:05 a.m. in the north east portion of the building. A vivid description is given in Sanderson's thesis and pictures of Tech in construction before and after the fire are shown as Figures 1 and 2.
Department of Civil and Environmental Engineering Infrastructure Technology Institute Northwestern University, Evanston, Illinois, 60208
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Figure 1. Partially Constructed Tech before the fire that shows the tarps in place for the curing of concrete
Figure 2. Remains of the reinforcing after the fire that consumed all formwork and led to the collapse of much of the concrete floors
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
The building construction was insured and after many weeks of conferences with plenty of "haggling," the price settlements were made. The total adjustment of loss was for approximately $608,000. The above delay and the subsequent delay caused by wrecking (completed April 8, 1941) and rebuilding set back the completion date so much that it pushed many items into war priority difficulties at the end of the job. The stone facing of Tech comes from Lannon Wisconsin and is a ripple marked, slightly fossiliferous dolomite, which is very weather resistant. Some 140,000 square feet or 5000 tons were placed by a crew of as many 100 masons. Since they might be working on the same wall at the same time, uniformity of appearance could be achieved only if the crew was accustomed to working together. Fortunately, Tech benefited by having its stone placed by the same crew of masons that had labored for the previous two years on Patton Gym and Scott Hall. To control the appearance, the specifications called for 50 percent of the stone to be rock face (parallel to the bedding or layering but not along a seam), 25 percent split face (perpendicular to the bedding), 15 percent bedding face, and 10 percent seam face (parallel to the bedding with a different color). In addition, there were specifications as to the percent of height and width ratios. The stone trim including the carved ornaments are Indiana Limestone from Bloomington Indiana, the quarries that were popularized by the "cutters" as the townies were called by the university students in the movie "Breaking Away." A total of 35,000 cubic feet or 1,800 tons were used. The carved figures of industrial and scientific processes and achievements near the main entrances were conceived and modeled by Edgar Miller for $7,500. Jon Johnson sculpted the figures after the stone was placed for $3,225. Replacement today would cost hundreds of times more. The interior tile along the corridors is made of a special vitricotta clay that is fired at ultra high temperatures to ensure vitrification of the clays. It was originally to be a manganese spot gray; however, the war-time footing curtailed the supply and the present iron spot buff was substituted. Before the vitricotta was selected it was required to withstand acids, alkalies, grease; ink and paint were required to be removed with ordinary solvents. Because of its abrasion resistance, all cutting required special carborundum saws. The building contained a number of special purpose facilities. A million pound testing machine was built in the north end of the Civil Engineering wing. It was the largest of its kind in the world at the time; not because of the capacity of the machine nor the three story height of the frame, but because of the length of the base. This base, which was in reality a huge concrete girder, was 54 feet long and 22 feet wide with the actual machine sitting in the center twelve feet. Some fourteen 3-1/2 in. diameter bolts anchored the base of the machine proper during a test. These bolts were approximately nine feet long and weighed about 300 pounds each. Next to the one million pound testing machine was a five million pound direct compression machine. The entire load of the machine was thrown on the reinforced concrete frame, which was 35 feet tall and could crush 10 feet tall specimens. Seven steel bands 2-3/4 in. thick and 8 in. wide, that withstood the tension, weighed about two tons apiece! This reinforcing had to be placed with a special derrick. One of the highlights of the building was the sound proof rooms, two of which were located in the sub-basement of the Physics Department and the third was on the third floor of the Electrical Engineering wing. While basically similar, the most interesting and perhaps the room most deserving of the title (never proven) "the quietest room on earth" was one of the rooms in the sub-basement. This room was a room within a room. The inner room was built of concrete block on a structural steel frame and weighed about 100,000 pounds and "floated" on 14 stacks of rubber cushions. The weight of the room was carefully checked and the rubber area calculated to give the optimum compression, so that the rubber- would be at its maximum absorption loading. U.S. Rubber Company engineers worked with the architect to determine this loading even to the extent of making up and testing sample stacks. The entire surface of the room inside and out was coated with Spray-O-Flake, which was ground up newspapers blown onto the surfaces with a bituminous binder. It is a very effective sound absorbent material and it was thought that it would absorb all stray exterior noises. Inside of the room the walls and ceilings were hung with 16 layers of muslin and flannel curtains supported from a pipe framework. These curtains were suspended on centers which varied from 1/2 to 3-1/2 in. on the theory that various length sound waves would be trapped in the spaces between curtains. The floor was covered with Blow-Knox Subway Grating under which was 3 in. of rock wool was covered with a layer of muslin. Dusting this room must have been quite a challenge.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
A shielded radio room was built on the second floor of the Electrical Engineering wing. This room had a complete copper lining grounded to the electrical conduit system. Over the floor, a linoleum rug was placed to protect the copper. The wooden door was lined on the back with copper, and a copper screen was even placed over the glass in the door. A high voltage laboratory was located on the north end of the Electrical Engineering wing. The most interesting feature of the room was its "ground-grid" system. Everything metallic in the room was bonded together in a grounded electrical grid. The wall grid was welded to the roof on one foot intervals at the ceiling line. In the floor grid, which was soldered to the wall grid, 1 in. x 1/8 in. copper bus bars, doubled, were laid out on approximately nine foot squares. Tinned copper lath was stretched over the entire floor, turned up at the edges and soldered at one foot intervals. At the intersections of the copper bus bars bronze sockets were soldered on to provide "ground" outlets -for experimental purposes. The copper bus bars were connected to large copper ground cables located in three corners of the room. These ground cables were also extended and connected to the roof. Each door frame, window frame, ladder or other metal object in the room was connected to the ground-grid. The southwest corner of the room was a specially prepared corner known as the "water test area" which was designed for testing of motors, insulators, etc. under severe water conditions.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
GEOTECHNICAL ENGINEERING AT NORTHWESTERN Originally published in “Tech Anthology II” M.E. Fine and M. Seniw Eds, McCormick, Northwestern University, Evanston, IL, Aug 2001. Raymond J. Krizek Department of Civil and Environmental Engineering, Northwestern University 2145 Sheridan Road, A236, Evanston IL, 60208-3109
[email protected]
Geotechnical engineering has been a strong and vibrant part of Civil Engineering at Northwestern University for the past six decades. When Jorj O. Osterberg (1943-1985) and Philip C. Rutledge (1943-1952) came in the middle 1940's, the newly created Technological Institute provided a home for the emerging discipline of soil mechanics, as it was then called, and early activities concentrated on the development of laboratory testing equipment. With the arrival of Robert L. Kondner (1960-1966) and Raymond J. Krizek (1961-present) in the early 1960's, the program assumed a distinct research flavor. Early projects dealt with the constitutive behavior of clays subjected to repetitive loading at high strain rates and the dynamic response of foundations. A highlight of that period was hosting the 1964 ASCE Specialty Conference on the Design of Foundations for Control of Settlement, which attracted about 600 attendees. Other major studies initiated in the middle to late 1960's were concerned with a variety of groundwater flow problems and the effect of micro-fabric (orientation and spacing of particles) on the macro-behavior (strength, compressibility, permeability, etc.) of clays. Geotechnical research during the 1970's impacted national design standards with several major projects in the areas of buried conduits, disposal of dredged materials, and blasting vibrations. Concrete pipe research involved two heavily instrumented field installations and the development of sophisticated finite element models in cooperation with the structures group. Field research in collaboration with the environmental group on the use of dredged materials for landfill and the investigation of effluent filtering systems for dredged material containment facilities involved many "vacations" to four disposal sites along the Maumee River in Toledo, as well as several other "resort areas" throughout the United States. Similar research was undertaken on flue gas scrubber sludge, a soil-like waste material produced by the removal of sulfur from the emissions of coal burning power plants. Rock mechanics and engineering geology were expanded by Charles H. Dowding (1976 - present) and after its initiation by Arley G. Franklin (1967-1974). Dowding's early work with the U.S. Bureau of Mines on the response spectrum analysis of blasting vibrations led to the adoption of a frequency-based criterion for allowable vibration levels in the United States. Continuation of this work led to his 1985 book entitled "Blast Vibration Monitoring and Control." He then expanded this work to cover vibrations produced by the entire constellation of construction activities and condensed the results in a book entitled "Construction Vibrations" published in 1996. In the mid1980's Dowding discovered that the deformation of cables grouted in earth materials produces wave reflections at the deformities that are proportional to the intensity of the localized shearing. This led to a new family of geomeasurement tools based upon Time Domain Reflectometry, and the results are summarized in a 1999 book entitled "GeoMeasurements by Pulsing TDR Cables and Probes." Along the way he worked with Ted Belytschko on several projects to develop a computational capability for modeling the three-dimensional response of caverns in jointed rock masses to earthquake shaking. This work led to the development of a rigid block model and the early use of parallel processors to calculate the earthquake response of million-block models of caverns. Subsequently, he developed a technique of vibration control by comparing environmental and blast-induced changes in crack width; this technique uses transducers and computers to continuously monitor and record crack widths caused by both long-term and short-term dynamic motions at the same location.
Department of Civil and Environmental Engineering Infrastructure Technology Institute Northwestern University, Evanston, Illinois, 60208
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
The arrival of Frank Somogyi (1981-1985) in the early 1980's coincided with the revitalization of the geotechnical laboratory facilities. The teaching laboratory built by Professor Osterberg in the 1940's was renovated and the multi-cell triaxial system was fitted with transducers to enable computer acquisition of data. A project with ALCOA afforded Professors Krizek and Somogyi an excellent opportunity to synthesize laboratory, field, and analytical efforts, tempered with a wealth of previous experience on the disposal of various wastes, to develop a dry stacking procedure for more effectively disposing of bauxite residue. Long-term interests in man-made soils and soil improvement techniques led to an intensive research program to assess the distribution of various grouts in soil and the resulting mechanical properties of the grouted mass; this work comprised the basis for several ASTM standards. A computerized robotic rock saw enabled Professor Dowding to replicate joints in real rock and opened a new era in laboratory testing of rock joints. Professor Osterberg continued his inventiveness by designing and commercializing a load cell for conducting relatively inexpensive in situ load tests on piles and drilled shafts. Other research projects dealt with subsidence over coal mines, acoustic emissions of rock fracture, blast densification of sand, permeability of clay liners, and behavior of slurrified wastes. From 1980 to 1992, Professor Krizek served as Chair of the Department and Professor Dowding contributed a substantial portion of his time and effort to head a committee established to oversee the complete reconstruction of Department space and facilities. After the untimely death of Professor Somogyi in 1985, Dr. Safdar Gill very capably assisted in the teaching of geotechnical engineering until the arrival of Richard J. Finno (1986 - present). Finno's work placed renewed emphasis on combining theory and practice to reconcile full-scale performance with analytical and numerical predictions. In a pile prediction event held in conjunction with the 1989 ASCE Foundation Engineering Conference at Northwestern, four piles were installed on the Lakefill, and predictions of performance under axial load were solicited from 23 practitioners and compared with the results of four load tests conducted over a span of 43 weeks before the Congress. The work at this test section led to the designation of the Lakefill site as an NSF-FHWA sponsored National Geotechnical Experimentation Site. In addition to the extensive site characterization work done at the site, two test sections were established; the first is a permanent non-destructive test section with foundations up to 90 feet deep, and the second is a grouted micropile test section where piles were installed, load tested, and then exhumed to evaluate load transfer mechanisms. Individual foundation elements in the first section were treated as a wave-guide and a theory was formulated to define the limits of conventional techniques and to develop new techniques for nondestructively sensing deep foundations. Other research has focused on the performance of supported excavations in soft clay. Detailed ground deformation and pore pressure responses were collected at the 1989 Howard-Dan Ryan subway extension project, a 40-foot deep excavation constructed with a flexible support system, and at the 1999 Chicago-State subway renovation project, a 40-foot deep excavation built with a stiff support system. As a result of capturing in the field the development of a shear band in the soft clays as excavation progressed, the results of the 1989 study led to the design and construction of a unique biaxial compression device wherein failure processes in soils can be studied in detail. Digital image analysis techniques have evolved so that the behavior inside thin zones of intensively sheared soils can be studied. The 1989 study led to a project that will extend the classical observational method to allow the numerical predictions of excavation performance to be updated in a timely fashion based on field observations. Our presence in the geoenvironmental arena was initiated by Professor Barbara-Ann Lewis (1979-present), who is primarily a member of the environmental engineering faculty, but has researched and taught in areas of strong interest to our faculty and students. Professor Joseph Feldkamp (1987-1993) worked extensively in the areas of geoenvironmental engineering, nonlinear consolidation, and the effects of electrokinetic phenomena on the permeability of clays. In addition, he invented an instrument to measure groundwater flow velocity and direction from a single well. Geoenvironmental research was continued by Howard Reeves (1994-2001), who emphasized numerical modeling of fluid flow and contaminant transport through unsaturated and saturated soils. His work included an extensive field and modeling study of shallow groundwater and salinity dynamics for a coastal saltmarsh, development and implementation of novel techniques for modeling soil vapor extraction though heterogeneous soils, and analysis of the transport of volatile liquids in the unsaturated zone. Professors Krizek and Reeves also led a consortium of Northwestern University, the Universities of Michigan and Wisconsin, and Argonne National Laboratories in a project to develop a curriculum consisting of twenty ten-hour modules for teaching graduate courses in geoenvironmental engineering. Professors Reeves, Dowding, and Igusa collaborated on an EPA-funded project to develop quantitative methods to direct exploration by combining three-dimensional geologic uncertainty with the sensitivity of three-dimensional finite element models; the approaches developed in this research have been successfully applied to building settlement, groundwater flow, and contaminant transport.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
The immense growth and stature of our program in Geotechnical Engineering during the past half of a century would not have been possible without the hard work and creative talent of many excellent graduate students. More than 110 Ph.D. and about 175 M.S. degrees have been awarded, and virtually all of our graduates have achieved a considerable measure of professional success. Approximately 50 are presently on the faculties of major academic institutions in over ten countries throughout the world, and many have advanced to high levels of administrative leadership. A number of very successful consulting and contracting companies have been established by Geotechnical Engineering alumni, and many others have risen to principal positions in other firms. Without doubt, much of the credit for our success story is attributable to "the guys in the trenches" who comprised and led the research teams that actually did the lab and field work. To all of these individuals, thanks for your contributions to our program, thanks for the espirit de corps you so unselfishly manifested, and thanks for your help in making Geotechnical Engineering at Northwestern University the success that it is today. Ray Krizek is Stanley F. Pepper Professor of Civil Engineering and Director of the MS program in Project Management.
Raymond J. Krizek Biography
From studying the microfabric of clays, to grouting with microfine cement, to determining the engineering behavior of various waste slurries, to measuring the soilstructure interaction of buried pipe, to calculating groundwater flows, to formulating constitutive relations for soils, Ray Krizek has integrated basic concepts from soil mechanics, engineering mechanics, and physico-chemical reactions to enhance our understanding of these phenomena. For more than forty years he helped hundreds of very talented graduate students from approximately 30 countries through the maze of academia to successful careers in teaching, research, and consulting. He is justly proud of their many achievements and feels privileged for the role he has played in their careers. Especially heartwarming is his close continuing relationship with many of them and the awareness that many have adapted elements of his ·philosophy· in their endeavors throughout the world. Born of Czechoslovakian heritage and raised in rural Maryland, Ray was strongly influenced by the depression-era hard-work ethics of his parents, teachers, and community. Memorable achievements during those early years were attaining the rank of Eagle Scout and hiking several hundred miles of the Appalachian Trail. An examinationbased scholarship allowed him to attend Johns Hopkins University, where he received his BE degree and an ROTC commission in the U.S. Army Corps of Engineers. Along the way he found time for his favorite pastime baseball at various amateur levels. Playing against Al Kaline and Johnny Podres at different stages of his rather lackluster career provided a strong incentive for him to redouble his interest in pursuing a career in academia. After a year with a computer company, two years in the Army, and four years as an instructor at the University of Maryland (during which time he received his MS), he came to Northwestern University in 1961 to earn his PhD. Upon completion, he joined the CE faculty and no one since has offered him a better position so he·s still at NU. Sports a ski trip with a Baltimore social club - even played a role in helping Ray to ultimately marry Claudia, his wife of 38 years. They have two sons Robert, a computer specialist and sports official, and Kevin, a professor in urban planning at the University of Minnesota and Iron-Man triathlete. In 1980 Ray became Department Chair, a position he was to hold longer than any other CE Chair. He inherited a 40-year-old building with hardening of the arteries and a university budget in the red, but his dogged emphasis on thriftiness and old fashion hard work led to a happy ending. Upon stepping down twelve
years later, he handed the new Chair the keys to a totally reconstructed Civil Engineering wing, a revitalized and refocused faculty with 50% new professors, a strong capability in cemetitious materials, and the nucleus of a first class environmental engineering group. Not to let any grass grow under his feet, Ray then established an entirely new Master of Project Management program, which has grown to more than 50 students. On the broader front he was a principal player in the establishment of the International Water Resources Association in the 1970s (serving on its charter Board of Directors) and the Geo-Institute within ASCE in the 1990s (serving as its second president). His major awards include Engineer of the Year (Illinois Section, ASCE), the Hogentogler Award, the Huber Prize, the Terzaghi Award, the Palmes Academiques (French Ministry of Education), election to the Spanish Academy of Engineering, and membership in the U.S. National Academy of Engineering.
PhD Students Raj Pal Khera ( India), 1967 Strength Response of an Anisotropically Consolated Clay Arley Graves Franklin (Unites States), 1968 Energy Dispensation and Nonlinear Mechanical Responses in a Karolin Clay Wallace Hayward Baker (United States), 1968 Mohr-Coulomb Strength Theory for Anisotropic Cohesive Soils Clarence Hubert Cavendish James (United States), 1968 Optimum Preload Rates for Compressible Normally Consolidated Soils Joseph Babalola Adeyeri (Nigeria), 1969 Multiple Integral Description of the Viscoelastic Response of Cohesive Soils Abdelhameed Ahmed Elnaggar (United States), 1970 Effect of Non-Darcian Flow on the Time Rate of Consolidation Francis Glen Mclean (United States), 1970 Analytical Investigation of Steady-State Flow Through Anisotropic and Layered Porous Media Robert Dean Holtz II(United States), 1970 Some Effects of Stress Path and Overconsolidation Ratio on the Shear Strength Properties of Georgia Kaolinite James Neil Kay (Australia), 1971 A Bayesian Approach to Soils Engineering Problems
Domalpally Babu Rao (India), 1971 Transient Seepage in Well Systems Enrique Castillo (Spain), 1972 Dispersion of a Contaminant in Jointed Rock Donald Earl Sheeran (United States), 1972 A Spectrophotometric Technique for the Fabric Analysis of Monomineralic Kaolin Soils Antonio Soriano (Spain), 1972 Application of Conformal Mapping to Transient Seepage Problems Antonio Santos-Moreno (Spain), 1972 Application of Summary Representaion to Certain Fluid Flow Problems Peter Karl Richard Krugmann (Germany) 1972 Placement Rates for Highway Embankments Paul Leslie Hummel (United States), 1973 Engineering Characteristics of Polluted Dredgings Dinesh Chandra Gupta (India), 1973 Coupled Sliding and Rocking of Harmonically Excited Continuous Foundations Tuncer Berat Edil (Turkey), 1973 Influence of Fabric and Soil-Water Potential on Stress-Strain Response of Clay Eduardo E. Alonso (Spain), 1973 Application of Random Function Theory to Settlement Problems in Soil Engineering Mohammed Hassan Farzin (Iran), 1973 Nonlinear Soil Behavior and its Effect on Soil-Structure Interaction Kanwarjit Singh Chawla (India), 1973 Effect of Fabric on Creep Response of Kaolinite Clay Ibrahim Kutay Ozaydin (Turkey), 1974 A Micro-Mechanistic Analysis for Creep Response of Kaolin Clay Max Wolfgang Giger (Switzerland), 1974 Application of Limit Analysis to Certain Problems in Geotechnical Engineering Vicente Cuellar (Spain), 1974 Rearrangement Measure Theory Applied to Dynamic Behavior of Sand
Manuel Casteleiro (Spain), 1975 Mathematical Model of One-Dimensional Consolidation and Desiccation of Dredged Materials Houssam El-Din Hafez El-Moursi (Egypt), 1975 Probabilistic Approach to One-Dimensional Consolidation Settlement Abdelsalem Muhammed Salem (Egypt), 1975 Behavior of Dredged Materials in Diked Containment Areas Mohamed Salah Abdelhamid (United States), 1975 At-Rest Earth Pressure of Clays During One-Dimensional Consolidation Jose Luis Monte (Spain), 1975 One-Dimensional Mathematical Model for Large Strain Consolidation Jack Leonard Rosenfarb (United States), 1975 Effect of Fabric on the Directional Shear Strength of a Kaolin Clay Dimitrios K. Atmatzidis (Greece), 1976 Methodology for Selection of Effluent Filtering Systems for Dredged Material Confinement Facilities Jau Scott Jin (Taiwan), 1976 Stabilization of Dredged Materials Federico Arrizabalaga (Spain), 1976 A Variational Lake Model with Depth-Dependent Eddy Viscosity Coefficient Riley M. Chung (United States), 1977 Directional Variation of Compressibility and Permeability in an Anisotropic Kaolin Clay Enrique Jose Socias (Spain), 1977 Experimental Study of Dispersion in a Jointed Rock Mass Atilla Mustafa Ansal (Turkey), 1977 An Endochronic Constitutive Law for Normally Consolidated Cohesive Soils Ali Abdussalam Elzaroughi (Libya), 1978 Application of Endochronic Constitutive Law to One-Dimensional Liquefaction of Sand Rafael Blazquez (Spain), 1978. Endochronic Model for Liqeufaction of Sand Deposits as Inelastic Two-Phase Media
Celal Sener (United States), 1979 An Endochronic Nonlinear Inelastic Constitutive Law for Cohesionless Soils Subjected to Dynamic Loading Roy H. Borden (United States), 1980 Time-Dependent Strength and Stress-Strain Behavior of Silicate-Grouted Sand Tal’at Abdul- Aziz Bader (Saudi Arabia), 1981 Injection and Distribution of Silicate Grouted Sand Mohamed Adel Benltayf (Libya), 1981 Effective Stress-Strain-Strength Behavior of Silicate-Grouted Sand Shi-Chih Chu (Taiwan), 1983 Geotechnical Properties and Disposal Considerations for Flue Gas Desulfurization Sludges Cumaraswamy Vipulanandan (Sri Lanka), 1984 Interactive Roles of Constituents on the Mechanical Behavior of Chemically Grouted Sand Saad A. Ghalib (Iraq), 1986 Development of Finite Element Program for Solution of Two-Dimensional Consolidation Hung-Jiun Liao (Taiwan), 1987 Role of Physical Components in Mechanical Behavior of Microfine Cement Sodium Silicate Grouted Sand Antonio A. Huerta (Spain), 1987 Numerical Modeling of Shurry Mechanics Luis L. Arenzana (Spain), 1987 An Experimental Investigation of the Properties and Behavior of Dilute Microfine Cement Grouts Bertrand Stephane Palmer (France), 1989 Model for Dry Stacking Thickened Slurries of Bauxite Residue Ahmed A. Hadavi (Iran), 1991 Improvement in Construction Productivity Through Goal Setting in a Unionized Environment Maan Helal (Syria), 1992 Microstructure of Microfine Cement Grouted Sand and its Anisotropic Behavior
Wei W. Lo (Taiwan), 1996 The Transfer of Construction Technology in a Newly Industrialized Country Khaled Sobhan (Bangladesh), 1996 Stabilized Fiber-Reinforcement Pavement Base Course with Recycled Aggregate Lois Geralyn Schwarz (United States), 1998 Role of Theology and Chemical Filtration on Injectability of Microfine Cement Grouts Rateb R. Sweis (United States), 1999 A Model to Assess Alternative Policies to Promote the Construction Industry in Developing Countries Richard William Sievert Jr. (United States), 2000 A Model for Managing Co-Marketing Alliances Paul D. Tennis (United States), 2001 Mass Transport Characteristics of Low-Permeability Materials using CounterDiffusion Chian Hsueng C Chao (Taiwan), 2001 Formulation of an E-Business Inter-Enterprise Collaboration for the Construction Industry
PUBLICATIONS of Dr. Raymond J. Krizek A Nondimensional Approach to the Static and Vibratory Loading of Footings, Highway Research Board, Bulletin Number 277, 1960, pp. 37-60 (with R. L. Kondner) Nondimensional Techniques Applied to Rigid Plate Bearing Tests on Flexible Pavements, Highway Research Board, Bulletin Number 289, 1961, pp. 80-90 (with R. L. Kondner) Maryland Engineering Soil Study, Highway Research Board, Bulletin Number 299, 1961, pp. 77-86 (with E. S. Barber and H. W. Piper) Correlation of Load Bearing Tests on Soils, Proceedings of the Highway Research Board, Volume 41, 1962, pp. 557-584 (with R. L. Kondner)
Lateral Stability of Rigid Poles Subjected to an Applied Couple, Soil and Foundation, Japanese Society of Soil Mechanics and Foundation Engineering, Volume 4, Number 1, August 1963, pp. 24-43 (with R. L. Kondner and B. B. Schimming) Settlement Response Caused by Footing Groups, Journal of the Soil Mechanics and Foundations Division, American Society of Civil Engineers, Volume 90, Number SM5, September 1964, pp. 269-287 (with R. L. Kondner) Application of the One-sided Fourier Transform to Determine Soil Storage and Dissipation Characteristics, Proceedings of the Symposium on Soil-Structure Interaction, Engineering Research Laboratory, University of Arizona, Tucson, Arizona, September 1964, pp. 625-633 Strength-Consistency Indices for a Cohesive Soil, Highway Research Board, Record Number 48, 1964, pp. 1-18 (with R. L. Kondner) A Vibratory Uniaxial Compression Device for Cohesive Soils, Proceedings of the American Society for Testing and Materials, Volume 64, 1964, pp. 934-943 (with R. L. Kondner) Use of a Miniature Specimen in Compression Tests of Cohesive Soils, Proceedings of the American Society for Testing and Materials, Volume 64, 1964, pp. 944-957 (with R. L. Kondner) Dynamic Response of Cohesive Soils for Earthquake Considerations, Proceedings of the Third World Conference on Earthquake Engineering, Volume 1, Part 1, New Zealand, January 1965, pp. 96-106 (with R. L. Kondner) Approximation for Terzaghi's Bearing Capacity Factors, Journal of the Soil Mechanics and Foundations Division, American Society of Civil Engineers, Volume 91, Number SM2, March 1965, pp. 1-3 Creep Compliance Response of a Cohesive Soil, Journal of The Franklin Institute, Volume 279, Number 5, May 1965, pp. 366-373 (with R. L. Kondner) Calculation of the Vertical Stress Distribution Beneath Groups of Footings, Geotechnique, The Institution of Civil Engineers, London, Volume 15, Number 4, December 1965, pp. 396-408 (with R. L. Kondner) Viscoelastic Principles and Techniques Applied to Soil Mechanics, Proceedings of the Symposium on Behavior of Soil under Stress, Volume 2, Bangalore, India, 1965, pp. 1-23 (with R. L. Kondner) Dynamic Clay Properties by Vibratory Compression, Proceedings of the International Symposium on the Effects of Repeated Loading of Materials and Structures, Volume 2, Mexico City, Mexico, September 1966, pp. 1-29 (with R. L. Kondner and H. J. Haas)
Rheologic Response Spectrum of a Soil, Proceedings of the Third Symposium on Earthquake Engineering, Volume 1, Roorkee, India, November 1966, pp. 241-262 (with R. L. Kondner) Correlation of Creep and Dynamic Response of a Cohesive Soil, Rheology and Soil Mechanics (edited by P. Kravtchenko and P. M. Sirieys), Springer-Verlag, 1966, pp. 333-342 (with R. L. Kondner) Factors Influencing Flexible Pavement Performance, National Cooperative Highway Research Program, Report 22, 1966, 69 pp. (with R. L. Kondner) Load-Deflection Response of Layered Flexible Pavement Sections under Rigid Bearing Plates, Soil and Foundation, Japanese Society of Soil Mechanics and Foundation Engineering, Volume 7, Number 2, March 1967, pp. 1-19 (with R. L. Kondner and E. Yamamoto) Stress-Strain Behavior of a Marine Clay, Proceedings of the Southeast Asian Regional Conference on Soil Engineering, Bangkok, Thailand, April 1967, pp. 95-102 Characteristics of Seepage into a Double-Wall Sheetpile Cofferdam, Proceedings of the Southeast Asian Regional Conference on Soil Engineering, Bangkok, Thailand, April 1967, pp. 393-399, (with W. H. Baker and A. G. Franklin) Assessment of Shear Strength Theories for Anisotropic Cohesive Soils, Proceedings of the Symposium on Pore Pressures and Shearing Resistance of Soils, New Delhi, India, May 1967, pp. 75-105 (with W. H. Baker) Soil Mechanics Demonstrations, Proceedings of the Third Panamerican Conference on Soil Mechanics and Foundation Engineering, Volume 2, Caracas, Venezuela, July 1967, pp. 375-389 (with E. S. Barber) Energy Dissipation in a Soft Clay, Proceedings of the International Symposium on Wave Propagation and Dynamic Properties of Earth Materials, Albuquerque, New Mexico, August 1967, pp. 797-807 (with A. G. Franklin) Measurement and Control of Radial Deformation in the Triaxial Test of Soils, Materials Research and Standards, American Society for Testing and Materials, Volume 7, Number 9, September 1967, pp. 392-396 (with R. P. Khera) Strain Rate Response of a Bangkok Clay, Proceedings of the Third Asian Regional Conference on Soil Mechanics and Foundation Engineering, Volume 1, Haifa, Israel, September 1967, pp. 289-292
Model Study of a Cohesionless Embankment on a Soft Soil, Proceedings of the Third Asian Regional Conference on Soil Mechanics and Foundation Engineering, Volume 1, Haifa, Israel, September 1967, pp. 333-337 (with J. 0. Osterberg and A. G. Franklin) Experimental Study of Pulse Velocities in Compacted Soils, Highway Research Board, Record Number 177, 1967, pp. 226-238 (with D. E. Sheeran and W. H. Baker) Strength Behavior of an Anisotropically Consolidated Remolded Clay, Highway Research Board, Record Number 190, 1967, pp. 8-18 (with R. P. Khera) Viscoelastic Shear Response of a Kaolinite, Clays and Clay Minerals, Volume 15, 1967, pp. 227-240 (with A. G. Franklin) Flow Around a Vertical Sheetpile Embedded in an Inclined Stratified Medium, Water Resources Research, American Geophysical Union, Volume 4, Number 1, February 1968, pp. 113-123 (with V. B. Anand) Effect of Principal Consolidation Stress Difference on Undrained Shear Strength, Soils and Foundations, Japanese Society of Soil Mechanics and Foundation Engineering, Volume 8, Number 1, March 1968, pp. 1-17 (with R. P. Khera) Unsteady Seepage Flow Between Fully-Penetrating Trenches, Journal of Hydrology, Volume 6, Number 4, August 1968, pp. 417-430 (with G. M. Karadi and H. A. Elnaggar) Constitutive Behavior of Clay Soils, Proceedings of the Fifth International Congress on Rheology, Volume 2, Kyoto, Japan, October 1968, pp. 471-491 Seasonal Variation in Flexible Pavement Deflections, Journal of the Highway Division, American Society of Civil Engineers, Volume 94, Number HW2, November 1968, pp. 219-228 Energy Dissipation of a Kaolinite at Different Water Contents, Clays and Clay Minerals, Volume 16, 1968, pp. 353-364 (with A. G. Franklin) Phenomenological Soil-Polymer Parallels, American Scientist, Volume 56, Number 3, 1968, pp. 279-287 Modified Theory for Seepage Characteristics of a Single Gravity Well, Highway Research Board, Record Number 223, 1968, pp. 63-71 (with G. M. Karadi) Permeability of Anisotropic Porous Media, Highway Research Board, Record Number 223, 1968, pp. 72-81 Nonlinear Dynamic Response of Soft Clay, Special Technical Publication 450, Vibration Effects of Earthquakes on Soils and Foundations, American Society for Testing and Materials, 1968, pp. 96-114 (with A. G. Franklin)
Torsional Shear Testing Technique for Dynamic Properties of Clay, Special Technical Publication 450, Vibration Effects of Earthquakes on Soils and Foundations, American Society for Testing and Materials, 1968, pp. 115-137 (with A. G. Franklin) Strength Anisotropy in Cohesive Soils, Journal of the Indian National Society of Soil Mechanics and Foundation Engineering, Volume 8, Number 1, January 1969, pp. 41-55 (with R. P. Khera) Pore Pressure Equation for Anisotropic Clays, Journal of the Soil Mechanics and Foundations Division, American Society of Civil Engineers, Volume 95, Number SM2, March 1969, pp. 719-724 (with W. H. Baker) Well Capacity for Continuous Permeability Variation, Journal of the Irrigation and Drainage Division, American Society of Civil Engineers, Volume 95, Number IR3, September 1969, pp. 409-414 (with W. H. Baker and A. G. Franklin) Optimum Preload Rates for Compressible Normally Consolidated Soils, Soils and Foundations, Japanese Society of Soil Mechanics and Foundation Engineering, Volume 9, Number 3, September 1969, pp. 58-81 (with J. O. Osterberg and C. H. C. James) MIMIC Source Approach to Viscoelastic Analysis, Proceedings of the Symposium on Application of Finite Element Methods in Civil Engineering, Nashville, Tennessee, November 1969, pp. 517-527 (with J. B. Adeyeri) Culvert Design in Some European Countries, Highway Research Board, Record Number 262, 1969, pp. 29-43 (with G. M. Karadi) Bearing Capacity of Purely Cohesive Soils with a Nonhomogeneous Strength Distribution, Highway Research Board, Record Number 282, 1969, pp. 48-56 (with W. H. Baker and C. H. C. James) Effectiveness of a Leaky Sheetpile, Highway Research Board, Record Number 282, 1969, pp. 57-62 (with G. M. Karadi) Complex Viscosity of a Kaolin Clay, Clays and Clay Minerals, Volume 17, Number 2, 1969, pp. 101-110 (with A. G. Franklin) Solutions to Boundary Value Problems of Stresses and Displacements in Earth Masses and Layered Systems, Highway Research Board, Bibliography Number 48, 1969, pp. 1-142 (with D. Hampton, B. B. Schimming, and E. L. Skok, Jr.) Mohr-Coulomb Strength Theory for Anisotropic Soils, Journal of the Soil Mechanics and Foundations Division, American Society of Civil Engineers, Volume 96, Number SM1, January 1970, pp. 269-292 (with W. H. Baker)
Behavior of Rectangular Footings on Dense Sand, Journal of the Indian National Society of Soil Mechanics and Foundation Engineering, Volume 9, Number 1, January 1970, pp. 51-62 Adaptation of Elastic Theory to the Design of Circular Conduits, Civil Engineering Transactions, The Institution of Engineers, Australia, Volume CE 12, Number 1, April 1970, pp. 85-90 (with J. N. Kay) Critical Evaluation of Certain Methods of Unsteady Groundwater Hydraulics, Water Resources Bulletin, Volume 6, Number 3, June 1970, pp. 424-438 (with G. M. Karadi and M. Rechea) Properties of Slightly Organic Topsoils, Journal of the Construction Division, American Society of Civil Engineers, Volume 96, Number CO1, June 1970, pp. 29-43 (with R. D. Holtz) Evaluation of Methods for Inspection and Quality Control of Compacted Earth Embankments, The Construction Specifier, Part I, Volume 23, Number 6, June 1970, pp. 23-28; Part II, Volume 23, Number 7, July 1970, pp. 47-53 (with R. D. Holtz and D. E. Sheeran) Lateral Forces on Rigid Unyielding Vertical Retaining Walls due to Surcharge Loading, Proceedings of the Fourth Brazilian Conference on Soil Mechanics and Foundation Engineering, Rio de Janeiro, Brazil, August 1970, Volume 2 (with J. O. Osterberg and H. C. Hall) Hysteresis in Soft Clay, Proceedings of the Fourth Symposium on Earthquake Engineering, Volume 1, Roorkee, India, November 1970, pp. 283-288 (with A. G. Franklin) Spring Analog Model for Flexible Culvert Behavior, Highway Research Board, Record Number 310, 1970, pp. 29-39 (with J. N. Kay) Statistical Approximation for Consolidation Settlement, Highway Research Board, Record Number 323, 1970, pp. 87-96 (with H. A. Elnaggar) Influence of Poisson's Ratio on the Surface Deflection of Layered Systems, Highway Research Board, Record Number 337, 1970, pp. 1-10 (with J. O. Osterberg and G. A. Ali) Multiple Integral Description of Nonlinear Viscoelastic Behavior of a Clay Soil, Transactions of the Society of Rheology, Volume 14, Number 3, 1970, pp. 375-392 (with J. D. Achenbach and J. B. Adeyeri) Seepage Characteristics of Imperfect Cutoffs, Journal of the Soil Mechanics and Foundations Division, American Society of Civil Engineers, Volume 97, Number SM1, January 1971, pp. 305-312 (with F. G. McLean)
Transient Seepage Characteristics of a Single Well Fully Penetrating an Anisotropic Artesian Aquifer, Proceedings of the Symposium on Water Resources Engineering, Part B, Section 12, Bangalore, India, May 1971, pp. 1-11 (with G. M. Karadi and D. B. Rao) Preparation of Homogeneous Soil Samples by Slurry Consolidation, Journal of Materials, American Society for Testing and Materials, Volume 6, Number 2, June 1971, pp. 356-373 (with D. E. Sheeran) Effect of Non-Darcian Behavior on the Characteristics of Transient Flow, Journal of Hydrology, Volume 13, Number 2, June 1971, pp. 127-138 (with G. M. Karadi and H. A. Elnaggar) Drawdown in a Well Group Along a Straight Line, Ground Water, Journal of the Technical Division, National Water Well Association, Volume 9, Number 4, July-August 1971, pp. 12-18 (with D. B. Rao and G. M. Karadi) Vibratory Densification of Damp Clayey Sands, Journal of the Soil Mechanics and Foundations Division, American Society of Civil Engineers, Volume 97, Number SM8, August 1971, pp. 1069-1079 (with J. I. Fernandez) Effects of Stress Path and Overconsolidation Ratio on the Shear Strength of a Kaolin Clay, Proceedings of the Fifth Regional Conference for Africa on Soil Mechanics and Foundation Engineering, Luanda, Angola, Volume 1, Part 3, August 1971, pp. 17-25 (with R. D. Holtz) On Certain Plasticity Solutions in Soil Mechanics, Journal of The Franklin Institute, Volume 292, Number 3, September 1971, pp. 153-167 (with M. Rechea) Influence of AASHO Road Test Local Factors on Present Serviceability Index for Flexible Pavement Systems, Proceedings of the Seminar on Strengthening of Existing Road Pavements, Part 1, Indian Roads Congress, Srinagar (Kashmir), September 1971, pp. 25-44 (with R. L. Kondner and N. Hasan) Statistical Evaluation of Soils Test Data, Proceedings of the First International Conference on Applications of Statistics and Probability in Soil and Structural Engineering, Hong Kong, September 1971, pp. 230-266 (with R. D. Holtz) Estimation of the Mean for Soil Properties, Proceedings of the First International Conference on Applications of Statistics and Probability in Soil and Structural Engineering, Hong Kong, September 1971, pp. 280-286 (with J. N. Kay) Unsteady Drawdown in Well Groups, Journal of the Hydraulics Division, American Society of Civil Engineers, Volume 97, Number HY10, October 1971, pp. 1625-1638 (with D. B. Rao and G. M. Karadi)
Analysis of Uncertainty in Settlement Prediction, Geotechnical Engineering, Journal of the Southeast Asian Society of Soil Engineering, Volume 2, Number 2, December 1971, pp. 119-129 (with J. N. Kay) Rheologic Behavior of Clay Soils Subjected to Dynamic Loads, Transactions of the Society of Rheology, Volume 15, Number 3, 1971, pp. 443-489 Rheologic Behavior of Cohesionless Soils Subjected to Dynamic Loads, Transactions of the Society of Rheology, Volume 15, Number 3, 1971, pp. 491-540 Rational Structural Analysis and Design of Pipe Culverts, National Cooperative Highway Research Program, Report 116, 1971, 155 pp. (with R. A. Parmelee, J. N. Kay, and H. A. Elnaggar) Probabilistic Approach to the Determination of Safety Factors for the Bearing Capacity of Cohesive Soils, Highway Research Board, Record Number 345, 1971, pp. 69-76 (with J. N. Kay) Gravity Flow to Excavations and Drainage Trenches in Layered Aquifers, Highway Research Board, Record Number 360, 1971, pp. 65-76 (with F. G. McLean) Hydrorheology of Clay Soils, Transactions of the Society of Rheology, Volume 15, Number 4, 1971, pp. 771-781 (with J. D. Achenbach and J. B. Adeyeri) Non-Darcian Flow in Clay Soils, Proceedings of the Symposium on Measurement and Control of Flow in Science and Industry, Instrument Society of America, Pittsburgh, Pennsylvania, 1971, pp. 53-61 (with H. A. Elnaggar and G. M. Karadi) Summary Representation Applied to Seepage Problems, Journal of the Engineering Mechanics Division, American Society of Civil Engineers, Volume 98, Number EM1, February 1972, pp. 141-157 (with G. M. Karadi and A. Santos) Unconfined Flow Through Jointed Rock, Water Resources Bulletin, Volume 8, Number 2, April 1972, pp. 266-281 (with E. Castillo and G. M. Karadi) Comparison of Dispersion Characteristics of Fissured Rock, Proceedings of the Second Symposium on the Fundamentals of Transport Phenomena in Porous Media, Volume 2, Guelph, Ontario, August 1972, pp. 778-797 (with G. M. Karadi and E. Castillo) Dispersion of a Contaminant in Fissured Rock, Proceedings of the International Symposium on Percolation Through Fissured Rock, Stuttgart, Germany, September 1972, Part T3-C, pp. 1-15 (with G. M. Karadi and E. Socias) Hydrodynamic Dispersion in a Single Rock Joint, Journal of Applied Physics, Volume 43, Number 12, December 1972, pp. 5013-5021 (with E. Castillo and G. M. Karadi)
Coupled Sliding and Rocking of Embedded Foundations, Journal of the Soil Mechanics and Foundations Division, American Society of Civil Engineers, Volume 98, Number SM12, December 1972, pp. 1347-1358 (with D. C. Gupta and R. A. Parmelee) Effects of Lattice Disorder on the Quantitative Determination of Kaolinite by X-ray Diffraction, Proceedings of the 1972 International Clay Conference, Madrid, Spain, 1972 pp. 827-835 (with T. B. Edil) Uncertainty of Settlement Analysis in Overconsolidated Clays, Highway Research Board, Record Number 405, 1972, pp. 143-151 (with J. N. Kay) Material Properties Affecting Soil-Structure Interaction of Underground Conduits, Highway Research Board, Record Number 413, 1972, pp. 13-29 (with J. N. Kay) Coupled Sliding and Rocking Vibrations of a Rigid Foundation on an Elastic Medium, Proceedings of the Symposium on Behavior of Earth and Earth Structures Subjected to Earthquakes and Other Dynamic Loads, Roorkee, India, Volume 1, March 1973, pp. 113-121 (with D. C. Gupta and R. A. Parmelee) Stability Charts for Inhomogeneous Soil Conditions, Geotechnical Engineering, Journal of the Southeast Asian Society of Soil Engineering, Volume 4, Number 1, June 1973, pp. 1-13 (with P. K. Krugmann) Unsteady Flow to Bottom Drain in Bounded Aquifer, Journal of the Irrigation and Drainage Division, American Society of Civil Engineers, Volume 99, Number IR2, June 1973, pp. 169-182 (with A. Soriano and I. Gyuk) Probabilistic Approach to Heave of Soft Clay Around Sheetpile Walls, Proceedings of the Eighth International Conference on Soil Mechanics and Foundation Engineering, Volume 1, Part 3, Moscow, U.S.S.R., August 1973, pp. 143-150 (with R. B. Corotis) Disposal of Polluted Dredgings from the Great Lakes Area, Proceedings of the First World Congress on Water Resources, International Water Resources Association, Chicago, Illinois, September 1973, Volume 4, pp. 482-491 (with G. M. Karadi) Effect of Non-Darcian Flow on Time Rate of Consolidation, Journal of The Franklin Institute, Volume 296, Number 5, November 1973, pp. 323-337 (with H. A. Elnaggar and G. M. Karadi) Consolidation Characteristics of Dredging Slurries, Journal of the Waterways, Harbors, and Coastal Engineering Division, American Society of Civil Engineers, Volume 99, Number WW4, November 1973, pp. 439-457 (with A. M. Salem) Unsteady Drawdown at a Partially Penetrating Well in a Transversely Isotropic Artesian Aquifer, Ground Water, Journal of the Technical Division, National Water Well
Association, Volume 11, Number 6, November-December 1973, pp. 44-49 (with D. B. Rao and G. M. Karadi) Bounds for Tunnel Roof Support, Proceedings of the Ninth Canadian Symposium on Rock Mechanics, Montreal, Canada, December 1973, pp. 299-322 (with M. W. Giger) Theoretical Study of Dispersion in a Fractured Rock Aquifer, Journal of Geophysical Research, American Geophysical Union, Volume 78, Number 3, 1973, pp. 558-573 (with E. Castillo and G. M. Karadi) Electrical Resistivity Survey of Two Dredging Disposal Sites, Bulletin of the Association of Engineering Geologists, Volume 10, Number 2, 1973, pp. 107-119 (with M. W. Giger and A. G. Franklin) Vertical Consolidation Due to Vertical and Radial Flow, Acta Technica, Hungarian Academy of Science, Volume 75, 1973, pp. 235-259 (with P. K. Krugmann) Effect of Particle Characteristics on Wave Velocity, Journal of the Geotechnical Engineering Division, American Society of Civil Engineers, Volume 100, Number GT1, January 1974, pp. 89-94 (with F. G. McLean and M. W. Giger) Sampling of Maintenance Dredgings, Journal of Testing and Evaluation, American Society for Testing and Materials, Volume 2, Number 3, May 1974, pp. 139-145 (with P. L. Hummel) Vibration Behavior of Foundations on Layered Media, Geotechnical Engineering, Journal of the Southeast Asian Society of Soil Engineering, Volume 5, Number 1, June 1974, pp. 1-19 (with R. A. Parmelee and D. C. Gupta) Precompression Analysis for Highway Embankments, Proceedings of the Specialty Conference on Analysis and Design in Geotechnical Engineering, American Society of Civil Engineers, Austin, Texas, June 1974, Volume 1, pp. 111-141 (with P. K. Krugmann) Collection and Analysis of Representative Samples of Dredging Slurries, Proceedings of the Symposium on Water Resources Instrumentation, International Water Resources Association, Chicago, Illinois, June 1974, Volume 1, pp. 400-408 (with L. A. Raphaelian) Comparative Determinations of Organic Matter in Polluted Maintenance Dredgings, Proceedings of the Symposium on Water Resources Instrumentation, International Water Resources Association, Chicago, Illinois, June 1974, Volume 2, pp. 492-504 (with B. M. Katz) Storage Capacity of Diked Containment Areas for Polluted Dredgings, Proceedings of the Sixth World Congress on Dredging, Taipei, Taiwan, August 1974, pp. 354-364 (with M. W. Giger)
Disposition of Pollutants During Dredging and Disposal Operation, Proceedings of the Sixth World Congress on Dredging, Taipei, Taiwan, August 1974, pp. 368-383 (with R. Y. Lai) A Spectrophotometric Technique for the Fabric Analysis of Monomineralic Kaolinitic Soils, Journal of Testing and Evaluation, American Society for Testing and Materials, Volume 2, Number 5, September 1974, pp. 323-330 (with D. E. Sheeran) Nonlinear Stress-Strain Formulation for Soils, Journal of the Geotechnical Engineering Division, American Society of Civil Engineers, Volume 100, Number GT9, September 1974, pp. 993-1008 (with R. B. Corotis and M. H. Farzin) Initial Distribution of Average Excess Pore Water Pressure Due to a Trapezoidal Load, Soils and Foundations, Japanese Society of Soil Mechanics and Foundation Engineering, Volume 14, Number 3, September 1974, pp. 67-79 (with P. K. Krugmann) Randomness of Settlement Rate Under Stochastic Load, Journal of the Engineering Mechanics Division, American Society of Civil Engineers, Volume 100, Number EM6, December 1974, pp. 1211-1226 (with E. E. Alonso) Evaluation of Stress Cell Performance, Journal of the Geotechnical Engineering Division, American Society of Civil Engineers, Volume 100, Number GT12, December 1974, pp. 1275-1295 (with M. H. Farzin, A. E. Z. Wissa, and R. T. Martin) Environmental Effects of Pumped Storage Development, Proceedings of the International Symposium on Multipurpose Storage-Pumping Schemes, International Water Resources Association, Madrid, Spain, 1974, pp. 291-303 (with G. M. Karadi) Experimental Study of Clay Deformability in Terms of Initial Fabric and Soil-Water Potential, Rheologica Acta, Volume 13, Number 4/5, 1974, pp. 803-813 (with T. B. Edil) Stress Propagation in a Frequency-Dependent Clay Soil, Transactions of the Society of Rheology, Volume 18, Number 3, 1974, pp. 411-429 (with R. N. Yong and J. C. Dutertre) Application of Conformal Mapping to Transient Tile Drainage, Acta Technica, Hungarian Academy of Science, Volume 79, 1974, pp. 203-223 (with A. Soriano and I. Gyuk) Behavior of Beams on Randomly Nonhomogeneous Bases, Transportation Research Board, Record Number 510, 1974, pp. 77-91 (with E. E. Alonso) Field Performance of Reinforced Concrete Pipe, Transportation Research Board, Record Number 517, 1974, pp. 30-42 (with R. B. Corotis) Seismic Survey of a Hydraulic Landfill Composed of Maintenance Dredgings, Bulletin of the Association of Engineering Geologists, Volume 11, Number 3, 1974, pp. 173-202 (with A. G. Franklin and A. Soriano)
Inverse Method for Determining Approximate Stress-Strain Behavior of Soils, Journal of Testing and Evaluation, American Society for Testing and Materials, Volume 3, Number 1, January 1975, pp. 51-61 (with R. B. Corotis and M. H. Farzin) Micromechanics Model for Creep of Anisotropic Clay, Journal of the Engineering Mechanics Division, American Society of Civil Engineers, Volume 101, Number EM1, February 1975, pp. 57-78 (with Z. P. Bazant and I. K. Ozaydin) The Water Resources Engineer As a Problem Solver, Proceedings of the International Seminar on Water Resources Education, International Water Resources Association, Paris, France, March 1975, pp. 257-264 (with G. M. Karadi) Effect of Grain Characteristics on Packing of Sands, Proceedings of the Istanbul Conference on Soil Mechanics and Foundation Engineering, Volume 1, Istanbul, Turkey, March-April 1975, pp. 46-54 (with J. S. Zelasko and T. B. Edil) Shear Behavior of Sands As a Function of Grain Characteristics, Proceedings of the Istanbul Conference on Soil Mechanics and Foundation Engineering, Volume 1, Istanbul, Turkey, March-April 1975, pp. 55-64 (with J. S. Zelasko and T. B. Edil) Organic Content and Engineering Behavior of Typical Maintenance Dredgings, Proceedings of the Fourth Southeast Asian Conference on Soil Engineering, Kuala Lumpur, Malaysia, April 1975, Part 3, pp. 6-15 (with M. W. Giger and P. L. Hummel) Synthesis of Soil Moduli Determined from Different Types of Laboratory and Field Tests, Proceedings of the Specialty Conference on In-Situ Measurement of Soil Properties, American Society of Civil Engineers, Raleigh, North Carolina, June 1975, Volume 1, pp. 225-240 (with R. B. Corotis) Stability Analysis of Vertical Cut with Variable Corner Angle, Soils and Foundations, Japanese Society of Soil Mechanics and Foundation Engineering, Volume 15, Number 2, June 1975, pp. 63-71 (with M. W. Giger) Saturated Sand As an Inelastic Two-Phase Medium, Journal of the Engineering Mechanics Division, American Society of Civil Engineers, Volume 101, Number EM4, August 1975, pp. 317-332 (with Z. P. Bazant) Stochastic Formulation of Soil Properties, Proceedings of the Second International Conference on Applications of Statistics and Probability in Soil and Structural Engineering, Aachen, Germany, September 1975, pp. 9-32 (with E. E. Alonso) Statistical Evaluation of Soil Index Properties and Constrained Modulus, Proceedings of the Second International Conference on Applications of Statistics and Probability in Soil and Structural Engineering, Aachen, Germany, September 1975, pp. 273-293 (with R. B. Corotis and A. S. Azzouz)
Statistical Optimization of Friction Pile Foundations, Proceedings of the Second International Conference on Applications of Statistics and Probability in Soil and Structural Engineering, Aachen, Germany, September 1975, pp. 523-544 (with S. P. Wagner) Empirical Liquefaction Index for Sands, Proceedings of the Fifth European Conference on Earthquake Engineering, Istanbul, Turkey, September 1975 (with S. W. Nunnally and T. B. Edil) Dewatering of Dredged Materials by Evaporation, Proceedings of the First International Conference on Dredging Technology, University of Kent, Canterbury, England, September 1975, Section F3, pp. 29-38 (with M. W. Giger and J. S. Jin) Assessment of the Dredging and Disposal Problem in the United States, Proceedings of the First International Conference on Dredging Technology, University of Kent, Canterbury, England, September 1975, Section H3, pp. 49-56 (with M. W. Giger and A. M. Salem) A Micromechanistic Formulation for Stress-Strain Response of Clay, International Journal of Engineering Science, Volume 13, Numbers 9/10, September/October 1975, pp. 831-840 (with T. B. Edil and T. Mura) Lake Shore Management and Dredging Waste Disposal, Proceedings of the Second World Congress on Water Resources, International Water Resources Association, New Delhi, India, December 1975, Volume 4, pp. 347-353 (with G. M. Karadi) Statistical Analysis of Constrained Soil Modulus, Transportation Research Board, Record Number 537, 1975, pp. 59-68 (with R. B. Corotis and J. H. Salazar-Espinosa) Evaluation of Modulus and Poisson's Ratio from Triaxial Tests, Transportation Research Board, Record Number 537, 1975, pp. 69-80 (with R. B. Corotis and M. H. Farzin) Consolidation of Randomly Heterogeneous Clay Strata, Transportation Research Board, Record Number 548, 1975, pp. 30-46 (with E. E. Alonso) Probabilistic Approach to Prediction of Consolidation Settlement, Transportation Research Board, Record Number 548, 1975, pp. 47-61 (with R. B. Corotis and H. H. El-Moursi) Preparation and Identification of Clay Samples with Controlled Fabric, Engineering Geology, Volume 9, 1975, pp. 13-38 (with T. B. Edil and I. K. Ozaydin) Quantitative Dependence of Strength on Particle Orientation of Clay, Bulletin of the International Association of Engineering Geology, Number 11, 1975, PP- 19-22 (with T. B. Edil) Water Reuse in Perspective, Proceedings of the International Symposium on Water for Arid Lands, International Water Resources Association, Tehran, Iran, 1975 (with G. M. Karadi)
Secondary Compression of Maintenance Dredgings, Proceedings of the Fifth Panamerican Conference on Soil Mechanics and Foundation Engineering, Buenos Aires, Argentina, 1975 (with A. M. Salem) Compressibility and Strength of Compacted Dredgings, Proceedings of the Seventh Dredging Seminar, Center for Dredging Studies, Texas A&M University, College Station, Texas, 1975, pp. 140-153 (with M. W. Giger) Permeability and Drainage Characteristics of Dredgings, Proceedings of the Seventh Dredging Seminar, Center for Dredging Studies, Texas A&M University, College Station, Texas, 1975, pp. 154-179 (with J. S. Jin and A. M. Salem) Stability of Vertical Corner Cut with Concentrated Surcharge Load, Journal of the Geotechnical Engineering Division, American Society of Civil Engineers, Volume 102, Number GT1, January 1976, pp. 31-40 (with M. W. Giger) Dredged Material Confinement Facilities As Solid-Liquid Separation Systems, Proceedings of the Specialty Conference on Dredging and Its Environmental Effects, American Society of Civil Engineers, Mobile, Alabama, January 1976, pp. 609-632 (with J. A. FitzPatrick and D. K. Atmatzidis) Stress-Deformation-Time Behavior of Dredgings, Journal of the Geotechnical Engineering Division, American Society of Civil Engineers, Volume 102, Number GT2, February 1976, pp. 139-157 (with A. M. Salem) Directional Drying Rates for Anisotropic Clays, Journal of the Geotechnical Engineering Division, American Society of Civil Engineers, Volume 102, Number GT2, February 1976, pp. 188-194 (with J. L. Rosenfarb and M. S. Abdelhamid) Endochronic Constitutive Law for Liquefaction of Sand, Journal of the Engineering Mechanics Division, American Society of Civil Engineers, Volume 102, Number EM2, April 1976, pp. 225-238 (with Z. P. Bazant) Water Quality Effects of a Dredging Disposal Area, Journal of the Environmental Engineering Division, American Society of Civil Engineers, Volume 102, Number EE2, April 1976, pp. 389-409 (with B. J. Gallagher and G. M. Karadi) Regression Analysis of Soil Compressibility, Soils and Foundations, Japanese Society of Soil Mechanics and Foundation Engineering, Volume 16, Number 2, June 1976, pp. 19-29 (with R. B. Corotis and A. S. Azzouz) Variational Model for Lake Circulation, Proceedings of the Second International Symposium on Finite Element Methods in Flow Problems, Rapallo, Italy, June 1976 (with G. M. Karadi and F. Arrizabalaga)
At-Rest Lateral Earth Pressure of a Consolidating Clay, Journal of the Geotechnical Engineering Division, American Society of Civil Engineers, Volume 102, Number GT7, July 1976, pp. 721-738 (with M. S. Abdelhamid) Spatial Nonhomogeneity of Dredged Materials in Confined Disposal Areas, Proceedings of the Seventh World Congress on Dredging, San Francisco, California, July 1976, pp. 779-797 Investigation of Effluent Filtering Systems for Dredged Material Containment Facilities, Contract Report D-76-8, U.S. Army Engineer Waterways Experiment Station, Vicksburg, Mississippi, August 1976 (with J. A. FitzPatrick and D. K. Atmatzidis) Soil-Structure Interaction of Concrete Pipe Systems, Proceedings of the Specialty Conference on Methods of Structural Analysis, American Society of Civil Engineers, Volume 2, Madison, Wisconsin, August 1976, pp. 607-643 (with R. B. Corotis and T. H. Wenzel) Characterization and Handling of Sulfur Dioxide Scrubber Sludge with Flyash, Proceedings of the International Symposium on New Horizons in Construction Materials, Lehigh University, Bethlehem, Pennsylvania, November 1976, pp. 67-81 (with M. W. Giger and L. K. Legatski) Engineering Properties of Sulfur Dioxide Scrubber Sludge with Flyash, Proceedings of the International Symposium on New Horizons in Construction Materials, Lehigh University, Bethlehem, Pennsylvania, November 1976, pp. 83-94 (with M. W. Giger and L. K. Legatski) Characterization and Usefulness of Dredged Materials, Proceedings of the International Symposium on New Horizons in Construction Materials, Lehigh University, Bethlehem, Pennsylvania, November 1976, pp. 95-109 (with A. M. Salem) One-Dimensional Mathematical Model for Large-Strain Consolidation, Geotechnique, The Institution of Civil Engineers, London, Volume 26, Number 3, 1976, pp. 495-510 (with J. L. Monte) Consolidation Around Sand Drains in Non-Darcian Soils, Acta Technica, Hungarian Academy of Science, Volume 82, 1976, pp. 99-119 (with H. A. Elnaggar and A. S. Azzouz) Influence of Fabric and Soil Suction on Mechanical Behavior of a Kaolinitic Clay, Geoderma, Volume 15, 1976, pp. 323-341 (with T. B. Edil) Probabilistic Analysis of Predicted and Measured Settlements, Canadian Geotechnical Journal, Volume 14, Number 1, February 1977, pp. 17-33 (with R. B. Corotis and H. H. El-Moursi)
Time-Dependent Development of Strength in Dredgings, Journal of the Geotechnical Engineering Division, American Society of Civil Engineers, Volume 103, Number GT3, March 1977, pp. 169-184 (with A. M. Salem) Leaching Characteristics of Polluted Dredgings, Journal of the Environmental Engineering Division, American Society of Civil Engineers, Volume 103, Number EE2, April 1977, pp. 197-215 (with J. S. Jin and G. L. Roderick) Densification and Hysteresis of Sand Under Cyclic Shear, Journal of the Geotechnical Engineering Division, American Society of Civil Engineers, Volume 103, Number GT5, May 1977, pp. 399-416 (with V. Cuellar, Z. P. Bazant, and M. L. Silver) Field Performance of a Dredgings Disposal Area, Proceedings of the Specialty Conference on Geotechnical Practice for Disposal of Solid Waste Materials, American Society of Civil Engineers, Ann Arbor, Michigan, June 1977, pp. 358-383 (with A. M. Salem) Chemical Stabilization of Dredged Materials, Proceedings of the Specialty Conference on Geotechnical Practice for Disposal of Solid Waste Materials, American Society of Civil Engineers, Ann Arbor, Michigan, June 1977, pp. 517-540 (with G. L. Roderick and J. S. Jin) Seismic Refraction Surveying in Soils with Variable Propagation Velocity, Soils and Foundations, Japanese Society of Soil Mechanics and Foundation Engineering, Volume 17, Number 2, June 1977, pp. 1-15 (with A. Soriano and A. G. Franklin) Site Use Evaluation for Dredged Material Landfills, Journal of the Urban Planning and Development Division, American Society of Civil Engineers, Volume 103, Number UP1, July 1977, pp. 39-51 (with J. S. deBettencourt and G. L. Peterson) Preparation of Isotropically Consolidated Clay Samples with Random Fabrics, Journal of Testing and Evaluation, American Society for Testing and Materials, Volume 5, Number 5, September 1977, pp. 406-412 (with T. B. Edil) Characteristics of Dredged Bottom Sediments, Journal of the Waterway, Port, Coastal, and Ocean Division, American Society of Civil Engineers, Volume 103, Number WW4, November 1977, pp. 471-486 (with D. K. Atmatzidis and J. A. FitzPatrick) Physical and Conceptual Simulation of Effluents from Dredged Material Confinement Facilities, Water Resources Bulletin, Volume 13, Number 6, December 1977, pp. 1107-1118 (with D. K. Atmatzidis and J. A. FitzPatrick) Indirect Determination of K0 from Multi-Stage Triaxial Compression Tests, Geotechnical Engineering, Journal of the Southeast Asian Society of Soil Engineering, Volume 8, Number 2, December 1977, pp. 31-52 (with M. S. Abdelhamid)
Desiccation and Consolidation of Dredged Materials, Journal of the Geotechnical Engineering Division, American Society of Civil Engineers, Volume 103, Number GT12, December 1977, pp. 1399-1418 (with M. Casteleiro and T. B. Edil) Effect of Particle Arrangement on Electrical Resistivity of Clay, Proceedings of the First National Symposium on Expansive Soils, Kanpur, India, December 1977, Part I-2, pp. 1-8 (with J. L. Rosenfarb and D. M. O'Shaughnessy) Fabric Effects on Strength and Deformation of Kaolin Clay, Proceedings of the Ninth International Conference on Soil Mechanics and Foundation Engineering, Tokyo, Japan, Volume 1, 1977, pp. 169-176 Composition of Polluted Bottom Sediments in Great Lakes Harbors, Chemistry of Marine Sediments (edited by T. F. Yen), Ann Arbor Science Publishers, Ann Arbor, Michigan, 1977, pp. 111-123 (with B. M. Katz and P. L. Hummel) Fate of Pesticides in Bottom Sediments During Dredging and Disposal Cycle, Chemistry of Marine Sediments (edited by T. F. Yen), Ann Arbor Science Publishers, Ann Arbor, Michigan, 1977, pp. 157-162 (with L. A. Raphaelian) Estimate for Bearing Capacity of a Prismatic Pillar, Rock Mechanics, Journal of the International Society for Rock Mechanics, Volume 9, Number 4, 1977, pp. 189-211 (with M. W. Giger) Soil Stresses and Displacements in a Concrete Pipe Trench Installation, Transportation Research Board, Record Number 640, 1977, pp. 52-58 (with R. B. Corotis and T. H. Wenzel) Uncertainty Analysis of Settlement Rate, Transportation Research Board, Record Number 616, 1977, pp. 81-84 (with R. B. Corotis and H. H. El-Moursi) Directional Creep Response of Anisotropic Clays, Geotechnique, The Institution of Civil Engineers, London, Volume 27, Number 1, 1977, pp. 37-51 (with T. B. Edil and K. S. Chawla) Assessment of Soil Constitutive Models for Numerical Analysis of Buried Concrete Pipe Systems, Special Technical Publication 630, Concrete Pipe and the Soil-Structure System, American Society for Testing and Materials, 1977, pp. 76-90 (with D. K. Atmatzidis) Analysis and Measurement of Soil Behavior Around Buried Concrete Pipe, Special Technical Publication 630, Concrete Pipe and the Soil-Structure System, American Society for Testing and Materials, 1977, pp. 91-104 (with R. B. Corotis) Primary Consolidation and Compressibility of Dredgings, Proceedings of the Ninth Dredging Seminar, Center for Dredging Studies, Texas A&M University, College Station, Texas, 1977, pp. 208-232 (with A. M. Salem and A. S. Azzouz)
Stabilization of Polluted Dredgings by Electro-Osmosis, Proceedings of the Ninth Dredging Seminar, Center for Dredging Studies, Texas A&M University, College Station, Texas, 1977, pp. 233-260 (with F. B. Gularte and P. L. Hummel) Flocculation and Sedimentation of Fresh-water Dredged Material Slurries, Proceedings of the Second International Conference on Dredging Technology, Texas A&M University, College Station, Texas, 1977 (with J. S. Jin and G. L. Roderick) Filter Systems for Dredgings Confinement Areas, Journal of the Environmental Engineering Division, American Society of Civil Engineers, Volume 104, Number EE1, February 1978, pp. 31-45 (with J. A. FitzPatrick and D. K. Atmatzidis) Constitutive Equation for Cyclic Behavior of Cohesive Soils, Proceedings of the Specialty Conference on Earthquake Engineering and Soil Dynamics, American Society of Civil Engineers, Pasadena, California, Volume 1, June 1978, pp. 557-568 (with Z. P. Bazant and A. M. Ansal) Migration of Contaminants during Electro-osmotic Dewatering of Polluted Dredgings, Proceedings of the Third World Congress on Water Resources, International Water Resources Association, Sao Paulo, Brazil, June 1978 Behavior of Buried Concrete Pipe, Journal of the Geotechnical Engineering Division, American Society of Civil Engineers, Volume 104, Number GT7, July 1978, pp. 815-836 (with P. V. McQuade) Micro-characteristics of Chemically Stabilized Granular Materials, Journal of the Geotechnical Engineering Division, American Society of Civil Engineers, Volume 104, Number GT7, July 1978, pp. 939-952 (with K. M. O'Connor and D. K. Atmatzidis) Directional Vane Strengths for Different Clay Fabrics and Stress Conditions, Soils and Foundations, Japanese Society of Soil Mechanics and Foundation Engineering, Volume 18, Number 3, September 1978, pp. 1-11 (with J. L. Rosenfarb) Estimate of Soil Compressibility from Standard Penetration Test, Geotechnical Engineering, Journal of the Southeast Asian Society of Soil Engineering, Volume 9, Number 2, December 1978, pp. 1-12 (with R. B. Corotis and H. H. El-Moursi) Disposition of Dredged Material, Reclamation of Drastically Disturbed Lands (edited by F. W. Schaller and P. Sutton), American Society of Agronomy, 1978, Chapter 35, pp. 629-644 (with D. K. Atmatzidis) Endochronic Constitutive Law for Soils, Proceedings of the Sixth European Conference on Earthquake Engineering, Dubrovnik, Yugoslavia, Volume 4, 1978, pp. 9-14 (with Z. P. Bazant and A. M. Ansal)
Granular Media Filtration of Dredging Effluents, Journal of the Waterway, Port, Coastal, and Ocean Division, American Society of Civil Engineers, Volume 105, Number WW1, February 1979, pp. 33-50 (with D. K. Atmatzidis and J. A. FitzPatrick) Viscoplasticity of Normally Consolidated Clays, Journal of the Geotechnical Engineering Division, American Society of Civil Engineers, Volume 105, Number GT4, April 1979, pp. 519-537 (with Z. P. Bazant and A. M. Ansal) Statics Matrix for Analysis of Pile Group with Rigid Cap, Proceedings of the Third International Conference on Numerical Methods in Geomechanics, Aachen, West Germany, April 1979, pp. 1061-1068 (with M. W. Giger) Landfill Management for Double Alkali Sulfur Dioxide Scrubber Sludge, Journal of the Energy Division, American Society of Civil Engineers, Volume 105, Number EY2, August 1979, pp. 229-239 Viscoplasticity of Transversely Isotropic Clays, Journal of the Engineering Mechanics Division, American Society of Civil Engineers, Volume 105, Number EM4, August 1979, pp. 549-565 (with Z. P. Bazant and A. M. Ansal) Strength and Stiffness of Silicate Grouted Sand with Different Stress Histories, Geotechnical Testing Journal, American Society for Testing and Materials, Volume 2, Number 4, December 1979, pp. 200-205 (with D. C. Diefenthal, R. H. Borden and W. H. Baker) Dredgings Containment Areas as Sedimentation Basins, Proceedings of the Eleventh Dredging Seminar, Center for Dredging Studies, Texas A&M University, College Station, Texas, 1979, pp. 69-89 (with D. K. Atmatzidis and B. J. Gallagher) A General Procedure for Simulating EPR Spectra of Partially Oriented Paramagnetic Centers, Journal of Magnetic Resonance, Volume 36, 1979, pp. 259-268 (with J. C. Swartz, B. M. Hoffman, and D. K. Atmatzidis) Engineering Properties of Three Double Alkali Scrubber Sludges, Journal of Civil Engineering Design, Volume 1, Number 1, 1979, pp. 69-93 (with R. H. Borden and B. R. Christopher) Buried Conduits, Structural Engineering Handbook (edited by E. H. Gaylord and C. N. Gaylord), McGraw-Hill Book Company, New York, Second Edition, Chapter 25, 1979,pp. 1-32 Site Factors Controlling Liquefaction, Journal of the Geotechnical Engineering Division, American Society of Civil Engineers, Volume 106, Number GT7, July 1980, pp. 785-801 (with R. M. Blazquez and Z. P. Bazant)
Analysis of Leaky Sheetpile by Summary Representation, International Journal for Numerical and Analytical Methods in Geomechanics, Volume 4, Number 3, July-September 1980, pp. 199-213 (with G. M. Karadi and E. DeMiguel) Stabilization of Double Alkali Scrubber Sludges, Journal of Civil Engineering Design, Volume 2, Number 3, 1980, pp. 279-304 (with B. R. Christopher and S. D. Scherer) Mathematical Model for One-dimensional Desiccation and Consolidation of Sedimented Soils, International Journal for Numerical and Analytical Methods in Geomechanics, Volume 5, Number 2, April-June 1981, pp. 195-215 (with M. Casteleiro and T. B. Edil) Prediction of Soil Behavior by Endochronic Theory, Limit Equilibrium, Plasticity and Generalized Stress-Strain in Geotechnical Engineering, American Society of Civil Engineers, 1981, pp. 286-327 (with A. M. Ansal and Z. P. Bazant) Critical Appraisal of Endochronic Theory for Soils, Limit Equilibrium, Plasticity, and Generalized Stress-Strain in Geotechnical Engineering, American Society of Civil Engineers, 1981, pp. 539-552 (with Z. P. Bazant and A. M. Ansal) Creep Behavior of Silicate-Grouted Sand, Proceedings of the Specialty Conference on Grouting in Geotechnical Engineering, American Society of Civil Engineers, New Orleans, Louisiana, February 1982, pp. 450-469 (with R. H. Borden and W. H. Baker) Effective Stress-Strain-Strength Behavior of Silicate Grouted Sand, Proceedings of the Specialty Conference on Grouting in Geotechnical Engineering, American Society of Civil Engineers, New Orleans, Louisiana, February 1982, pp. 482-497 (with D. K. Atmatzidis and M. A. Benltayf) Injection and Distribution of Silicate Grout in Sand, Proceedings of the Specialty Conference on Grouting in Geotechnical Engineering, American Society of Civil Engineers, New Orleans, Louisiana, February 1982, pp. 540-563 (with T. A. Bader) Seismic Analysis of an Earth Dam Based on Endochronic Theory, Proceedings of the International Symposium on Numerical Models in Geomechanics, Zurich, Switzerland, September 1982, pp. 559-576 (with Z. P. Bazant and A. M. Ansal) Modelo Variacional para Circulacion de Lagos, Proceedings of the Symposium on Applications of the Finite Element Method in Engineering, Barcelona, Spain, December 1982, Part III, pp. 1-15 (with F. Arrizabalaga and G. M. Karadi) Erosion Model for Reclamation Areas, Proceedings of the Symposium on Surface Mining, Hydrology, Sedimentology, and Reclamation, Lexington, Kentucky, December 1982, pp. 339-348 (with C. Vipulanandan)
Endochronic Models for Soils, Soil Mechanics - Transient and Cyclic Loads (edited by G. N. Pande and O. C. Zienkiewicz), John Wiley and Sons, 1982, pp. 419-438 (with Z. P. Bazant and A. M. Ansal) Design and Control of Chemical Grouting, Report Number FHWA/RD-82-037, U.S. Department of Transportation, Federal Highway Administration, Volume 2, April 1983 (with W. H. Baker) Hysteretic Endochronic Theory for Sand, Journal of the Engineering Mechanics Division, American Society of Civil Engineers, Volume 109, Number EM4, August 1983, pp. 1073-1095 (with Z. P. Bazant and C. L. Shieh) Quality of Run-off Water from Soil-covered Reclamation Site, Proceedings of the National Symposium on Surface Mining, Hydrology, Sedimentology, and Reclamation, Lexington, Kentucky, December 1983 (with C. Vipulanandan) Perspectives on Modelling Consolidation of Dredged Materials, Sedimentation-Consolidation Models, American Society of Civil Engineers, San Francisco, California, October 1984, pp. 296-332 (with F. Somogyi) Permanence of Chemically Grouted Sands, Issues in Dam Grouting, American Society of Civil Engineers, Denver, Colorado, April 1985, pp. 1-26 (with M. G. Madden) Chemical Grouting in Soils Permeated by Water, Journal of Geotechnical Engineering, American Society of Civil Engineers, Volume 111, Number 7, July 1985, pp. 898-915 (with T. Perez) Evaluation of Adhesion in Chemically Grouted Geomaterials, Geotechnical Testing Journal, American Society for Testing and Materials, Volume 8, Number 4, December 1985, pp. 184-190 (with C. Vipulanandan) Tensile Properties of Chemically Grouted Sand, Transportation Research Board, Record 1008, 1985, pp. 80-89 (with C. Vipulanandan) Control of Chemical Grout Injected in Seepage Domain, Proceedings of the Eleventh International Conference on Soil Mechanics and Foundation Engineering, San Francisco, California, Volume 3, 1985, pp. 1217-1220 (with T. Perez) Mechanical Behavior of Chemically Grouted Sand, Journal of Geotechnical Engineering, American Society of Civil Engineers, Volume 112, Number 9, September 1986, pp. 869-887 (with C. Vipulanandan) Modelling Grouted Sand Under Torsional Loading, Transportation Research Record 1104, National Research Council, Washington, D.C., 1986, pp. 33-42 (with C. Vipulanandan)
Large-Strain Consolidation of Fine-Grained Slurries, Proceedings of the Symposium on Consolidation and Disposal of Phosphatic and Other Waste Clays, Florida Phosphate Institute, Lakeland, Florida, May 1987, Paper No. 3, pp. 1-22 (with J. P. Trin and B. Palmer) Geotechnical Properties and Landfill Disposal of FGD Sludge, Proceedings of the Specialty Conference on Geotechnical Practice for Waste Disposal, American Society of Civil Engineers, Ann Arbor, Michigan, June 1987, pp. 625-639 (with S. C. Chu and D. K. Atmatzidis) Thickened Slurry Disposal Method for Process Tailings, Proceedings of the Specialty Conference on Geotechnical Practice for Waste Disposal, American Society of Civil Engineers, Ann Arbor, Michigan, June 1987, pp. 728-743 (with B. Palmer) Modelling Cyclic Elastic Behaviour of Sands, Soil Dynamics and Earthquake Engineering, Volume 6, Number 2, 1987, pp. 90-99 (with A. M. Ansal and H. K. Ansal) Evaluating Ground Response to Tunnelling with FEM, Proceedings of the International Symposium on Tunneling for Water Resources and Power Projects, Delhi, India, January 1988 (with R. J. Finno) Permeability and Compressibility of Slurries from Seepage-Induced Consolidation, Journal of Geotechnical Engineering, American Society of Civil Engineers, Volume 114, Number 5, May 1988, pp. 614-627 (with A. Huerta and G. Kriegsmann) Full-Scale Load Test of Caisson on Chicago Hardpan, Proceedings of the Second International Conference on Case Histories in Geotechnical Engineering, St. Louis, Missouri, June 1988, Volume 2, pp. 1303-1308 (with S. A. Bucher and D. K. Atmatzidis) Properties of Sedimented Double Alkali FGD Slurries, Proceedings of the Specialty Conference on Hydraulic Fill Structures, American Society of Civil Engineers, Fort Collins, Colorado, August 1988, pp. 778-794 (with S. Tatioussian and D. K. Atmatzidis) Stability Analysis of a Slurry Deposited Site, Proceedings of the Specialty Conference on Hydraulic Fill Structures, American Society of Civil Engineers, Fort Collins, Colorado, August 1988, pp. 591-605 (with F. Masse and D. K. Atmatzidis) Laboratory Testing of Chemically Grouted Sand, Geotechnical Testing Journal, American Society for Testing and Materials, Volume 12, Number 2, pages 109-118, June 1989 (with B. R. Christopher and D. K. Atmatzidis) Injection of Dilute Microfine Cement Suspensions into Fine Sands, Proceedings of the Twelfth International Conference on Soil Mechanics and Foundation Engineering, Rio de Janeiro, Brazil 1989 (with L. Arenzana)
Injection of Fine Sands with Very Fine Cement Grout, Journal of Geotechnical Engineering, American Society of Civil Engineers, December 1989, pp. 1717-1733 (with S. Zebovitz and D. K. Atmatzidis) Buried Conduits, Structural Engineering Handbook (edited by E. H. Gaylord and C. N. Gaylord), McGraw-Hill Book Company, New York, Third Edition, Chapter 29, 1990, pp. 1-33 Ethics in Engineering, Proceedings of the National Forum on Education and Continuing Professional Development for the Civil Engineer, American Society of Civil Engineers, Las Vegas, Nevada, April 1990, pp. 1007-1012 (with Bradley J. Hulbert) Effects of Mixing on Rheological Properties of Microfine Cement Grout, Proceedings of the Conference on Grouting, Soil Improvement, and Geosynthetics, American Society of Civil Engineers, New Orleans, Louisiana, February 1992, pp. 512-525 (with L. G. Schwarz) Preferred Orientation of Pore Structure in Cement-Grouted Sand, Proceedings of the Conference on Grouting, Soil Improvement, and Geosynthetics, American Society of Civil Engineers, New Orleans, Louisiana, February 1992, pp. 526-540 (with M. Helal) Anisotropic Behavior of Cement-Grouted Sand, Proceedings of the Conference on Grouting, Soil Improvement, and Geosynthetics, American Society of Civil Engineers, New Orleans, Louisiana, February 1992, pp. 541-550 (with M. Helal) Microfine Cement/Sodium Silicate Grout, Proceedings of the Conference on Grouting, Soil Improvement, and Geosynthetics, American Society of Civil Engineers, New Orleans, Louisiana, February 1992, pp. 676-687 (with H. J. Liao and R. H. Borden) Mechanical Properties of Microfine Cement/Sodium Silicate Grouted Sand, Proceedings of the Conference on Grouting, Soil Improvement, and Geosynthetics, American Society of Civil Engineers, New Orleans, Louisiana, February 1992, pp. 688-699 (with H. J. Liao and R. H. Borden) Engineering Properties of Acrylate Polymer Grout, Proceedings of the Conference on Grouting, Soil Improvement, and Geosynthetics, American Society of Civil Engineers, New Orleans, Louisiana, February 1992, pp. 712-724 (with D. F. Michel, M. Helal, and R. H. Borden) Permanence of Grouted Sands Exposed to Various Water Chemistries, Proceedings of the Conference on Grouting, Soil Improvement, and Geosynthetics, American Society of Civil Engineers, New Orleans, Louisiana, February 1992, pp. 1403-1419 (with J. M. Siwula) Short Term Goal Setting for Construction, Journal of Construction Engineering and Management, American Society of Civil Engineers, Volume 119, Number 3, September 1993, pp. 622-630 (with A. Hadavi)
Bleed and Rheology of Cement Grouts, Proceedings of the International Conference on Grouting in Rock and Concrete, Salzburg, Austria, October 1993, A. A. Balkema, Rotterdam, pp. 55-64 (with L. G. Schwarz) Influence of Grout Pressure on Capacity of Boreinjected Piles and Anchors, Proceedings of the International Conference on Grouting in Rock and Concrete, Salzburg, Austria, October 1993, A. A. Balkema, Rotterdam, pp. 159-165 (with I. M. Kleyner) Compaction Fabrics of Pelites: Experimental Consolidation of Kaolinite and Implications for Analysis of Strain in Slate, Journal of Structural Geology, Volume 15, Number 9/10, 1993, pp. 1123-1137 (with D. W. Baker and K. S. Chawla) Properties of Cement-Grouted Sand with Distance from Injection, Proceedings of the Thirteenth International Conference on Soil Mechanics and Foundation Engineering, New Delhi, India, 1994 (with L. G. Schwarz) Coupled Numerical Solutions to Plane Strain Consolidation Problem, Proceedings of the Eighth International Conference on Computer Methods and Advances in Geomechanics, Morgantown, West Virginia, May 1994 (with S. Ghalib) Difficulties With Implementation of Goal Setting for Construction, Journal of Management in Engineering, American Society of Civil Engineers, Volume 10, Number 5, Sept./Oct. 1994, pp. 48-54 (with A. Hadavi) Activation of Microfine Slag Cement Grout, Proceedings of the First International Conference on Alkaline Cements and Concrete, Kiev, Ukraine, Volume 2, October 1994, pp.1009-1019 (with L. G. Schwarz) Effect of Preparation Technique on Permeability and Strength of Cement-Grouted Sand, Geotechnical Testing Journal, American Society for Testing and Materials, Volume 17, Number 4, December 1994, pp. 434-443 (with L. G. Schwarz) Test Procedures to Evaluate Absorption and Swelling of Grout, Geotechnical Testing Journal, American Society for Testing and Materials, Volume 17, Number 4, December 1994, pp. 511-515 (with R. H. Borden) Grouting Gasoline-Contaminated Sand with Microfine Cement, Proceedings of the Conference on Characterization, Containment, Remediation, and Performance in Environmental Geotechnics, American Society of Civil Engineers, New Orleans, Louisiana, February 1995, pp. 1366-1380 (with L. G. Schwarz) Effects of Gasoline Contamination on Microfine Cement Grouts and Grouted Sands, Proceedings of the Tenth Panamerican Conference on Soil Mechanics and Foundation Engineering, Guadalajara, Mexico, October 1995 (with L. G. Schwarz)
Mathematical Model for Bore-Injected Cement Grout Installations, Journal of Geotechnical Engineering, American Society of Civil Engineers, Volume 121, Number 11, November 1995, pp. 782-788 (with I. M. Kleyner) Fiber-Reinforced Recycled Concrete as a Stabilized Base Course for Highway Pavements, Fiber Composites in Infrastructure, Proceedings of the First International Conference on Composites in Infrastructure, Tuscon, Arizona, January 1996, pp. 9961011 (with K. Sobhan) Lessons Learned From a Multi-Phase Reconstruction Project, Journal of Construction Engineering and Management, American Society of Civil Engineers, Volume 122, Number 1, March 1996, pp. 44-54 (with W. Lo and A. Hadavi) Multi-Phase Flow Effects During Injection of Gasoline Contaminated Sands with Microfine Cement Grout, Proceedings of the Third International Symposium on Environmental Geotechnology, San Diego, California, June 1996 (with L.G. Schwarz) Anisotropy of Gasoline-Contaminated Sands Grouted with Microfine Cement, Proceedings of the Second International Congress on Environmental Geotechnics, Osaka, Japan, November 1996, pp. 1067-1072 (with L. G. Schwarz) Waste Fibers in Cement-Stabilized Recycled Aggregate Base Course Material, Transportation Research Record No. 1486, Transportation Research Board, National Research Council, Washington, D.C., 1997, pp. 97-106 (with J. K. Cavey, K. Sobhan, and W.H. Baker) Repeated Loading of Stabilized Recycled Aggregate Base Course, Geotechnical Special Publication 79, Geo-Institute, American Society of Civil Engineers, Reston, Virginia, 1998, pp. 180-194 (with K. Sobhan) Contractor Selection Process for Taipei Mass Rapid Transit System, Journal of Management in Engineering, American Society of Civil Engineers, Volume 14, Number 3, May/June 1998 (with W. Lo, C. H. Chao, and A. Hadavi) Resilient Properties and Fatigue Damage in a Stabilized Recycled Aggregate Base Course Material, Transportation Research Record No. 1611, Transportation Research Board, National Research Council, Washington, D.C., 1998, pp. 28-37 (with K. Sobhan) Effects of High Prequalification Requirements, Construction Management and Economics, Volume 17, Reading, England, 1999, pp. 603-612; (with W. Lo and A. Hadavi) Fatigue Behavior of Fiber-reinforced Recycled Aggregate Base Course, Journal of Materials in Civil Engineering, American Society of Civil Engineers Volume 11, Number 2, May 1999, pp.124-130 (with K. Sobhan)
Geotechnics of High Water Content Materials, Special Technical Publication 1374, American Society for Testing and Materials, February 2000, pp. 3-28 Spatial and Directional Variations in Engineering Properties of an In-Situ SilicateGrouted Sand, Geotechnical Special Publication 104, Advances in Grouting and Ground Modification, American Society of Civil Engineers, August 2000, pp. 139-154 (with M. Spino) Evolving Morphology of Early Age Microfine Cement Grout, Geotechnical Special Publication 104, Advances in Grouting and Ground Modification, Geo-Institute, American Society of Civil Engineers, August 2000, pp. 181-199 (with L. G. Schwarz) Construction Industry Business Processes Modeled with Object-Oriented Approach, Proceedings of the Seventeenth International Symposium on Automation and Robotics in Construction, Taipei, Taiwan, September 2000, pp. 1059-1064 (with C. H. Chao, A. Hadavi, and W. Lo) Toward a Supply Chain Collaboration in the E-Business Era for the Construction Industry, Proceedings of the Seventeenth International Symposium on Automation and Robotics in Construction, Taipei, Taiwan, September 2000, pp. 1183-1188 (with C. H. Chao and A. Hadavi) Strength Changes in Dredged Materials Due to Aging, Proceedings of the Fifteenth International Conference on Soil Mechanics and Geotechnical Engineering, Istanbul, Turkey, August 2001, Volume 1, pp. 55-58 (with T. Calmeau)
Symposium Papers Measurement & Performance Acceptance of New Technology in Geotechnical Practice K. O'Connor, GeoTDR Inc., USA Field and Laboratory Techniques for Measurement of Dynamic Behavior of Soils V. Cuéllar, A. Santos, J. L. Monte, Laboratorio de Geotecnia (CEDEX), Spain Performance Observations of the World Trade Center Slurry Wall J. Moskowitz, Mueser Rutledge Consulting Engineers, USA Load Testing High Capacity Drilled Shafts J. O. Osterberg, Professor Emeritus Northwestern University,Consultant, Aurora CO., USA
Materials & Behavior Behavior of Polyurethane Grouts Used for Leak Control in Wastewater Systems C. Vipulanandan, University of Houston, USA Scaling Laws for Sea Ice Fracture Z. P. Bazant, Northwestern University, USA Particle Size Effects on Rockfill Compressibility E. E. Alonso and D. Montobbio, Universitat Politècnica de Catalunya, Spain and Asdoconsult Ingenieros S.L., Spain Vibrations Caused by Chiselling in Deep Foundations Construction C. Bouniol, H. Duplaine, Balineau SA, France Chemical Transport Issues in Modern Landfill Liners T. B. Edil, University of Wisconsin – Madison, USA
Analysis & Design Design and Construction of Reinforced Earth Walls on Marginal Lands A. Abraham, R. E. Allen, J. E. Sankey, K.Truong, H. Tran, The Reinforced Earth Co, USA A Generalized Probabilistic Approach to the Stability of Cut Slopes J. N. Kay, Griffith University, Australia Innovative High Rise Foundation Design in Chicago C. Barker, T. D. Bushell, T. A. Kiefer, R. Diebold, STS Consultants, USA, and ThorntonTomasetti Engineers, USA Lessons Learned from a Bermed Excavation in Soft Clay H. J. Liao, National Taiwan University of Science and Technology, Taiwan, R.O.C. A New Slope Stability Approach Using Calculus of Variations, and Safety and Sensitivity Analysis E. Castillo and R. Mínguez, University of Cantabria, Spain, and University of Castilla La Mancha, Spain Uncertainties in Simplified Liquefaction Analysis of Soil Deposits R. Blázquez, S. López, V. Navarro, J. Sánchez, E. González, Universidad de Castilla – La Mancha, Spain
Teaching & Management Early 1970’S Engineering Education at Northwestern M. H. Farzin, EES Corporation, USA Teaching Research – The Successful Legacy of RJK R. D. Holtz, University of Washington, USA Role of University Educators in Developing Future Leaders of Engineering Enterprise S. Abdelhamid and T. B. Edil, CH2M Hill, USA, and University of Wisconsin-Madison, USA
Measurement & Performance Acceptance of New Technology in Geotechnical Practice K. O'Connor, GeoTDR Inc., USA Field and Laboratory Techniques for Measurement of Dynamic Behavior of Soils V. Cuéllar, A. Santos, J. L. Monte, Laboratorio de Geotecnia (CEDEX), Spain Performance Observations of the World Trade Center Slurry Wall J. Moskowitz, Mueser Rutledge Consulting Engineers, USA Load Testing High Capacity Drilled Shafts J. O. Osterberg, Northwestern University, USA
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
ACCEPTANCE OF NEW TECHNOLOGY IN GEOTECHNICAL PRACTICE Kevin O’Connor GeoTDR Inc. 720 Greencrest Drive, Westerville, Ohio 43081
[email protected]
ABSTRACT New technologies provide new insights which may not be readily accepted by the community of geotechnical practitioners. Rather than considering the wide diversity of applications of TDR technology, this paper concentrates on monitoring of mine subsidence. In particular, it concentrates on monitoring movement within the rock and soil over underground mines which has been perceived as having a very low benefit for the cost involved. Two case histories are used to illustrate this application as a more proactive approach since it provides monitoring of precursor movement compared with measurement of surface movement which is reactive. Displacements along the rock mass discontinuities cause deformation of embedded coaxial cables which is monitored using time domain reflectometry (TDR). A significant implication of these TDR measurements is the finding that precursor subsurface movement occurs ahead of the mine face, and outside the edges of the panel being mined. Deformation has occurred within the overburden over 1000 ft in front of the mine face. Furthermore, data show that movement has occurred at locations more than 400 ft laterally from active mining. It is important to emphasize that the precursor movement is shear deformation along discontinuities within the rock mass. This behavior is not taken into consideration with the conventional angle of draw concept that only accounts for the vertical displacement and surface subsidence. Subsurface movement occurs well beyond the limits of the angle of draw. The second example involves monitoring subsidence over abandoned underground mines. In this case, the precursor movement indicates that the site is actively experiencing subsidence and movements are detected in a time frame which allows for rational assessment and implementation of a plan of action. The liability and economic implications of these findings will stimulate both acceptance and resistance to TDR technology. INTRODUCTION A Brief Chronology Geotechnical applications of TDR technology have been stimulated through the interest, imagination, and efforts of many people with a diversity of backgrounds. This is a case in which clever people became aware of a principle utilized by the electronic and electrical community and started playing with applications. It began with applications using coaxial cylinders to measure dielectric properties (Felner-Feldegg, 1969; De Loor, 1968) and volumetric soil water content (Davis and Chudobiak, 1975; Cole, 1976; Topp et al, 1980) then researchers were intrigued by the possibility of monitoring rock mass deformation by embedding coaxial cable in caving rock (Bartel et al, 1980; Wade and Conroy, 1980; Panek and Tesch, 1981; O’Connor and Dowding, 1984). The progression of TDR applications continued as it was compared with other types of instrumentation and users had a basis for comparison. It became apparent that there are applications in which TDR could provide information that is not available using other techniques. Also, there are applications in which it was very valuable when used in combination with other tools. Department of Civil and Environmental Engineering Infrastructure Technology Institute Northwestern University, Evanston, Illinois, 60208
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Acceptance in the soil science and agricultural community has been more advanced than the geotechnical applications, and there are a number of manufacturers producing a variety of commercially available systems specifically for soil moisture measurements. They are operating in the world marketplace with distributors in several countries. Professional organizations such as the Soil Science Society of America (Advances in the Measurement of Soil Physical Properties: Bringing Theory into Practice, G.C. Topp et al, editors, SSSA Sp. Pub 30, 1992) and the American Geophysical Union (Water Resources Research) have provided venues for users to collaborate and have provided printed media to make technological advancements and case histories available to users worldwide. Acceptance in the geotechnical community has been enhanced by the worldwide availability of coaxial cable for communications and commercially available TDR units. In 1982, there were on the order of two dozen installations in the United States for monitoring rock deformation. Today there have been several hundred installations around the world to monitor deformation in both rock and soil. The off-the-shelf do-it-yourself nature of this technology (CommScope, 1994, Cablewave Systems, 1985) has motivated users to try it. Development by Tektronix (1989) and Campbell Scientific (1991, 2001) of quality hardware and software which is compatible with geotechnical applications. In particular, this has increased the availability and reliability of TDR technology for automated remote monitoring using robust systems. A symbiotic gathering of more that 100 researchers and practitioners was convened at Northwestern University in 1994 for the First International Workshop and Symposium on TDR in Environmental, Infrastructure, and Mining Applications. The TDR List Server and web site (http://www.iti.northwestern.edu/tdr/) maintained by the Infrastructure Technology Institute were both spawned by this meeting and continue to serve the TDR community as well as the excellent SOWACS web site (http://www.sowacs.com/sensors/tdr.html) that is dedicated to measurement of soil water content. The second international conference TDR2001: Innovative Applications of TDR (http://www.iti.northwestern.edu/tdr/tdr2001/index.html) was convened at Northwestern University in 2001. Time Domain Reflectometry Applied to Measurement of Rock Mass Displacements Rock mass displacements within the strata overlying several mines have been monitored using TDR technology. Basically, a solid-aluminum coaxial cable is crimped at equally-spaced locations then grouted into a borehole drilled from the surface (Figure 1). A TDR cable tester is connected to the cable and sends voltage pulses down the cable. At every location where the cable is crimped or rock movement has caused cable deformation, a reflection is sent back to the tester which displays and records a TDR waveform. Initially, the monitoring was limited to quantitative analysis of the location of cable breaks. For example, reflections from the crimps provide distance reference markers in a TDR waveform so that deformation due to rock movement can be located accurately. A major development began with quantitative analysis of TDR records indicating that the magnitude of movement could be monitored (Dowding et al, 1988; O’Connor and Dowding, 1999) The shape and magnitude of a reflection at each location where deformation is occurring correlates with the type and magnitude of cable damage at that location. Based on laboratory calibrations, it is possible to distinguish shear deformation from tensile deformation and to quantify shear displacement (Dowding et al, 1988).
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Figure 1. Schematic of coaxial cable installation for TDR monitoring of rock and soil deformation. Becoming a Commodity Since 1996, GeoTDR Inc and Kane Geotech Inc have been the principal geotechnical consulting firms engaged in TDR applications while there is also a large community of users in State DoT’s, USGS, mining companies, and universities. It is emerging as a mainstream technology utilized by several consulting firms. The progress is incremental, and the possibilities increase as the community of users grows larger. As TDR becomes a mainstream technology used on high profile projects, it attracts more users and will progress to the status of a commodity. As with all research grants and consulting jobs, it’s not what you know, it’s who you know (owners, contractors, agencies, researchers, etc) in terms of creating a customer base. Marketing and networking are the life blood of any business. It is a continual process of making people aware that the technology exists and how it can be utilized to as a cost-effective solution for a variety of problems. Credibility does matter- not only for the technology but also the installation of cable, acquisition of data, and presentation of data. One other thought - education and marketing are essential but you will never get business with anyone to whom you provide free information which is indicative of a transition from status as a research tool to status as a tool in the practice of geotechnical and geoenvironmental engineering. The following case histories are presented to illustrate how the application of TDR to mine subsidence monitoring and risk assessment has evolved, how it is causing a paradigm shift, and how there will be technical, liability, and cost ramifications. FIRST EXAMPLE Effects of Undermining I-70, Washington County, PA November 1999 TO October 2000 The first example project was located on I-70 east of Washington, PA as shown on Figure 2 and Figure 3. The Eighty-Four Mining Company extracted coal from two panels at a depth of 559 to 651 ft (170 to 198 m) beneath I-70 using the longwall mining technique. This high extraction technique involved removal of two large blocks of coal approximately 1000 ft wide, 6000 to 8700 ft long, and 6 ft thick. Figure 4 shows a plan view of the highway and limits of underground coal mine panels. The northern panel (3 South LW Panel) was mined between November 22, 1999 and March 2, 2000 and averaged 70 ft (18 m) of advance per day (Appendix A). 4 South LW Panel was mined between March 9, 2000 and October 16, 2000 and averaged 40 ft/day (12 m/day).
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Figure 2. View looking east along I-70 east of Washington, Pennsylvania; overpass in Figure 3 is located near the overhead sign.
Figure 3. Overpass along I-70; the concrete blocks and wood cribbing were placed for temporary support of the bridge deck. TDR7 is located at the cover to the right of the stop sign. Project Background In the longwall method of high extraction coal mining, a shearer moves across the full width of a panel making a cut about 3 ft (1 m) deep and loads the coal onto a conveyor that transports it to another loading point (Figure 4). Hydraulic roof supports are advanced behind the shearer so the mine roof and overlying rock fractures and collapses into the void behind the supports. Caving and fracturing propagates up through the overlying rock mass as shown in Figure 7. With this loss of support, subsidence of the overlying rock mass is a certainty and the ground surface ultimately deforms into a trough with maximum subsidence of 3 to 5 ft (1.0 to 1.5 m) as shown by the transverse subsidence profile in Figure 7.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Figure 4. Plan view of mine panels, I-70, and instrument locations. PennDoT Engineering District 12 assumed responsibility for precautionary measures as I-70 was undermined and repairs after mining was completed. To ensure the safety of the traveling public, PennDoT took several precautions including temporary support of an overpass (Figure 3), reduction of speed limits, provision for lane closures and detours, visual monitoring patrols, and real time monitoring of ground movement with a call back alarm capability. Innovative monitoring of ground deformation was accomplished with TDR to interrogate coaxial cables installed in seven deep holes, and an array of thirty-two tiltmeters installed along the highway shoulder. Surface monitoring was also conducted with Global Positioning System measurements at more than five hundred locations. The tiltmeters were connected to a remote data acquisition system that automatically recorded and stored measurements. When specified tilt values were exceeded, the system initiated a phone call to key PennDOT personnel who then monitored tiltmeter measurements in real time via a phone line connection. Based on this information they could alert other agencies if necessary, and intensify visual reconnaissance to determine if lane closures were necessary. Measured Subsurface Movement The angle of draw shown in Figure 5 represents a mathematical model of the limits of movement within the overburden and on the ground surface. Conventional experience is based on vertical movement at the surface, so the angle of draw concept is consistent with intuition. However, there are other modes of deformation, especially subsurface movement within the overburden, that are precursors of surface subsidence and not consistent with intuition.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Figure 5. Conventional subsidence engineering terminology. Based on TDR measurements that have been made over several longwall mines, it is known that, in fact, vertical movement is only one component of the actual behavior. TDR measurements have made it possible monitor shear deformation within the overburden. This is represented in Figure 6 by the horizontal double-headed arrows intersecting vertical cables within the overburden. Based on TDR measurements it is known that this deformation not only occurs as a precursor in advance of mining but also occurs well beyond the limits of surface subsidence and the angle of draw.
Figure 6. Shear within the overburden ahead of the active mine face. When the overlying rock collapses into the mined-out void, a large amount of energy is transmitted throughout the overburden. It travels as compressive stresses and shear stresses that propagate away from the collapsing rock in all
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
directions. When the shear stress at any location is greater than the shear strength of the rock mass at that location, fracture and slip occur. Typically, the weakest component of shear strength is the resistance to slip along discontinuities such as joints and bedding planes. This type of localized shearing was detected and measured using TDR.
Figure 7. Subsidence profile over the two longwall panels.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Figures 8a and 8b show TDR waveforms acquired at location TDR4. The horizontal discontinuities defined by the histogram on the left side of the figure are discussed below. The waveforms were recorded as the longwall face approached and advanced past the borehole. The regularly spaced spikes identified with asterisks are associated with crimps made in the cable prior to placement in the borehole and are used as distance reference markers. The reflection spikes which increased in magnitude indicate where cable deformation occurred due to rock movement.
Figure 8. Waveform Acquired at location TDR4. A) 12/99 through 6/3/00
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Figure 8 (continued). B) 6/05/00 through 6/19/00 Confirmation from Experience at Other Sites The overburden response was consistent with behavior observed at other sites where TDR has been used to monitor subsurface deformation over longwall coal mines (O'Connor et at, 1995). Deformation was concentrated at "significant horizontal discontinuities" which are bedding planes between strata with a large difference in strata stiffness. This is
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
represented graphically in Figure 3 by the stiffness histogram and was evident in all the TDR cables. Deformations were consistently concentrated at specific discontinuities in specific strata. We have used TDR to monitor rock mass behavior at active coal mines (O’Connor et al, 1995), potash mines (O’Connor and Zimmerly, 1991), and Frasch sulphur mines (O’Connor and Norland, 1995). In all cases, shear displacement along horizontal discontinuities was a predominant component of rock deformation. Dowding and Huang (1994) found that by cumulatively summing all the subsurface shear displacements measured with TDR, the total was close to the horizontal displacement measured by surveying on the surface above a longwall panel. Recent experience even in monitoring slope movement in both stiff soils and softer soils has demonstrated that displacements tend to occur along contacts between strata which have a high contrast in stiffness. Whether the variation is a consequence of fracture density or intact shear strength, these contact planes are weak links as the redistribution of shear stress occurs. Ramifications The plot in Figures 8a and 8b illustrates that deformation was occurring within the overburden over 1000 ft in front of the mine face. Furthermore, the waveform for May 1, 2000 in Figure 8a shows that movement had occurred at location TDR4 during mining of Panel 3 South that was more than 443 ft north of the cable location. This indicates the lateral extent of the influence of mining. A summary of depth locations (a) in front of the active mine face and (b) laterally outside the mine panel where shearing was detected by TDR for all cables is presented in Figure 9. Note subsurface shearing occurred along significant discontinuities (a) ahead of the mine face, and (b) outside the edges of the panel being mined. It is important to emphasize that the precursor movement is shear deformation along discontinuities within the rock mass. This behavior is not taken into consideration with angle of draw concept shown in Figure 5 which only accounts for the lateral extent of vertical subsidence on the ground surface. For purposes of comparison, a line representing an angle of draw of 30 degrees is superimposed in Figure 9 to illustrate that subsurface movement occurs well beyond the limits of mining (and beyond the limits of vertical surface subsidence)..
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Figure 9. Locations of shearing within the overburden (a) ahead of the active mine face and (b) outside the mine panel. SECOND EXAMPLE. Subsidence Risk Assessment for Dorris Elementary School, Collinsville, IL April 1995 to May 1998 In situations where important structures are located over subsurface cavities, decision makers need a quantitative measure of the likelihood of subsidence occurring. In one case, automated TDR monitoring was an integral part of the site selection process for a new school building (O’Connor and Murphy, 1997). Coal mining was active within Collinsville, Illinois from 1870 to 1964 and the area is underlain by a network of mine openings. Support for the overlying rock is provided by pillars and blocks of coal that have begun to fail or possibly punch
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
into the underlying claystone floor. As a result of deep failures within the abandoned mines, there have been occurrences of localized surface subsidence throughout the city. Movement of the overlying rock, and ultimately the surface, have subjected structures, streets, and utilities to strains and stresses that have caused damage. Project Background The Dorris Elementary School had been undergoing damage due to subsidence since 1989 (Gibson and Schottel, 1990) and the School Board needed information to determine a location for a new building. One of the sites under consideration was the athletic field north of the existing building (Figure 10) which would make it possible to keep the school in its current neighborhood. A second site was an athletic field adjacent to another school several blocks away. Subsurface investigations, which included the installation of TDR monitoring cables, were done at both sites.
Figure 10. Drilling at the Dorris School site. Note the dip in the school building roof line where maximum subsidence has occurred. The mine is approximately 200 ft below the surface, overlain by 100 ft of glacial material and 100 ft of Pennsylvanian Age rock (Figures 11 and 12). Cables were installed at three locations (TDR1, TDR2 and TDR3) to maximize the value of data received and to maximize the likelihood of detecting subsurface movements before subsidence occurred at the surface. The locations were selected based on the following criteria: 1) historical subsidence data (O’Connor et al, 1996), 2) proposed location for a new school, 3) subsidence currently developing northeast of Dorris School, and 4) mine geometry. It was originally planned that all holes would penetrate mine entries in the area of high extraction ratio (TDR3) and in the area proposed for a new building (TDR1 and TDR2), but this was accomplished only at TDR2.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Figure 11. Plan view of school building, abandoned underground mine, and instrument locations. At the surface, lead wires were connected to each of the three coaxial cables and brought to a central location where a utility pole was installed. A TDR data acquisition system (Campbell Scientific, 1991) was installed on this pole within enclosures. The lead wires were connected to a multiplexer which was connected to a Tektronix 1502B TDR cable tester which in turn was connected to a storage module and modem. The datalogger was programmed to turn on the cable tester, interrogate each cable, store data in the storage module, and then turn off the cable tester. Data is downloaded from the storage module via a phone line.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Figure 12. Cross section along TDR hole locations. Measured Subsurface Displacements and Surface Subsidence Representative data for TDR3 are shown in Figure 13. The histogram of stiffness versus depth on the left side of the figure was produced using the technique developed by O’Connor et al (1995) to estimate the stiffness of each rock stratum based on its rock mass classification. The significant features are the two stiffer limestone strata. The upper one is about 4 ft (1.2 m) thick, and the lower one is 20 to 23 ft (6 to 7 m) thick. This lower unit forms the immediate mine roof. No fractures were observed in cores obtained in these strata.
Figure 13. Waveforms acquired at location TDR3. The reflections at a depth of 138 ft correspond with shear and tensile movement at a "clay seam." The change in TDR waveforms at a depth of 138 ft (42 m) is characteristic of cable shear and tensile deformation (Figure 13). An increase in magnitude is associated with shear deformation and a decrease is associated with tensile deformation (Dowding et al, 1988). The periods of deformation can be conveniently identified by the rate of change in reflection magnitude as summarized in Figure 14b. During the period from April 1995 to August 1996 when the cables experienced
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
tension, probably due to strata separation, surface subsidence of 7 to 20 mm occurred as shown in Figure 14a. When the mode of cable deformation is predominately shear, it is possible to convert TDR reflection magnitude to shear displacement (Dowding and Huang, 1994; O’Connor and Zimmerly, 1991). In this case, the predominant mode has been tensile deformation so a correlation with the magnitude of surface subsidence is not possible although a correlation with rate of movement will be possible as more survey data becomes available. WITHIN TROUGH
SUBSIDENCE (mm)
0
-5
-10
-15
-20 Oct-89
Nov-90
Dec-91
Jan-93
BM15
Mar-94
BM16
Apr-95
May-96
Jun-97
BM P-9
Figure 14. Measured deformation. A) surface subsidence. TDR3 RATE OF CHANGE
CHANGE IN MAGNITUDE (mrho/month)
10
5
0
-5
-10 Apr-95
Oct-95
May-96
Dec-96
Jun-97
Jan-98
Figure 14 (continued). B) rate of change in reflection magnitude at a depth of 138 ft at location TDR3. Confirmation from Experience at Other Sites TDR has been used to monitor stability of crown pillars over abandoned underground mines in Nova Scotia and Ontario (Charette, 1993; Aston et al, 1994) as well as strata movements over abandoned underground mines in Illinois, Ohio and Virginia (Dowding and O’Connor, 2000; O’Connor et al, 2002). At many sites data were acquired locally by personnel using a TDR unit and laptop computer but the more effective applications involved automated remote monitoring such as at the Dorris School site. Not only has been data been acquired and downloaded over land line phone connections but it has been possible to implement real time monitoring and comparison with baseline readings. This has made it possible to implement call back alarms when movement exceeds preset levels. Ramifications TDR technology makes it possible to do automated, remote monitoring of changes in subsurface conditions over abandoned mines. At the Dorris School site, subsurface strata separation and shearing occurred intermittently during the period from April 1995 to June 1996 along horizontal discontinuities represented by clay seams and a fractured shale stratum which are less stiff than adjacent strata. Surface subsidence, which is a consequence of these subsurface movements, was verified by
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
periodic survey measurements. Based partly on the results of this monitoring program, it was apparent that this site was experiencing subsidence that would be a chronic problem. The School Board was able to justify to faculty, students and parents that this site was not a viable location for the new school building and the alternative site was selected. At other sites TDR monitoring has been used as a real time monitoring and alarm system (O’Connor et al, 2002). This has been particularly cost-effective when the cable was installed in a trench. Individual cables extended 1000 ft from a remote data acquisition system and arrays were combined to cover an area 2000 feet long beneath an Interstate highway. In this application the objective was to automatically detect movement that severed the cable then download data for assessment of the changes in the TDR waveforms since noise levels could trigger a false alarm. CONCLUSION AND SUMMARY - THE PARADIGM SHIFT Acceptance versus Resistance It is apparent that TDR technology can be used to reliably monitor subsurface effects of high extraction mining. Measurements have demonstrated that movement occurs extend well beyond the limits that would be expected on the basis of the state-of-the-practice for subsidence engineering. As this paradigm shift occurs within the realm of environmental regulation and legal disputes, there will be both enthusiasm and reluctance to embrace TDR technology. Liability Ramifications The TDR measurements make it possible to document that high extraction mining is affecting the subsurface and is affecting wells beyond limits of angle of draw. This type of information may be embraced by regulators, but there will be resistance by mining companies. This is understandable since, on one hand, it is bad for mining companies and, on the other hand, it substantiates observations that have been made by regulators and landowners. These measurements have not yet been used in court. However, subsurface movement measured with TDR was used in Collinsville as basis of School Board to justify their decision for constructing a new school at a site outside the neighborhood to minimize liability due to continuing movement. Economic Ramifications From a negative perspective, mining companies will be held responsible for damage to wells outside the limits of the angle of draw. From a positive perspective, these measurements can be used as a component of subsidence risk assessment. They will allow action plans to be formulated that will allow development of undermined areas. Typically, developers don’t want to disclose such information, but their attitude may shift if technology allows them to develop land if monitoring is accepted by insurance companies and regulators as being a prudent and responsible approach. REFERENCES 1.
Aston, T.R.C., M.C. Betournay, J.D. Hill, and F. Charette Application of TDR for Monitoring the Long term behavior of Canadian Abandoned Metal Mines. Proceedings of the Symposium on Time Domain Reflectometry in Environmental, Infrastructure, and Mining Applications, Evanston, Illinois, Sept 7-9, U.S. Bureau of Mines, Special Publication SP 19-94, 1994, NTIS PB95-105789, pp. 518-528.
2.
Bartel, E. W., Fox, M., Borgoyn, E., and Wingfield, P., TDRM Testing (contract J0377021), BuMines OFR 60-82, 1980, 135 pp.
3.
Campbell Scientific, Inc. (Logan, UT), Time Domain Reflectometry for Measurement of Rock Mass Deformation. Product brochure, July, 1991, 2pp.
4.
Campbell Scientific, Inc. (Logan, UT), TDR100 Time Domain Reflectometer. http://www.campbellsci.com/tdr.html#Tdr100.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
5.
Charette, F., Results of the Monitoring of Three Crown Pillar Sites in Cobalt, Ontario, CANMET Div. Rep. 92-101 (CL), Min. Res. Lab., Ottawa, Mar., 1993, 31 pp.
6.
Cole, R.H., Time Domain Spectroscopy of Dielectric Materials, IEEE Transactions I&M, Vol. IM25, No. 4, Dec., 1976, pp. 371-375.
7.
Cablewave Systems (North Haven, CT), Antenna and Transmission Line Systems, Catalog 600, 1985, http://www.rfsworld.com/RFSGlobal.
8.
CommScope, Inc. of North Carolina (Claremont, NC), Cable Products Catalog, 1994. http://www.commscope.com
9.
Dalton, F.N., Herkelrath, W.N., Rawlins, D.S., and Rhoades, J.D., Time Domain Reflectometry: Simultaneous Measurement of Soil Water Content and Electrical Conductivity with a Single Probe, Science, Vol. 224, June,1984, pp. 989-990.
10.
Davis, J.L. and Chudobiak, W.J., In-situ Meter for Measuring Relative Permittivity of Soils, Pap 75-1A, Geol. Surv. Can., Ottawa, 1975, pp. 75-79.
11.
De Loor, G.P., Dielectric Properties of Heterogeneous Mixtures Containing Water, J. Microwave Power, Vol. 3-2, 1968, pp. 67-73.
12.
Dowding, C. H. and F.-C. Huang. Telemetric Monitoring for Early Detection of Rock Movement With Time Domain Reflectometry. J. Geot. Eng., Am. Soc. Civ. Eng., v. 120, No. 8, 1994, pp. 1413-1427.
13.
Dowding, C.H. and K.M. O’Connor. Real Time Monitoring of Transportation Infrastructure with TDR Technology. Proc, Structural Materials Technology IV, Atlantic City, New Jersey, March, 2000.
14.
Dowding, C.H., M.B. Su, and K.M. O'Connor. Principle of time domain reflectometry applied to measurement of rock mass deformation. Int. Journal of Rock Mech., Mining Sci, and Geomechanical Abst., Vol. 25, No. 5, 1988, pp 287-297.
15.
Fellner-Feldegg, J., The Measurement of Dielectrics in the Time Domain, J. Phys. Chem., Vol. 73, 1969, pp. 616623.
16.
Gibson, R. D., and B. C. Schottel. A Case History Illustrating the Application of Computerized Modeling of Coal Mine Subsidence Profiles and the Development of a Settlement Prediction Technique. Proceedings, Third Conference on Ground Control Problems in the Illinois Coal Basin, Mt. Vernon, IL, August, 1990, pp. 369-381.
17.
O'Connor, K.M. and C.H. Dowding. Application of Time Domain Reflectometry to Mining. Proceedings, 25th Symposium on Rock Mechanics, Northwestern University, Evanston, Illinois, June, 1984, pp 14-17.
18.
O'Connor, K.M. and C.H. Dowding. GeoMeasurements using TDR Cables and Probes. CRC Press LLC, Boca Raton, 1999.
19.
O'Connor, K.M. and E.W. Murphy. TDR Monitoring as a Component of Subsidence Risk Assessment Over Abandoned Mines. Int. Journal of Rock Mechanics & Mining Science. Vol. 34, Nos. 3-4, Paper 230. 1997.
20.
O’Connor, K.M. and M.R. Norland. Monitoring Subsidence Mechanisms Using Time Domain Reflectometry. Proceedings, Joseph F. Poland Symposium on Land Subsidence, Sacramento, October, 1995, pp. 427-434.
21.
O’Connor, K.M., J.A. Siekmeier, and L.R. Powell. Using a Computer Spreadsheet to Characterize Rock Masses Prior to Subsidence Prediction and Numerical Analysis. U. S. Bureau of Mines RI 9581, 1995, 69pp.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
22.
O’Connor, K.M., J.A. Siekmeier, and J. Stache. Using GIS and Numerical Modeling to Assess Subsidence Over Abandoned Mines. Proceedings, 13th Annual National Meeting of the American Society for Surface Mining and Reclamation, Knoxville, May, 1996.
23.
O’Connor, K.M., L.R. Ruegsegger, and K. Beach. Real Time Monitoring of Subsidence, I-77 Summit County, Ohio. Proceedings, 10th Great Lakes Conference on Geotechnical and Geoenvironmental Engineering, Univ of Toledo, Toledo, Ohio, May, 2002, pp. 20-30.
24.
O'Connor, K. M. and T. Zimmerly. Application of Time Domain Reflectometry to Ground Control. Paper in Proceedings of the 10th International Conference on Ground Control in Mining (Morgantown, June, 1991). WV Univ., 1991, pp. 115-121.
25.
Panek, L.A. and Tesch, W.J., Monitoring Ground Movements Near Caving Stopes-Methods and Measurements, BuMines RI 8585,1981, 109 pp.
26.
Tektronix, Inc. (Redmond, OR), TDR products. Http://www.tempo.textron.com/products
27.
Topp, G.C., Davis, J.L., and Annan, A.P., Electromagnetic Determination of Soil Water Content: Measurement in Coaxial Transmission Lines, Water Resources Research, Vol. 16, No. 3, June, 1980, pp. 574-582.
28.
Wade, L.V. and Conroy, P.J., Rock Mechanics Study of a Longwall Panel. Mining Engineering, Dec., 1980, pp. 1728-1734.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
FIELD AND LABORATORY TECHNIQUES FOR MEASUREMENT OF DYNAMIC BEHAVIOR OF SOILS Vicente Cuéllar, Antonio Santos and José L. Monte Laboratorio de Geotecnia (CEDEX), Madrid, Spain
[email protected]
ABSTRACT This paper describes the developments achieved in the field of soil dynamics at the Laboratorio de Geotecnia of CEDEX in the last thirty years. Due to the deep involvement of the Laboratorio in special projects of civil engineering and building construction or restoration, those techniques have been applied to the response analysis of real systems to vibrations or seismic actions. A synthesis is given for the main equipments and techniques available at the Laboratorio de Geotcnia, in dynamic studies, following the chronology of their adoption and development at the institution. The main periods considered have been, the seventies, eighties, nineties and in present time. INTRODUCTION In Spain, geotechnical studies for dynamic behavior of soils have been related mainly to design and construction of singular structures in which earthquake risk had to be considered. As it is shown in Fig. 1, Eurasian and African tectonic Plates have a "shear contact" along Gibraltar Straits; associated to this main tectonic feature, historical destructive earthquakes have shaked the South of the Iberian Peninsula and the North of Africa.
losses due to damaged houses, public buildings and civil structures, take place. In Spain, the last large destructive quake took place on december 25, 1884, while more recently, destructive earthquakes have been suffered in Algeria (Orleansville, september 9, 1954, El-Asnam, october 10, 1980, and Constantine, october 27, 1986) and Morocco (Agadir, february 29, 1960).
From a recurrent destructive aspect, both areas can be considered as moderately seismic. This historical fact could be motivated because the energy accumulation, associated with a "shear contact" zone, up to the required level to originate a fracture (geological fault) in the earth crust, and the corresponding seismic wave generation process, require a time period larger than in zones where Plates shift produces a more direct "collision" between them. Historical records of the Iberian Peninsula since the XIV century establish a "recurrent" sequence of seismic activity together with long "quiet" periods; a "similar schedule" could be accepted for the Maghreb countries. During each active seismic period, destructive earthquakes with significant losses in human lives and important economic
Figure 1: Tectonic plate distribution and contact types between plates From data obtained through the Portuguese-Spanish Seismic Network, two earthquakes of large magnitude have been recorded; on March 29, 1954 (magnitude = 7) and on february 28, 1969 (magnitude = 7.3). Both of these,
Department of Civil and Environmental Engineering Infrastructure Technology Institute Northwestern University, Evanston, Illinois, 60208
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
originated in the Atlantic at the shear contact between Plates, but were registered, at the Iberian Peninsula, as moderate intensity quakes, due to the several hundred kilometers depth the focus. Historical records point out that destructive earthquakes take place in two "land strips" north and south of Gibraltar Straits, while at the Southern Corner of the Iberian Peninsula and at the Northern Corner of Africa no destructive earthquakes take place. In order to investigate this favorable engineering behavior, laboratory tests on soil samples from Algeciras (Cadiz) have been run.
introduced into Geotechnical Engineering practice, mainly through research developed at Texas University, Austin, by Stokoe, Roesset and their co-workers, on the theoretical basis previously established by Kausel. (see Kausel, 1981, Kausel-Roesset, 1982, Roesset et al. 1991, Stokoe et al. 1994). The established procedure implies a wave source directly applied to the ground surface, and two receivers placed at preselected locations on this surface, as shown in Fig. 6; the method takes advantage of the "dispersive" nature of surface waves as they travel through layered soils, which implies that the propagation velocity (phase velocity) depends on the frequency of the wave.
This paper summarizes our actual soil dynamics "state of the art" and dynamic data obtained in Spanish soils. LABORATORY AND FIELD EQUIPMENTS It should be noted that acquisition and implementation of laboratory and field equipment for soil dynamic test has followed a continuous but irregular development during the last 30 years. Resonant Column and Crosshole Tests Like in many other countries, resonant column and crosshole tests were the first methods used in the 70's to obtain dynamic soil parameters in geotechnical practice. In the low dynamic strain range they are kept on use, but since both tests are widely applied over the world no further comments will be made. (ASTM, D-4015, Standard Test Method for Modulus and Damping of Soil by the Resonant Column Method, and D-4428/4428M, Standard Test Method for Crosshole Seismic Testing) Cyclic Triaxial Test and Spectral Analysis of Surface Waves In the early 80's, equipments for dynamic triaxial tests and Spectral Analysis of Surface Waves (SASW) were available and progressively implemented at the Laboratorio de Geotecnia (CEDEX). The dynamic triaxial equipment has three hydraulic servocontroller presses with capacities up to 16 kN, 100 kN and 2500 kN, depending upon the sample strength requirement (Cuéllar et al. 1993); lower press capacity with greater equipment sensibility allows dynamic tests on soils of soft to medium stiffness; the largest press capacity is used for rock samples. Typical results of tests on cohesive and cohesionless soils under cyclic load are shown in Fig. 2 and 3. Irregular (earthquake) loading is shown in Fig. 4, and the stress-strain behavior of a cement-bentonite grout is depicted in Fig. 5. At that time, Spectral Analysis of Surface Waves was
Figure 2: Cohesive material. Load control
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Figure 3: Granular material. Load control
In practice, the generated transient signal F(t) is captured in the time domain by the receivers, as X1(t) and X2(t). For each frequency, the coherence between both signals can be obtained through its spectral analysis, and from their phase difference, the time lag; knowing the distance between receivers, the celerity of the Rayleigh wave can be determined. A "GR-810/40 Surface Wave System" was acquired for SASW field test; its main components are vibration control and processing unit, power amplifier, generator, vibration exciter, calculator, spectral analyzer, X-Y plotter and vibration detectors. As it is shown in Fig. 7, SASW method was combined with other "in situ" techniques, such as static plate loading tests, to define the geomechanical behavior of a superficial hard soil on which the Alhambra Palace of Granada is founded. Fig. 8 recopilates the characteristic curves of shear modulus versus shear strain, obtained from the "in situ" tests. The results obtained with the SASW technique at the location of the Comares tower were extrapolated to deeper ground by means of cross-hole testing (see Fig. 9) (Cuéllar, V. 1998).
Figure 4: Design earthquake
Figure 5: Unconfined compression. Loading-unloading cycles
Figure 6: Input and receivers signals in SASW tests
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Figure 7: Synthesis of the “in situ” static and dynamic testing along with the selected hyperbolic model for simulating the dynamic behavior of the “Alhambra formation”.
Cyclic simple shear equipment allows to apply to the soil sample a vertical and shear stress path similar to the more significant stress path which earthquake induces in the soil. The test can be programmed for a cyclic shear stress or strain history; since, a larger range cyclic strain than in the triaxial equipment can be applied to the soil sample, a high non-linear soil behavior can be studied. Fig. 10. shows the dynamic response of Algeciras clay under strain control for a series of 100 cycles, 1 Hz. frequency, and increasing deformation. Equipment allows to combine a quasistatic (two days) loop with cyclic (1 Hz.) stresses departing from the residual stress state induced by the static loop, as indicated in Fig. 11.
Figure 8: Characteristic curves of shear modulus vs. shear strain.
Figure 10: Site: Algeciras Stress-Strain cycles
Figure 9: Zonation of the structure and foundation of the Comares Tower. Cyclic Simple Shear and P-S Logging In the 90's, two devices for laboratory and field testing were implemented.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Figure 11: Monotonic simple shear test Although apparently Oyo P-S logging seems to be a test similar to sonic log, its practical performance requires the generation of body waves (P and S waves) within a fluid, which must be transmitted through the soil and recorded at the two receivers included in the suspended device surrounded by the fluid. The basic difference remains in the wave length generated by the two devices; sonic log generates a wave length considerably smaller than the borehole diameter, while P-S logging generates a wave length considerably larger than the borehole diameter. Besides, in the sonic test, dilatational fluid wave celerity must be smaller than ground shear wave celerity, while in the P-S test the determination of both, P and S wave velocities in the ground is independent of the dilatational fluid wave celerity. In practice, sonic log is usually applied only to rock formations while P-S logging can be used either in rock or soil tests. Fig. 12 (a & b) shows schematically the physical device performance.
Figure 12a: Suspended Oyo probe for "in-hole" testing
Figure 12b: P and S waves generated by the suspended Oyo probe The analysis of the recorded signal data, in both time and frequency domains is illustrated in Fig. 13 for uncased borehole at Barajas, (Madrid airport). The right hand side shows the profile for the particle horizontal peak velocity, while the Vs profile is given on the left hand side; both profiles show that two points in the ground can have similar values of Vs, (670 m/s at depths of 12 and 25 m. in
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
the left hand figure), but the values of the particle peak velocity could be very different (4 µm/s and 43 µm/s in the right hand figure), indicating totally different elastic response capacities (Valerio-Cuellar, 2001).
Figure 13: Vs profile (left) and particle peak velocity profile (right) obtained in uncased borehole B4 at Barajas site. Torsional Test and Impact Equipment In the present decade two new devices have been developed at the Laboratorio de Geotecnia, (CEDEX): a new torsional simple shear apparatus, able to perform both monotonic and cyclic loading on solid cylindrical soil specimens, and a new impact equipment capable to generate the range of low frequencies required in the field to investigate with sufficient energy the behavior of high rock-fill embankments. With the new torsional simple shear equipment stress-strain characteristics of soils can be accurately and reliably obtained for a wide range of shear strains. The device is able to perform tests on solid cylindrical specimens of standard dimensions 38 mm diameter and 76 mm high. The shear strain measurement system, consisting of two pairs of LVDT's, allows to obtain a continuous record of shear strains from 10-4 % to 3 %.Torque is applied by an electromagnetic loading system which can give up to 1.2 Nm and is measured and registered by means of a torsional load cell with resolution of 5*10-5 Nm. Fig. 14 (a & b) shows a schematic section and a view of the torsional device (Manzanas-Cuellar, 2001).
Figure 14 a: Schematic section of the torsional device.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Figure 14 b: View of the torsional device. Fig. 15 shows the type of impulse force generated with the new SASW device. Using that impulse source, wavelengths up to 40 m.have been generated in rock.fill embankments with particles up to 1 m. in size. To gain mobility, the equipment was mounted on a special carriage as can be seen in Fig. 16 (a & b).
Figure 16 a & b: Equipment for dynamic impact generation CONCLUSIONS In this paper a summary has been made of the laboratory and in situ test equipments developed at CEDEX's Geotechnical Laboratory to analyse the dynamic behavior of soils in the last 30 years. Special attention has been devoted to the new divices available at the present time. Among their characteristics and applications the three following main features are worth to outline:
Figure 15: Impulse force generated with the new SASW device
- The possibility to analyse with only one sample the whole non-linear range of behavior of a soil type using the quasistatic new simple shear apparatus. - The capacity to characterise the mechanical behavior of high rock-fill embankments generating large surface wavelengths with new impulse source equipment.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
- The possibility to explain the high elastic capacity of very plastic materials (e. g. Mexico clay) through the P-S logging peak particle velocity records. REFERENCES 1.
Cuellar, V., Martin, M.E. and Olalla, C. 1993, Comportamiento dinámico de muestras remoldeadas y talladas de arcilla frente a solicitaciones sinusoidales e irregulares. Simposio sobre Geotecnica de Presas de Materiales Sueltos, pp. 69-75, 20-22 octubre, Zaragoza.
2.
Cuellar, V., 1998, Trabajos realizados y coordinados por el Centro de Estudios y Experimentación de Obras Públicas. Cuadernos de La Alhambra. Vol. 33-34, pp. 9-16, Granada 1997-1998.
3.
Kausel, E., 1981, An explicit solution for the Green functions for dynamics loads in layered media. Re-search Report R81-13. Departement of Civil Engineering, Massachusetts Insitute of Techology, Cambridge, MA.
4.
Kausel, E. and Roesset, J.M., 1981, Stiffness matrices for layered soils. Bulletin of the Seismological Society of America. Vol. 71, Nº 6, pp. 1743-1761.
5.
Manzanas, J. and Cuellar, V., 2001, A solid cylinder torsional shear device. Advanced Laboratory Stress-Strain Testing of Geomaterials, pp. 275-278, Balkema, Rotterdam.
6.
Roesset, J.M.; Chang, D.W. and Stokoe II, K.H., 1991,Comparison of 2-D and 3-D models for analisis of surface wave tests. 5th International Conference on Soil Dynamic and Earthquake Engineering Karlsruhe: 111-126.
7.
Stokoe, K.H.II; Wright, J.A.; Bay, J.A. and Roesset, J.M., 1994, Characterization of geotechnical sites by SASW method. Geophysical characterization of sites R.D. Woods, ed, Oxford Press, New Delhi.
8.
Valerio, J, and Cuellar, V., 2001, Experiences gained in Spain using the suspension method to determine Vs profiles. Earthquake Geotechnical Engineering Satellite Conference, 23-25 August, Istambul, Turkey.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
PERFORMANCE OBSERVATIONS OF THE WORLD TRADE CENTER SLURRY WALL Joel Moskowitz Mueser Rutledge Consulting Engineers 225 West 34th Street - 14 Penn Plaza New York, NY 10122
[email protected]
INTRODUCTION Construction of the World Trade Center towers took nearly six years. Destruction occurred in a matter of hours on September 11, 2001. Installation of the original basement slurry wall, excavation of about 1,000,000 cubic yards of soil and the drilling of more than 1,400 high-strength steel strand tiebacks took about two years. Recovery of the World Trade Center (WTC) basement over a period of roughly nine months included the removal of an estimated 1.2 million tons of debris and the installation and testing of 980 new tieback anchors in rock. This paper summarizes some of the observations of slurry wall performance during the recovery effort. Comparisons are made to observations during the original construction. RESCUE AND RECOVERY Figure No. 1 shows the 16 acre WTC site. The complex was dominated by the twin 1350 ft tall towers, but included a hotel (WTC 3); three office buildings designated WTC 4, 5 & 6; and the adjacent 47-story WTC 7. The two towers were utterly destroyed, collapsing to a pile of debris about 10 stories high. Miraculously, the North Tower, No. 1, collapsed nearly vertically while the South Tower, No. 2, fell only modestly to the southeast. Facades from both towers peeled away as the floors fell. Damage caused by the collapses and flying debris from the two towers ultimately resulted in the need to demolish WTC 4, 5 and 6 and the hotel, and the resulting fire in WTC 7 caused its total collapse. A number of nearby buildings including the Bankers Trust Building, a Greek Orthodox Church, and the World Financial Center also suffered significant damage from the falling debris.
Department of Civil and Environmental Engineering Infrastructure Technology Institute Northwestern University, Evanston, Illinois, 60208
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
WTC 7 VESEY STREET
WTC 5
WTC 6
NYC Subways
WTC Slurry Wall
CHURCH STREET
World Financial Center
WEST STREET
WTC 1
GREENWICH STREET
N
WTC 3 PATH commuter Railroad from NJ
WTC 2
WTC 4
LIBERTY STREET
Figure 1 - General Site Plan In the immediate aftermath of the collapse, the New York City Department of Design & Construction (DDC) established a team of engineers and contractors to assist the New York Fire Department (FDNY) in the early rescue and recovery efforts. Those efforts dominated the early work at the site, and performance monitoring was limited to visual observations. Engineers, including the staff of Mueser Rutledge Consulting Engineers (MRCE), provided assistance to FDNY in the staging of the massive equipment which began arriving at the site within hours of the collapse. The six story deep basement was enclosed by a 36-inch thick reinforced concrete slurry wall socketed a minimum of two feet into bedrock at depths of 60 to 75 feet. Groundwater is about 5 ft below street grade on the west side of the site, some 170 ft from the Hudson River. MRCE compiled subsurface information for the area immediately around the slurry wall and began evaluation of its stability. There was an urgent need to prevent collapse of the slurry wall and inundation of the basement from the nearby Hudson River because of concerns that survivors might be in the basement space and because of the potential for flooding the commuter railroad and tunnels that entered the WTC basement. Drawings were compiled for FDNY showing the locations of the PATH commuter railroad, and 60-inch diameter intake pipes which supplied cooling water from the Hudson River to the WTC air conditioning system and numerous other subsurface utilities present in the streets between the WTC and the River and the NYC subway tunnel just east of the slurry wall. Several old bulkhead and platform structures near the riverfront were also of concern given the size of the cranes, massive grapplers and heavy trucks used in the debris removal shown in Figure No.2.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Figure 2 – Debris Removal SITE DESCRIPTION AND GEOLOGIC SETTING The WTC deep basement is bounded by Greenwich Street and West Street on the east and west and Liberty and Vesey Streets on the south and north sides respectively. Early surveys of Manhattan show that the pre-colonial shoreline corresponded approximately to the present alignment of Greenwich Street, and that the WTC site is all within filled land as shown in Figure No.3.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Pre-Colonial High Water Line Colonial High Water Line (1767)
Old Bulkhead Line
N
Battery Park City Bulkhead
World Trade Center West St.
World Financial Center PATH commuter railroad
Vesey St. Tijger burned and sank here in 1614 Broadway
WTC Disposal Area
Liberty St.
Battery Park City Landfill
Trinity Church Wall St.
Battery
HUDSON RIVER
Figure 3 – Historical Site Plan The western shoreline of Manhattan was advanced incrementally some 600 feet between the late 18th and early 20th centuries. The shoreline was extended an additional 700 feet west in the 1970s with the land reclamation project now known as Battery Park City. Excavation spoil from the WTC basement was used to form a portion of that site as seen in Figure No. 4.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
The Hudson River originates in the Adirondack Mountains and flows southward to the Atlantic Ocean becoming tidal near Troy, New York some 100 miles north of Manhattan. In general, the river separates the older, Precambrian rocks on the east, from the Triassic rocks on the west. On the east side, the Precambrian rocks are part of the New York City group of the Manhattan prong. The New York City group comprises a series of metamorphic rocks of which the mica schist found at the site is a member. This formation is known to contain igneous intrusions which erode or weather more rapidly than the surrounding rock, producing topographic valleys and troughs. The most significant influence on the Figure 4 – WTC Basement Excavation Disposal drainage patterns of the present Hudson River was the continental glaciation which gouged the river channel down to the unweathered rock. As the glaciers advanced and retreated, they left behind a sequence of dense glacial till often including boulders, and outwash silt and sands. Sea levels rose during the final glacial retreats and a brackish environment developed leaving behind a substantial layer of river bottom organic silt and peat. Manmade fills overlie the natural deposits over the entire WTC site. The typical preconstruction profile, on the east side of the WTC at Greenwich Street as shown in Figure No. 5 comprised 10 to 20 ft of fill over 40 ft of outwash sand, and up to about 10 ft of glacial till overlying rock. The profile at West Street comprised up to 30 ft of fill overlying 30 ft of organic material, 20 ft of outwash sand and a thin veneer of till overlying bedrock at depths of roughly 60 to 70 ft. Groundwater is about 5 ft below existing street grade at West Street.
Figure 5 – Preconstruction Profile WTC CONSTRUCTION AND SLURRY WALL The slurry wall method was selected for the WTC basement because of the need to excavate through heterogeneous fill which was known to contain obstructions from prior construction including bulkhead walls, old foundations and ship ballast, a high groundwater table, the need to provide cutoff from the top of rock and the necessary utility penetrations through the wall including the existing PATH commuter railroad. The nearly 3,000 ft long wall, 36 inches thick, comprised 152 panels approximately 22 ft in length. Reinforcing cages, as shown in Figure No. 6, weighing nearly 25 tons, were socketed to a
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
minimum of 2 ft into bedrock at depths of 60 to 70 ft. Both positive and negative bending reinforcing steel were provided. Joints between panels were half-round in plan, formed by the placement of pipes during primary panel installation. The end stop pipes were removed for construction of the intermediate panels. During construction, temporary support of the slurry walls was provided by 1430 tieback tendons angled at 45 degrees and socketed 30 to 35 ft into bedrock. Each tendon consisted of up to 24, 2 -inch diameter high tensile strength strands using seven 0.165-inch diameter wires. Tendons ranged in lengths from 40 to 115 ft and were anchored in 7-inch diameter holes. Tiebacks were placed in four to six tiers with two tiebacks per tier per panel. Design loads were between 400 and 600 kips. The tendons were tested up to 800 kips. Following placement of the basement floors, which provided permanent lateral support, tiebacks were detensioned, cut and sealed. The west side of the slurry wall was penetrated in two locations by the PATH commuter railroad which consisted of 17-foot diameter cast-iron tunnels constructed around 1900, that crossed beneath the Hudson River to New Jersey. Tunnel inverts are at the lowest basement levels. In addition, there were four locations where the slurry wall was penetrated by 6 ft diameter water lines and by automobile entry ramps to the below-grade parking areas, one of which was used as access for the 1993 terrorist bombing. On the east side of the slurry wall, a New York City subway tunnel paralleled the deep basement. PERFORMANCE OBSERVATIONS - ORIGINAL CONSTRUCTION Saxena (1974) reported on the results of an instrumentation program undertaken by the Port Authority of New York and New Jersey during the original basement excavation. The instrumentation program comprised electric resistivity load cells, slope inclinometers, electric strain gages attached to reinforcing bars, piezometers and wellpoints. Four of the panels had load cells on tiebacks and three of those were summarized by Saxena (1974). Only two of the panels are summarized herein. Panel W35, situated about 1/3 of the way south along West Street penetrated a subsurface profile comprising 15 ft of fill, 25 ft of organic silt, 8 ft of outwash sand and about 14 ft of hardpan overlying rock. Data summarized by Saxena (1974) are shown in Figure No.7. Six levels of tiebacks were installed and locked off at 100% of Figure 6 – Reinforcing Cage their design load except for the upper tier which was preloaded to 90%. In so doing, the ties had been tested at installation to about 80% of their ultimate strength in the steel and locked off at about 55%. During excavation, the wall moved toward the soil that is, away from the excavation as each additional level of anchors was installed. Load cells showed a decrease from lock-off load during excavation, as the next lower tie was installed. Maximum deflection at the top of the wall was about 2.5 inches. Total horizontal pressures were somewhat higher than at-rest plus water pressure and well below undrained passive pressure in the organic silt. Panel G21, approximately centered along the Greenwich Street wall alignment had a substantially different profile and performance. That panel was constructed at a location where the existing subway structure was present and a temporary brace against the slurry wall adjacent to the subway structure was used until the excavation reached the uppermost tie level. In order to avoid overstressing the subway wall, the upper tieback was locked off at only 40% of design load. Below the upper tie level, the remaining ties were installed at 100% of design load.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Figure 7 – Observations at Panel W35 The data recorded are summarized in Figure No. 8. At Panel G21, continuous deformation toward the excavation was observed. Anchor loads at the upper level increased slightly, probably due to the elastic elongation, while the lower ties showed a slight decrease in load despite the elongation which would have occurred along with the deformation of the wall. This implies slight slippage between the tie and the grout, or creep, or a combination of both.
Figure 8 – Observations at Panel G2
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
INITIAL OBSERVATIONS AND PERFORMANCE Within days of the collapse, MRCE foundation engineering teams, along with others, began preparation of damage assessment diagrams based on results of visual inspections made with the various emergency services personnel. Those diagrams, one for each level of the basement, showed the locations of debris piles, collapsed floors, and locations where intact basement slabs appeared to be providing necessary lateral support for the slurry wall. The diagrams, as shown in Figure No. 9, were needed for assessment of the stability of the slurry wall and were useful in providing guidance to contractors performing debris removal so as not to comprise the stability of the slurry wall.
Backfill Collapsed
Safe
Figure 9 – Damage Assessment Diagram
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
The immediate concern was the possible inundation of the basement Abathtub@. The 60-inch diameter cooling water intake pipes leading directly from the Hudson River to the WTC basement were sealed immediately. The outfalls were sealed later in the first week and additional utility penetrations were investigated over the next several weeks. Nevertheless, large quantities of water entered the PATH tunnels, in particular, the northerly tunnel, completely flooding the center 200 ft of the 5400 ft long tunnel. The source of that water was not immediately clear. To avoid possible flooding of the entire PATH system, the Port Authority of New York and New Jersey chose to construct concrete bulkheads at the New Jersey side in each tunnel. Those bulkheads were designed by MRCE in collaboration with the Port Authority and its contractor, to withstand an 80 ft water head in the event of total flooding but included pipe sleeves for ongoing pumping and access for personnel in the future. Prior to installation of the bulkheads, some 3000 gallons per minute were being pumped from the tunnels up to 12 hours per day. Ultimately, it was evident that the majority of the water in the tunnels was from the huge volume of water being pumped into the WTC basement to suppress the fires which burned for months. Initial inspections of the slurry wall, where accessible, did not reveal significant damage nor was it evident that large quantities of water were seeping into the Abathtub@. Nevertheless, it became apparent early on, that a large unsupported area of wall was present along the south wall at Liberty Street. Collapse of the south tower on September 11th caused the collapse of significant areas of basement floors leaving an unsupported section, seen in Figure No. 10, some 150 ft long by 45 ft deep. That area needed to be immediately protected from the surcharge of heavy rescue and recovery equipment in addition to 60 ft of water and soil lateral load acting on the unsupported wall.
Street Above
Void Slurry Wall Loose Debris at Bottom Figure 10 – Unsupported Slurry wall along Liberty Street
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Damage to the subway tunnel along Greenwich St., east of the slurry wall, which included the crushing of a portion of the tunnel and penetrations through the roof by falling projectiles, suggested the possibility of damage to the slurry wall. However, the top of the wall could not be examined for some time until debris was removed. INITIAL EVALUATIONS AND RESPONSE Although it was immediately clear that the Liberty Street wall was in danger, it was not until some three weeks after the collapse that backfilling operations could begin when it was safe to work in that area. Before backfilling began, a 150 to 200 ft long tension crack appeared in Liberty Street about 25 feet from the basement wall. The crack grew from 2 inch wide to 2 to 4 inches wide in 48 hours, with subsidence on the wall side of about 2 inches. Nearly 30,000 cubic yards of fill were placed in four days, some of that fill merely dumped into basement void space to attempt to provide passive support of the wall. Subsequent measurements of the top of the slurry wall in this area documented that as much as 9 inches or more of lateral movement must have occurred. Shortly thereafter, installation of dewatering wells began and slope inclinometers were installed to evaluate ground movements. During the ongoing evaluation of the basement, it became clear that it would be impossible to recover old tiebacks and that in the selection of design loads for new tiebacks, a somewhat out of the ordinary design philosophy was necessary. Lateral support was being provided by basement floors in various conditions of stability, by piles of debris, or by end-dumped and largely uncompacted backfill. Further, conditions were changing on a daily basis as localized floor collapses occurred and large equipment moved about the site. Collapses occurred mainly as a result of Acontrolled@ demolition of remaining superstructure in the early days which seemed to be considered more serious or immediate danger. TIEBACK DESIGN MRCE used two basic design methods: a continuous beam method in which Rankine earth and water pressures are used outside the wall and pressures inside the wall are modeled as a subgrade reaction; and a finite element model in which earth pressures and deformations are computed from effective stress-based parameters. Each stage of the excavation/tieback installation sequence can be modeled in the latter method. Measurements made during original WTC construction were used to calibrate the model. Although the computer model successfully replicated the principal observations of the original monitoring program, it tended to overestimate bending moments near the bottom of the wall. Design of a conventionally excavated slurry wall is typically performed by MRCE using a staged analysis in which initial conditions are known, passive support at each stage can be calculated and a certain sense of confidence is provided in having reasonable control during construction of those conditions at each stage. In this instance, particularly where floors had collapsed or were in uncertain condition, and debris remained in place, little was known about the level of support provided at each stage of the excavation. A series of scenarios was evaluated to select the appropriate tieback design. The essence of the anchor design was to consider the full bending capacity of the wall to be mobilized. Hence, the anchors would be sized and pretensioned so as not to be the limiting component of the wall system as the debris was removed. The analysis started with a cantilevered wall section, continued with one anchor in place assuming no resistance from the debris, and accounted for three possible floor configurations with only one basement floor in place, as shown on Figure No. 11. As the design developed for the complete basement, checks were made using the conventional staged analysis in which support is assumed at each level below the tiebacks to verify that none of those loads exceeded the critical case load. Upper level tiebacks were designed for 700 kips at a 45E inclination, spaced roughly at 11 ft horizontally. A 600 kip anchor lock-off was selected. Anchors were tested to 800 kips. In general, the new tieback loads could be increased compared to the original tieback loads because of the availability of higher strength tendons. Consequently, one level of tiebacks was eliminated.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Figure 11 – Schematic of Simplified Analysis for Tieback Loads Calculation INSTRUMENTATION PROGRAM The instrumentation program comprised monitoring wells, inclinometers and optical monitoring points. Instruments were installed as soon as monitoring locations were cleared of debris and equipment could safely operate in those areas. Monitoring wells had screens set in the upper fill and in the lower glacial granular materials just above top of rock. Deep wells were installed to depress groundwater levels outside the slurry wall. Pumping began on October 8, 2001, one day after the tension crack was observed. Piezometric levels dropped rapidly from about Elev. 300, roughly equal to local mean sea level, to about Elev. 290 in the upper aquifer on Liberty Street and West Street and to about Elev. 275 in the glacial deposits along Liberty Street but as low as Elev. 260 to 270 along West St. Groundwater levels were permitted to recover in most areas after large segments of the slurry wall were completely resupported. Discharge rates were typically between 3 and 10 gpm for each well, with a total site discharge, surprisingly low at only 200 gpm with approximately 40 wells on line. The wells consisted of 6-inch diameter full length slotted pipes set in a 12-inch diameter borehole with a granular filter pack. Ground movements were observed with the use of 2 -¾ inch diameter slope indicator casings installed between 1 and 15 ft outside the wall. Inclinometer casings were socketed a minimum of 10 ft into rock to assure stable origin. In the case of Liberty Street, inclinometer casing installation was after the initial tension cracks appeared and substantial wall displacements had already occurred. Optical monitoring points were set in the top of the slurry wall or a grade beam above the slurry wall. Survey readings were taken and plotted on an almost hourly basis during the early recovery period. The intent was to monitor trends rather than specific small movements so as to avert large catastrophic movements. The inclinometers were monitoring ground movements whereas the optical monitoring points permitted observation of wall performance. Figure No. 12 shows results of inclinometer readings in Liberty Street at Panel No. L9 or L10 located 150 west of Greenwich Street. Although the inclinometer was installed during backfilling, movements inward toward the excavation continued until at least the first tier tieback was installed when the trend was reversed and the wall moved into the soil away from the excavation. Total movement at the ground surface approached 3 inches, and at the third and fourth level tiebacks about ¼ to 2 inch, suggesting recovery of some of the movement that occurred prior to backfilling.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Tieback
Figure 12 – Inclinometer Readings near Slurry Wall panel L9 & L10 Figure No. 13 provides results of optical monitoring versus time and construction milestones at Panel No. L-13. The initial deceleration of movements toward the excavation is seen from the flattening of the curve followed by movement away from the excavation after the first tiers of ties were installed. Deflection toward the excavation is seen in each stage of excavation or debris removal followed by stabilization of negative deflection as each lower tier tieback was installed. Note that the movement which preceded instrumentation is estimated. Inclinomete Baseline 10/17/0 12. 11. 10.
WTC - Liberty Slurry Wall Panel L-13 Lateral Displacements Subsequent response of slurry wall to re-excavation and Tieback installation and tensioning.
Backfillin
9.
Lateral 8. Displacement 7. 6. 5. 4. 3. 2.
Response of slurry wall to top Tieback tensioning after lateral wall Initial pumping wells placed onOctober 8,
Upper Level Tiebacks 2nd Level Tiebacks 3rd Level Tiebacks 4th Level Tiebacks
1. 0. -
Estimated movements prior to monitoring (4")
10/6/200
10/20/200
11/3/200
11/17/200
Tension Crack Occurred Oct.
12/1/200
12/15/200
12/29/200
1/12/200
1/26/200
Dat
Figure 13 – Optical Monitoring at Liberty St. Slurry Wall Panel L-13
2/9/200
2/23/200
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Results of slope inclinometer observations in West Street are shown in Figures Nos. 14 through 16. Observations in Panel W9, near the south end of the wall, showed displacements away from the excavation in the order of 3/4 inch. When plotted as a function of time, the rate of movement was approximately 2 inch in 80 days slowing to about half the rate at the latest observations.
Tieback
Figure 14 – Inclinometer readings near Slurry Wall panel W9
Tieback
Figure 15 – Inclinometer readings near Slurry Wall panel W35
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Tieback
Figure 16 – Inclinometer readings near slurry wall panel W31 An inclinometer casing was installed at Panel W35/W36 near the same location as the data summarized by Saxena (1974). The original instrumentation was placed in the slurry wall rather than in the soil behind the wall as during that early construction there was the luxury of planning ahead and placing the casings within the panels prior to concreting. As time was critical during recovery efforts, MRCE opted to install the inclinometers in soil to eliminate the need to drill through concrete, potentially further compromising existing unknown conditions of the wall. Initial readings showed the wall movement inward during approximately the first 60 days of observations. Upper tier tiebacks were tensioned between November 3rd and December 27, 2001, at which time movement away from the excavation began. Total movement away from the excavation exceeded the movement observed toward the excavation by about 1/4 inch. Inward movements in the organic layer 20 to 50 ft below grade were recovered to only about 50 percent. No recovery is observed in the very compact hardpan. Observations at Panel W31, four panels to the south, provide performance curves similar to those observed during original construction at Panel W35, although in the lower portions of the curve there is minor inward movement toward the excavation. Outward creep of the wall following tieback tensioning was in the order of 2 inches in 85 days near the top of the wall and 1 ½ inches in 80 days near the bottom of the fill layer. Observations in inclinometer SI-12 along Vesey Street (Figure No. 17) were strikingly different. Movements up to 1 inch at the bottom of the inclinometer were observed. Little or no movement was recorded in most of the portion of the profile where support was derived from the intact floors and little or no movement had occurred prior to tieback installation. Substantial loss of ground occurred during tieback drilling at the lower level, where water pressure was not relieved outside the wall, and nonplastic fine sand and silt was present. The inclinometers recorded ground movement toward the wall, but little or no wall movement is believed to have occurred. The inward soil movements could not be recovered by the tieback tensioning process.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Tieback
Figure 17 – Inclinometer readings near slurry wall panel V17 The most significant damage to the slurry wall was sustained along Greenwich Street where the wall was struck by the collapse of Tower No. 2. The top of the wall could not be examined until substantial debris removal was completed and the truck access ramp into the site was removed from that location. A dense debris pile from Tower No. 2 remained in place in that location until the last months of the recovery process. When access to a lower portion of that wall became available in late winter/early spring, the front face reinforcing steel shown in Figure No. 18 was observed to have been separated from the concrete, and badly distorted. The vertical impact on the wall as well as the lateral movement caused complete crushing of the bottom portion of the wall and a hinge was formed where only reinforcing steel remained in placed. The upper portion of the wall was so badly damaged that the top tier anchors were not installed. The second tier anchors were lowered 3 feet from the design location. Oversized steel plates were used with the tieback trumpets in order to spread the load over a wider area of the badly damaged concrete. Top of the wall was displaced 3 to 4 ft inward when compared to the straight line of the original slurry wall. Damage extends over approximately 5 to 6 panels in that area.
Figure 18 – Crushed wall at Greenwich Street
A new reinforced concrete liner wall is being installed inside the slurry wall to restore lateral stability to
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
that area. The new liner wall will be between 2.5 and 6.5 ft thick to account for the current deviation of the wall and to form a vertical inside face. In conclusion, it should be emphasized that the instrumentation was installed under less than ideal conditions, that numerous instruments were damaged and replaced, and that the instrumentation was intended for early warning of large scale wall movements rather than as a research project. ACKNOWLEDGMENTS MRCE work was performed under the direction of Senior Partner, Mr. George J. Tamaro. Tieback design was under the supervision of Senior Associate, Mr. Theodore Popoff, and computer analyses were performed under the direction of Senior Associate, Dr. Robert Semple. Much of the instrumentation field work and data compilation was performed under the direction of Mr. Joel Volterra. Thanks to Mr. Tamaro, Dr. Semple and Mr. Volterra, each of whom reviewed this manuscript. Field damage assessments and coordination of tieback work were under the direction of Mr. Pablo Lopez and Mr. Andrew Pontecorvo. Many members of the staff of MRCE worked tirelessly under the most difficult of circumstances at the WTC, for which the writer expresses admiration and a debt of gratitude. REFERENCES 1.
Gould, J.P., ALateral Pressures On Rigid Permanent Structures,@ Lateral Stresses in the Ground and Design of Earth-Retaining Structures, American Society of Civil Engineers, NY, pp 247-253.
2.
Kapp, M.S., ASlurry-trench construction for basement wall of World Trade Center,@ Civil Engineering, American Society of Civil Engineering, April, 1969, pp 36-40.
3.
Moskowitz, J., ARecovery of the World Trade Center Basement,” Fulcrum, Deep Foundations Institute, Englewood Cliffs, NJ, winter 2001-2002, pp 34-35.
4.
Saxena, S.K., A Measured Performance of a Rigid Concrete Wall at the World Trade Center,@ Diaphragm Walls & Anchorages, Proceedings of the Conference organized by the Institution of Civil Engineers and held in London, 1820 September 1974, Institution of Civil Engineers, London, 1975.
5.
Volterra, J.L., AOverview of Site Conditions and Instrumentation at the World Trade Center Collapse,@ Geotechnical News, BiTech Publishers, Vol. 20, No.1, March 2002, pp 29-33.
6.
Engineering News-Record, McGraw-Hill Publishing. References and other articles of interest: July 9, 1964 Page 36 September 23, 1965 Page 29 April 13, 1967 Page 62 June 1, 1967 Page 21
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
LOAD TESTING HIGH CAPACITY DRILLED SHAFTS Jorj O. Osterberg Department of Civil and Environmental Engineering, Northwestern University 2145 Sheridan Road, A236, Evanston IL, 60208-3109
[email protected]
SUMMARY Until ten years ago, the largest load tests on drilled shafts (bored piles) were about 3,000 tons (27MN). These tests were made with kentledge (large weights, usually concrete blocks) stacked up as a reaction to the applied downward load of a hydraulic jack. In some cases, the jack load is resisted by a load frame held down by piles or anchors connected to the frame. When the test loads exceed 3,000 tons (27MN), these methods become cumbersome expensive and time consuming. The Osterberg (O-cell) test method does not require any reaction load. It has become the only cost effective method for static load testing of high capacity shafts. Over 650 load tests have been performed with this method, 86 of which were over 3,000 tons (27 MN) and 40 over 5,000 tons (44 MN), and the largest test load was 17,000 tons (150 MN). Tests have been made on sands below the water table, shafts ending in rock sockets, and cemented sands and gravels. Test results and details for selected load tests are discussed. INTRODUCTION In recent years the availability of much larger and more powerful drilling equipment and the development of drilling muds has made it possible to drill shafts of larger diameter and to greater depths than was formerly possible. Shafts 10 ft (3m) diameter of depths over 300 ft. in granular soils below the water table have become practical and economical to construct. These shafts have become competitive with clusters of driven piles for bridge piers.
is applied to the device, an equal upward and downward force is applied to the shaft. The load is obtained from a pre-determined calibration curve. The downward force is resisted by the side shear (often called skin friction) and the upward force is resisted by the end bearing. Therefore no kentledge or hold down frame is needed.
The use of a single drilled shaft to support a bridge pier instead of a cluster of driven piles must have adequate capacity to take the pier load safely and with limited allowable settlement. Hence load testing such a drilled shaft becomes very important. Using the O-cell method for such a shaft makes it possible to test to very high loads over water to determine the capacity and expected settlement of the shaft. HOW THE O-CELL WORKS The Osterberg Load Cell (O-cell) is a jack-like hydraulic device placed at the bottom or/and some distance above the bottom (Fig. 1). When fluid pressure (usually water) Department of Civil and Environmental Engineering Infrastructure Technology Institute Northwestern University, Evanston, Illinois, 60208
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
upward movement occurs until just before the ultimate shear resistance is reached. The downward movement, representing compression in end bearing, is much larger than the side shear at equal upward and downward loads. This means that at a load of one half the ultimate load in either side shear or end bearing, which is roughly the working load of a typical properly designed shaft, most of the load is taken in side shear. Ironically, many shafts which are designed using only end bearing and no shear actually have most of the load taken in side shear. And in cases where there is appreciable bottom disturbance, some shafts could take the entire load in side shear!
Fig 1 The Osterberg Load Test Setup The test is continued until either the ultimate side shear, the ultimate end bearing or the capacity of the O-cell is reached, whichever occurs first. Since the upward and downward forces are always equal, the tested capacity is equal to two times the load activated by the load cell. Tests are ordinarily made using the “quick method” specification of the Federal Highways division of the Department of Transportation that requires the load to be held for four minutes. However, loads can be applied for any interval. A number of tests have been made by testing over long intervals of time and in two cases as long as three years. As shown in Fig. 1, all movements are measured in relation to a fixed reference beam and are recorded on a data logger. Measured are the applied pressure, the downward movement of the bottom of the load cell, the upward movement of the top of the cell, and the upward movement of the top of the concrete. A computer records and plots the data as the test proceeds. Thus, the loadupward movement curve and the load-downward movement curve can be observed on the monitor screen as the test proceeds. The difference between the upward movement of the top of the load cell and the upward movement of the top of the concrete measures the elastic compression of the concrete above the load cell. Fig.2 shows the results of a typical O-cell test. The shapes of the two curves shown are typical of almost all soils and rock types even though the magnitudes of the loads are quite different for different soils and rocks. Note from the curves that as the load is applied, very little
Fig. 2 Typical O-Cell Test Results An equivalent top-down curve can be constructed from the test results shown in Fig. 2. This can be done by first assuming that the shaft is incompressible and that the ultimate side shear in an upward direction is the same as the ultimate shear in a downward direction. To do this, the movement is determined at an arbitrary load on the loaddownward movement curve. Then the load for this movement on the upward movement curve is determined. The sum of these upward and downward loads at this movement is the equivalent top-down load for this movement. The process is continued until enough points are determined to draw the equivalent top-down curve. The top-down curve can then be modified to take into account the compressibility of the shaft. The equivalent top-down curve derived from the test data in Fig.2 is shown in Fig.3. Verification of Shear Assumption To determine the validity of assuming that the upward ultimate shear is equal to the ultimate downward shear, tests were
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Fig.3 Equivalent Top-Down Curve Derived from Fig. 2 Test Curve Performed in Japan and reported by Ogura (1996). Piles were pushed upward with the O-cell and then with the Ocell open and depressurized (no end bearing), the piles were pushed down from the top by means of a load frame and a jack with a load cell to measure the load. It was found that the ultimate side shear in both directions were the same. Tests in Singapore also demonstrated the validity of the O-cell test. An O-cell test was made on a drilled shaft 1.2m (4 ft.) in diameter and 37m (121 ft.) long. A kentledge test was then made on a drilled shaft of the same dimensions 10m (33 ft,) away (2) (1999 Peng et.al.) A comparison of the results of the two tests is shown in Fig. 4, The only significant difference between the load curves is the large creep measured in the kentledge test beyond 1.5 times the working load, The creep was larger because the 1 hour holding time at each load interval compared to the 4 minute holding time for each interval in the O-cell test. From the above tests it is seen that there is no significant difference in ultimate side shear in a drilled shaft whether it is pushed up from the bottom or pushed down from the top.
Fig.4 Kentledge Test vs. O-cell Equivalent Top-Down LoadSettlement Curve Positioning the O-cell for Various Soil and Rock Conditions. The O-cell can be placed in various locations in the shaft as shown in Fig, 5 5a. This position is appropriate where the estimated side shear and the end bearing are approximately equal or the end bearing is much larger than the side shear when only the side shear is to be determined. 5b. To determine both the ultimate end bearing and ultimate side shear, the cell is placed at a predetermined distance above the bottom. If the distance is determined correctly, the ultimate side shear above the cell will be reached when the side shear below the cell plus the end bearing reaches its ultimate value. Strength values or blow counts can be used to determine the cell location. Good results have been achieved as shown by several of the following case histories.
Fig. 5 Alternate Positions for Placement of O-Cells 5c. For a rock socket, the concrete can be placed to the top of rock. As an alternative the entire hole can be filled with concrete and strain gages place at intervals above the socket to differentiate the load that is taken by the socket and by the side shear above the socket. 5d When it is estimated that the end bearing will be less than the side shear and the side shear is to be determined, a bell can be excavated below the planned bottom level
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
and the cell placed on top of the bell and concreting the rest of the hole. 5e Where shafts are to be installed from the existing ground surface and excavation for the basement is to be made later, the shaft is excavated to the planned depth, the cell placed and the concrete filled only to the planned cutoff elevation. The rest of the hole can be filled with sand or other fill material 5f. Where the ultimate side shear of two layers are to be determined, the concrete Is filled only to the top of the lower layer and tested. After the remainder of the shaft is filled with concrete, it is tested again to obtain the ultimate side shear of the entire shaft. By subtracting the side shear obtained for the lower layer from the total, the side shear of the upper layer is determined.
Fig. 6 illustrates the rock profile for which a rock socket is to take an upward and downward load of 200 tons (180 metric tons). After casing the overburden, a hole 10 ft. (3m) in diameter was drilled into hard limestone. Rock cores from the limestone indicated compressive strengths of 15,000 to 20,000 psi. The lower 5 ft. (1.5m) was filled with concrete and tested using an O-cell. It was planned to fill the upper 5 ft. with concrete after the test and then to test the full 10 ft. However, the test indicated that the side shear of the lower 5 ft. was much larger than expected and therefore the second test of the 10 ft. concrete filled socket was cancelled. The test of the lower 5 ft. indicated that at 900 tons or 4.5 the design load of 200 tons, the upward deflection was only 0.06 inches (0.3mm). The test showed that even with a 5 ft. socket, the upward and downward design load would cause almost no measurable movements.
5g. Using two cells with one at or near the bottom and the other at a predetermined distance from the bottom, the ultimate side shear above the upper cell, the ultimate side shear below the upper cell, and the end bearing can be determined. This requires the cell loads to be applied in stages. One procedure is to pressurize the lower cell to obtain the end bearing below the cell. Then with the lower cell depressurized, the upper cell is re-pressurized to obtain the side shear below the upper cell. Then with the pressure line to the lower cell closed, the upper cell is pressurized to obtain the side shear above the upper cell. Other sequences of applying pressure can be used depending on the soil profile and the location of the cells in the shaft. Tests on Rock Sockets For drilled shafts which are socketed in rock, it has been found from the results of O-cell that in all but a relatively few cases that the shafts have been too conservatively designed. It is frequently not recognized that the side shear in rock sockets can be very large, even for fractured and highly laminated rocks. The following cases are examples of much larger side shear capacity than anticipated by the designer.
Fig. 7 Soil and Rock Profile and Load-Deflection Curves for Bridge over Ohio River
Fig. 6 Limestone Rock Socket to be designed for Tension and Compression
Fig. 7 shows the soil and rock profile and load deflection curves for a test shaft for a bridge over the Ohio River in Kentucky. The rock consists largely of shale with seams of sandstone, limestone, and coal seams. The compressive strength of the rock cores varied from 350 to 500 psi
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Fig. 8 Test Results for Rock Socket Load Tests
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
(2,500 to 3,500 Kpa). Because of possible deep scour which might occur in the future, only the load capacity of the19 ft (5,8m rock socket) in the shale below the sand was to be considered in the design. As seen in Fig. 7, concrete was placed some distance above the rock socket. However, strain gages in the concrete indicated that the load taken in the shaft above the rock socket was negligible. The test was to be carried to three times the design load of 1000 tons (8.85MN). The load was actually carried to six times the design load (3000 up and 3000 down) before it reached the capacity of the O-cell. At the design load the total deflection was 0.2 inch (5m). No effort was made to redesign the working shafts to achieve considerable cost savings.
occurred. For each test the ratio of the maximum test load to the design load is indicated except in cases where the design load was not known to the writer. Of the 22 test, there are 13 cases in which the ultimate loads in both side shear and end bearing were not reached and yet the loads were carried to from 7 to 9 times the design loads. For the Chicago test on hard limestone, end bearing was reached at a deflection of 0,13 in. (3.3mm), considerably below the ultimate load. The Chicago code allows 100 tons/sq. ft. on hard limestone and an increase of 20% for each foot of penetration into the rock up to a maximum of 200 tons/sq. ft. The results shown in Fig. 8 indicate that design loads are in almost all cases too conservative. CASE HISTORIES
TEST RESULTS FOR ROCK SOCKETS
World Record 17,000 ton Test
Many O-cell tests have been made on rock sockets and virtually all of them went to high test loads which indicated both side shear and end bearing values were much greater than was assumed in the design. The writer chose results from 22 O-cell load tests on
A new world record was set in 2001 when a drilled shaft was loaded to a test load of 17,000 tons (150 MN) in Tucson, Arizona in a test for the Arizona DOT. The soil profile is shown in Fig. 9. The soil consists of layers of stiff to very stiff clays, dense to very dense sand, and the bottom 50 feet, slightly cemented very dense clayey and silt sand with gravel. No water table was encountered. The shaft was 8 ft.(2.4 m) diameter and 136 ft. (41m) deep. The Arizona Highway Department felt that at this depth the total capacity was greater than actually needed but no load tests had been performed to determine the actual load bearing capacity of shafts in this formation. The purpose of the test was therefore to learn if high capacity shafts could safely be installed at shallower depths.
Fig. 9 Soil Profile for Arizona Test rock sockets for study and evaluation. The results are tabulated in Fig. 8. For each test the unit maximum side shear and unit maximum end bearing were calculated and is indicated in Fig. 8. Also indicated below the maximums are the deflections at which these maximums
Fig. 10 shows the load-downward movement of the portion of the shaft below the O-cell assembly and the load-upward movement of the portion above. It is seen that the end bearing plus the side shear of the shaft below the O-cell assembly reached ultimate at about 8,500 tons (75MN). The side shear on the shaft above the cells reached about 8,500 tons (75MN) at an upward movement of only 0.2 inches (5 mm). This small movement and the fact that the first and second cycle of loading yielded almost identical curves indicates that upward movement was virtually all elastic. Thus large capacity drilled shafts of other diameters and with the same average side shear reached on this shaft would take almost all the load in side shear with very small downward movement. Strain gages readings were used to determine how the applied O-cell load was dissipated along the shaft which made it possible to determine how the unit shear varied along the shaft. From this information, rational designs for shafts of various load capacities can be obtained. Fig.11 shows the equivalent top down load-movement curve.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Equivalent Top Load
Fig. 11 Equivalent Top Load- Movement Curve World Record – 150MN (17,000 tons)
Fig. 10 Movement Curves – Arizona Test SHAFT TESTED FOR PULLOUT RESISTANCE. Bridge piers for a California bridge were to be retrofitted to withstand future earthquakes. For this purpose a drilled shaft (Fig.12) was constructed with O-cells placed above the bottom. A 6 ft. (1.83m) hole was drilled to the top of rock. A steel casing was set into the hole and a short distance into the rock to seal off soil and water from entering. A smaller hole was then drilled inside the casing to 209 ft. (66m) below the water level. The rock consisted of siltstone, sandstone, and shale ranging from very soft to hard. Reinforce concrete was tremmied to the top of rock level. This 73 ft. (22.3m) reinforced 5.5 ft diameter (1.7m) shaft was to be tested for pullout capacity. Four 21 inch O-cells were placed between two rigid plates were set 27 ft. (8.3m) above the bottom. To eliminate end bearing, a 2 ft. (0.61m) thick foam cushion was installed below the bottom of the shaft. (See Fig. 12). Installing such a cushion just below the shaft bottom presented a challenge. A jack at the base supported a plate which held the rebars and newly poured concrete until the concrete set. A rubber bladder was placed around the foam and then pressurized to keep newly poured concrete out of the annular space around the foam. After the concrete set, the jack was released and the bladder depressurized. Three levels of strain gages with two at each level were placed on top of and below the Ocells as shown in Fig. 12. Telltales, LVDT’s and tremie pipes are not shown.
Fig. 12 Tension Test Shaft for Bridge to Resist Earthquakes Testing was performed by pressurizing the O-cells to mobilize the downward shear resistance above the cells and the upward shear resistance below. Since it has been demonstrated that the ultimate upward shear is equal to the ultimate downward shear, and since there was no end bearing (verified by the strain gage readings), the pullout resistance is equal to the sum of the side resistance above and below the cells. The loadupward movement and the load-downward movement curves are shown in Fig. 13. The equivalent top pullout load- movement curve is shown in Fig. 14.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Fig. 13 Load-Upward and Load Downward Movement Curves (Tons)
Fig. 14 Equivalent Top Pullout Curve (Tons)
Fig. 15 World Records for Static Load Tests using O-Cell Method
Materials & Behavior Behavior of Polyurethane Grouts Used for Leak Control in Wastewater Systems C. Vipulanandan, University of Houston, USA Scaling Laws for Sea Ice Fracture Z. P. Bazant, Northwestern University, USA Particle Size Effects on Rockfill Compressibility E. E. Alonso and D. Montobbio, Universitat Politècnica de Catalunya, Spain and Asdoconsult Ingenieros S.L., Spain Vibrations Caused by Chiselling in Deep Foundations Construction C. Bouniol, H. Duplaine, Balineau SA, France Chemical Transport Issues in Modern Landfill Liners T. B. Edil, University of Wisconsin – Madison, USA
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
BEHAVIOR OF POLYURETHANE GROUTS USED FOR LEAK CONTROL IN WASTEWATER SYSTEMS C. Vipulanandan Center for Innovative Grouting Materials and Technology (CIGMAT), University of Houston, 4800 Calhoun, Engineering Building 1, Houston Texas, 77204 cvipulanandan @uh.edu
ABSTRACT Water leaks into civil infrastructure facilities are being controlled by grouting using polyurethane based grouts. Hence polyurethane grouts are used in wide range of environmental conditions and there is increased interest in better characterizing the behavior of hydrophilic polyurethane grouts. In order to investigate the behavior a typical polyurethane grout was selected for this study, which had a free volume expansion of over 800% for water-to-grout (W/G) ratio of 0.5. In this study, hydrophilic polyurethane grout was mixed with varying amounts of water (up to 8 times the net resin volume) and the curing parameters (pressure, temperature, and volume change) and the compressive behavior of cured (solidified) grouts were investigated. Samples were prepared in a specially designed molds by controlling the volume change. Pressuretemperature-time relationships during curing have been developed. Unit weights of cured polyurethane so prepared varied from 1.1 kN/m3 (7 pcf) to 10 kN/m3 (63 pcf). The compressive stress-strain relationships of the grouts were influenced by the water-to-grout ratio and the properties were less depended on the unit weight of the cured grouts. Polyurethane grout bonded to the siliceous surface (representing sand surface) and the bonding strength also varied with water-to-grout ratio. The behavior of cured polyurethane grout was studied under wet and drying cycles. Each cycle had a week of immersion in water followed by a week in air. During the test, weight change and volume change were measured up to forty wet-dry cycles. Parameters influencing the swelling and shrinkage of the grout have been identified. Using multiple regression analysis a relationship has been developed to represent the maximum swelling in terms of cured grout parameters. INTRODUCTION When faced with leaking problems in wastewater systems that are structurally sound, grouting is an effective method of rehabilitation and polyurethane based grouts are frequently used. The ideal grout for these applications should have low viscosity to minimize pumping pressure, good gel time, and the ability to make the soil around the leak impermeable [Karol 1990; Bodocsi et al. 1991; CIGMAT News 1995; Vipulanandan et al. 1996 a & b; Concrete Construction, 1998]. Hydrophilic polyurethane grouts when mixed with water will expand and seal the leaks. But the grout behavior during and after curing is not well understood. Polyurethane chemistry is extremely versatile and is under development all the time [Goods, 1982; Aronld, 1995; Arefmanesh, 1995; Yasunaga et al. 1997]. Polyurethane forming is a continuous process and the cell formation is considered to be an important stage since it determines the durability of the material. By varying the proportions of the components in the system it allows foams to be produced which have a range of densities of less than 8 pcf up to 70 pcf and with extremely useful chemical and mechanical properties [Vipulanandan et al., 2000]. The porosity and unit weight of the polyurethane foam could influence the performance of the materials. Only limited information is available on polyurethane grout performance for civil infrastructure related applications.
Department of Civil and Environmental Engineering Infrastructure Technology Institute Northwestern University, Evanston, Illinois, 60208
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
History During the past four decades, polyurethane chemistry has matured into a sophisticated industry despite the initial slow-start, which is typical of so many ultimately successful materials except asbestos. Nearly 14 years elapsed before sufficient commercial interest developed to invest capital in the diisocyanate polymerization reaction first discovered by German chemist Otto Bayer in 1937. Only in 1954, polyurethane chemistry with innovative methods of "tailor making" plastics, gels and foams was transferred to the U.S.[CIGMAT News, 1995] It should be noted that ASTM has no standard method for testing polyurethane grouts and limited data reported in the literature. Lack of standard testing procedures for polyurethane foams and rigids in civil engineering applications makes it difficult for the design engineer to select this material. It is not clear how the pressure, temperature and microstructure of the grout is affected if the free volume expansion is restricted, which is the case in most the civil engineering applications. Also the behavior of the cured foam and gel grouts produced under confinement has not been quantified. It is not clear how pressure, temperature, and microstructure of the grout are affected if the free volume expansion is restricted, which is the case in most leakage control applications. When polyurethane is used to grout sandy materials (out side pipe joints), the grout not only fills the voids but also bond the sand. It is not clear how the bonding strength of the grout is affected by the water-to-grout ratio. Characterization of the curing process and the behavior of cured polyurethane grout including wet-dry cycles will lead to better grout mix design based on the application. OBJECTIVES The objective of this study was to investigate the effect of curing conditions and the performance of polyurethane grout under wet-dry cycles. The specific objectives are as follows: (1)
Develop pressure-temperature-time relationships for curing polyurethane grout specimens under controlled volume change.
(2)
Quantify the compressive stress-strain relationships for the cured polyurethane grout mixes.
(3)
Determine the bonding strength of the selected polyurethane grout mixes.
(4)
Quantify the changes in the polyurethane foam grouts under wet-dry cycles after selecting the wet and dry testing conditions.
(5)
Determine the relationship between maximum swelling and selected polyurethane grout properties.
All the testing was performed at the CIGMAT Research Laboratory at the University of Houston. MATERIALS AND TESTING PROGRAM (a) Net Grout: A commercially available AV-202 multigrout (Avanti International, Webster, Texas) was selected for this investigation since it reacts with water in any proportion to form a foam or gel. The grout resin was dark brown in color with a viscosity of 2500 cps (at 30oC) and a specific gravity of 1.15. Depending on the proportioning of the resin to water, a range of products from very porous foam to gel can be obtained. (b) Specimen Preparation Using specially instrumented 100 mL molds and various grout –to-water mixes, specimens were prepared with controlled volume change [Vipulanandan et al., 2000]. The amount of resin and water to be added was determined by the allowable volume change needed. Resin and water were mixed at room condition (temperature (23 ± 2 °C) and at relative humidity (50 ± 5 %)). Using the thermocouple, temperature increments were monitored during the curing process. These measurements were taken until temperature returned to its initial temperature (or temperature increment of zero). The change in load was measured with time to an accuracy of 2 lb. The pressure was monitored for a period of 3 to 6 hours, or as needed. The specimens were then removed from the molds and stored in airtight zip lock bags till the time of testing.
Department of Civil and Environmental Engineering Infrastructure Technology Institute Northwestern University, Evanston, Illinois, 60208
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
(c) Unit Weight For each cured specimen length and diameter of specimen was measured with a Vernier caliper (least square of 0.01mm). When measuring the specimen, special care was taken to not apply pressure on the foam specimen. The specimens were weighted to an accuracy of 0.1 g. The results reported are average of three readings. (d) Volume Expansion In both free expansion tests and controlled volume tests, polyurethane grout was allowed to expand in one direction. The change in volume of an expanding material may be expressed as
∆V = ( where
V f − V1 V1
) * 100
[1]
Vi = Initial volume of unreacted material (resin + water) Vf = Final volume of reacted material ∆V = Percentage Volume Change due to foaming
The volume expansion (∆V) was thus defined as the ratio of the change in volume of reacted (hardened) with respect to unreacted materials (resin+water). Volume expansion describes how many times the material has expanded with respect to original volume of the material mixed and is equal to the void ratio. (e) Mechanical Properties Unconfined Compressive Stress-Strain Relationship (CIGMAT GR 2): There is no standard method to test the polyurethane grout in compression under monotonically increasing load. ASTM D 3574 and ASTM C 109 were used to select the loading rates and specimen size for the unconfined compression test [Vipulanandan et al., 1996, 1997]. Compression tests were done using a screw-type machine with a capacity of 5 kips and specimens were loaded at a rate of 1%/min. Specimens 38 mm in diameter and 30 to 35 mm in length were tested. At least two specimens were tested under each condition. Bonding Strength (CIGMAT GR 5): This test consisted of sandwiching a layer of polyurethane grout between flat surfaces of quartz rocks and then loaded in tension. Flat rock surfaces were obtained by sawing a 2-inch diameter cored rock specimen, which was glued with epoxy to the aluminum plates. This test configuration has been used earlier in determining the bonding strength of silicate and cement grouts (Vipulanandan et al. 1992; CIGMAT 1998). (f) Microstructure Scanning electron microscope (SEM) was used to study the microstructure of the cured polyurethane with different unit weights to determine the distribution of the voids and their sizes and shapes. Polyurethane specimens were coated with conducting materials before using for SEM analysis. (g) Wet-Dry Cycle Test Water absorption during wet and dry cycles for the foam grouts was determined by measuring the change in both weight and volume when the specimen was immersed in tap water (drinking water, pH of 8) for a week and the following week placed at room condition (temperature of 23±2 °C and relative humidity of 50±5 %). The room condition was selected to represent the dry condition in this study. Volume and weight were recorded on a daily basis during the test period. At the end of the immersion period, the difference between the initial weight and final weight was the weight gain in the specimen. Similarly change in volume was determined. At the end of the dry period, the difference between the initial and the final weight was due to the loss of water. Results recorded were the mean of a minimum of three measurements. Tests were conducted continuously for 40 cycles. The detailed procedure used for this test is described in CIGMAT Standard GR 3-00.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
TEST RESULTS AND DISCUSSION In this study, a hydrophilic polyurethane resin was mixed with water up to 8 times the volume of the resin. During curing under controlled volume change, temperature and pressure were monitored. The unit weights of the specimens varied from 1.1 kN/m3 (7 pcf) to 10 kN/m3 (63 pcf). (a) Curing Process Pressure (P)-Temperature (T)-Time (t) Relationships: The changes in pressure and temperature with curing time for a grout mix with 150% volume change is shown in Fig. 1. In all the cases investigated the temperature rise was faster than the pressure. For gout mixes with water-to-grout ratios of 1 and 6 the peak temperature were reached in 2 and 8 minutes respectively. While the temperature continued to decrease rapidly the pressure remained almost unchanged after reaching the peak for longer period of time. According to the grout supplier's (Avanti International) brochure the foam-time for the waterto-grout ratios of 1 and 6 are 1.5 and 8 minutes respectively. These foam-times better corresponded to the time to reach maximum temperature than maximum pressure. In this study pressure as high as 1.3 MPa (180 psi) has been measured. Increasing the water-to-grout ratio and allowable volume change during curing resulted in reduced pressure increase. Increasing the water-to-grout ratio also reduced the maximum temperature raise during curing. For the variables investigated the maximum temperature rise (above ambient) was about 30oC. (b) Unit Weight versus Water-to-Grout Ratio (W/G) The relationship between the polyurethane foam unit weight and the W/G ratio (volume) was determined through a series of tests and the variation is shown in Fig. 2. The relationship between the unit weight (γ in pcf) and water-to-grout ratio (W/G) in terms of volume (αv) can be represented as follows: γ = 13.3 + 5.15αv
[2]
The coefficient of correlation was 0.83 and αv varied from 0.5 to 8. It must be noted that 10 pcf is equal to 1.6 kN/m3. (c) Mechanical Properties Stress-Strain Relationship: The compressive relationships were nonlinear and are affected by the water-to-grout ratio and void ratio. The strength of grout at comparable strain was depended on the water-to-grout ratio. Loading and unloading showed that the material was not elastic at the rate of testing. The material is becomes stiffer with increase in strain and is explained by collapsing of the thin cell wall and hence the material becomes more dense during the loading. The average modulus in the initial 20% strain for the water-to-grout mix ratios of 1 and 6 were 0.86 MPa (120 psi) and 0.086 MPa (12 psi) respectively. Bonding Strength: Test results indicate that the polyurethane grout bonds to the quartz surface. Bonding tests were done after at least two days of curing. For W/G of o.5 the bonding strength was 26 psi (180 kPa). Bonding strength was 7 psi (50 kPa) for W/G ratio of 2. Bonding strength reduced with increasing water-to-grout ratio. (d) Microstructure The microstructure of polyurethane grout (water-to-grout ratio of 0.5) with a void ratio of 1.5. was such that the voids were spherical and were not connected. Non connectivity of the voids indicate closed cell structure for the grout. Also the void7s were uniform in size and had diameters of the order of 400 µm. The microstructural information will be very valuable to model the polyurethane grout behavior. (e) Wet-Dry Cycles When the W/G ratio was 0.5 and ∆V=150% (unit weight of 34 pcf/5.4 kN/m3) the maximum weight gain during the wetting (swelling) was 190% (Fig 3(a)). During drying, the minimum weight gain was 0%. The grout repeated its performance over 40 cycles without any degradation. The maximum and minimum volume changes were 100% and 20% respectively. Although weight returned to the original value the volume did not and had residual swelling. Performance of another grout mix with W/G ratio of 2 and ∆V= 0% (unit weight of 56 pcf / 9 kN/m3) the maximum weight gain during the
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
wetting (swelling) was 60% (Fig 3(b)). During drying, the maximum weight loss was -45%. Additional weight loss is possibly due the loss of water from the grout. 80
50
40
Pressure (psi)
60 50
30
40 30
20
Pressure (psi) o
Temperature ( C)
20
10
o
∆T max=42 C (4min; 68 psi)
10
o
Max P= 80 psi (6 min; 40C)
0
0 0
2
4
6
8
10
12
14
16
Time (min) Pressure-Temperature-Time relationships for curing polyurethane grout mix with water-to-gout of 0.5:1 and volume change of 150%. 70
60
Unit Weight, γ, (pcf)
Figure 1.
Temperature Increment, ∆ T (°C)
Poly#11/0.5:1/150%
70
50
40
30
20 y = 13.322 + 5.1516x R= 0.83 10
0 0
2
4
6
8
Water-to-Grout Ratio,W/G (α ) V
Figure 2.
Effect of water-to-grout ratio on the unit weight of cured grouts.
10
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
250
140
(a)
Weight Fluctuation (%)
100 150
80
100
60 40
50
20
Poly 9 W/G = 0.5 ∆V = 43 % γ = 34 pcf
0
Weight (%)
0
Volume (%)
-50 0
Volume Fluctuation (%)
120
200
1000
2000
3000
4000
-20 5000
Time (hrs)
80
80
(b)
60
Weight Fluctuation (%)
40
40
20 20 0 0 -20 -20
-40 Poly 14 W/G = 2 ∆V = 0% γ = 56.3 pcf
-60 -80 0
1000
2000
3000
Weight (%)
Volume Fluctuation (%)
60
-40
Volume (%) 4000
-60 5000
Time (hrs) Figure 3.
Wet-dry cycle test with swelling and shrinkage (a) water-to-gout of 0.5:1 and volume change of 43%; (b) water-to-gout of 2:1 and volume change of 0%.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
The grout repeated its performance over 40 cycles without any degradation. The maximum and minimum volume changes were 50% and -30% respectively. In this case neither weight nor volume returned to the original state and there was residual shrinkage. Inspecting the test results showed that when W/G was increased the swelling capacity (weight based quantity) of polyurethane grouts decreased. Also the swelling capacity of the grout mixes increased with the volume expansion (∆V) of the grout (Fig. 4). Higher the ∆V, the higher the voids in the specimens and hence the grout can hold greater amount of water during the water immersion. (f) Multiple Regression Analysis When analyzing the maximum swelling capacity (by weight) of the polyurethane grout, the following equation was assumed to predict the behavior of the grout.
Swell max = Aswell + Bswell *
W + C swell * ∆V . G
[3]
Using the test data and performing the multiple regression analysis leads to the following relationship
Swell max = 114.3 − 13.2 *
W + 1.3 * ∆V G
[4]
Model parameters Bswell and Cswell suggest that the swelling capacity of the grout is more dependent on the water-to-grout ratio than it is on the volume expansion ∆V. The maximum swelling capacity decreased with increase in W/G ratio. On the other hand, the maximum swelling capacity increased with ∆V. This correlation can be seen in Fig. 5 where the model prediction is compared to the experimental results. Of the data considered, 25% had higher than predicted swelling capacity and the measure-to-predicted swelling ratio was 1 with a correlation factor of 0.99. CONCLUSIONS This study focused on developing a method to prepare polyurethane grouts under controlled volume change and monitor the pressure and temperature during curing. The water-to-grout ratio was varied from 0.5 to 8 and the volume change (controlled volume test) during curing varied from 0 to 150% (based on initial liquid volume). The compressive behavior, microstructure and bonding strength of the grout were studied. The behavior of cured polyurethane grout specimens was studied under wet and dry cycles. Each cycle had a week of immersion in water and another week in air. The unit weight of cured grouts varied from 1.1 kN/m3 (20 pcf) to 10 kN/m3 (63 pcf). During the tests, weight change and volume change were measured up to forty wet-dry cycles. Based on the test results, the following observations can be advanced:
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
1200
Maximum Swelling, SwellMax’
(%)
1000
800
600
400
y = 69.124 + 1.3649x R= 0.99
200
0 0
200
400
600
800
1000
Volume Expansion, ∆V (%) Relationship between maximum swelling and void ratio (processing volume change (∆V)).
Figure 4.
400
Maximum Swelling, SwellMax’
, (%)
∆V = 0% ∆V = 43% ∆V = 150% ∆V = 0% Exp ∆V = 43% Exp ∆V = 150% Exp
350 ∆V = 0%
300 250 200
∆V = 43%
150
∆V = 150%
100 50 0 0
1
2
3
4
5
6
7
Water-to-Grout Ratio, W/G, (αv) Figure 5.
Predicted and measured maximum swelling in polyurethane grout.
1.
Pressure-Temperature-Time Relationships: This is affected by the water-to-grout ratio and the volume change allowed for the curing grout mix. Maximum temperature and pressure are out of phase and the temperature peaked first. Increasing the water-to-grout ratio reduced the maximum pressure and temperature.
2.
Stress-Strain Relationship: The compressive stress-strain relationships of cured grouts were nonlinear and inelastic at a strain rate of 1%/min. The grouts became stiffer with increased strain. Compared to the unit weight of cured
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
polyurethane, the initial water-to-grout ratio and volume change during curing were more important in influencing the compressive behavior of the grout. 3.
Bonding Strength: Polyurethane grout bonds to the quartz (simulating siliceous sand) surface. The tensile bonding strength reduced with increasing water-to-grout ratio.
4. Swelling depends on the water-to-grout ratio and ∆V. Swelling (weight based) decreased with increased W/G and decreased with ∆V. When the specimens were immersed in water, they reached the maximum swelling, which decreased with increased W/G ratio. 5.
Shrinking was related to water-to-grout ratio. In general, the shrinking capacity of the grout decreased with increased W/G ratio. When W/G ratio was 2 or higher, there was negative residual shrinkage during the dry cycles.
ACKNOWLEDGMENT This work was supported by the Center for Innovative Grouting Materials and Technology (CIGMAT) under grants from various industries and the National Science Foundation (NSF) (CMS-9526094). REFERENCES 1.
Annual Book of ASTM Standards (2001), Section 4 (Construction) and Section 8 (Plastics), ASTM, Philadelphia, PA.
2.
Arefmanesh, A. and Advani, G. (1995), "Nonisothermal Bubble Growth in Polymeric Foams”, Polymer Engineering and Science, Vol. 35, No. 3, pp. 252-259.
3.
Arnold J. C. (1995), "The Effect of Physical Aging on Brittle Fracture of Polymers”, Polymer Engineering and Science, Vol. 35, No. 2, pp. 165-169.
4.
Bodocsi, A. and Bowers, M. T. (1991), "Permeability and Acrylate, Urethane and Silicate Grouted Sands with Chemicals, Journal of Geotechnical Engineering, Vol. 117, No. 8, pp.
5.
CIGMAT News and Literature Review, Vol. 1, No. 3 (1995), Center for Innovative Grouting Materials and Technology (CIGMAT), University of Houston, November 1995 (htttp://gem1.uh.cive.edu)
6.
CIGMAT Standard GR 3-00 (2000) "Standard Test Method for Wet and Dry Cycle Resistance of Grouts and Grouted Sands," Center for Innovative Grouting Materials and Technology (CIGMAT), University of Houston, Texas. 4 p.
7.
CIGMAT Standard GR 2-02 (2002) "Unconfined Compressive Strength of Grouts and Grouted Sand," Center for Innovative Grouting Materials and Technology (CIGMAT), University of Houston, Texas. 6p.
8.
CIGMAT Standard GR 5-00 (2000) "Tensile Bonding Strength for Grouts," Center for Innovative Grouting Materials and Technology (CIGMAT), University of Houston, Texas. 8p.
9.
Concrete Construction (Oct. 1998), "Repair, Protection and Rehabilitation, pp. 898-890.
10.
Goods, G. (1982), "Flexible Polyurethane Foams, Chemistry and Technology," Applied Science Publishers, London, England.
11.
Karol, R. H. (1990), Chemical Grouting, Marcel Dekker Inc., New York, NY, 465 p.
12.
Klempner, D. and Frisch, K. C. (1991), "Handbook of Polymeric Foams and Foam Technology," Hanser Publisher, New York, New York.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
13.
Lowther, J. and Gabr, M. A. (1997), "Permeability and Strength Characteristic of Urethane-Grouted Sand," Proceedings, Grouting, Geotechnical Special Publication No. 66, ASCE , pp. 197-211.
14.
Malyshev, L. I., Korolev, V. M. and Skokov, V. G. (1995), "Use of Synthetic Resin for Grouting When Repairing Structures and Performing Antiseepage and Strengthening Works," Hydrotechnical Construction, Vol. 29, No. 12, pp. 690-693.
15.
Tonyan, T. D., and Gibson, L.J. (1992), "Structure and Mechanics of Cement Foams, " Journal of Materials Science, Vol. 27, pp. 6272- 6378.
16.
Vipulanandan, C. Jasti, V., Magill, D. and Mack, D. (1996a), "Shrinkage Control in Acrylamide Grouts and Grouted Sands," Proceedings, Materials for the New Millennium, ASCE, Washington D.C., pp.840-850.
17.
Vipulanandan, C. and Jasti, V. (1996b) "Development and Characterization of Cellular Grouts for Sliplining," Proceedings, Materials for New Millennium, ASCE, pp. 829-839.
18.
Vipulanandan, C. and Jasti, V.(1997) "Behavior of Lightweight Cementitious Cellular Grouts," Proceedings, Grouting, Geotechnical Special Publication No. 66, ASCE , pp. 197-211.
19.
Vipulanandan, C., Mattey, Y., Magil, D. and Mack, D.(2000) "Characterizing the Behavior of Hydrophilic Polyurethane Grout, Geotechnical Special Publication No. 104, ASCE , pp. 235-245.
20.
Yasunaga, K., Zhang, X. D., and Macosko, C. W. (1997), "Skin Development in Free Rise, Flexible Structure and Mechanics of Cement Foams, " Journal of Celluar Plastics, Vol. 33, pp. 528- 544.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002 Authorized republication of Z.P. Bazant’s article (2000). “Scaling laws for brittle failure of sea ice.” {preprints, IUTAM Symp. on Scaling Laws in Ice Mechanics} (Univ. of Alaska, Fairbanks, June), J.P. Dempsey, H.H. Shen and L.H. Shapiro, eds. Paper No. 3, pp. 1—23.
Department of Civil and Environmental Engineering Infrastructure Technology Institute Northwestern University, Evanston, Illinois, 60208
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Department of Civil and Environmental Engineering Infrastructure Technology Institute Northwestern University, Evanston, Illinois, 60208
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Department of Civil and Environmental Engineering Infrastructure Technology Institute Northwestern University, Evanston, Illinois, 60208
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Department of Civil and Environmental Engineering Infrastructure Technology Institute Northwestern University, Evanston, Illinois, 60208
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Department of Civil and Environmental Engineering Infrastructure Technology Institute Northwestern University, Evanston, Illinois, 60208
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Department of Civil and Environmental Engineering Infrastructure Technology Institute Northwestern University, Evanston, Illinois, 60208
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Department of Civil and Environmental Engineering Infrastructure Technology Institute Northwestern University, Evanston, Illinois, 60208
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Department of Civil and Environmental Engineering Infrastructure Technology Institute Northwestern University, Evanston, Illinois, 60208
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Department of Civil and Environmental Engineering Infrastructure Technology Institute Northwestern University, Evanston, Illinois, 60208
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Department of Civil and Environmental Engineering Infrastructure Technology Institute Northwestern University, Evanston, Illinois, 60208
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Department of Civil and Environmental Engineering Infrastructure Technology Institute Northwestern University, Evanston, Illinois, 60208
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Department of Civil and Environmental Engineering Infrastructure Technology Institute Northwestern University, Evanston, Illinois, 60208
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
PARTICLE SIZE EFFECTS ON ROCKFILL COMPRESSIBILITY E.E. Alonso1 and D. Montobbio2 1
Universitat Politècnica de Catalunya (UPC) c/ Jordi Girona 1-3 Módulo D-2 08034 Barcelona Spain
[email protected] 2
Asdoconsult Ingenieros, S.L. c/ Sant Pere Més Alt, 1, 1º 2ª 08003 Barcelona Spain
[email protected]
ABSTRACT Rockfill behavior depends on grain size distribution, because particle breakage, a dominant deformation mechanism, depends on particle size. A constitutive model, developed within the framework of hardening plasticity, uses some basic results of the theory of subcritical crack propagation to incorporate particle breakage phenomena into the formulation. Several compressibility tests on samples of different gradation were performed in a specially designed large diameter oedometer cell in which the ambient relative humidity could be controlled. Constitutive parameters were identified for the different grain size distributions. In this way, particle effects were quantified through this effect on macroscopic material parameters. Decreasing grain size implies a decrease in the delayed compressibility index and a decrease of the parameter governing the change in rockfill compressibility with relative humidity. The rest of material parameters are slightly affected by grain size. INTRODUCTION Rockfill is a common construction material in Civil Engineering works. Experimental research and field observations of rockfill structures, such as rockfill dams, lead to a good engineering understanding of the main factors controlling rockfill behavior. It was soon realized (Terzaghi, 1960) that particle breakage was a key phenomenon to explain rockfill compressibility. In a rockfill, a number of mechanisms leading to the macroscopic deformations could be isolated: contact breakage, particle breakage, structural rearrangement (sliding, rotation of particles) and, to a lesser extent, particle deformation. During loading, these mechanisms interact in a process that leads to a progressive degradation of the initial grain size distribution and to updated rockfill fabrics. The grain breakage process is controlled by the presence of water. Test reported by Nobari and Duncan (1972) demonstrated that the rockfill compressibility increased as the compaction water content increased. Rockfill collapse was explained as a transition from the equilibrium void ratio at a given confining stress and water content to a new (lower) void ratio at the same confining stress but increased water content. This transition implied the rupture of particles and the subsequent rearrangement of the rockfill structure. Some compression tests on irregular rock particles help to understand the nature of particle breakage (Fig. 1). In a first stage the contact zone is crushed and the contact area between the rock and the steel plattens increases. As loading increases, local fractures are induced in the vicinity of contacts and this mechanism results in the transient reduction of applied load. At higher loads the contact area has increased and local failures do not occur any more but a diametral fracture eventually develops along one plane containing the applied load vector. Tests of this kind have been reported by Marsal (1973) and Lee (1992). A rock tension strength, σf, may be defined on the basis of this type of test (“Brazilian” type of test) as:
Department of Civil and Environmental Engineering Infrastructure Technology Institute Northwestern University, Evanston, Illinois, 60208
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Figure 1 : Tests on rock particles (McDowell and Bolton (1998), quoting Lee (1992)) a) Particle tensile strength test set-up; b) Typical load-deflection plot; c) Mean tensile strength as a function of particle size.
σ f = Ff / d 2
[1]
where Ff is the (point) rupture load and d is the average particle diameter. Lee (1992) found that σf decreases as d increases (Fig. 1c):
σ f ∝db
[2]
where b ≅ (-0.34 to –0.42) for the materials that he tested. Marsal (1973) reached a similar conclusion. McDowell and Bolton (1998) found that Expression [2] is consistent with the hypothesis that the survival probability of a rock particle depends on the particle volume and on prevailing confining stress and follows a Weibull distribution. The important point, however, is that the previous findings justify the existence of a scale effect for the rockfill behavior in the sense that a homothetic change of the particle size distribution of a given “sample” will introduce changes in the observed stress-strain behavior. This effect was shown by Marachi et al. (1969), when isotropic compression tests on samples of crushed claystone from Pyramid dam were tested (Fig. 2). Samples had geometrically similar grain size distributions and were tested in cells of increasing diameter in order to maintain a constant ratio between cell diameter and maximum grain size. Scale effects are expected to be large when particle breakage dominates the deformation mechanism. Other mechanisms are also present, however (frictional sliding and particle rotation) and it is not clear how to design laboratory experiments to account for scale effects.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
An alternative is developed in this paper. The idea is to investigate scale effects through their effect on constitutive parameters. If a given constitutive model is consistent with observed behavior (which includes breakage phenomena) then all the information provided by the model is included in the set of constitutive parameters. If scaled samples of rockfill are tested and their model parameters are identified, then scale effects will be translated in variations of (some) constitutive parameters with the grain size distribution (which may, in turn, be characterized by one or more “free” parameters). Hopefully, scale effects will be reflected in a limited number of constitutive parameters: those properly describing breakage phenomena. Perhaps, other constitutive parameters will remain stable as the grain size distribution changes. It will be interpreted that these parameters describe other fundamental mechanisms (such as friction) not affected by scale effects.
Figure 2: Isotropic compression tests on geometrically similar specimens of rockfill (φspecimen/dmax = 6). Pyramid dam claystone : cu = 7.7; angular particles, basalt (from Marachi et al. 1969). This paper reports the results of several oedometer compressibility tests of samples of scaled grain size distribution. They were performed in a 300 mm diameter oedometer cell designed for relative humidity control. Test results were interpreted within the framework of the elastoplastic model developed by Oldecop and Alonso (2000). Material parameters were derived for the different grain size distributions. Changes in material parameters were then related to the initial sample grain size distribution. This comparison has provided a new insight into scale effects. The applied methodology may also be used to approximate the behavior of large-scale rockfill on the basis of laboratory tests conducted on scaled samples. A CONSTITUTIVE MODEL FOR ROCKFILL COMPRESSIBILITY The fundamentals of subcritical crack propagation in the brittle materials offers a convenient framework to analyze the fracture process of rock particles. Oldecop and Alonso (2000) demonstrated that time-dependant deformations of rockfill may be conveniently interpreted from this perspective. Subcritical crack propagation leads to time dependant deformation mechanisms (TDM) of the rockfill. TDM is activated beyond a threshold confining stress (σy, in a one-dimensional compression model). σy may be called “clastic yield stress” because it marks the beginning of particle breakage. Delayed deformation mechanisms are strongly controlled by the action of water. In addition, instantaneous deformation mechanisms (IDM) were identified. They are present at any stress level and they seem to be independent of the rockfill water content. A compressibility model was formulated on the basis of these basic ideas. An incremental compression line was defined as follows: for σ ≤ σy only IDM is present. This deformation is characterized by a linear law: dε = dε i = λi dσ
[3]
For σ > σy, both IDM and TDM contribute to rockfill deformation. A simple linear formulation leads to:
[
]
dε = dε i + dε d = λi + λd (w) dσ
[4]
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
where dεi, dεd are the instantaneous and delayed components of total deformation, λi is the (constant) compressibility index associated with IDM and λd is the delayed compressibility index when σ > σy. λd is a compressibility index associated with the breakage of particles induced by crack propagation phenomena. λd was shown to depend o the water content of the rock particles or, alternatively, of the total suction ψ (or relative humidity) prevailing at the large rockfill voids. It may be shown that collapse behavior is included in this formulation. Collapse is interpreted as an acceleration of deformations attributed to the increased velocity of particle fracture propagation in the presence of water, which acts as a corrosive agent. For unloading-reloading an elastic behavior was assumed: dε e = κ dσ
[5]
where κ is the elastic compressibility coefficient, which was assumed independent of water action. Tests reported in Oldecop and Alonso (2000) in a Cambrian quartzitic shale, have shown that changes in water content (or suction) lead to (small) expansion or shrinkage deformations. To account for this effect an index, κw or alternatively, κψ, if water action is modeled through suction ψ were introduced as follows: dw w d ψ dε ψ = −κψ ψ + patm dε w = −κ w
[6] [7]
where patm is the atmospheric pressure. In collapse tests reported in Oldecop and Alonso (2000) a linear relationship between collapse deformations and the logarithm of water content changes was observed. This observation provided a clue on the type of function λd(w), which relates the delayed compressibility index λd and the water content, w. It was proposed: wo w
[8]
ψ + patm patm
[9]
λd (w) = λdo − α w ln or, if the model is formulated in terms of suction:
λd (ψ ) = λdo − αψ ln
αw (or αψ) and λdo are model parameters. λdo is interpreted as the maximum (delayed) compressibility index of a given rockfill. It corresponds to a saturated state (w ≥ wo). λd is therefore bounded as follows: 0 < λd < λdo
[10]
The variation of the delayed compressibility index λd(w) with humidity (or suction) explains also the collapse phenomena observed in tests when a rockfill sample is wetted under constant load. If a collapse test is run by controlling the change of relative humidity, at a given confining stress, σo, the collapse strain may be plotted against the change in water content (or suction). The slope of this relationship (χw or χψ) is related to other material parameters (Oldecop and Alonso, 2000):
(
)
[11]
(
)
[12]
χw = αw σ o −σ y −κ w or, in terms of suction variables:
χψ = αψ σ o − σ y − κψ
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
The ideas outlined lead to the formulation of an elastoplastic constitutive model. Yield surfaces and associated relations were derived in Oldecop and Alonso (2000). The derivation of material parameters is a simple task and has been summarized in Figure 3. Consider a compression test on an initially very dry sample. The stress is increased until a reference stress σo, is reached. The sample is then wetted at constant σo. Once saturated the sample is further loaded and finally unloaded. In a second test the specimen is first wetted and then loaded and unloaded in a classical way. Both tests have been schematically shown in Figure 3a. Also indicated in the figure are the slopes of the virgin compression indexes of the dry and saturated stages (λi and λi + λdo respectively). The slope of the unloading path provides κ and the change in compressibility observed in the saturated sample provides the threshold value σy. The remaining model parameters (κw and αw; alternatively κψ and αψ) may be determined if the strains changes upon wetting at constant stress are interpreted. The swelling/shrinkage index κw (or κψ) is directly found when the swelling strain at low confining stress is plotted against the change in water content (or ψ) (Fig. 3b). Finally, the value of αw (or αψ) may be derived from the plot relating collapse strains with the change in water content (or ψ). The slope of this plot (χw or χψ) is related to (αw or αψ) through equations (11 and 12). The set of tests reported later were interpreted in this way. MATERIAL AND TESTING PROCEDURE The material tested was taken from an outcrop of quartzitic shale of Cambrian origin. This material will be used in the construction of an earth and rockfill dam (Lechago dam in the province of Teruel, Spain). The rock in situ is highly fractured (mean RQD : 42). Unconfined compression strengths determined on intact rock cores varied between 14.2 and 31.9 MPa. The rock porosity determined on cores varied between 6.3% and 11.8%. The rock durability as determined through the Slake durability test was high (99%). Los Angeles abrasion test gave a value LA = 24.8%. The following minerals were identified on X-ray diffraction tests on rock powder: quartz, muscovite, chlorite and dolomite.
Figure 3: Identification of model parameters: a) Compression curves for dry and wet condition; b) Variation of swelling strain with increasing water content; c) Collapse strains as a function of increasing water content. Rock fragments were taken from a trial excavation pit. They were further crushed in the laboratory in order to prepare samples for the testing program. The coarser sample was defined having in mind the diameter of the oedometer cell (300 mm). A maximum grain size of 40 mm was accepted. Four grain size distributions were then defined as shown in Figure 4. The coarser material may be described as a rather uniform gravel with a mean diameter D50 = 24 mm. The finer material is a coarse sand (D50 = 1 mm). The scaling adopted maintains a parallelism of the grain size distributions. The coefficient of uniformity increased from 2.9 for the coarser gravel sample to 7 for the coarse sand.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Figure 4 : Particle size distribution of tested samples . The raw material was crushed and then sieved. The nominal sizes of the screens used are 20, 15.08, 12.5, 10, 7, 5, 2.5, 0.63, 0.32 and 0.08 mm. Once the samples were mixed to the desired grain size distribution they were stored for at least a week at the ambient relative humidity (40-50%) and temperature (21º ± 1º) of the laboratory in Barcelona at the time of testing. The initial state of all tested samples corresponds therefore to a rather dry state. The material was compacted inside the oedometer ring, in four layers. A Marshall type hammer, which delivers 20.33 Joules per blow was used. A total number of 101 blows/layer, uniformly distributed were applied. The compaction energy supplied corresponds to the compaction energy of the Standard Proctor test. The blow was applied through a steel footing to reduce initial breakage. Table 1 provides the characteristics of the samples/tests performed. It may be seen that the compaction energy remained close to the Standard Proctor nominal value (584.3 Joules/l). The change in grain size distribution from the coarser to the finer material resulted in a progressive increase of the total specific weight. The variation of void ratio of compacted samples as the mean grain diameter decreases is shown in Figure 5. Figure 6 shows photographs of the compacted samples inside the oedometer ring for three of the grain sizes tested (D50 = 24, 10 and 1 mm).
a)
b)
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
c) Figure 5 : Compacted samples inside the 300 mm diameter oedometer ring: a) D50 = 24.6 mm; b) D50 = 10 mm; c) D50 = 1mm. The suction controlled oedometer cell used in all tests is shown in Figures 7 and 8. The cell is designed for samples 300 mm in diameter and 200 mm in height. Upper and lower platens allow a fast flooding of the sample if required. Load is applied to the upper platen by means of a Bellofram-type of membrane, which receives the air pressure. Radial strains are measured in the cylindrical confining ring in specially designed measuring points. They are transformed into radial stresses by means of the appropriate calibration. The vertical load can also be measured by means of three loading cells, which support the lower platen. In this way the total lateral friction can be estimated. The lateral friction was minimized, however, by means of an internal double polyethylene lining, appropriately lubricated. Air with a controlled relative humidity can be circulated through the sample. Relative humidity is measured by means of hygrometer at the lower outlet. The cell is connected to a computer controlled system for load application and instrument recording. Table 1. Characteristics of the tests performed. D50 H(1) E(4) γo(2) eo(3) Test HRo(5) wo(6) 3 (mm) (mm) (kN/m ) (joule/l) L1 24.2 195.6 17.35 0.587 594 50(*) 0.75(*) L2 24.2 195.6 17.40 0.583 594 30 0.61 E1 10 205.5 18.25 0.509 565 40(*) 0.822 E2 10 208.0 18.20 0.528 559 39.84 0.810 E3 5 191.0 18.37 0.499 608 30.50 0.845 E4 5 187.5 18.71 0.472 620 43.20 0.707 E5 10 200.0 18.75 0.469 581 41.75 0.673 E6 1 197.5 19.47 0.414 588 49.90 1.718 E7 1 197.0 19.52 0.411 590 46.90 0.777 (1) H : Sample thickness; (2) γ : Initial specific weight; (3) Initial void ratio; (4) E : Compaction energy; (5) HRo : Initial relative humidity; (6) wo : Initial water content; (*) Estimated value
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Figure 6 : Variation of void ratio of compacted samples with the grain size distribution, represented by D50. TEST RESULTS Samples were subjected to relatively simple stress paths (Fig. 9). In some of the tests (L2, E3 and E3) the sample was first wetted under a small vertical confining stress (0.01 MPa) and then loaded under a high relative humidity. In other cases (L1, E4, E5, E6, E7) the sample was loaded at the constant initial water content and wetted at a maintained vertical stress (0.6 MPa). Once wetted, loading continued under saturated conditions. The initial suction (100 MPa) was similar in all samples tested except in sample E6 (53 MPa). In test E1 the sample was never wetted. Once the maximum vertical stress was reached, samples were unloaded in steps. The maximum applied vertical stress was 1 MPa in most cases. In two tests (E1 and E2) the maximum stress was increased (see Figs. 15a, b).
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Figure 7 : Suction-controlled oedometer cell used in the tests reported in the paper.
Figure 8 : Assembled testing cell.
Figure 9 : Stress paths applied to samples. Stress paths for tests E1 and E2 reached vertical stresses in excess of 1 MPa.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Figure 10 : Measured compressibility curves in terms of void ratio. Vertical stress in natural scale.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Figure 11 : Compressibility curves in terms of vertical strain. Vertical stress in natural scale. Compression curves in terms of void ratio for all samples tested are given in Figure 10. The differences in initial void ratio may be in part attributed to the increase in uniformity coefficient, Cu, as the main grain size decreases. Loading and unloading compression lines tend to be linear if stress is plotted in natural scale. Equivalent information is given in terms of measured strains in Figures 11 (stress in natural scale) and 12 (stress in logarithmic scale).
Figure 12 : Compressibility curves in terms of vertical strain. Vertical stress in logarithmic scale. Grain size distribution was determined at the end of the tests. The change in sizes is shown in Figure 12. The curve represented for “after test” conditions corresponds to the maximum breakage recorded for each particular grain size distribution. As it could be expected, the amount of breakage decreases as the initial grain size decreases. Note that the change is grain size plotted in Figure 13 includes compaction effects as well as subsequent changes induced by the test itself.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
The amount of breakage was quantified by means of the parameters defined by Hardin (1985). In Figure 14, the Total Breakage parameter Bt is plotted against the Potential Breakage Bp. The relative breakage (Br = Bt / Bp) is essentially constant for all the tests performed (Br = 6.58 %).
Figure 13 : Measured grain size distribution curves before and after testing.
Figure 14 : Breakage parameters Bt, Bp and Br, as defined by Hardin (1987), of tested samples. GRAIN SIZE EFFECTS ON CONSTITUTIVE PARAMETERS Model parameters were determined as explained before for each of the tests performed. Comparisons between model performance and actual test results are given in Figures 15, 16 and 17, for average grain sizes of 10, 5 and 1 mm respectively.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
A single set of constitutive parameters was obtained for each grain size distribution. They are given in Table 2. Parameters for the material with D50 = 24.2 mm were reported and analyzed in Oldecop and Alonso (2000).
Figure 15 : Tests on samples having D50 = 10 mm. Open circle indicates the beginning of wetting. Comparison between measured results and model predictions.
Figure 16 : Tests on samples having D50 = 5 mm. Open circle indicates the beginning of wetting. Comparison between measured results and model predictions.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Figure 17: Tests on samples having D50 = 1 mm. Open circle indicates the beginning of wetting. Comparison between measured results and model predictions. Parameter λi (× 10-2)
λdo (× 10-2) κ (× 10-3) κw (× 10-3) κψ (× 10-4) αw (× 10-2) (MPa-1) αψ (× 10-3) (MPa-1) σy (MPa)
D50 = 24.2 mm 2.31 5.03 2.67 2.62 7.86 1.39 4.2 0.29
Table 2. Constitutive parameters. D50 = 10 mm D50 = 5 mm 2.15 2.23 3.64 3.30 2.898 2.77 5.96 0.70 2.1 -
3.44 2.10 4.92 0.28 0.85 -
D50 = 1 mm 2.30 3.13 3.00 1.40 5.79 0.13 0.40 -
Some comments can be made. The value of the yield stress, σy, is not well defined in the series of tests reported here. For instance, in Figures 15b and 16a, the saturated compression lines for D50 = 10 mm (test E2) and D50 = 5 mm (test E3) are plotted and compared with model predictions for two values of the clastic yield stress. Experimental points are well reproduced with σy ≅ 0. However, in tests reported in Oldecop and Alonso (2000) on the same material, but having D50 = 24.2 mm, a well defined value σy = 0.29 MPa was identified. The determination of σy for the materials analyzed in this paper probably requires additional testing and a reduced stress increment during the early stages of loading. In test E2 (Fig. 15b), which was loaded to a maximum vertical stress of 2.6 MPa, a change in deformation mechanism is apparent beyond σv ≅ 1.4 MPa. Following McDowell and Bolton (1998) the sample has entered a “clastic hardening” regime characterized by an upward curvature of the stress (natural scale)-strain relationship. The elastoplastic model outlined before is valid for stresses below the onset of clastic hardening. The agreement between model predictions and experimental results is good for the set of tests reproduced in Figures 15, 16 and 17. The instantaneous and delayed compression indexes λi and λdo have been plotted in terms of D50 in Figures 18 and 19. λi is seen to be independent of the grain size distribution. However, the delayed component, λdo , which is strongly linked with the mechanisms of particle breakage, decreases in a significant way as D50 decreases. A linear relationship between λdo and the logarithm of D50 seem to hold. The elastic compression index is plotted in Figure 20 against D50. Elastic deformations are in any case very low. A very limited variation with D50 is found. The values of the swelling index κw (or κψ) are also very small and scarcely relevant for the purposes of constitutive modeling of rockfill. Figure 21 indicates that κw (or κψ) decrease substantially as D50 reduces. Finally, the variation of coefficient αw (or αψ) with D50 is plotted in Figure 22. αw describes the rate of change of the delayed compressibility index as the humidity changes. It is again a parameter linked to the mechanisms of particle breakage. Figure 22 indicates that this rate index decreases as the particle size decreases. A similar trend, not so well marked, is observed in αψ.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Figure 18 : Variation of virgin instantaneous compressibility index, λi, with sample gradation, as given by D50.
Figure 19 : Variation of maximum delayed compressibility index, λdo , with sample gradation, as given by D50.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Figure 20 : Variation of elastic compressibility index, κ, with sample gradation, as given by D50.
Figure 21 : Variation of shrinkage-swelling coefficients, κw and κψ, with sample gradation, as given by D50. Results on the measured Ko values at different stages of the loading-wetting-unloading stress paths applied to the samples are shown in Figure 23 as a function of D50. Three distinct values of Ko are shown. They correspond to the following conditions: virgin loading under dry conditions; immediately after wetting induced collapse and at the final unloading stage. Virgin dry loading is characterized by a low Ko value (0.25-0.2). Collapse (under 0.6 MPa) causes an increase in Ko to values close to 0.33. Upon unloading, Ko increases to values in the range 0.9-1.2. Sample gradation seems to have a limited effect on Ko. A slight decrease in Ko with decreasing D50 may be observed in Figure 22. SUMMARY AND CONCLUSIONS A fundamental mechanism of rockfill deformation is the breakage of particles under load. Particle breakage is also controlled by the prevailing relative humidity at the rockfill pores. Tests on rock particles show also that the stress intensity leading to particle breakage depends on the size of the particle. This result introduces a scale effect on the constitutive behavior of rockfill and, in general, in granular media whose particles break under the imposed stress path. In rockfill, breakage is significant at relatively low stress levels, well within the stresses expected in engineering structures. Scale effects are difficult to handle in testing because the usual sizes of blocks imply very large testing cells. On the other hand particle breakage is not the sole mechanism leading to rockfill deformation (sliding and rotation of particles occur) and therefore it is not expected that scaling laws applicable to individual particles will be directly applicable to the granular mix.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Figure 22 : Variation of rate parameter αw (or αψ) with sample gradation, as given by D50.
Figure 23 : Recorded values at rest coefficient, K0, plotted in terms of sample gradation, as given by D50. The approach followed in this work to identify scale effects is to investigate the effects of grading into the constitutive parameters of an elastoplastic constitutive model for rockfill compressibility. The model uses, as a physical framework, some results of subcritical crack propagation phenomena. However, it was cast in terms of hardening plasticity on the basis of experimental results. Several tests on gravels of parallel gradation were performed in a suction control oedometer cell 300 mm in diameter. Material parameters were derived for the grain size distributions tested. The variation of these parameters (which have a physical interpretation), with a representative parameter of the grain size distribution (the mean equivalent diameter of particle, D50) was investigated. It was found that the delayed compressibility index ( λdo ) and the parameter governing the rate of change of rockfill compressibility index with relative humidity (αw) decreased as the mean grain size decreased. This is consistent with the mechanisms of rockfill compressibility implied by the model. The remaining compressibility indexes defined in the model are essentially independent of gradation. Inconclusive results were found for one of the constitutive parameters: the clastic yield stress σy, which marks the beginning of clastic yielding (particle crushing).
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Measured Ko values were quite insensitive to rockfill gradation. However, rockfill collapse leads to an increase in horizontal stresses (Ko increased from a value Ko = 0.2 in virgin loading to Ko = 0.33 after flooding). The analysis described provides a procedure to estimate rockfill parameters at large scale. In fact, the extrapolation of the relationships observed between model parameters and average grain size provides such an estimate. Ideally, however, this technique should be validated through field records of rockfill behavior. REFERENCES 1.
Hardin, B.O., 1985, Crushing of soil particles: J. Geotech. Engng., 11, 10, 1177-1192.
2.
Lee, D.M., 1992, The angles of friction of granular fills: Ph.D. Dissertation, University of Cambridge.
3.
Marachi, N.D., Chan, C.K., Seed, H.B. and Duncan, J.M., 1969, Strength and deformation characteristics of rockfill materials, report No. TE-69-5, Department of Civil Engineering, University of California.
4.
Marsal, R.J., 1973, Mechanical properties of rockfill: Embankment Dam Engng. Casagrande Volume.
5.
McDowell, G.R. and Bolton, M.D., 1998, On the micromechanics of crushable aggregates: Géotechnique, 48, 5, 667-679.
6.
Nobari, E.S. and Duncan, J.M., 1972, Effect of reservoir filling on stresses and movements in earth and rockfill dams, report No. TE-72-1, Department of Civil Engineering, University of California.
7.
Oldecop, L.A. and Alonso, E.E., 2000, A model for rockfill compressibility: Géotechnique, 51, 2, 127-139.
8.
Terzaghi, K., 1960, Discussion on salt springs and lower bear river dams: Trans. ASCE, 125, 2, 139-148.
Geotechnical Materials: Measurement and Analysis : R.J. Krizek Commemorative Symposium: 3 Aug 2002
VIBRATIONS CAUSED BY CHISELLING IN DEEP FOUNDATIONS CONSTRUCTION Hervé Duplaine1 and Christophe Bouniol1 1,2
Balineau Sa 18 Ave Gustave Eiffel, 33608 Pessac Cedex, France 1
[email protected],
[email protected]
ABSTRACT Lots of techniques used in civil engineering and more particularly in deep foundations construction generate vibrations : pile driving, dynamic compaction, use of explosives for excavation or compaction, chiselling during pile or diaphragm wall excavation. For some techniques like use of explosives or dynamic compaction for example, levels of vibrations can be predicted because lots of data have been collected and analyzed over the years. It is therefore possible for Owners and Engineers to decide whether the considered technique is acceptable or not for their particular application in terms of nuisances and potential damages to existing structures. Pile drilling or diaphragm wall excavation often requires chiselling in order to go through blocks or hard layers but almost no vibrations data are available for this technique. A diaphragm wall site near Bordeaux has been closely monitored in terms of peak particle velocity measurement during chiselling operations. The set of data obtained follows the general law for peak particle velocity dissipation as a function of the distance from the source of vibrations INTRODUCTION Diaphragm wall or bored pile construction often requires the use of a chisel to go through blocks, rock beds or for socket excavation in the bedrock. A chisel can be described as drilling tool which weight can reach up to 12 tons ( Figure N°1 ) .It is made of steel and various shapes can be observed but its base is always composed of blades . A succession of free falls from a height varying between 1 and 3 meters in the excavation will allow these blades to break the hard materials into pieces and then remove it with more classical tools such as buckets or augers.
Department of Civil and Environmental Engineering Infrastructure Technology Institute Northwestern University, Evanston, Illinois, 60208
Geotechnical Materials: Measurement and Analysis : R.J. Krizek Commemorative Symposium: 3 Aug 2002
Figure N°1 : Chisel These free falls end up in shocks generating vibrations which can be a source of nuisance for the people living nearby or damages to the neighbouring structures. Vibrations can be characterized by three inter-related parameters : amplitude, peak particle velocity and acceleration. Peak particle velocity is the parameter which is commonly used in the construction field. There are almost no national or international standards or codes for transient vibrations in the construction industry. Existing rules are made for steady-state vibrations and cannot be applied to transient vibrations . However, it is commonly admitted that a safe limit in terms of peak particle velocity is about 10 mm/s for resistant structures and 5 mm/s for sensitive structure. Vibrations generated by the use of chisels in deep foundation construction have almost never been measured and the few existing results have not been neither analyzed nor communicated. This lack of data has often let to simply prohibit the use of chisels for a large number of urban sites. SITE PRESENTATION Recalibration of the river bed plus rock slope protection combined with diaphragm walls in the most constructed part of the project will help to prevent the Eau Bourde river in Villenave d’Ornon ( near Bordeaux ) from regularly flooding the constructed areas nearby. The diaphragm walls are 500 mm thick and about 10 m deep. The different layers encountered are described in the geological profile given in Figure N°2.
Figure N°2 Géological Profile Going through the limestone often required the use of a 5.5 tons steel chisel dropped from a 2 to 3 meters height.
Geotechnical Materials: Measurement and Analysis : R.J. Krizek Commemorative Symposium: 3 Aug 2002
Samples from one of the boreholes of the preliminary site investigation gives the following characteristics for the limestone which is going to be excavated with the use of a chisel : Apparent unit weight : 22 à 22.8 KN/m3 Unconfined compressive strength : 7.9 à 18.5 MPa Laboratory tests have been carried out on samples which have been extracted after diaphragm construction from depths where chiselling had occurred and have led to higher values Apparent unit weight : 24.8 KN/m3 Unconfined compressive strength : 32.2 MPa The contract had allowed for some vibration measurements if and when necessary. This specification implied the continuous presence of a vibration analyzer on site. Keeping the vibration analyzer on site proved very helpful in order to collect the largest possible set of data over the duration of the works with the hope of being able to establish a reliable dissipation law for the vibrations induced by the chiselling operations just like it had been done for dynamic compaction (Liausu, 1981) METHODOLOGY USED FOR THE VIBRATION MEASURES The vibration analyzer was a 2513-WH2328 model from BRUEL & KJAER with a 4391 accelerometer. It can measure acceleration or peak particle velocities for frequencies between 4 à 1000 Hz ( after a filter modification). Peak values along the vertical axis have been systematically recorded. The analyzer looks for this value over a one second time interval , which is more than necessary since the vibration level goes back to zero after 100 to 300 milliseconds. Vibrations measurements were sometimes taken on houses but most of the measures were taken on the guidewalls of the diaphragm wall. The guide walls are reinforced concrete ( see figure N°3 ) structures which are used to set the position of the diaphragm wall and guide the rope suspended grab during the first meters of excavation. They are 0.7 to 1 m high and 0.3 m thick . They constitute a continuous structure over a length of more than 200 meters in the present case, which allows them to be considered as a representative structure.
Figure N°3 Guide Walls section
Geotechnical Materials: Measurement and Analysis : R.J. Krizek Commemorative Symposium: 3 Aug 2002
The interesting side of vibration measurements over the guide walls is that it is possible for the same sequence of chiselling to get very quickly measures at various distances from the source on the same structure which is not the case for residential structures. Figure N°4 shows the arrangement for vibrations measures.
Figure N°4 Bruel &Kjaer Equipment Some measures were taken by an external laboratory, EXAM BTP , with a LE3D equipment in order to check the cohérence of the values obtained with the two equipments and also to perform a frequency analysis on the first values. RESULTS The equipment used by EXAM BTP was able to perform frequency analysis. The dominant frequency was found to be between 11,5 and 14 Hz with one value reaching 22 Hz. The three components of the peak particle velocity show values which are equivalent in the three directions, the vertical component being slightly greater than the two others. The BRUEL & KJAER equipment was used to take more than 380 peak particle velocity values with distances from the source ranging between 5 and 50 meters. Some of the values measured at a distance of 25 m were taken on residential structures. All other values were taken on the diaphragm wall guide-walls. All the couple of values (peak particle velocity , distance) have been plotted on the same graph ( Figure N° 5 ). Limestone heterogeneity, small variations in the fall height and the diminution of the level of vibrations associated with the breakage of the rock under the action of the chisel lead naturally to scattered values. It is interesting to observe that the two lines for minima and maxima are almost parallel. Representative values are the maxima observed for each one of the considered distances.
Geotechnical Materials: Measurement and Analysis : R.J. Krizek Commemorative Symposium: 3 Aug 2002
peak particle velocity (mm/s)
10
1
0,1
0,01 1
10
100
distance ( m )
Figure N°5 Attenuation law. The attenuation law for the peak particle velocity as a function of the distance from the source for a 5.5 tons chisel falling from a 2 to 3 m height on limestone is :
V = 10 × d −1.5 CONCLUSION The attenuation law for chisel induced vibrations obeys the general relationship between peak particle velocity and distance ( Wiss, 1981 ). The dominant frequency for vibrations generated by the use of a chisel ranges between 11 and 14 Hz and the peak particle velocity is less than 4 mm/s at a distance of 5 meters. This level of vibration cannot be detrimental to adjacent structures but is perceptible by human beings and therefore can be considered as a nuisance for the neighbours of the site. The attenuation law presented here has been established for a specific chisel and rock combination. It should not be extrapolated blindly for other sites where heavier chisels and/or stronger rock conditions might be encountered. It gives anyway an order of magnitude and is the starting point for data collection which hopefully will lead in the future to a better assessment of chisel induced vibrations.
Geotechnical Materials: Measurement and Analysis : R.J. Krizek Commemorative Symposium: 3 Aug 2002
REFERENCES 1.
LIAUSU, P, 1981. Vibrations engendrées par le Compactage Dynamique. Revue Française de Géotechnique , N°14 bis
2.
WISS, John F., 1981, Construction Vibrations : State-of-the-Art, Journal of the Geotechnical Engineering Division, Vol 107, N° GT2, February 1981, p. 3
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
CHEMICAL TRANSPORT ISSUES IN MODERN LANDFILL LINERS Tuncer B. Edil Civil & Environmental Engineering and Geological Engineering, University of Wisconsin-Madison 1415 Engineering Drive, Madison, WI 53706
[email protected]
ABSTRACT Leachates from municipal solid waste (MSW) and hazardous waste landfills contain a wide range of volatile organic compounds (VOCs) in addition to inorganic compounds. VOCs have been shown to migrate and contaminate the surrounding environment and impair the use of groundwater. Therefore, the effectiveness of modern landfill liner systems to minimize migration of VOCs is of concern. Most modern landfills employ a composite liner consisting of a geomembrane overlying a compacted clay liner or a geosynthetic clay liner. The geomembrane is often believed to be the primary barrier to contaminant transport. However, for VOCs, the clay component usually controls the rate of transport since VOCs are shown to diffuse through geomembrane at appreciable rates. Therefore, transport of VOCs through clay liners and modeling of transport through composite liners merit scrutiny. This paper presents a review of recent research by the author and his associates on these two topics. A systematic and comprehensive approach to determine mass transport parameters for transport of VOCs through compacted clay liners and geomembranes has enabled to develop realistic models to predict mass flux of VOCs through modern composite liners and have a quantitative basis to evaluate the equivalency of different composite liners. INTRODUCTION The United States produces over 200 million tons of solid waste per year and 75% of the solid waste is landfilled. Landfilling has gone through significant development in the last several decades. Solid waste landfills have evolved from uncontrolled city dumps to highly engineered structures. The most important requirement of a landfill is that it does not pollute or degrade its environment. In this effort the liner system is the most significant component as it is intended to provide a barrier against advective (hydraulic) and diffusive (chemical) transport of leachate solutes. Until about 1982, the main liner material used in landfills was compacted clay. Federal Hazardous and Solid Waste Amendments of 1984 required that all landfills have composite and/or double liners and leachate collection and removal systems. The provision for alternative liner designs in Subtitle D of the Resource Conservation and Recovery Act (RCRA) and the development of geosynthetic clay liners (GCLs) for use in composite liner systems have resulted in a need for a rational method for comparing alternative landfill liner systems. The current approach is to assume that an alternative liner is equivalent to the prescriptive liner if it discharges less liquid (e.g., Richardson 1997). However, Park and Nibras (1993) report that volatile organic compounds (VOCs) can diffuse through geomembranes in significant quantities in less than one day. Hence, just as diffusion can be a significant mode of contaminant transport in soil liners (e.g. Shackelford 1990, Rowe 1987, and Crooks and Quigley 1984), diffusion can also be a significant mode of transport in composite liners. Consequently, comparing composite liner systems based on leakage rate may not be sufficient. Early concern on clay liners focused on their hydraulic characteristics and the ability to limit advective transport. In early 1980s there was significant concern regarding the interaction of organic chemicals with clay liners and the potential for increased hydraulic conductivity. It became clear that the typical municipal solid waste (MSW) leachate contains organic Department of Civil and Environmental Engineering Infrastructure Technology Institute Northwestern University, Evanston, Illinois, 60208
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
chemicals but at relatively low concentrations to impact the hydraulic conductivity of clay significantly. As the understanding of compaction parameters and environmental factors on hydraulic conductivity of clay liners improved, construction of uniformly high quality clay liners with a hydraulic conductivity of 1x10-7 cm/s or less has become a common practice. Introduction of GCLs made even lower hydraulic conductivity and high quality a reality. With the requirement of composite liners involving a geomembrane and a compacted clay or GCL along with leachate collection, the advective transport was significantly reduced. It also became apparent that diffusive transport in such systems becomes significant. Further studies showed that the transport of inorganic chemicals, though they exhibit a higher concentration in the MSW leachate than the organic compounds, is less critical than the transport of VOCs. Furthermore, VOCs are toxic at much lower concentration than many inorganic chemicals. Therefore, liner systems for MSW landfills should usually be compared based on transport of VOCs (Foose et al. 2002). More importantly, evaluating the performance of composite liners based solely on leakage rate does not include the major pathway for transport of VOCs, which is diffusion through the intact composite liner. Failure to include an analysis of diffusion of VOCs through intact liners can lead to incorrect conclusions regarding equivalency. Leachates from municipal solid waste (MSW) and hazardous waste landfills contain a wide range of volatile organic compounds (VOCs) in addition to inorganic compounds (Friedman 1988; Gibbons et al. 1992; Krug and Ham 1995). VOCs have been shown to migrate and contaminate the surrounding environment and impair the use of groundwater (Plumb and Pitchford 1985; Nelson and Book 1986; Friedman 1988; Battista and Connelly 1994; Rügge et al. 1995). Therefore, the effectiveness of modern landfill liner systems to minimize migration of VOCs is of concern. Most modern landfills employ a composite liner consisting of a geomembrane overlying a compacted clay liner or a geosynthetic clay liner. The geomembrane is often believed to be the primary barrier to contaminant transport. However, for VOCs, the clay component usually controls the rate of transport since VOCs diffuse through geomembrane at appreciable rates (Park and Nibras 1993; Sakti 1993). For example, a recent study by Foose (1997) has shown analytically that breakthrough and mass flux of VOCs from composite liners under realistic field conditions depends on the characteristics (e.g. thickness, sorptive capacity) of the clay liner component. Therefore, transport of VOCs through clay liners and modeling of transport through composite liners merit scrutiny. This paper will present a review of recent research by the author and his associates on these two topics. TRANSPORT OF VOCS THROUGH COMPACTED CLAY To evaluate VOC transport through compacted clay liners, mass transport parameters such as retardation factor, hydrodynamic dispersion coefficient, degradation rate, and seepage velocity must be known. Limited laboratory tests have been conducted to characterize and quantify the mass transport parameters of VOCs in natural clay deposits (Barone et al. 1992; Myrand et al. 1992). Previous laboratory tests have been performed using small-scale columns with specimen lengths of 25 mm (Barone et al.) to 100 mm (Myrand et al.). These tests were short-term, covering a period of less than a month. These small-scale column tests may not represent the conditions in the field and preclude an understanding of the long-term behavior of VOCs in clay liners. A series of small-scale column tests and large-scale tank tests were conducted at the University of Wisconsin-Madison to estimate the mass transport parameters of VOCs under hydraulic gradients like those occur in the field and in long-term conditions. A typical landfill liner soil that classifies as low-plasticity clay (CL) according to the Unified Soil Classification System was used in investigating the transport of VOCs (Heim 1992, Wambold 1993). It had 494 mg/kg of exchangeable cations (i.e., Ca+2, Mg=2, K+, and Na+), and 1.06% organic carbon mass fraction. VOCs that were investigated for transport included chloroform (CF), ethylbenzene (EB), methylene chloride (MC), toluene (TOL), 1,1,1-trichloroethane (1,1,1-TCA), trichloroethylene (TCE), and m-xylene (m-XYL) and they were selected for testing based on the availability of laboratory analytical methods, relatively high detection frequency in contaminated environments, e.g., landfill leachate, as well as their reasonably broad range of solubilities and molecular weights (Kim et al. 2001). Two types of tests were conducted: batch isotherm tests and column/tank tests. Batch isotherm test is a standard test used to determine the sorptive capacity of a soil for a chemical in a contact solution. It is conducted to estimate the partition coefficients for the VOCs from water to soil. In this test, a ground sample of soil is mixed with solutions of the chemical of interest in water in a vial and vigorously agitated. Solution concentrations are measured initially and periodically thereafter using a gas chromatograph. Measured solution concentrations typically stabilize in seven days indicating that equilibrium is reached. As a VOC gets sorbed on the soil, its concentration in the solution decreases. The ratio of the equilibrium solid-
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
phase concentration of mass sorbed on the soil divided by the solution-phase concentration is expressed as partition coefficient. It is an indicator of the sorptive capacity of the soil with respect to a given chemical. The conditions of the soil and the test procedure do not replicate the conditions occurring in the field during exfiltration of leachate through a clay liner. Therefore, the findings based on batch isotherms have to be corroborated in column tests where permeation through a compacted column of soil is simulated. In the past, column tests, in general, were not scrutinized to meet this fundamental geotechnical testing principle. Therefore, experimental apparatus and relevant procedures were developed specifically to test VOCs in an environment simulating the conditions in the field as closely as possible and overcome numerous difficulties of testing with such chemicals (Edil et al. 1994, 1995; Kim et al. 1997 and 2001). Schematics of the tank and the column apparatus are shown in Figure 1. The column apparatus is a smaller version of the tank. The influent and the effluent were supplied and collected by means of reservoirs above and below the specimen, respectively. This means of collection permits the water to flow evenly across the entire soil layer. All parts used were made of stainless steel and brass. Teflon® bags were connected to the influent and effluent reservoirs to control head and rate of flow with minimal loss of VOCs to atmosphere. All tubing was made of Teflon®
610 mm I.D.
Influent Bag
152 mm I.D.
Influent Reservoir
Glass Fiber Filter Stainless Steel Screen
Compacted Clay Layer
Influent Reservoir
Glass Fiber Filter Stainless Steel Screen
Sampling Ports
Sampling Ports Compacted Clay Layer
Effluent Reservoir
Effluent Reservoir Effluent Bag
Threaded Tie-rod
200 mm
Driving Head
Nitrile Gasket
914 mm
Driving Head (900 mm)
.
Influent Bag
Effluent Bag
Sand Foundation
Tank
Column
Figure. 1. Design of the Tank and the Column Used in the Tests (Kim et al. 2001) In a one-dimensional column test, mass transport of a non-decaying solute through a porous medium can be expressed as follows using the retardation factor, Rf and assuming a first order degradation reaction for VOCS (Hashimoto et al. 1964; Freeze and Cherry 1979):
∂C(z,t) ∂t
2
∂ C(z,t) vz ∂C(z,t) kd = Dh ⋅ - ⋅ - ⋅ C(z,t) Rf 2 Rf Rf ∂z ∂z
Rf = 1 +
ρ Kp d
n
[1]
[2]
where C(z,t) = aqueous concentration of a non-decaying solute [M/L3]; z = distance along the direction of mass transport [L]; t = elapsed time [T]; Dh = hydrodynamic dispersion coefficient [L2/T]; Rf = retardation factor; vz = seepage velocity [L/T]; n = total porosity of the porous medium; ρd = dry density of the soil [M/L3]; Kp = soil-water partition coefficient [L3/M];
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
and kd = first order degradation constant [1/T]. In the one-dimensional flow system, the hydrodynamic dispersion coefficient is represented mathematically as follows (Freeze and Cherry 1979): D h = D* + DL
[3]
DL = αL vz
[4]
where D* = effective molecular diffusion coefficient [L2/T]; DL = longitudinal mechanical dispersion coefficient [L2/T]; and αL = longitudinal dispersivity [L]. Subject to the boundary conditions in the influent (upper) and effluent (lower) reservoirs, solution for Eqn. 1 was obtained numerically by the implicit finite difference scheme (Kim, 1996). Therefore, the results of column/tank tests could be analyzed and the operating transport parameters could be estimated. Typically, in these tests the concentration of the influent was kept nearly constant and the breakthrough (i.e., increasing concentration) in the effluent reservoir was monitored over time. These tests typically took a long-time (300 to 900 days). At the end of the test, the clay layer in the tank tests was sampled using a thin-wall tube sampler. The samples (the whole specimens in the case of smaller column tests) were sectioned and the sorbed chemicals in the pore fluid were extracted to determine a concentration-depth profile. Some tests were pure diffusion tests with no hydraulic gradient. In addition to VOCs, lithium bromide was added to the solution as a tracer. Bromide is one of the most frequently used ionic tracers in groundwater studies. Bromide was selected as a nonreactive tracer in this study because of its biological stability and lower background concentration than other common nonreactive anions such as chloride. Bromide breakthrough data provided a direct estimation of seepage velocity in the tank and column tests. The results of these tests were carefully analyzed using the numerical model to estimate the various transport parameters and unlike most previous investigations it relied on multiple and independent measurements in determining the transport parameters. Some of the significant questions resolved are presented. EFFECTIVE POROSITY Seepage velocity in Eqn. 1 can be estimated from the measured permeant flux, i.e., Darcy velocity if the effective porosity (i.e., the percent pore volume that conducts flow) is known. Therefore, the effective porosity assumed in the analysis of column tests impacts the seepage velocity and in turn the other transport parameters such as hydrodynamic dispersion coefficient. Past studies reported effective porosities equal to 25 to 100% of the total porosity. Analysis of the tracer bromide indicated that effective porosity ranged from 89 to 104 % of the total porosity. In view of the analytical and experimental errors involved, this range of the effective porosities implies that the effective porosity is essentially the same as the total porosity and the difference does not impact the estimated transport parameters significantly. SOIL-WATER PARTITION COEFFICIENT Soil-water partition coefficient is the main parameter in determining the retardation factor given in Eqn. 2 and used in Eqn. 1. Partition coefficients were determined from batch isotherm tests and also directly from the tube samples of clay extracted, sectioned and analyzed at the end of the tank and column tests. In these tests both single and multiple species were used. The soil-water partition coefficients of VOCs were not significantly affected by the presence of other VOCs in the range of concentrations tested. The soil-water partition coefficients from batch tests were not statistically different from those determined by direct measurement on the end-of-test soil specimens in the column/tank tests, thus establishing the simpler and well-established batch isotherm test as an acceptable means of estimating the retardation of VOCs in soils. Table 1 provides a summary of the partition coefficients for the VOCs tested. Most significantly, it was shown that the soil-water partition coefficient of a VOC could be predicted using the available octanol-water partition coefficient of the VOC and organic carbon mass fraction of the soil (up to 6% organic carbon mass fraction). This provides a convenient means of estimating soil-water partition coefficient of a VOC since octanol-water partition coefficients of VOCs are readily available. For soils with varying organic carbon content, foc, soil-water partition coefficient, Kp for a given VOC can be obtained from
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Kp = foc Koc
[5]
where Koc is the organic compound/organic carbon partition coefficient. Koc was shown to be a reasonably constant quantity (±8%) for a large number of soils with foc in the range of 0.1 to 6 % (Kile et al. 1995); thus Eqn.5 can be used to estimate Kp. Figure 2 shows the relationship between the Koc and the octanol-water partition coefficient, Kow of the VOCs tested (in a range of log Kow from 1.25 to 3.25) which is expressed as: (R2 = 0.928)
log Koc = 0.920 + 0.360·log Kow
[6]
Since Kow of most organic compounds are readily available (USEPA RREL, Database Program, 1997), Eqns. 5 and 6 provide a very convenient basis for estimating Kp for soils with foc in the range of 0.1 to 6 %. VOC
Soil-Water Partition Coefficient (L/kg)
Molecular Diffusion Coefficient (D*) (cm2/sec)
Apparent Tortuosity (τa)
Longitudinal Dispersion Coefficient (αL) (cm)
First Order Degradation Constant (kd) (1/day)
CF
0.39
8.60 × 10-6
0.92
0.92
0.030
EB
1.22
6.02 × 10-6
0.75
0.75
0.038
MC
0.27
1.38 × 10-6
0.13
0.13
0.021
TOL 1,1,1-TCA
0.80 0.55
4.13 × 10-6
0.49
0.49
0.019
2.37 × 10-5
2.92
2.92
0.244
TCE
0.94
4.39 × 10-6
0.51
0.51
0.043
m-XYL
1.26
4.86 × 10-6
0.71
0.71
0.045
Table 1. Transport Parameters (Adapted from Kim et al. 2001) 2.5
log K
oc
2.0
1.5
1.0 0.5
1.0
1.5
2.0 log K
2.5
3.0
3.5
4.0
ow
Figure. 2. Relationship between Soil Organic Carbon-Organic Compound Partition Coefficients (log Koc) and Octanol-Water Partition Coefficients (log Kow) of VOCs Tested (Kim et al. 2001)
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
DEGRADATION OF VOCS In the tank and column experiments, the upper (influent) reservoir concentrations decreased as a function of time even though the influent bag concentrations were kept nearly constant. For most of the VOCs, the concentration decreased faster in the influent reservoir as the soil-water partition coefficient increased. In particular, the concentration of 1,1,1-trichloroethane and to an extent chloroform in the influent reservoirs of all column tests decreased unreasonably fast when compared with the other compounds. This happened despite continual addition of the solution at a constant concentration in the influent bags and use of inorganic disinfectants, mercuric chloride and sodium azide, in the reservoirs of the column/tank tests and their feeding solutions to prevent biodegradation of VOCs. Several factors suggested that VOC degradation was significant even though disinfectants were added to the influent. Degradation was found to follow a first order reaction, which is reasonable since most processes are first order, i.e., concentration dependent (Howard et al. 1991) and appeared to happen only within the clay specimen, not in the reservoirs. The average hydraulic conductivity after the spike of VOCs, the effective porosities as obtained from the bromide tracer tests in the same columns (Kim et al. 1997), and the soil-water partition coefficients obtained from the batch tests under multisolute condition were used as inputs to the model based on Eqn. 1. The hydrodynamic dispersion coefficient and the first order degradation constant were estimated based on both the influent and effluent reservoir concentration-time data using a least squares method. The estimated first order degradation constants for the VOCs are listed in Table 1. The estimated first order degradation constants varied depending on the compound. The estimated half-lives ranged from about 2 to 116 days. Of the seven VOCs tested, 1,1,1-trichloroethane, m-xylene, trichloroethylene, ethylbenzene, chloroform, toluene, methylene chloride degraded in descending order and roughly in order of decreasing molecular weight. The degradation of VOCs observed in tank tests seems to be an anaerobic biodegradation, which is one of the most common degradation processes in soil-water systems (Howard et al. 1991). The estimated first order degradation half-lives given in Table 1 are generally comparable to the range reported for anaerobic degradation (Howard et al. 1991) for most of the compounds except 1,1,1-TCA, and to a lesser extent for TCE and EB. It should be noted that the degradation rate values reported herein may include potential loss, chemical/physical decomposition as well as biological degradation. HYDRODYNAMIC DISPERSION AND DIFFUSION COEFFICIENTS The hydrodynamic dispersion coefficients were in the range of 10-6 to 10-5 cm2/sec except for 1,1,1-TCA and CF, which were generally larger than 10-5 cm2/sec. The hydrodynamic dispersion coefficients estimated without considering degradation are greater than those with degradation, which is expected. Because of the strong evidence of degradation in the tests, the hydrodynamic dispersion coefficients estimated by considering degradation were accepted and analyzed. The effective molecular diffusion coefficient and longitudinal dispersivity can be estimated using the hydrodynamic dispersion coefficient determined in the column/tank tests and the seepage velocity using Eqns. 3 and 4 (i.e., the intercept and slope of the hydrodynamic dispersion coefficient versus seepage velocity, respectively, for a given VOC). The results are listed in Table 1. The longitudinal dispersivities of VOCs estimated in this study ranged approximately from 2 to 4 cm except 1,1,1-trichloroethane. The magnitude of estimated longitudinal dispersivities of VOCs implies that the seepage velocity in the testing range does not significantly affect the hydrodynamic dispersion coefficients of VOCs; in other words, mechanical dispersion was not significant in the tests and the molecular diffusion governed. Apparent tortuosity, τa (Shackelford and Daniel 1991) is the ratio of the effective molecular diffusion coefficient, D* to the free solution diffusion coefficient, Do. The apparent tortuosity is primarily a characteristic of the porous medium. The apparent tortuosities estimated in this study are also given in Table 1. Unreasonably high apparent tortuosities for 1,1,1-TCA and CF indicate that the diffusion estimates may not be reliable for these two compounds. CONTAMINANT TRANSPORT THROUGH INTACT GEOMEMBRANES VOC transport occurs over the entire surface of a geomembrane. Thus, mass transport can be modeled one-dimensionally if the transport process is assumed to be spatially invariant. Transport of VOCs through geomembranes can be described as a three step process (Mueller et al. 1998, Park and Nibras 1993, Park et al. 1996): (1) partitioning between the leachate and geomembrane, (2) diffusion through the geomembrane, and (3) partitioning between the geomembrane and the pore water at
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
the down gradient surface of the geomembrane (Figure 3). The diffusion and partition coefficients for a variety of geomembranes were measured by Park and Nibras (1993) and Park et al. (1996) using a confined double-compartment apparatus in which the test geomembrane separated the two compartments. Mass flux of VOCs was shown to depend on the initial concentration of VOCs, thickness and type of geomembrane, and to a certain extent on stretch resulting from tension (Park et al. 1996). They also showed that the time for breakthrough increased in proportion to the square of geomembrane thickness. The transport parameters for a 1.5-mm high-density polyethylene (HDPE) geomembrane (commonly used in landfills) are given in Table 2. Park et al. (1996) calculated that the mass flux by diffusion through a HDPE geomembrane to be more than 2 orders of magnitude greater than the mass flux through the typical number and size of holes found in geomembranes installed in landfills. 1
C/Co Partitioning into Geomembrane C/Co= Kd,gm
Leachate
Ci
Z=0
tgm
Geomembrane Ls Soil Liner
Depth
Concentration Jump at Interface Concentration Profile
Figure 3. Transport Process of VOCs in intact composite liner (Foose et al. 2002) A one-dimensional finite-difference model was constructed by Foose et al. (2002) for analyzing diffusive transport of VOCs through intact composite liners consisting of contiguous layers of finite thickness: (1) the geomembrane, (2) the clay barrier, and (3) the subgrade soil. The governing equation for diffusive transport of VOCs through geomembranes is the same as Eq. 1 with R=1, kd and v=0. Additionally, the diffusion coefficient of the geomembrane (Dgm) is substituted in place of the hydrodynamic dispersion tensor.
VOC Partition Coefficient Diffusion Coefficient (cm2/s)
TCE 115
TOL 148
m-XYL 6.2
MC 450
5.2x10-9
4.3 x10-9
7.9 x10-9
2.9 x10-9
Note: Geomembrane: 1.5-mm thick. VOC initial concentration 100 mg/L Table 2. Mass Flux and Partition Coefficient of High Density Polyethylene Geomembrane with VOCs (Adapted from Park et al. 1996) VOC TRANSPORT IN COMPOSITE LINERS Geomembranes are essentially impervious to diffusion of inorganic solutes compared to many organic solutes (Haxo and Lahey 1988, Rowe et al. 1995). Hence, the most significant pathway for inorganic solutes to travel through a composite liner is through defects in the geomembrane (holes, defective seams, etc.) and subsequently through the soil liner via advection, diffusion, or a combination thereof. Organic solutes can also be transported along the same pathways and by the same mechanisms. In contrast to inorganic solutes, organic solutes can diffuse through intact geomembranes at appreciable rates (Mueller et al. 1998, Park and Nibras 1993). Thus, there are two pathways for solute transport through composite liners: (1)
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
advection and diffusion of inorganic and organic solutes through defects in the geomembrane and subsequently through the soil liner and (2) diffusion of organic solutes through the intact geomembrane and subsequently through the soil liner (Figure 4). Foose et al. (2002) developed models based on MODFLOW (a three-dimensional finite-difference groundwater flow model) and MT3D (a three-dimensional block-centered finite-difference computer model for solving the three-dimensional advection-dispersion equation) for analyzing contaminant transport in composite liners taking into account the pathways described above and for assessing equivalency of alternative liner designs. The size and shape of the geomembrane defects and geomembrane-soil liner contact conditions were also modeled. The alternative liner designs were compared by Defect Leachate Geomembrane VOC Clay or GCL
Inorganic and/or VOC Contaminant
Figure 4. Pathways of Contaminant Transport in Composite Liners (Foose et al. 2002) conducting analyses with the models. Three composite liners were analyzed and compared in terms of equivalency: (1) the liner prescribed in Subtitle D of RCRA, (2) a composite liner having a GCL, and (3) the composite liner prescribed in the Wisconsin Administrative Code Section NR500. The Subtitle D and Wisconsin NR500 liners consist of a thick layer of compacted clay (≥ 61 cm and ≥ 122 cm, respectively) overlain by a 1.5-mm-thick high-density polyethylene (HDPE) geomembrane. The GCL composite liner is similar, except the compacted clay component is replaced with a 6.5-mm-thick GCL. A comparison of the three composite liner systems was conducted for a one-hectare section of liner (Foose et al. 2002). The frequency of geomembrane defects was assumed to be 2.5 defects/ha, which is consistent with recommendations by Giroud and Bonaparte (1989) for installations having a high level of quality control. The area of defects was assumed to be 0.66 cm2, which is within the range of sizes of defects recommended by Giroud and Bonaparte (1989) for analysis of composite liners. The depth of leachate was set at 30 cm, which is the common maximum design depth for leachate collection systems in municipal solid waste (MSW) and hazardous waste landfills. The length of the simulation period was 100 years. For the transport of toluene, the pertinent soil and geomembrane transport parameters were obtained from the research described above. As shown in Figure 5, the mass flux of toluene through intact composite liners ranges from 1 to 185 g/ha/year. That is, the mass flux of VOCs through the intact liner is four to six orders of magnitude greater than that through defects. Thus, mass flux of VOCs through defects can be ignored. Also shown in Figure 3 is that mass flux in the GCL composite liner is nearly independent of the partition coefficient for the GCL and toluene. The mass flux of toluene after 100 years for the GCL composite liner is 1.5 orders of magnitude greater than that for the Subtitle D liner and 2.1 orders of magnitude greater than that through the Wisconsin NR500 liner. Similar observations can be made based on the cumulative mass discharged (Figure 6). At 100 years, the cumulative mass discharged for the GCL composite liner is 18,365 g/ha compared to 385 g/ha for the Subtitle D liner and 43 g/ha for the Wisconsin NR500 liner. The cumulative mass of toluene discharged was as much as 23 million times higher than the cumulative mass of cadmium discharged even though the concentrations of both solutes in the leachate were assumed to be equal. The reason for this
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
difference in contaminant flux and mass discharged is that transport of toluene occurs across the entire surface of the liner, whereas transport of cadmium primarily occurs through a small area beneath defects in the geomembrane. The mass flux of toluene after 100 years for the GCL composite liner is 1.5 orders of magnitude greater than that for the Subtitle D liner and 2.1 orders of magnitude greater than that through the Wisconsin NR500 liner. Similar observations can be made based on the cumulative mass discharged (Figure 6). At 100 years, the cumulative mass discharged for the GCL composite liner is 18,365 g/ha compared to 385 g/ha for the Subtitle D liner and 43 g/ha for the Wisconsin NR500 liner. The cumulative mass of toluene discharged was as much as 23 million times higher than the cumulative mass of cadmium discharged even though the concentrations of both solutes in the leachate were assumed to be equal. The reason for this difference in contaminant flux and mass discharged is that transport of toluene occurs across the entire surface of the liner, whereas transport of cadmium primarily occurs through a small area beneath defects in the geomembrane. 10
8
10
7
10
6
10
5
10
4
10
3
10
2
10
1
10
0
GCL Composite Liner GCL Composite Liner K =2.6 mL/g K =5.2 mL/g d d
10
-1
10
-2
10
-3
Subtitle D Subtitle D, Steady-State Flux of Toluene through Defects C =100 µg/L o
Wisconsin NR500 C=0 µg/L Toluene 0
20
40 60 Time (years)
80
* (a) 100
Figure. 5. Transport of Toluene in Three Composite Liners-Mass Flux (Adapted from Foose et al. 2002) The thicker liners have greater sorptive capacity for toluene than the GCL composite liner because they are at least 94 times thicker than the GCL composite liner even through the factor ρdKd, which is the ratio of mass of contaminant sorbed to mass of soil, is 2.05 for the GCL composite liner and 1.24 for the thicker composite liners. In particular, the sorptive capacity for toluene for the Subtitle D liner is expected to be 57 times greater than that for the GCL composite liner (i.e., 94 x 1.24/2.05 =57).
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
10
8
10
7
10
6
10
5
10
4
10
3
10
2
10
1
10
0
GCL Composite Subtitle D
C =100 µg/L o
Wisconsin NR500
10
-1
10
-2
C=0 µg/L Toluene 0
20
40 60 Time (years)
80
(b) 100
Figure. 6. Transport of Toluene in Three Composite Liners-Cumulative Mass Discharged (Adapted from Foose et al. 2002) Results of the analyses show that equivalency assessments based on leakage rate (i.e., purely on advection) can lead to incorrect conclusions. Assessments based on contaminant transport are more appropriate. Analyses showed that the variation in mass flow rates of inorganic solutes for different types of liners is less than an order of magnitude. For VOCs, composite liners having thicker soil barriers had lower mass flux and greater sorptive capacity than the GCL composite liner. An analysis based on leakage rate would lead to the opposite conclusion. A limitation of these conclusions is that laboratory or field data are unavailable regarding contaminant transport from composite liners. Thus, the accuracy of the models in replicating field conditions is unknown. Additionally, these conclusions are based solely on solute transport. Other factors such as resistance to environmental degradation, potential for catastrophic puncture, quality of construction, and mechanical stability should also be considered. In addition, the type and size of defects were assumed to be the same regardless of whether the composite liner had a compacted soil liner or GCL, and only a limited range of material and solute properties was used. Degradation of the contaminants was also ignored and the leachate source was assumed to have a constant concentration and depth. However, for most of the evaluation criteria, performance of the liners varied by several orders of magnitude. Thus, variations in the material properties or contaminants, which typically are less than an order of magnitude, and changes in the boundary conditions, which would have similar effects on solute transport in all three liners, should not vary enough such that the conclusions reached in this analysis will change significantly. SUMlMARY VOCs are encountered in the leachates generated in municipal and hazardous waste landfills and the transport of these chemicals out of containment systems is of concern due to the hazardous nature of the materials. Limited data exists for the transport parameters of VOCs through compacted clay liners and geomembranes. Test conditions used (large sample sizes, long testing times, realistic boundary conditions, and solution of the problems associated in testing volatile materials), independent determination of transport parameters unlike most other previous investigations, and analysis of degradation of the VOCs in the compacted clays provided data that was not extensively available in the past and sets a benchmark for future investigations of VOCs. A systematic and comprehensive approach to determine mass transport parameters for transport of VOCs through compacted clay liners and geomembranes has enabled to develop realistic models to predict mass flux of
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
VOCs through modern composite liners and have a quantitative basis to evaluate the equivalency of different composite liners. ACKNOWLEDGMENT This paper is dedicated to Dr. Raymond J. Krizek, author’s doctoral advisor and friend and to Nortwestern University where the author received his Ph.D. and acquired the necessary tools for a satisfactory academic life. The material presented is a summary of research conducted by the author and his colleagues and students over the last decade. The author did not know most of this material when he was a graduate student at Northwestern. The hallmark of excellent education is in the tools and attitudes imparted not in the specific knowledge learned. The author is indebted for all that he was given during the time he spent at Northwestern. Heartfelt thanks Ray. REFERENCES 1.
Barone, F. S., Rowe, R. K., and Quigley, R. M. (1992). “A laboratory estimation of diffusion and adsorption coefficients for several volatile organics in a natural clayey soil,” Jour. Contaminant Hydrology, 10(3), 225-250.
2.
Battista, J. R., and Connelly, J. P. (1994). “VOCs at Wisconsin landfills: recent findings,” Proc. 17th International Madison waste conf., Madison, WI, 67-86Berens, A. R. Prediction of Organic Permeation through PVC Pipe, Jour. American Water Works Association, Vol. 77, No. 11, pp. 57-65, 1985.
3.
Edil, T. B., Park, J. K., and Heim, D. P. (1994). “Large-size test for transport of organics through clay liners,” ASTM STP 1142, 353-374.
4.
Crooks, V. E. and Quigley, R. M. (1984), “Saline leachate migration through clay: a comparative laboratory and field investigation,” Canadian Geotechnical Journal, 21, 349-362.
5.
Edil, T. B., Wambold, W. S., and Park, J. K. (1995), “Partitioning of VOCs in clay liner materials,” Geoenvironment 2000, ASCE GSP 46, 775-790.
6.
Freeze, R. A., and Cherry J. A. (1979). Groundwater, Prentice-Hall, Englewood Cliffs, N.J.
7.
Friedman, F. A. (1988). Volatile organic compounds in groundwater and leachate at Wisconsin landfills, Wisconsin Department of Natural Resources, Rep. PUBL-WR-192-88, Madison, WI.
8.
Foose, G. J. (1997), "Leakage rates and chemical transport through composite liners," Ph.D. Dissertation, Dept. of Civil and Environmental Engineering, University of Wisconsin Madison.
9.
Foose, G. J., Benson, C. H., and Edil, T. B. (2002), “Comparison of solute transport in three composite liners,” Journal of Geotechnical and Geoenvironmental Engineering, ASCE, 128(5), 391-403.
10.
Gibbons, R. D., Dolan, D., Keough, H., O’Leary, K., and O’Hara R. (1992). “A comparison of chemical constituents in leachate from industrial hazardous waste & municipal solid waste landfills,” Proc. 15th annual Madison waste conf., Madison, WI, 251-276.
11.
Giroud, J. P. and Bonaparte, R. (1989), "Leakage through liners constructed with geomembranes - parts I and II," Geotextiles and Geomembranes, 8, 27-67, 71-111.
12.
Hashimoto, I., Deshpande, K. B., and Thomas, H. C. (1964). “Peclet numbers and retardation factors for ion exchange columns, Industrial & Engineering Chemistry Fundamentals, 3(3), 213-218.
13.
Haxo, H. E. and Lahey, T. P. (1988), "Transport dissolved organics from dilute aqueous solutions through flexible membrane liners," Hazardous Waste and Hazardous Materials, 5(4), 275-294.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
14.
Heim, D. P. (1992). Advective and diffusive transport of three volatile organic compounds through a compacted clay, M.S. thesis, University of Wisconsin at Madison.
15.
Howard, P. H., Boethling, R. S., Jarvis, W., F., Meylan, W. M., and Michalenko, E. M. (1991). Handbook of environmental degradation rates, Lewis Publishers, Chelsea, MI.
16.
Kile, D. E., Chiou, C. T., Zhou, H., Li, H., and Xu, O. (1995). “Partitioning of nonpolar organic pollutants from water to soil and sediment organic matters,” Environ. Sci. & Technol., 29(5), 1401-1406.
17.
Kim, J. Y. (1996) “Retardation of volatile organic compound movement in landfills using scrap tires, Ph. D. thesis, Department of Civil and Environmental Engineering, University of Wisconsin at Madison.
18.
Kim, J. Y., Edil, T. B., and Park, J. K. (1997), “Effective porosity and seepage velocity in column tests on compacted clay,” Journal of Geotechnical and Geoenvironmental Engineering, ASCE, 123(12), 1135-1124.
19.
Kim, J. Y., Edil, T. B., and Park, J. K. (2001), “Volatile organic compound (VOC) transport through compacted clay,” Journal of Geotechnical and Geoenvironmental Engineering, ASCE, 127(2), 126-134.
20.
Krug, M. N., and Ham, R. K. (1995), “Analysis of long-term leachate characteristics in Wisconsin landfills,” Proc. of 18th International Madison Waste Conference, Dept. of Engr. Professional Development, University of Wisconsin-Madison, Madison, WI, 168-184.
21.
Mueller, W., Jakob, R., Tatzky-Gerth, R., and August, H. (1998), “Solubilities, diffusion and partition coefficients of organic pollutants in HDPE geomembranes: experimental results and calculations,” Proc. of the Sixth International Conference on Geosynthetics, Atlanta, Industrial Fabrics Association International, St. Paul, MN, 239248.
22.
Myrand, D., Gillham, R. W., Sudicky, E. A., O’Hannesin, S. F., and Johnson, R. L. (1992). “Diffusion of volatile organic compounds in natural clay deposits: Laboratory tests,” Jour. Contaminant Hydrology, 10(2), 159-177.
23.
Nelson, B. R., and Book, P., R. (1986). “Monitoring for volatile organic hydrocarbons at Minnesota sanitary landfills,” Proc. 9th annual Madison waste conf., Madison, WI, 72-84.
24.
Park, J. K. and Nibras, M. (1993), "Mass flux of organic chemicals through polyethylene geomembranes," Water Environment Research, Washington, D.C., 65(3), 227-237.
25.
Park, J. K., Sakti, J. P., and Hoopes, J. A. (1996), “Transport of organic compounds in thermoplastic geomembranes I: mathematical model,” Journal of Environmental Engineering, ASCE, 122(9), 800-806.
26.
Plumb, R. H., and Pitchford, A. M. (1985). “Volatile organic scans: implications for groundwater monitoring,” Proc. NWWA/API conf. on petroleum hydrocarbons and organic chemicals in ground water-prevention, detection and restoration, Houston, TX, 207-222.
27.
Richardson, G. (1997), “GCL’s: alternative subtitle D systems,” Geotechnical Fabrics Report, Industrial Fabrics Association, St. Paul MN, 15(5), 36-42.
28.
Rowe, R. K. (1987), "Pollutant transport through barriers," Geotechnical Practice for Waste Disposal, GSP No. 26, ASCE, R. Woods, ed., 159-181.
29.
Rowe, R. K., Hrapovic, L., and Kosaric, N. (1995), “Diffusion of chloride and dichloromethane through an HDPE geomembrane,” Geosynthetics International, 2(3), 507-536.
30.
Rügge, K., Bjerg, P. L., and Christensen, T. H. (1995). “Distribution of organic compounds from municipal solid waste in the groundwater downgradient of a landfill (Grindsted, Denmark),” Environ. Sci. & Tech., 29(5), 13951400.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
31.
Sakti, J. P. (1993). Transport of organic chemicals in dilute aqueous solutions through intact geomembranes, Ph.D. dissertation, University of Wisconsin at Madison.
32.
Shackelford, C. D. (1990), "Transit-time design of earthen barriers," Engineering Geology, Elsevier Publishing, Amsterdam, 29, 79-94.
33.
Shackelford, C. D. and Daniel, D. E. (1991), “Diffusion in saturated soil I: background,” Journal of Geotechnical Engineering, ASCE, 117(3), 467-484.
34.
U.S. EPA (1997). RRRL treatability database program, Perry’s Chemical Engineers’ Handbook, 7th Ed., McGrawHill, N.Y.
35.
Wambold, W. S. (1993). Large and small scale laboratory tests of VOC transport through compacted clay, M.S. thesis, University of Wisconsin at Madison.
Analysis & Design Design and Construction of Reinforced Earth Walls on Marginal Lands A. Abraham, R. E. Allen, J. E. Sankey, K.Truong, H. Tran, The Reinforced Earth Co, USA A Generalized Probabilistic Approach to the Stability of Cut Slopes J. N. Kay, Griffith University, Australia Innovative High Rise Foundation Design in Chicago C. Barker, T. D. Bushell, T. A. Kiefer, R. Diebold, STS Consultants, USA, and ThorntonTomasetti Engineers, USA Lessons Learned from a Bermed Excavation in Soft Clay H. J. Liao, National Taiwan University of Science and Technology, Taiwan, R.O.C. A New Slope Stability Approach Using Calculus of Variations, and Safety and Sensitivity Analysis E. Castillo and R. Mínguez, University of Cantabria, Spain, and University of Castilla La Mancha, Spain Uncertainties in Simplified Liquefaction Analysis of Soil Deposits R. Blázquez, S. López, V. Navarro, J. Sánchez, E. González, Universidad de Castilla – La Mancha, Spain
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
DESIGN AND CONSTRUCTION OF REINFORCED EARTH WALLS ON MARGINAL LANDS Alexander Abraham, Hieu Tran, Richard E. Allen, Kim Truong, and John E. Sankey, The Reinforced Earth Co. 1444 North Farnsworth Ave, Suite 505, Aurora, IL 60505
[email protected]
ABSTRACT Reinforced Earth (RE) walls described in this paper are a composite system consisting of alternating layers of granular backfill and discrete steel reinforcements connected to precast concrete facing panels. This paper discusses technical features that make RE walls a better alternative to rigid cast-in-place structures for roadway applications over marginally stable soils. The design of the facing panels, steel reinforcements, and backfill used in RE walls supporting bridge loads and constructed over marginal soils is provided. The effect of bridge loadings on the behavior of RE walls is also discussed. It is noted that RE walls accommodate variations in field conditions, which further benefit construction. A case history is presented to emphasize the technical and practical factors that influence the design of RE walls over marginal lands. The paper concludes by providing a summary of the design and construction practices that will improve the aesthetics, cost, safety and durability of RE walls. INTRODUCTION Reinforced Earth (RE) walls consist of alternating layers of granular backfill and galvanized steel reinforcements connected to precast concrete facing panels. This paper discusses the design features of RE walls in relationship with their use over marginal lands. A case history that emphasizes the technical and practical factors that influence the design of RE walls over marginal lands is also provided. DESIGN OF RE WALLS ON MARGINAL LANDS In a RE wall, the mechanical properties of a soil mass - in this case composed of granular backfill - are improved by discrete steel reinforcing elements aligned parallel to the orientation of its greatest strains. The friction between the granular soil and the steel reinforcements allows the backfill, which can withstand only compressive and shear stresses, to transfer tensile stresses to the reinforcements. The combination of soil backfill, steel reinforcing elements and a durable facing produces a composite material that is both strong and flexible. A description of the components with an emphasis on the modification to these components for construction on marginal lands is provided in the following sections.
Department of Civil and Environmental Engineering Infrastructure Technology Institute Northwestern University, Evanston, Illinois, 60208
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Facing Elements RE walls for permanent applications (defined as a design life of 50 to 100 years) predominantly use precast concrete facing panels with open joints separating the panels. The flexibility of a RE wall depends on maintaining the integrity of the open joints between the precast modular facing elements. The open joints allow the panels to move relative to each other, thereby making it possible to undergo significant deformation. The joints generally utilize ‘bearing pads’ made of elastomeric materials to provide a bearing surface between the top of one panel and the bottom of the panel. The bearing pads are typically 20 or 25 mm in thickness. These pad materials are highly resistant to ozone oxidation, resistant to corrosive attack, retain resiliency and do not crack. The weight of the overlying panels as well as the effect of long-term creep usually results in a 10 to 15 percent compression of the bearing pads. The bearing pads provide the needed balance between compressibility under increased loads and ability to maintain an open joint during the design life of the RE wall. Backfill The current practice in the design and construction of RE walls is to use durable granular materials as backfill with no aggregate greater than 3-inch diameter or more than 15 percent particles finer than 0.075 mm (US Sieve No. 200) [American Association of State Highway and Transportation Officials (AASHTO) 1996a]. The granular composition of the backfill within the reinforced volume enhances drainage, as well as ease-of-construction considerations at the work site. In the range of loading normally associated with RE walls, the granular materials behave elastically. Post-construction movements associated with internal yielding or readjustments of the backfill are not typically generated. Furthermore, as compression of the supporting subsurface materials below the embankment occurs, the compacted granular backfill used in a RE wall deforms as an integral mass Steel Reinforcements and Connection of Facing Elements to Reinforcements Discrete steel strips used in RE walls, as discussed in this paper, are classified as inextensible reinforcements. The steel reinforcements deform much less readily than the backfill that envelops the reinforcements. Detailed procedures for the design of inextensible (steel) reinforcements are provided in the Standard Specifications referenced by AASHTO (1996b). For certain applications, the global stability may be improved slightly by using longer reinforcements at the base and by varying the density (the quantity of reinforcement per unit volume of the embankment). The connection between the facing elements and the reinforcements is selected based on a detailed evaluation of the anticipated stresses at the connection and the ease of connecting the reinforcing elements to the facing panels during construction. A well-designed positive connection will not allow separation to occur between the facing and the reinforcing elements. A positive connection also prevents movement between the facing and the backfill in the event of deformation due to external loads. EFFECT OF BRIDGE LOADINGS ON THE BEHAVIOR OF RE WALLS A RE wall is capable of supporting a bridge abutment spread footing bearing directly on top of the reinforced soil. The footing bears only on the reinforced soil and is not supported by piles or other structural members. Abutment bearing pressure is transferred directly into the reinforced soil and, depending on the height of the RE wall, either is fully dissipated within the reinforced soil or in the case of low height walls is distributed through it to the site foundation soil below. The pressure under the abutment footing is dissipated with depth through the RE volume according to Boussinesq pressure distribution. For computational simplicity, a 1:2 (Horizontal: Vertical) linear distribution is conservatively used to envelop the Bousinesq pressure contribution. In order to limit the pressure applied directly to the wall facing panels, the abutment’s centerline of bearing must be at least 1 m behind the facing. Where the 1:2 distribution intersects the face of the RE wall, the load from the abutment is transferred through the reinforcing strips back to the reinforced soil mass as horizontal stress. This increased horizontal stress from the bridge loads may require additional reinforcing strips to be added to the RE walls design. When the RE wall height exceeds three times the width of the abutment footing, the bearing pressure from the abutment is almost completely dissipated within the RE volume according to the 1:2 pressure distribution discussed above. Therefore the foundation soil does not receive significant additional bearing pressure due to the bridge loads (The Reinforced Earth Co. 2000). This is an important advantage in the case of marginal foundation soils that can accept the distributed load of a RE
Department of Civil and Environmental Engineering Infrastructure Technology Institute Northwestern University, Evanston, Illinois, 60208
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
wall but not the additional concentrated load of an abutment footing; resulting in not having to install a more costly pile foundation system. For abutment walls of lesser height (those less than twice as high as the abutment footing is wide), the total bearing pressure at the foundation will be the sum of the undissipated portion of the 1:2 distribution (the portion which extends below the base of the reinforced volume) and the bearing pressure determined by conventional RE design calculations. Therefore, the allowable bearing pressure for the site must be sufficient to support this increased load. GEOTECHNICAL ASPECTS OF RE WALLS ON MARGINAL LANDS The provision of a comprehensive geotechnical evaluation of subsurface conditions plays an integral part in determining the future performance of a RE wall (Anderson, 1991). Follow-up construction should include provisions for field observation and monitoring to verify the evaluations made during design. Differential Settlement of Foundation Soils Based upon monitoring results from a number of RE walls with precast facing panels that have undergone significant differential settlement, a maximum allowable value of 1 m of differential settlement per 100 m of wall length (1 percent) is recommended to maintain structural and aesthetic integrity. It is noted however that several structures have undergone differential settlement in excess of the recommended limit without structural distress. Butler (1980) documents the case history of a ramp constructed at the Sprain Brook Parkway in New York that settled differentially 530 mm per 30.5 m of wall length (1.7 percent) without adverse structural effects. For RE walls constructed on marginal lands, where the soil conditions are highly irregular (non-uniform), the potential exists for differential settlement in excess of 1 percent. Differential settlements in excess of 1 percent will result in closing and opening of the joints between the panels. Spalling or cracking of the concrete panels may in turn occur at some points along the closed joints depending on the magnitude of the settlement. At locations where the predicted differential settlement is beyond the 1 percent limit, the design and construction procedures should include provisions to accommodate the anticipated differential settlement (Abraham, 1999). Total Settlement There is no formal definition of the tolerable total settlement for a RE wall. Some RE walls have experienced as much as 0.6 m of total settlement. While this might be acceptable for a retaining wall not having a roadway or other elevation-sensitive structure on top, it would be unacceptable for an abutment or a wall connecting to another structure. Bearing Capacity When analyzing foundation soils for bearing, the RE walls are modeled as continuous strip footings with a width and magnitude equivalent to that of the Meyerhof bearing pressure diagram (AASHTO 1996b). The resulting bearing pressure diagram is rectangular, with a width equal to the reinforcement length minus two times the eccentricity of the embankment. In the case of RE walls, the use of a lower factor of safety against bearing failure is justified, since the uniform bearing pressure of the RE volume will not tend to increase if deformation of the embankment occurs (due to flexibility of the RE walls). Factors of safety in the range of 2.0 to 3.0 are generally used based on an evaluation of the geotechnical information and the external loads applied on the RE wall. The bearing capacity evaluations are generally based on the assumption that the critical mode of failure analysis is in general shear of the subsurface materials. However, in the case of highly compressible soils, emphasis also should be placed on an evaluation of the punching shear as well.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Global Stability A check on global stability looks not only at the RE volume and its relationship to the adjacent soils, but also at the characteristics of the deeper soil strata that will affect the stability of the whole structure, embankment and/or hillside. For certain projects, the placement of a layer of granular material (interlayered with tensile reinforcements, as appropriate) of sufficient thickness and width over marginal soils is a viable approach to improve the global stability. As an alternative, RE walls may also be economically constructed after improving the soft soils with stone columns to increase the bearing capacity and rotational stability of RE walls. Sliding Along the Base In the case of RE walls on marginal soils, it must be determined that there is no risk of horizontal sliding along the base of the RE volume. The criteria for horizontal sliding is more critical when the RE wall is supporting a heavy surcharge; for example, when supporting a high, sloping embankment. When sliding along the base controls the design due to the characteristics of the bearing soils along the base, a lengthening of the reinforcements may be required to improve the sliding resistance. Undercutting the weak soils and replacing with granular materials is another common technique used to improve the resistance to sliding along the base. Undercutting may be uneconomical when the required depth exceeds 2 m or the depth intercepts ground water. CASE-HISTORY OF A RE WALL SUPPORTING BRIDGE LOADS AND CONSTRUCTED OVER MARGINAL LANDS This section provides a brief summary of a RE wall constructed on marginal soils. The case history also demonstrates the load carrying capacity of RE walls, and suggests techniques for designers and contractors to adapt RE wall design and construction to prevailing subsurface conditions. Original Design Provided in the Contract Drawings This case history describes the design and construction of the abutments for a single span bridge across a tributary to the Kickapoo River in Vernon County, Wisconsin. The initial design provided in the contract drawings incorporated Cast-InPlace concrete abutments supported on H piles. The contract documents required that each abutment should be supported on 47 HP 310 x 79 steel piles, of which 29 of the piles were driven at a 4 to 1 batter. Piles were to prebored 1 m into bedrock prior to driving the piles to a minimum bearing value of 625 kN per pile. Additionally, the 1.5 Horizontal: 1 Vertical slopes in front of the abutment were reinforced with geosynthetics. The steeper 1.5 H: 1 V slopes were selected instead of flatter, unreinforced slopes to minimize the area disturbed by construction activities. Value-Engineering Proposal After the award of the contract, the contractor (Edward Kraemer & Sons, Inc., Plain, Wisconsin) proposed a valueengineering proposal that replaced the pile-supported abutments with abutments supported on RE walls. The RE walls were 14.4 m tall from the base of the wall to the top of pavement. It is noted that the actual RE wall did not extend full height; rather cast-in-place concrete abutments and wing walls supported on the reinforced soil zone were extended the final 6.3 m. Subsurface Conditions The subsurface profile revealed by the borings consisted of fill, clay and/or silt layers at the surface overlying sand strata, which in turn were underlain by weathered to sound sandstone bedrock (CGC, Inc. 2001). While the layers and soil types were similar on both sides of the valley, the soil layers overlying bedrock were substantially thicker on the south side. Groundwater was generally not encountered in the borings on the north side. Below the south approach embankment, groundwater was encountered in a coarse sand /gravel stratum immediately overlying bedrock.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
External Stability Analyses The external stability evaluations for the RE walls included overturning, sliding, bearing capacity, settlement and global stability calculations. The global stability evaluations indicated that the soil layers below the RE embankment should be undercut to expose the underlying sandstone bedrock. The undercut excavation was backfilled with breaker run stone (50 mm diameter nominal size) to the base of the RE embankment. The contact pressure at the base of the RE embankment was 403 kPa. This is the sum of the undissipated portion of the bridge loads (38 kPa) and the eccentric soil loads (365 kPa). Contact pressure remained well below the allowable bearing capacity of the foundation materials. Internal Stability Analyses The internal stability evaluations included the design of the reinforcing strips for soil loads as well as bridge loads in accordance with AASHTO guidelines (AASHTO 1996b) and Wisconsin Department of Transportation (DOT) Special Provisions for Reinforced Earth Concrete Panel Walls (Wisconsin DOT). The capacity of the reinforcement elements against pullout and tensile failure was determined by superimposing the bridge loads over the earth pressure loads. Reinforcement elements were also provided behind the abutment back wall to resist the lateral loads. RE Wall Construction The construction of the RE walls were in general accordance with the bridge shop drawings prepared by Westbrook Associated Engineers, Inc. (structural engineer for the value-engineering proposal) and the wall shop drawings prepared by the Reinforced Earth Co. The facing panels consisted of cruciform-shaped Reinforced Earth® (RE) panels with a nominal area of 2.25 m2. Prior to the construction of the RE embankment, the surficial soils were undercut until reasonably a competent weathered rock stratum was encountered. The depth of undercut at the south abutment generally agreed with the top of bedrock elevations estimated during the subsurface explorations. However, at the north abutment where extreme weathering of the sandstone rock strata was present, the excavation had to be extended a substantial distance (some to 4 to 5 meters) below the leveling pad to reach reasonably competent weathered bedrock. At both abutments, the excavation was backfilled with compacted breaker rock material. The breaker rock was placed in relatively thin lifts and compacted until no further deflection was evident before placing the next lift. The breaker rock was placed as well-compacted, dense–graded material with no obvious open voids. The breaker rock extended below the entire reinforced zone behind the wall to restore grades. The breaker rock also extended 0.6 m beyond the front face of the leveling pad and then at a 1H:1V slope downward to meet the excavation sidewall. Cohesive fill materials were placed above the breaker run to prevent the potential for erosion. A drainage pipe was used to intercept seepage. CONCLUSION RE walls using modular facing panels, steel reinforcements and granular backfill are well suited for the construction of retaining walls on marginal lands. The design of RE walls should include a comprehensive geotechnical evaluation of the subsurface materials. During construction, monitoring and inspection should be used to verify the evaluations made during design and to modify the construction procedures based on the field data. The case history demonstrates how RE walls carrying heavy loads were constructed over marginal lands based on an evaluation of the soil conditions. A cost-effective solution was developed based on an understanding of the site soil conditions as well as the design features of a RE wall. Field modifications to the design were made based on variations that were observed in the field. A successful solution to the changed field conditions was based on open communication between the RE wall engineers, site structural and geotechnical engineers, contractor and owner. Every RE wall built on marginal soils constitutes a new and unique project. In each project, the optimum solution is obtained by taking into account the geotechnical information, the scheduling requirements and deadlines, the nature and magnitude of external loads and the flexible nature of the RE wall.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
ACKNOWLEDGMENTS The authors wish to acknowledge the support of Mr. Robert P. Arndorfer, P.E., Mr. Bruce J. Pfister, P.E., and Mr. Gerald H. Anderson, P.E., Wisconsin Department of Transportation, Mr. Jeffrey J. Koch, P.E., Westbrook Associated Engineers, Inc., Mr. William W. Wuellner, P.E., CGC, Inc., and Mr. Larry Nagle and Mr. Merlin A. Leifker of Edward Kramer & Sons, Inc. in the design of the RE walls described in the case history. Hoffman Construction Co. constructed the RE walls mentioned in the case histories and their contributions are hereby acknowledged. LIMITATIONS Publication of this paper does not constitute an endorsement of the technology described in this paper by the owners, consultants and/or contractors mentioned herein. REFERENCES 1.
Abraham, A. and Sankey, J. E., “Design and Construction of Reinforced Earth Walls on Marginal Lands,” Geotechnics of High Water Content Materials, ASTM STP 1374, T. B. Edil and P. J. Fox, Eds., American Society for Testing and Materials, West Conshohocken, PA, 1999.
2.
AASHTO (American Association of State Highway and Transportation Officials), 1996a, Standard Specifications for Highway Bridges, Division II - Construction, Section 7 - Earth Retaining Systems, Sixteenth Edition, Washington, DC.
3.
AASHTO (American Association of State Highway and Transportation Officials), 1996b, Standard Specifications for Highway Bridges, Division I - Design, Section 5 - Retaining Walls, Sixteenth Edition, Washington, DC.
4.
Anderson, P.L., October 1991, “Subsurface Investigation and Improvements for MSE Structures Constructed on Poor Foundation Soils,” Proceedings of the 34th Annual Meeting, Environmental and Geological Challenges for the Decade, Association of Engineering Geologists.
5.
Butler, B. E., 1980, “Use of Reinforced Earth for Construction of the SM Ramp on the Sprain Brook Parkway,” Earth Support Systems, Seminar Sponsored by the New York City Metropolitan Section of the American Society of Civil Engineers.
6.
CGC, Inc., August 2001, Subsurface Investigation, STH 131 Mechanically Stabilized Earth Walls, Vernon County, Wisconsin, West Allis, Wisconsin.
7.
The Reinforced Earth Co., 2000, Design Manual for Reinforced Earth® Walls, Vienna, Virginia.
8.
Wisconsin Department of Transportation, Special Provisions for Mechanically Stabilized Earth Concrete Panel Wall, Madison, Wisconsin.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
A GENERALIZED PROBABILISTIC APPROACH TO THE STABILITY OF CUT SLOPES J.Neil Kay School of Engineering, Griffith University, Gold Coast Campus Griffith University, Gold Coast Campus, Parklands Drive, Southport QLD 4215, Australia
[email protected]
ABSTRACT In this approach to slope safety, use is made of data compilations on slope performance in Hong Kong where results have been interpreted in terms of probability of failure. The method has been extended previously by the writer and his colleague to examine the influence of slope improvement through the use of soil nailing. To apply the method to soil nailing, it was necessary to separate the slope failure probability distribution into two parts, one associated with the rainfall intensity and duration, the other associated with variability and uncertainty in the soil and site geology. While the rainfall component, would, in general, be peculiar to a given location, there is evidence to indicate that the site and soil uncertainties, may well be applicable to other sites where relatively similar conditions occur. The fact that, in Hong Kong, similar distributions were obtained for sites with different soil types, adds weight to this evidence. Accordingly, at a different location outside of Hong Kong, provided similar soils exist, the rainfall characteristics for that location could be combined with the Hong Kong site and soil data to obtain a failure characteristic for the new location. A generalized graph for such locations is proposed. INTRODUCTION Following an extensive compilation of the performance data related to the relatively frequent occurrences of slope failure in Hong Kong a proposal was made earlier for evaluation of slope performance based on the probability of failure. This was later extended to provide a probabilistic method for evaluation of Hong Kong slopes improved by soil nailing. The method used for the latter case was to separate the probability characteristic into two components, one associated with rainfall intensity and duration and a second associated with the site geology and soil. The rainfall characteristic corresponding to water penetrating to the increased potential failure depth near the back of the soil nailing was then reintroduced with the soil and site geology data to produce a design graph. In this paper, the method is extended to sites outside of Hong Kong. Owing to the fact that rainfall characteristics appear to be the dominant part of the uncertainty and variability, it is suggested that the part associated with the soil and general site geology might be transferable to other locations where similar decomposed rock and residual soil conditions occur. The rainfall characteristics for the new locations could be combined with the Hong Kong soil and site geology data to provide a probability of failure for slopes for that location. This would be particularly valuable for locations where it is not practicable to examine failure frequencies. The frequencies may be too low because of lower rainfall intensity or because land development may not have had a sufficiently long history.
Department of Civil and Environmental Engineering Infrastructure Technology Institute Northwestern University, Evanston, Illinois, 60208
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
UNDERLYING PREMISES Extensive observation of the performance of man made slopes has been carried out in Hong Kong over several years. Based on the results, certain conclusions have been reached: (Chen, 1994, Kay, 1998, and Brand, Premchitt and Phillipson, 1984). (a) for slopes that have been constructed in decomposed basalt or in volcanic deposits, (these were the majority of slopes – some others were in colluvium) a consistent correlation exists between probability of failure and slope angle (see Fig.1). (b) for slopes associated with heights greater than 10 m, probability of failure is sufficiently well described by slope angle alone. (for slopes of height less than 10 m, slope height also plays a significant roll but, to maintain simplicity, this range is not included in this study) (c) the characteristic failure type for the Hong Kong area for slopes greater than 10 m in height is that of a translational type failure that has a body depth in the range of 2 to 3 metres. (d) when rainfall exceeds 70 mm/hr for the duration of one hour, the probability of landslide occurrence is extremely high. PROBABILITY OF FAILURE FOR HONG KONG SLOPES For the earlier proposal for probabilistic assessment of slopes in Hong Kong, emphasis was placed on information derived from previously measured data. In Hong Kong, a great deal of information has been compiled in relation to failure occurrence together with the readily obtainable soil and profile characteristics for both failed and unfailed slopes. The various measurable parameters have been examined and plotted against propensity to failure (Chen and Kay ). This information permitted construction of a slope failure probability relationship for the area in terms of the most relevant variables. Although for slopes less than 10 m in height, both slope height and slope angle were significant, for those slopes higher than 10 m, only the latter was relevant. As slopes with heights greater than 10 m were the dominant types as well as to maintain simplicity, only those slopes will be addressed. The probability of failure versus slope angle relationship obtained from the previous study is reproduced in Fig. 1. 0.005 0.004 Annual 0.003 Probability of Failure 0.002 0.001 0.000 20
30
40
50
60
70
Slope Angle (degrees)
Fig.1 Probability of Failure versus Slope Angle for Cut Slopes in Hong Kong DEVELOPMENT OF CONDITIONAL PROBABILITY It is emphasized that, realistically, the 70 mm/hr for the duration of one-hour criterion is not a hard boundary below which there are no failures and above which there are many. The figure is only a guide and is quite fuzzy. Other factors play a role,
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
especially the antecedent rainfall or the amount of rain that falls in the days and weeks leading up to the time of potential failure. For the purposes of the present proposal (as was the case for the soil nailing proposal) this criterion will be considered to be a hard boundary. The probability of occurrence of rainfall in excess of 70 mm/hr for the duration of one hour will be represented as p(R ≥ 70). The concept of the probability of failure given that R ≥ 70 has occurred can then be considered. This will be written as a conditional probability, p(f | R ≥ 70). The overall probability of failure, p(f), is obtained by combining the two as follows: p(f) = p(f | R ≥ 70) p(R ≥ 70)
[1]
Because the probability of failure, p(f), for Hong Kong has been described in terms of slope angle and because the rainfall characteristic p(R ≥ 70), has been obtained from the Hong Kong Observatory data, that part of p(f) that can be ascribed to the variation in site geology and soil conditions, p(f | R ≥ 70), can be determined from: p(f | R ≥ 70) = p(f) / p(R ≥ 70)
[2]
This probability value can be computed for specific data points at various slope angles and, if desired, the continuous probability distribution may be constructed therefrom over the relevant range. COMPUTATIONS From compilation of rainfall data over a period of 16 years at the Hong Kong Observatory, the average number of occurrences of rainfall in excess of 70 mm/hr for a period of one hour was 3.80 or, on average, a frequency of 0.238 times per year. If N is the annual frequency and the assumption is made that such an event may be described as a Poisson process, the probability that the rainfall exceeds 70 mm/hr for one hour, p(R ≥ 70), may be obtained from: p(R ≥ 70) = 1 - eN = 0.211
[3]
As demonstrated in Fig.1, for the respective slope angles 30, 40, 50 and 60 degrees, the annual probabilities of failure for Hong Kong slopes are 0.0005, 0.0014, 0.0024 and 0.0036, respectively. Consequently, the probability of failure given that R ≥ 70 has occurred, p(f | R ≥ 70), may be obtained from Eq.2. Results are reproduced for the four slope angles in Table 1. 1 Slope Angle 30 40 50 60
2 p(f) Hong Kong 0.0005 0.0014 0.0024 0.0036
3 p(R ≥ 70) Hong Kong 0.211 0.211 0.211 0.211
4 p(f | R ≥ 70) 0.0025 0.0068 0.0116 0.0169
Table 1 APPLICATION TO OTHER LOCATIONS It is suggested that this distribution may be applied to other locations around the world where the soil conditions are similar to those found in Hong Kong. That this may be reasonable is supported by the fact that, for Hong Kong, where relatively consistent average rainfall conditions may be assumed over the region, similar p(f) results were obtained for two groups of sites even though those two groups were different on the basis of their classification in relation to particular soil types and particular geological origins. One group consisted of predominantly decomposed basalt and the other of predominantly decomposed rocks of volcanic origin. If the rainfall intensity and duration is obtained for a different location, the average annual frequency of rainfall exceeding 70 mm/hr for a period of 1 hr could be converted to the annual probability of experiencing such rainfall, pN(R ≥ 70) using Eq.3. This could then be combined with the distribution of the site and soil variability and uncertainty, p(f | R ≥ 70), to give the probabilistic slope failure characteristic for the new site, pN(f), according to:
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
pN(f) = p(f | R ≥ 70) pN(R ≥ 70)
[4]
As the value of rainfall frequency of 70 mm/hr for the period of one hour for Hong Kong was 0.238, a range of values either side of this value is used to enable development of a general data set that might be useful elsewhere. The values of average frequency chosen are 0.10, 0.20, 0.30 and 0.40 occurrences per year. The ranges of probability of failure are calculated as given in Table 2. Fig.2 illustrates the probability distributions derived therefrom. 1
2 Annual Frequency p(R ≥ 70)
3
4
5
6
7
0.1
0.2
0.238
0.3
0.4
0.095
0.181
0.211
0.259
0.330
Slope Angle
p(f | R ≥ 70)
30
0.0025
0.0002
0.0004
0.0005
0.0006
0.0008
40
0.0068
0.0006
0.0012
0.0014
0.0018
0.0022
50
0.0116
0.0011
0.0021
0.0025
0.0030
0.0038
60
0.0169
0.0016
0.0031
0.0036
0.0044
0.0056
Table 2 SAMPLE COMPUTATIONS At a location where the rainfall is considered to be similar to that in Hong Kong, it is desired to evaluate the acceptable slope angle for two extensive slopes, one, in the vicinity of a residential subdivision where it is considered that a 2 percent chance of failure in 50 years is acceptable and a second, adjacent to a highway where a 10 percent chance of failure in 50 years is considered reasonable. (a) Two percent per 50 years is equal to an annual probability of failure of 0.0004. From the Hong Kong probability curve this indicates a required slope angle of 28 degrees. (It should be noted that installation of soil nailing sufficient to increase a potential slide depth from 2 m to 4 m [from Kay and Li, 1999] would permit construction of slopes at 62 degrees). (b) Ten percent per 50 years is equal to an annual probability of failure of 0.002. From the Hong Kong probability curve this indicates a required slope angle of 46 degrees. (As for part (a) installation of soil nailing on slopes at 62 degrees would limit the probability of failure to 2 percent in 50 years). SUMMARY At locations where extensive land development has taken place in steeply sloping terrain and where intense rainfall occurs, it is likely that disastrous landslides will be relatively frequent. Such is the case for Hong Kong. The extensive compilation of details on landslide data available there has enabled the preparation of a relationship between probability of failure and slope angle. Consideration of the component parts of the failure uncertainty and variability has suggested that a separation is possible into a part that is site related and a part that may be applicable to slopes at other locations. The site related part is the rainfall intensity and duration. It is proposed that the rainfall component might be determined at other locations and combined with the soil and site variability component from Hong Kong for determination of the probability of failure at the other locations. Results have been produced that provide ready access to the probability of landslide failure when the average annual frequency of rainfall events in excess of 70 mm/hr for the duration of one hour are available for the new location. The development provided should be particularly valuable for locations where there is little history of previous land development.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
0.006 Numbers on curves indicate average annual frequency of rainfall events equal to or greater than 70 mm/hr for one hour duration
0.4
0.005
0.3 0.004 Annual Probability 0.003 of Failure
0.2
0.002 0.1 Indicates Hong Kong Case
0.001 0.000 20
30
40
50
60
70
Slope Angle (degrees) Fig.2 Probability of failure versus slope angle for cut slopes at various locations subject to location rainfall history REFERENCES 1.
Brand, E.W., Premchitt, J., and Phillipson, H.B., 1984, Relationship Between Rainfall and Lanslides in Hong Kong, Fourth International Conference on Landslides, 1. p377-384.
2.
Chen, T, 1994, Slope failure probability based on performance history in Hong Kong, M.Phil Thesis, University of Hong Kong, Hong Kong, 149 p.
3.
Kay, J.N., 1998, Slope stability assessment for the residual soil slopes in Hong Kong, Soils and Foundations, Journal of the Japanese Geotechnical Society, 38 4, 95-103.
4.
Kay, J.N., and Li, K.S., 1999, A Probabilistic Approach to Soil Nailing for Hong Kong Slopes, Geotechnical Seminar of the Hong Kong Institution of Engineers in Engineering.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
INNOVATIVE HIGH RISE FOUNDATION DESIGN IN CHICAGO Clyde N. Baker, Jr1, Ted. D. Bushell2, Tony A. Kiefer3, and Rob Diebold4 1,2,3
STS Consultants Inc 750 Corporate Woods Parkway. Vernon Hills, IL 60061-3153
[email protected],
[email protected],
[email protected] ` 4 Thornton-Tomasetti Engineers 14 E. Jackson Boulevard, Suite 1100, Chicago IL 60604
[email protected]
The history of high rise building foundation design and construction in Chicago prior to World War II is described by Ralph B. Peck in Reference 1 and after that in Reference 2. Prior to about 1895, most buildings, even the tallest (the Monadnock at 16 stories and Auditorium at 19 stories) were supported on footing foundations in the thin desiccated crust over the soft Chicago clay. However, experience with very large settlement under the heavier buildings which reached a reported 23 inches differential settlement under the Auditorium building by 1900 (10 year period) caused a change in design philosophy with increasing numbers of designers requiring deep foundation support for the taller buildings. This trend was accelerated by the shallow foundation settlements observed due to ground squeeze occurring during the construction of Chicago's freight tunnel system beginning in 1904. The above experience with large unpredictable settlements occurred before the development of modern soil mechanics including the theory of consolidation. The University of Illinois Bulletin by Peck and Uyanik on the "Observed and Computed Settlements of Structures in Chicago" (Reference 3) demonstrates that the settlement of foundations built over normally consolidated clay soils can be reasonably well predicted with modern investigation and testing tools. However, settlement prediction in over-consolidated soils is much less predictable. Settlement in over-consolidated soils can range from as little as 2 percent to as much as 20 percent of the calculated settlement in normally consolidated soils depending upon how close the foundation bearing pressure is to the preconsolidation pressure in the soil (Reference 4, page 108). To avoid any questions with regard to the possibility of excessive differential settlement, most designers historically have tried to support their structures on the same type of foundation system and not attempt to maximize the cost/performance efficiency of their foundations based on magnitude of loading. Conventional practice until fairly recently has been to support the entire structure on either hardpan or rock caissons (but not both under the same structure), if any portion of the structure was heavy enough to require deep foundations. IMPROVED SETTLEMENT PREDICTION IN OVER-CONSOLIDATED SOILS During the past thirty years, there has been some modification in design thinking resulting from our ability to better predict settlement in over-consolidated soils using in-situ pressuremeter testing. These developments have been used in Chicago to facilitate economical use of mixed foundations for a number of high rise buildings constructed in downtown Chicago over the past twenty years. In a number of cases, the structural engineers have found it advantageous to support the core of some of the heavier buildings on rock with the lesser loaded (but still very heavy) non-core caissons on the hardpan or very dense silt immediately under the hardpan, with the primary question being the magnitude of differential settlement expected between the rock caissons and the hardpan caissons. Typical examples would be the 50-story office towers at 35 and 77 West Wacker Drive, as well as 1 North Wacker Drive. Reference 5 describes the use of the pressuremeter in mixed high rise foundation design in Chicago.
Department of Civil and Environmental Engineering Infrastructure Technology Institute Northwestern University, Evanston, Illinois, 60208
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
SOME CURRENT INNOVATIVE DESIGN CONCEPTS Two innovative design concepts which are currently being used in Chicago are: 1) re-use of existing foundations in combination with a different type of new foundations, and 2) use of a limited number of highly stressed piles or caissons to reduce settlement of a mat or strip footing foundation system. Since both these approaches involve combining different foundation systems, it is essential to be able to predict how the different systems will perform. In the first category, there are really two different approaches that have been used. In one approach, a mat foundation has been used to transfer the load from columns to old existing caisson foundations that are not located directly under the new structure columns. In this case, no load is assumed taken by the mat itself. Examples of this are the Associates Building at the northwest corner of Randolph and Michigan and the office tower at 181 West Madison. The other design concept involves using a mat over existing caissons in which the mat and supporting soil share the load with existing caissons; i.e., part of the load is transferred to the existing caissons and part of the load is carried by the soil under the mat based on strain compatibility and comparable settlements. ILLUSTRATIVE CASE HISTORIES The following case histories are intended to illustrate how different foundation systems can sometimes be cost effectively designed utilizing in-situ pressuremeter testing to help predict ground deformation under load. Dearborn Center Dearborn Center is a case history that illustrates a mixed foundation system in which existing caissons which previously supported an 11-story building (and has been demolished down to street level) share the load with a mat constructed in the lowest basement level on top of the existing caissons to support a new 38-story office building. The geotechnical program for this project consisted of performing seven new soil borings denoted B-101 through B-107. These borings supplement ten earlier borings, nine of which were performed outside of the existing building perimeter. Five of the seven new borings were performed from the existing lowest basement elevation at -23 Chicago City Datum (CCD) with two borings performed at the first basement level at elevation -4 CCD. A location plan showing all borings, as well as the existing caissons, is included as Figure 1. Borings B-101, B-103 and B-106 were performed adjacent to existing columns 36, 56 and 125 to confirm the presence of the bells and to access the soil immediately below the bells for testing. These borings were blank drilled to the top of the caisson bell at which point the concrete caisson bell was cored with a diamond bit core barrel. These three borings were then extended below the bottom of the caisson bell to elevations ranging from -79 CCD to -85 CCD. Pressuremeter tests were performed below the caisson bell in all three of these borings. Borings B-102, B-104, B-105 and B-107 were extended through the lowest level basement slab to elevations ranging from -57 CCD to -60 CCD. Pressuremeter tests were also performed in these borings through the floor slab. Unconfined compression tests were performed on selected samples of the caisson bell concrete and indicated strengths ranging from 6300 to 7800 psi. These results were similar to those obtained in earlier investigation performed by others in 1984. A summary soil profile, along with a graphical plotting of the key pressuremeter test results is shown on Figure 2. A complete tabulation of pressuremeter test results is included in Table 1. The water content and unconfined compressive strength data, including penetrometer data, are shown graphically in Figure 3. Analysis and Design Geotechnical Analysis The design concept for the Dearborn Center project is to make cost effective use of the existing substructure at the site, while at the same time permitting development of the maximum practical number of office floors above the existing substructure (including two levels of retail at ground level). Substructure levels would be utilized primarily for car parking. To accomplish this, the design concept involves re-using the existing belled caisson foundations which are supported on the hard
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
clay stratum approximately 33 feet below basement level, or approximate elevation -56 CCD, and then developing additional load carrying capacity by using a mat placed on top of the bottom basement slab connecting to all of the existing columns and caissons. The new building load would be carried by the combination of the caisson foundations and mat foundation with the load distribution between the two foundation types based upon the compressibility of the subsoils. Because of the approximately 40 feet of basement excavation resulting in stress unloading of the subsoils below mat level, it is anticipated that significant loads (up to the weight of the soil removed) could be applied at the mat level with only a modest settlement for a subsoil deformation based on the elastic or pseudo-elastic properties of the subsoil. The pressuremeter test results which measure the pseudo-elastic properties of the soil up to the creep pressure, indicate an average creep pressure of approximately 9 tons per square foot (tsf) in the very stiff to hard silty clay zone beneath the caissons. The drop off in unconfined compressive strength and increase in water content noted in the zone from -68 to -75 CCD (Figure 3) did not result in significantly reduced modulus or creep pressures value indicating a fairly consistent preconsolidation pressure. It is likely that the higher water content indicates greater plasticity and moisture retention under comparable loads. In order for the settlement predictions to be reliable using pressuremeter data, the dead load bearing stress plus the overburden pressure should not exceed the average creep pressure. Thus, allowing for an existing overburden pressure in the hard clay just below caisson bearing level of approximately 2 tsf relative to top of mat level, the maximum dead load pressure should not exceed 7 tsf to keep the combined total less than the average creep pressure of 9 tsf. If the bearing pressure under the caissons exceeds this value, there would be a tendency towards increasing settlement and load transfer back to the mat. Caisson springs for use in a mat finite element analysis were developed assuming approximately 1 inch deflection under a pressure of 18 ksf on a representative 14 foot diameter belled caisson. Illustrative calculations are shown in Figure 4. With regard to the mat, utilizing the pressuremeter data obtained in the subsoils beneath the mat, the average mat pressure required to produce a 1 inch settlement comparable to the caisson settlement is approximately 2000 pounds per square foot (psf). This data can be used to calculate spring constants under the mat for use in a finite element analysis. This pressure/deflection estimate is based upon an elastic analysis using a Young's modulus for the soil zone beneath mat level of two times the pressuremeter rebound modulus. This is an empirically derived relationship based upon monitoring of large scale projects (Reference 6). Foundation Structural Analysis and Design The foundation design for the Dearborn Center project was driven by two major project requirements. First, the new structure would be maximized in terms of height and size while being founded on the existing foundations. Because of the high cost of installing deep foundations in an existing 3-story basement, no new deep foundation elements could be added to support the new building. Second, the existing basement walls and lower level 3 slab-on-grade must both be maintained, but the 3 basement levels must be replaced with 3 new basement levels. Figure 5 contains a foundation plan illustrating various elements of the structure. The existing caissons were regularly spaced throughout the site on approximately an 18 X 22 foot grid. With the exception of the caissons along the north property line that extended to rock, all of the caissons were belled and supported on hard pan clay. The new building columns were somewhat irregularly placed, with bays ranging from 20 to 38 feet. Obviously, the new columns did not align with the existing caissons. Furthermore, the caissons located around the perimeter of the site were positioned directly beneath the 4-foot thick basement walls and inaccessible from the basement. In order to maximize the new building's size, all of the caissons must be loaded to their capacity. Additionally, as described under geotechnical analysis, it was determined that the soil directly below the lower level three was adequate to support building loads. A thick concrete mat foundation would be the logical choice for distributing the new column loads to the existing caissons and the soil, but two project requirements prevented this. First, a thick, heavy concrete mat would use foundation capacity, thus decreasing the allowable building size. Second, fitting three basements in the existing excavation would leave very little depth for structure. A relatively thin, heavily reinforced, 10,000 psi concrete mat that varied from 54 inches to 42 inches was chosen. Preliminary analysis of the mat proved that a mat of this thickness would not be stiff enough to adequately distribute the high column loads to the existing caissons. To stiffen the mat, a series of concrete walls were introduced. The wall locations were coordinated with the architectural requirements for parking and mechanical space so that no parking spaces were sacrificed.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Two computer analyses were used in designing the concrete mat. A 3-dimensuional SAP model was built to determine the overall building behavior. Soil spring values generated by the geotechnical engineer were utilized as supports. Each caisson was assigned a spring value based on its bell size, while the caisson shaft was input as a concrete column. The soil springs directly beneath the slab-on-grade were arranged in a 2-foot grid. The caissons that extended to rock were given an extremely stiff spring, allowing no more that 1/16 inch settlement. (Caisson shaft side friction was ignored because the soil under the mat was being considered for bearing.) The caissons, soil, mat, existing basement walls, new walls, new columns and the entire buildings lateral support system were included in this model. The location, thickness, height, and exact location of the stiffening walls were refined using this SAP computer model. Both the soil and caissons capacities were determined by geotechnical analysis to generate 1 inch settlement. Therefore, strategically locating and sizing the stiffening walls achieved a uniform settlement of 1 inch maximum under full load. Accurate soil settlement predictions combined with an exact representation of the building loads and an accurate model of the building structure is critical in designing a highly refined and integrated foundation system such as this. Decreasing the weight of the braced core was key in maximizing the height of the building. Clearly, a full height concrete core was far too heavy, and the glassy exterior of the building eliminated using columns spaced closely enough to create a tube structure. Therefore, a braced steel core was chosen as the lateral force resisting system for the building. The SAP analysis indicated that differential foundation settlements generated enormous forces in the core bracing. To minimize the forces in the steel bracing, and to help distribute the loads from the heavy core columns, shear walls were introduced in the core area. These walls extended from the mat at lower level three up to lower level one. These walls optimized the load distribution while minimizing the building weight. New shear walls were added at the perimeter of the building, perpendicular to the existing basement walls. These walls performed three functions. First, the SAP analysis indicated that the existing caissons that landed between the core and the exterior columns were not receiving enough load because few new columns landed in this zone. These new shear walls helped to shift loads from the exterior columns to these under-utilized caissons. The second function of these shear walls was to provide a temporary site retention system. As mentioned, the existing basement walls were to remain, but the basement slabs would be demolished and replaced. The mat was placed directly on the lower level three slab-on-grade and the new shear walls were constructed before the existing basement slabs were removed. The shear walls that were perpendicular to the existing basement walls were designed to cantilever up from the mat with sufficient strength to resist the lateral soil pressure. Therefore, the three basements could then be completely cleared, and most of the retention system was also part of the permanent building structure. The third function of these walls was to connect the new mat to the existing caissons at the perimeter of the site. As mentioned, these perimeter caissons were directly beneath the existing basement walls, and so the new mat did not rest on them. These exterior caissons were engaged by creating a concrete girder at lower level one that rested on the existing columns that were supported on the perimeter caissons. This concrete girder supported the new building columns. To decrease the differential settlements between these perimeter caissons and the mat, the two were connected vertically by these shear walls that were dowelled into the mat at lower level three and framed into the concrete girder at lower level one. A second computer model of the mat was generated for design purposes as shown in Figure 6. A SAFE model of the complete mat was used for the design of flexural and shear reinforcement. To ensure that the mat had adequate strength under all possible soil conditions, load cases were run that varied the support of the soil directly under the mat. These load cases generated an envelope of shears and moments in the mat that were used for design. Since the soil directly under the existing slab-on-grade was being considered as part of the foundation system, a series of explorations were conducted to determine that no voids were present under the S.O.G. A series of trenches through the S.O.G. were required for the installation of a new sub-soil drainage system. Observations of these trenches provided evidence that there were no significant voids under the slab. At the time of this writing, Dearborn Center is nearing completion. The entire superstructure is erected, as well as the majority of the superimposed dead loads such as the exterior wall, raised floor system and mechanical systems. Tenants have not yet begun to move in, so live loads, partitions, etc. are not in place. It is estimated that that approximately 65% to 70% of the full design load is presently being supported by the foundations. Given the current load, and the 1 inch anticipated settlement under full load, the anticipated present settlement would be approximately 5/8 inch to 3/4 inch.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Settlement reference marks set on the building walls and mat at the start of construction and used during construction were checked at this time (those that could be found and were not covered). The readings indicate reported settlement that varied from 0 on the north wall (reported to be on rock caissons) to 1/2 inch on the west wall, 5/8 inch settlement on the south wall and 5/8 inch settlement on the interior mat. Allowing for survey accuracy of 1/8 inch, we estimate settlements ranging from 1/8 inch at the rock supported caissons to 3/4 inch elsewhere. This agrees with predictions used in the design and confirms the adequacy of the basic assumptions made and analyses performed. Acknowledgements The Project Developer on Dearborn Center is Prime-Beitler, the Structural Engineer is Thornton Tomasetti, the Architect is De Stefano + Partners and the Contractor is AMEC. STS Consultants, Ltd. was the Geotechnical Engineer.
Geotechnical Materials: Measurement and Analysis : R.J. Krizek
Figure 4
Dearborn Center Chicago, Illinois Deflection and Spring Calculations Ref: The Menard Pressuremeter, Interpretation and Application of Pressuremeter Test Results, Sols Soils, Nov. 26, 1975.
Material Properties Design Conc. Comp. Strength:
f'c := 5000psi
Modulus of Elasticity of Conc.:
Ec := 57000 psi
Length of Shaft:
L := 30ft
Shaft Diameter:
d := 6ft
Reference Length:
R0 := 1ft
Design Caisson Bell Diameter:
D := 14ft
The Caisson Bell Area:
A := π ⋅ R
(
)⋅
0.5
6
Ec = 4.031 × 10 psi
f'c
R :=
D 2
2
A = 153.9 ft
2
Units tsf :=
2000 psi 144
psf :=
1 psi 144
ksf := 1000psf
Pressumeter Moduli of the Soil (From Test Data) E1 := 232tsf E2 := 232tsf E3.4.5 :=
16 5 ⋅ E1 + ⋅ 5000 ⋅ tsf 21 21
E3.4.5 = 1367 tsf
α := 0.58 Determine Equivalent Moduli Spherical: Deviatoric:
EA := E1 EB :=
3.2 1 1 1 + + E1 0.85 ⋅ E2 E3.4.5
EA = 232 tsf EB = 316 tsf
Geotechnical Materials: Measurement and Analysis : R.J. Krizek
Determine Estimated Settlement (Ignores Elastic Shortening) p := 9tsf
Axial Load:
Shape Coefficients:
λ 2 := 1
λ 2⋅R 1.33 ⋅ p w2 := ⋅ R0 ⋅ 3 ⋅ EB R0 w3 :=
α
(4.5 ⋅ E1)
λ 3 := 1
α
⋅p ⋅λ 3⋅R
w2 = 0.039 ft
or,
w2 = 0.468 in
w3 = 0.035 ft
or,
w3 = 0.42 in
w = 0.074 ft
or,
w = 0.888 in
w := w2 + w3
Determine Elastic Shortening of Caisson Shaft δ :=
Elastic Shortening Due to an Applied Vertical Load is Defined as:
The Ratio of the Bell (Bearing) to the Caisson Shaft Area is:
δ :=
p⋅L 2⋅R ⋅ Ec d
2
δ = 5.066 × 10
−3
ft
or,
W := w2 + w3 + δ
The Total Deflection is:
ABell AShaft
P ⋅L A ⋅E
2 ⋅ RBell := DShaft
δ = 0.061 in W = 0.949 in
Determine the Pressure Beneath a Mat to Cause an Equivalent Deflection ( 400 + 157) tsf 2
Assume Young's Modulus:
E := 2 ⋅
Define a Geometry Factor:
i := 0.8
For:
H := 60ft If:
Settlement , W :=
E = 557 tsf
H ⋅p⋅i E
Then:
p :=
E ⋅W H ⋅i
p = 0.917 tsf
For a Settlement of 1":
p :=
E ⋅ 1in H ⋅i
p = 0.967 tsf or, p = 1934 psf
2
Geotechnical Materials: Measurement and Analysis : R.J. Krizek Commemorative Symposium: 3 Aug 2002
New Perimeter Wing Walls (TYP.)
New Core Walls
New Columns (TYP.)
Existing Caissons (TYP.)
Existing Foundation Wall
FIGURE 5 - DEARBORN CENTER FOUNDATION PLAN
Geotechnical Materials: Measurement and Analysis : R.J. Krizek Commemorative Symposium: 3 Aug 2002
High Stress RED
ORANGE
Low Stress YELLOW
BLUE
GREEN
PURPLE
GREY
FIGURE 6 - CONCRETE MAT SHEAR STRESS DIAGRAM
Geotechnical Materials: Measurement and Analysis : R.J. Krizek Commemorative Symposium: 3 Aug 2002
Table 1
Dearborn Center
Pressuremeter Test Results
BORING NUMBER
DEPTH (ft)
Po (tsf)
Pf (tsf)
Pl (tsf)
Ed (tsf)
E+ (tsf)
Ed /E+
Ed /Pl
Pl /Pf
101
33.3-35.8 38.0-40.5 43.5-46.0 48.5-51.0 54.5-57.0
2.0 2.5 2.5 2.5 4.0
7.5 8.5 8.5 7.5 25.0
15.4 15.3 15.7 16.5 47.0
103 215 470 110 324
177 295 624 276 1107
0.58 0.73 0.75 0.40 0.29
6.7 14.1 29.9 6.7 6.9
2.1 1.8 1.8 2.2 1.9
102
5.5-8.0 10.0-12.5 15.0-17.5 20.0-22.5 25.0-27.5 30.0-32.5
1.0 1.0 1.8 1.8 1.8 2.0
2.8 3.5 3.8 4.5 5.5 7.0
6.0 6.6 8.0 11.0 12.9 14.1
57 82 75 89 113 79
73 174 160 125 136 224
0.78 0.47 0.47 0.71 0.83 0.35
9.5 12.4 9.4 8.1 8.8 5.6
2.2 1.9 2.1 2.4 2.3 2.0
103
33.5-36.0 38.5-41.0 43.5-46.0 48.5-51.0 53.5-56.0
2.5 3.5 3.5 3.5 4.5
7.5 9.5 9.5 8.5
17.7 19.1 17.6 18.9
0.54 0.59 0.80 0.68 0.48
2.4 2.0 1.9 2.2
-
237 260 240 376 920
7.2 8.1 10.9 13.6
-
127 154 192 257 443
-
-
104
5.0-7.5 10.0-12.5 15.0-17.5 20.0-22.5 25.0-27.5 30.0-32.5
1.0 1.0 1.3 1.3 2.5 2.0
3.0 3.3 3.5 3.5 6.5 8.0
6.1 6.6 8.2 7.7 11.4 16.6
62 137 139 103 89 216
125 214 291 149 182 331
0.50 0.64 0.48 0.69 0.49 0.65
10.2 20.8 17.0 13.4 7.8 13.0
2.0 2.0 2.3 2.2 1.8 2.1
105
5.0-7.5 10.0-12.5 15.0-17.5 20.0-22.5 25.0-27.5 30.0-32.5
1.0 1.0 1.3 2.0 2.5 2.5
3.3 3.0 3.5 4.5 5.5 5.5
5.7 6.0 7.9 9.6 13.0 13.5
74 75 90 118 99 94
130 125 115 231 134 191
0.57 0.60 0.78 0.51 0.74 0.49
13.0 12.5 11.4 12.3 7.6 7.0
1.8 2.0 2.3 2.1 2.4 2.5
106
52.0-54.5 57.0-59.5 62.0-64.5 67.0-69.5 72.0-74.5 77.0-79.5
3.0 3.5 3.5 3.5 4.5 4.5
8.0 10.0 9.5 10.0 8.0
16.8 20.5 19.3 18.0 15.3
0.66 0.61 0.66 0.47 0.73 0.32
2.1 2.1 2.0 1.8 1.9
-
218 327 418 475 358 1418
8.6 9.7 14.4 12.3 17.1
-
144 198 277 222 261 452
-
-
27.0-29.5 32.0-34.5 37.0-39.5 42.0-44.5 47.0-49.5 52.0-54.5
1.5 2.3 2.0 2.5 3.5 3.0
3.3 4.0 4.5 5.0 6.5 6.5
6.3 6.7 8.5 10.0 11.0 12.5
52 46 96 158 82 93 AVERAGE
91 111 176 188 103 175
0.57 0.41 0.55 0.84 0.80 0.53 0.59
8.3 6.9 11.3 15.8 7.5 7.4 11.1
1.9 1.7 1.9 2.0 1.7 1.9 2.0
107
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Chicago Southside Office Building Construction of a 10-story (with provision for two more stories) combination parking structure and office building was completed in 1996 at 1911 South Indiana Avenue in Chicago, Illinois. The new structure is of reinforced concrete design with 24 foot x 40 foot bays. The lower floor levels are parking with upper floor levels office. The lowest floor over half of the structure is at grade with the other half depressed approximately four feet. Initial construction was 10 stories with two additional floors to be added at a later date as the need arises. Maximum design column loads are 2,700 kips. The soil profile at the site consists of medium dense to dense sand and sandy silt to a depth of approximately 16 feet followed by a stiff clayey crust underlain by soft compressible clay. The soft compressible clay gradually increases in strength to stiff and extends to a depth of approximately 65 feet where a thin (sometimes non-existent) very stiff to hard silty clay layer exists underlain by layers of dense to very dense water-bearing sandy silt to limestone bedrock at 90 feet. Because of the potential for squeezing of soft clays and the relative thinness of an adequate bearing layer at depth, a preliminary geotechnical report prepared for the site recommended against the use of conventional belled caissons for this project as too risky and expensive. STS Consultants, Ltd. was retained to further evaluate a shallow foundation solution and provide cost effective methods for reducing the anticipated settlement. A supplementary field exploration program was performed consisting of five (5) borings including in-situ pressuremeter tests conducted within the upper sands just below anticipated footing level, pressuremeter testing within the lower silty sands just below the potential deep caisson bearing level, in-situ vane shear testing within the soft clay below footing level and selective, undisturbed three inch diameter piston sampling of soft clay for consolidation testing, as well as shallow and deep water table measurements. Geotechnical parameters for the site are summarized on Figure 7. Shallow Foundation Analysis Because of the presence of the upper dense sand layer and stiff clay crust, strip footings were a possibility for support of the structure as they act in effect like a mat when combined with the dense sand layer. Because of the stress spreading effect of the dense sand and stiff clay layer, the actual bearing pressure design of the footings has little influence on the ultimate settlement since it is the average stress increase in the underlying soft clay resulting from the total weight of the building that causes the settlement. Calculated maximum settlement for this equivalent mat case was eight inches with two to three inches occurring during construction and five to six inches thereafter. This was considered excessive and ruled out shallow foundation only solutions. Deep Foundation Analysis Various deep foundation solutions were considered including rock caissons, piles, and straight-shaft caissons, but cost estimates on all solutions were outside of the project budget. Combination System Analysis To take advantage of the lower cost of the strip footing solution, while trying to reduce the settlement to an acceptable range, a combination system was designed. The combination consists of 14-foot wide continuous strip footings supported on the surface dense sand layer and five to six-foot diameter straight-shaft caissons extended down to the dense water-bearing sand and silt layer. It was anticipated that the straight shafts could be excavated and filled with concrete before water seepage became a problem (not possible for belled caissons). The design contemplated approximately 60% of the building load being initially supported by the strip footings with 40% carried by the straight shafts with this ratio reversing with long-term consolidation of the soft clay reacting to the strip footing pressures. The combination of strip footings and straight shaft caissons reduced the projected settlement to less than onethird that predicted for the strip footing or mat foundation solution alone. Since the straight shafts were considered primarily as settlement reducers, a higher than normal bearing pressure could be accepted consistent with the desired settlement limitation. The design approach was relatively unique in the sense that the settlement reducing elements carry the load primarily in end-bearing rather than in side friction, which is the common system where a mat supported on settlement reducing piles is normally utilized.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
The strip footings were designed to withstand a range of soil pressures since it was not possible to guarantee the exact load distribution between footing and shaft, particularly with time, as the underlying soft clay consolidates. Ultimate projected settlement for the combination system was in the range of two to three inches compared to eight inches for the strip footings only. The settlement reducing elements were designed to have a structural factor of safety of at least 2 at the point of calculated soil failure and a soil factor of safety greater than one assuming all the load is taken by the settlement reducer. To determine how load actually gets distributed into the ground, strain gages were placed in two representative shafts and first floor columns. The strain gages monitor the load sharing between the shafts and the strip footings. The soil profile, foundation schematic and instrumentation are shown in Figure 8. Strain gage data for both the columns and the caisson shafts taken over a 3-1/2 year period are shown in Figures 4 through 16. The strain gage data on the columns is relatively consistent and similar whereas the strain gage data in the caisson shafts differs drastically from one side of the shaft to the other indicating possible bending. However, the average values appear consistent and reasonable. The initial tension readings could be due to shrinkage in the concrete in the shaft being restrained by the large overlying strip footing to which the shafts were connected shortly after construction of the shafts and while cement hydration was undoubtedly still occurring. It is also interesting to note that there has been little load increase since the building was completed in early 1996. The small load increase as noted may be due to live load changes or possibly due to small concrete creep. Measured settlements have also been very small since completion of the building with a total measured settlement ranging from one inch at column C2 to 1-1/4 inches at column B6. Column B6 also has the greatest percentage of the load carried by the caisson shaft as compared to the strip footing. This is probably due to the fact that the column is at the end of the footing and does not get the same stress spreading influence that the massive footing provides for interior columns. The B6 caisson appears to be carrying 76% of the column load whereas the C2 caisson appears to be carrying 59% of the column load. It should be noted that the structure was designed for two additional floors so the current loading is only approximately 83% of the ultimate design loading. A summary of the instrumentation results is shown in Table 2. From the data obtained to date, it appears that the caissons are behaving slightly stiffer than anticipated and the ultimate settlement will be slightly less than predicted. In making the original calculations for load sharing between footing and shaft and settlement of footing and shaft, the actual test pressuremeter data obtained in the bearing stratum below the shafts was adjusted to account for possible disturbance and loosening upon shaft excavation (modulus values were reduced in half to allow for this possible loosening effect), and it would appear from the settlement data that no such loosening effect occurred and that a better correlation of prediction and performance would have been obtained by using the pressuremeter data without adjustment. ACKNOWLEDGEMENTS The Project Developer was the Yusuf Partnership, with Kolbjorn Saether, Structural Engineer and Gerald Slawin of Gordon Levin Slawin, Architect. Walsh Construction Company was the Contractor. STS Consultants, Ltd. was the Geotechnical Engineer. CONCLUSION Innovative cost effective solutions to foundation design problems are sometimes possible using combinations or mixtures of foundation elements provided that ground deformation and response to structure loading can be reasonably predicted within allowable tolerances.
Geotechnical Materials: Measurement and Analysis : R.J. Krizek Commemorative Symposium: 3 Aug 2002
FIGURE 9 COLUMN STRAIN GAGES 1911 SOUTH INDIANA COLUMN B-6 INSTALLED NOVEMBER 19, 1994 CONCRETE STRENGTH OF COLUMN = 8,000 PSI
DATE
TEMPERATURE DEGREES (F)
19-Nov-94 30-Nov-94 11-May-95 19-May-95 18-Jul-95 26-Sep-95 14-Mar-96 19-Aug-96 21-May-97 6-Nov-97 14-Apr-98
42 25 50 60 89 63 51 71 60 45 63
READINGS IN MICROSTRAINS GAGE NO. GAGE NO. SETT. BLACK-1 GREY-1 (INS) READING CHANGE READING CHANGE
0.00 -0.63
-1.00 -1.00 -1.00 -1.00 -1.12 -1.14
2468 2477 2193 2150 1963 1912 1900 1812 1763 1748 1733
0 9 -275 -318 -505 -556 -568 -656 -705 -720 -735
3384 3369 3135 3102 2965 2906 2898 2809 2769 2770 2745
0 -15 -249 -282 -419 -478 -486 -575 -615 -614 -639
AVERAGE CHANGE
CONSTRUCTION ACTIVITY
0 -3 -262 -300 -462 -517 -527 -616 -660 -667 -687
Before the column was poured After the column was poured Eight floors poured Starting to form the 9th floor Precast installed Windows installed Building completed Building completed Building completed Building completed Building completed
Geotechnical Materials: Measurement and Analysis : R.J. Krizek Commemorative Symposium: 3 Aug 2002
Figure 10 Column Strain Gage Data 1911 South Indiana, Column B-6 Date 3-Apr-95
3-Oct-95
- = Compression Microstrains + = Tension
0
2-Apr-96
2-Oct-96
2-Apr-97
2-Oct-97
2-Apr-98
2-Oct-98 2.0
0.0
-200
-2.0
-400
-4.0
-600
-6.0
-800
-8.0
Black - 1 Gray-1 Average Settlement -1000
-10.0
Settlement (Inches)
3-Oct-94 200
Geotechnical Materials: Measurement and Analysis : R.J. Krizek Commemorative Symposium: 3 Aug 2002
FIGURE 11 CAISSON - SISTER BAR STRAIN GAGES 1911 SOUTH INDIANA COLUMN B-6 INSTALLED OCTOBER 3, 1994 BOTTOM OF CAISSON = -68.10' B.D. *NOTE: CONCRETE TEMPERATURE
DATE
3-Oct-94 03-Oct-94 10-Nov-94 16-Nov-94 19-Nov-94 30-Nov-94 11-May-95 19-May-95 18-Jul-95 26-Sep-95 14-Mar-96 19-Aug-96 21-May-97 6-Nov-97 14-Apr-98
READINGS IN MICROSTRAINS TEMPERATURE GAGE 8745 GAGE 8746 DEGREES SETT. ELEV -21.50 B.D. ELEV -21.50 B.D. AVERAGE (F) (INS) READING CHANGE READING CHANGE CHANGE
55 82* 50 40 30 25 50 60 89 68 51 51 60 48 63
0.00 -0.63
-1.00 -1.00 -1.00 -1.00 -1.12 -1.14
6391 6395 6415 6366 6444 6484 6533 6534 6540 6563 6560 6553 6549 6552 6539
0 4 24 -25 53 93 142 143 149 172 169 162 158 161 148
6325 6351 6375 6405 6345 6339 6189 6171 6108 6041 6059 5980 5928 5902 5884
0 26 50 80 20 14 -136 -154 -217 -284 -266 -345 -397 -423 -441
0 15 37 28 37 54 3 -6 -34 -56 -49 -92 -120 -131 -147
CONSTRUCTION ACTIVITY
No concrete in the caissons Fresh concrete in the caissons Before cap poured After cap poured Before column poured After column poured Eight floors poured Starting to form the 9th floor Precast installed Windows installed Building completed Building completed Building completed Building completed Building completed
Geotechnical Materials: Measurement and Analysis : R.J. Krizek Commemorative Symposium: 3 Aug 2002
Figure 12 Caisson - Sister Bar Strain Gages 1911 South Indiana, Column B-6 Date 3-Oct-94 200
3-Apr-95
3-Oct-95
2-Apr-96
2-Oct-96
2-Apr-97
2-Oct-97
2-Apr-98
2-Oct-98 2.0 1.5 1.0 0.5
0
0.0 -0.5
-100
-1.0 -1.5
-200
-2.0 -2.5
-300
-400
-500
-3.0
Gage 8745 Gage 8746 Average Settlement
-3.5 -4.0 -4.5 -5.0
Settlement (Inches)
- = Compression Microstrains + = Tension
100
Geotechnical Materials: Measurement and Analysis : R.J. Krizek Commemorative Symposium: 3 Aug 2002
FIGURE 13 COLUMN STRAIN GAGES 1911 SOUTH INDIANA COLUMN C-2 INSTALLED NOVEMBER 7, 1994 CONCRETE STRENGTH OF COLUMN = 8,000 PSI
DATE
7-Nov-94 16-Nov-94 30-Nov-94 11-May-95 19-May-95 18-Jul-95 26-Sep-95 14-Mar-96 19-Aug-96 21-May-97 6-Nov-97 14-Apr-98
READINGS IN MICROSTRAINS TEMPERATURE GAGE NO. GAGE NO. GREY-2 AVERAGE DEGREES SETT. BLACK-2 (F) (INS) READING CHANGE READING CHANGE CHANGE
42 40 25 50 60 89 68 51 89 60 62 63
0
-1 -0.75 -0.88 -0.88 -0.88 -0.95 -0.95
3439 3457 3470 3145 3101 2950 2904 2892 2868 2856 2827 2832
0 18 31 -294 -338 -489 -535 -547 -571 -583 -612 -607
2980 3094 3103 2785 2740 2564 2508 2499 2448 2438 2376 2375
0 114 123 -195 -240 -416 -472 -481 -532 -542 -604 -605
0 66 77 -245 -289 -453 -504 -514 -552 -563 -608 -606
CONSTRUCTION ACTIVITY
Before the column was poured After the column was poured No change Eight floors poured Starting to form the 9th floor Precast installed Windows installed Building completed Building completed Building completed Building completed Building completed
Geotechnical Materials: Measurement and Analysis : R.J. Krizek Commemorative Symposium: 3 Aug 2002
Figure 14 Column Strain Gage Data 1911 South Indiana, Column C-2 Date 3-Apr-95
3-Oct-95
- = Compression Microstrains + = Tension
0
2-Apr-96
2-Oct-96
2-Apr-97
2-Oct-97
2-Apr-98
2-Oct-98 2.0
0.0
-200
-2.0
-400
-4.0
-600
-6.0
-800
-1000
Black - 2 Gray-2 Average Settlement
-8.0
-10.0
Settlement (Inches)
3-Oct-94 200
Geotechnical Materials: Measurement and Analysis : R.J. Krizek Commemorative Symposium: 3 Aug 2002
FIGURE 15 CAISSON - SISTER BAR STRAIN GAGES 1911 SOUTH INDIANA COLUMN C-2 INSTALLED OCTOBER 3, 1994 BOTTOM OF CAISSON = -65.66' B.D. *NOTE: CONCRETE TEMPERATURE
DATE
TEMPERATURE DEGREES (F)
3-Oct-94 03-Oct-94 01-Nov-94 07-Nov-94 16-Nov-94 30-Nov-94 11-May-95 19-May-95 18-Jul-95 26-Sep-95 14-Mar-96 19-Aug-96 21-May-97 6-Nov-97 14-Apr-98
68 83* 40 45 40 25 50 60 89 68 51 51 60 48 63
READINGS IN MICROSTRAINS GAGE 8747 GAGE 8748 SETT. ELEV -26.27 B.D. ELEV -26.27 B.D. AVERAGE (INS) READING CHANGE READING CHANGE CHANGE
0.00 -0.50 -0.75 -0.88 -0.88 -0.88 -0.91 -0.91
6508 6485 6438 6440 6459 6472 6499 6501 6509 6518 6510 6495 6471 6463 6455
0 -23 -70 -68 -49 -36 -9 -7 1 10 2 -13 -37 -45 -53
6308 6310 6433 6436 6438 6453 6346 6326 6271 6229 6223 6192 6181 6159 6173
0 2 125 128 130 145 38 18 -37 -79 -85 -116 -127 -149 -135
0 -11 28 30 41 55 15 6 -18 -35 -42 -65 -82 -97 -94
CONSTRUCTION ACTIVITY
No concrete in the caissons Fresh concrete in the caissons Before cap poured After cap poured Before column poured After column poured Eight floors poured Starting to form the 9th floor Precast installed Windows installed Building completed Building completed Building completed Building completed Building completed
Geotechnical Materials: Measurement and Analysis : R.J. Krizek Commemorative Symposium: 3 Aug 2002
Figure 16 Casisson - Sister Bar Strain Gages 1911 South Indiana, Column C-2 Date 3-Apr-95
3-Oct-95
2-Apr-96
2-Oct-96
2-Apr-97
2-Oct-97
2-Apr-98
2-Oct-98 2.00
150
1.50
100
1.00
50
0.50
0
0.00
-50
-0.50
-100
-1.00
Gage 8747 -150
Gage 8748 Average
-1.50
Settlement -200
-2.00
Settlement (Inches)
- = Compression Microstrains + = Tension
3-Oct-94 200
Geotechnical Materials: Measurement and Analysis : R.J. Krizek Commemorative Symposium: 3 Aug 2002
Table 2 1911 South Indiana Instrumentation Results as of April 14, 1998 Column No.
B6 (16”x36” col. 8000 psi conc. Over 5’ diam 4000 psi shaft) C2 (16”x48” col. 8000 psi conc. Over 6’ diam 4000 psi shaft)
Column Gage Avg (microstrains)
Calculated Column Load (kips)
Shaft Gage Avg (microstrains)
Calculated Shaft Load (kips)
Measured Settlement (inches)
687
1978
147
1496 (76 ksf)
1.25
606
2327
94
1377 (48 ksf)
0.9
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
REFERENCES 1.
Peck, R.B., "History of Building Foundations in Chicago", U of I Bulletin Series No. 373, January, 1948.
2.
Peck, R.B., and Uyanik, E.E., "Observed and Computed Settlements of Structures in Chicago", U of I Bulletin Series No. 429, 1954.
3.
Baker, C.N. Jr., Pfingsten, C.W., Gnaedinger, J.P., "History of Chicago Highrise Building Foundations 1948-1998".
4.
Baker, C.N., Jr., "Use of Presssuremeter in Mixed High Rise Foundation Design", ASCE Special Bulletin No. 38, 1993.
5.
Terzaghi, K., Peck, R.B., Mesri, G., "Soil Mechanics in Engineering Practice", John Wiley & Sons, 1996.
6.
Baker, C.N., Jr., Drumright, E., Joseph, L.M., Azani, T., "Foundation Design and Performance of the World's Tallest Building" Proceedings at Fourth International Conference on Case Histories in Geotechnical Engineering, St. Louis, MO, March 9-12, 1998.
7.
Baker, C.N., Jr., "Measured vs Predicted Long Term Load Distribution in Drilled Shaft Foundations", 6th Annual Great Lakes Geotechnical and Geo-environmental Conference, Indianapolis, IN, 1998.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
LESSONS LEARNED FROM A BERMED EXCAVATION IN SOFT CLAY H. J. Liao Department of Construction Engineering, National Taiwan University of Science and Technology 43, Section 4, Keelung Road, Taipei, Taiwan
[email protected]
ABSTRACT A two-level underground car park was to be constructed in soft clay. Due to its large area and irregular shape, the commonly used internal bracing method was considered inadequate to support this excavation project by the designer. Instead, a bermed excavation method together with sheet pile wall were chosen to support this excavation. In other words, neither bracing struts nor tieback anchors were used to counter balance the earth and groundwater pressures from around the excavated area. However, this aggressive supporting system was unable to control the ground movement during excavation. As a result, the supporting system for this excavation were changed twice to complete this project. Grout piles were added to strengthen the berm during the first change. But it still failed to arrest the ground movement. Then, it was decided to reduce the size and depth of excavation area of the second basement level. Although this project was finally completed, it was about six months behind schedule. This paper will discuss the problems encountered and lessons learned from this bermed excavation project. INTRODUCTION A two-level underground car park under a gymnasium was to be constructed in a thick soft clay layer with high groundwater table. A tight schedule was set for this project because the county government wanted this project to be completed before the next county governor election. To shorten the construction time and take into account the large excavation area and irregular shape, an aggressive bermed excavation method together with sheet pile wall were chosen to support this underground excavation by the designer. However, local slope failures were soon observed on the 1:1 (horizontal : vertical) berm after a torrential rainfall when the excavation carried to a depth of 3 m. A maximum of 95 mm lateral ground movement was recorded by the inclinometer installed just outside the sheet pile wall. After reassessing the suitability of this bermed excavation method and the time left to complete this project, it was decided to minimize the change of the excavation supporting system by adding grout piles at improvement ratios (Ir) of 20% and 40% (Fig. 1). The ground improvement covered 8m to 10m wide area from the excavation wall and the depth of grout piles extended from GL. –2.5m to GL. –20m (Fig. 2). Totally, about 2000 grout piles with a diameter of 0.8m were constructed by deep mixing method (Fig. 3). The ground
Department of Civil and Environmental Engineering Infrastructure Technology Institute Northwestern University, Evanston, Illinois, 60208
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Fig. 1 Plain view of this excavation project after the first design change
8 ~10 m -2.5 m
-11.4 m
Ground improved area I r = 20% or 40%
Sheet pile wall
-18 m -20 m Fig. 2 Schematic diagram of grout pile reinforcement for this bermed excavation
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Fig. 3 Deep mixing method used in this project movement around this grout piles strengthened bermed excavation was estimated in advance with FEM analysis. According to the calculated results, the maximum lateral wall movements of this project were about 8.45 cm (Ir = 40%) and 10.6 cm (Ir = 20%). The estimated maximum ground surface settlement occurred right outside the excavation wall was 5.94 cm (Ir = 40%)and 2 cm across a 20 m wide street. Such surface settlement was not expected to cause damage to the adjacent buildings. However, internal failure within the grout piles reinforced berm was observed some time after the excavation reached the depth of GL –7.5m. In addition, progressive toppling failure of grout piles within the berm was observed also. Gradually, the berm lost its stability and caused excess surface settlement around the excavation. The lateral ground movement monitored on one side of the excavation increased to an amount of 390 mm. Excavation was terminated and backfilled to GL –4m. The suitability of the excavation supporting system was evaluated again. To arrest the inward movement of excavation wall, the following measures were taken: (1) reduce the area of second basement level to leave sufficient soil around the excavation wall in place, (2) raise the bottom of the mat foundation from GL –10.15 m to GL -8.65 m at the center area (second basement level) and to GL –4.15 m around the center area (first basement level) (Fig. 4), and (3) install soldier piles to support the excavation of second basement level. After these major changes on the excavation plan, the ground movement around the excavation was controlled and the basement excavation was completed safely without causing any damage to the adjacent buildings. But it was finished about six months behind schedule. This paper will summarize the problems encountered and the lessons learned from this project.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Fig. 4 Plain view of this excavation project after the second design change LESSON 1: EXPERIENCE COULD BE MISLEADING The designer of this underground car park was a structural engineer who was relatively inexperienced in geotechnical engineering design. But the designer just had some experience in designing a rather successful bermed excavation. Neither stability problem nor excess ground movement was reported from that project. The recent experience led the designer to choose bermed excavation method for this project because the size and shape of this excavation project made braced excavation difficult and expensive to carry out. Despite there were similarities between these two projects, there also existed some differences which were overlooked by the designer: the subsoil conditions and the watertightness of excavation wall. As shown in Fig. 5 , this basement excavation was almost entirely carried out in the soft clay layer (Su = 20 ~ 30 kPa). In comparison, only a 3 m thick soft clay layer (Su ~ 35 kPa) was located above the bottom of the excavation and a 3 m thick silty sand layer underlying it for the other project. And there was located right above the. Moreover, the slope of the berm was 1:1 (horizontal : vertical) for this project compared with 1.5 : 1 for other project.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
first stage excavation (shotcreted) center post
1 3.0m
backfill
center post
GL.+0m
1 sturt GL.-3.5m
8.6m
H-pile@1m L=20m silty clay s u = 20 ~ 30 kPa
GL.-7.0m
jet grouted pile sheet pile L=18m 2.1m
26.3m
silty sand N = 10
second stage excavation
final stage excavation (shotcreted)
gutter
GL.-10m
gutter
silty clay
This excavation project first stage excavation (shotcreted) 3.5m
backfill
3.0m
silty clay
1 1.5
s u ~ 35kPa 3.0m
silty sand N = 6± 4
diaphragm wall L=19m t=0.7m
GL.-10m
silty clay 9.4m
s u ~ 40 kPa
Other excavation project Fig. 5 Comparison on subsoil and construction conditions of this excavation project and other project Due to time and budget limitations of this project, sheet piles were installed at two different spacings to support soil around the excavation site. When there were buildings nearby, sheet piles were interlocked; otherwise sheet piles were installed at 1 m spacing. To enhance the watertightness of the excavation wall, one row of jet grout pile was installed along the perimeter of the site. However, any experienced geotechnical engineers would realize that the watertightness of one row of jet grout pile is not accountable. Groundwater was likely to seep through sheet piles from the surrounding area into the berm and caused internal failure within the berm. In comparison, a watertight diaphragm wall was used as the excavation wall on the
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
other project and the surface of berm was covered with shotcrete also. No water was able to seep into the berm. The stability of the berm was secured. Obviously, the experience learned from other bermed excavations could be misleading if the differences between projects were not noticed and carefully studied. LESSON 2: IS NUMERICAL ANALYSIS A VIRTUAL DREAM? The development of numerical analysis and its application to geotechnical problems has provided geotechnical engineers with a powerful analysis tool. But is it a virtual dream or practical reality (Potts, 2002)? To evaluate the stability of the berm and the lateral movement of the sheet pile wall in response to excavation, the performance of grout piles reinforcement of this bermed excavation was also evaluated numerically. The stability of berm was analyzed by the PCSTABL program; the lateral movement of sheet pile wall was analyzed by finite element method. However, the strength parameters of grout piles reinforced clay to be input to the numerical analysis are not quite certain. For convenience, the composite strength (Sum) of grout piles reinforced soil was approximated by the following equation:
S um = (1 − I r )S u + αI r qu / 2
[1]
where Ir = improvement ratio, Su = undrained shear strength of soil, qu = unconfined compressive strength of grout pile, α=strength reduction factor for grout pile (= 0.35 for this case). Substitute the composite strength determined from Eq. 1 to PCSTABL program, the calculated safety factors for the grout piles strengthened soil body shown in Fig. 2 against sliding were equal to 2.7 (Ir = 20%) and 4.7 (Ir = 40%) respectively. The lateral wall displacement of this excavation project was estimated based on a two-dimensional plane strain finite element program which was modified from Zienkiewicz’ program (1977). The process of excavation was simulated by adding the equivalent nodal forces on the excavated boundary. The equivalent nodal forces resulted by the removal of soil were computed following the method proposed by Ghaboussi and Pecknold (1984). Beam element was used to simulate the behavior of excavation wall (sheet pile). Both the soil and the excavation wall were simulated with eight-noded quadrilateral isoparametric (Q8) elements. Excavation wall was assumed to behave as a linear-elastic material, for which both Young’s modulus and Poisson’s ratio were assumed to be constant. Soil was assumed to behave as an elasto-plastic material. The hyperbolic model proposed by Duncan and Chang (1970) and the modified hyperbolic model proposed by Hsieh and Ou (1997) were adopted to represent the stress-strain behaviors of sandy soil and clayey soil, respectively. Hsieh and Ou’s model can take into account the effect of Ko consolidation, strength anisotropy, and rotation of principle stresses during excavation. Undrained behavior (φu = 0) was assumed for the clayey soil. The soil parameters input to the numerical study are listed in Table 1. Among them, the ratio of Ei/Suc for the clayey layers of this excavation site was assumed to be equal to 1100, which was close to the lower bound of Ei/Suc ratio for Taipei silty clay. Ei is the initial tangential modulus; Suc is undrained shear strength obtained from the CIUC triaxial test. The elastic modulus of sandy layer (K) is determined empirically from the shear wave velocity:
K=
ρVs2 (1 + ν ) Pa
[2]
where ρ is the mass density of soil (t/m3); Pa is the atmospheric pressure (= 101.4 kPa); v is the Poisson’s ratio (= 0.3 for this case); Vs is the shear wave velocity in soil (m/sec). For the Taipei silty soil, the shear wave velocity can be calculated empirically from the SPT-N value : Vs = 65.58 N0.502
[3]
During the finite element analysis, the depth of meshes was extended to 40 m. A hinge boundary was used as the boundary constraint condition at the bottom and a roller boundary was used at the sides. The side boundary of the meshes extended to a distance 100 m from the excavation wall.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Soil Depth (m) Layer
γ (t/m3)
φ (deg)
Ka
Kp
Su/σv'
Active Passive Side Side Friction Ratio Friction Ratio
1
3.0
1.80
29.8
1.0
1.0
_
0.5
0.5
2
11.6
1.88
0
1.0
1.0
0.32
0.5
0.5
0.5
0.5
3
13.7
1.96
29.6
1.0
1.0
_
4
20.6
1.97
0
1.0
1.0
0.32
0.5
0.5
5
22.9
2.05
0
1.0
1.0
0.32
0.5
0.5
Table 1 Soil parameters input to numerical study Note:
Friction ratio = cw/Su for clayey layer; = δ/φ for sandy layer cw = adhesion between clay and wall, Su = undrained shear strength of clay, δ = friction angle between sandy soil and wall, φ = friction angle of sandy soil.
The construction sequence simulated in the finite element analysis was as follows: Stage 1. Excavate to the depth of 3 m. Stage 2. Grout piles installed around the excavation wall. Stage 3. Excavate to the depth of 5 m. Stage 4. Excavate to the depth of 8 m. Stage 5. Excavate to the final depth of 11.40 m. The lateral movement of the excavation wall (Ir = 40%) calculated from the above FEM analysis was compared with inclinometer readings taken at different dates (Fig. 6). Since there is no internal bracing used to support to the excavation wall, a cantilever type inward wall movement was observed. Although the calculated results were close to the inclinometer readings taken right after the excavation reached GL –7.5m, the inclinometer readings increased further to an amount of 390 mm over a period of 3 and a half months. They were much greater than the calculated values.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Displacement (cm) 40
30
20
10
0
0
5
Depth ( m)
10
15 99/03/06 20
99/03/28 99/04/05 99/05/22
25
99/05/31 99/06/21 Prediction(Ir=40%)
30
Fig. 6 Comparison of FEM calculated and field monitored lateral wall movements at different dates The parameters input to the FEM analysis were based on the field monitored data gathered from several braced excavations in Taipei Basin. Obviously, these parameters were not applicable to the bermed excavation if the stability of berm was not secured. As observed from the field, clay and grout piles did not response as a real composite material to the lateral unloading. Instead, clay on the outer surface fell off first due to unloading induced tension cracks and the effect of groundwater seepage; then the grout piles toppled. This phenomenon triggered the progressive failure within the grout pile reinforced berm and caused the berm to lose its ability to resist the inward movement of excavation wall. As a result, a large wall movement and excess surface settlement occurred. LESSON 3: CAN THE STABILITY OF BERM BE INCREASED BY GROUT PILES? Frequently, the strength of soft clay is increased by consolidation. But consolidation takes long time. To increase the strength of soft clay in a shorter time, grout pile reinforcement method is one of the most commonly used methods in geotechnical practice. But to effectively increase the overall strength of grout piles reinforced soft clay, grout piles and clay must behave as a composite material. In other words, grout piles must not separate from soft clay when subjected to external loading or unloading. In general, this requirement is usually met when the grout piles reinforced clay is subjected to vertical and/or lateral loading, such as the grout piles installed under the foundation or beneath the bottom of excavation (Fig. 7). However, because of the difference in deformability between grout pile and clay, grout pile may be separated from clay when the grout pile reinforced clay is subjected to lateral unloading, such as the berm for an excavation (Fig. 8). When the grout piles are separated from clay (Fig. 9), they may be toppled under lateral loading. Groundwater seepage through the berm may worsen the situation. So, grout pile reinforcement method does not seem to be an effective measure to increase the stability of an unstable berm in soft clay. If it is to be applied to a stable berm, a suitable (say 30% or less) strength reduction factor α should be used to obtain a reasonable composite strength (Sum) for grout piles reinforced berm from Eq. 1.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Struts
Excavation wall Grout piles to be excavated Grout piles left in the ground
Fig. 7 Schematic diagram of grout pile reinforcement in braced excavation (axial unloading and lateral loading)
Berm Grout piles Excavation wall Fig. 8 Schematic diagram for grout piles reinforcement in bermed excavation (lateral unloading)
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Fig. 9 Separation of clay and grout pile due to difference in movement
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
LESSON 4: ARE THE CORED SPECIMENS REPRESENTIVE TO GROUT PILES? The integrity and strength of grout piles installed in this project were examined by the cores drilled from the grout pile at a frequency of 1 out of 100. Totally, 20 grout piles were cored and up to 400 core specimens were tested. The nominal diameter of cored specimens was 5 cm. Typical test results for a grout pile are shown in Fig. 10. In general, the unconfined compressive strengths of all the core specimens were higher than the design strength (2 MPa) and with an average of 4.8 MPa. Usually, the specimens tested for the unconfined compressive strength were chosen from the core specimens which had sufficient length (more than 10 cm) and with better integrity. The average RQD value for 20 grout piles was 78.4%. In general, the compressive strengths obtained from these “good” core specimens tended to be higher than other core specimens. So, the unconfined compressive strengths shown on the test report would be higher than the compressive strength. which can be actually mobilized in the field, and should only be treated as the upper bound compressive strength of the grout pile. Field experience has shown that it will not be easy to recover core specimens by the core drilling method if the strength of grout pile is less than 1 MPa. For all the core specimens drilled from this site, the average core recovery ratio was about 93%. It indicated that about 7% of the cored specimens were lost during coring and they might have the unconfined compressive strength less than 1 Mpa. For grout piles without reinforcement, these weak spots within the pile will result a much lower resistance to lateral loading or shearing than axial loading. In other words, the resistance of grout plies to the bermed excavation induced lateral ground movement will be much less than that estimated from the unconfined compressive strength of cored specimens. 100
100
RQD
80
80
60
strength
60
40
40 water content
20
ω ( %) RQD (%)
strength (kg/cm 2 )
120
20
0
0 1
2
3
4
5
6
7
8
9
10
11
core speci men Fig. 10 Typical test results determined from cored specimens of a grout pile CONCLUSIONS Based on the lessons learned from this bermed excavation project, the following conclusions can be advanced: 1.
Berm excavation method is not suitable for excavation in soft clay, where the stability of berm cannot be secured. To enhance the stability of berm, a watertight excavation wall is needed to cut off groundwater from seeping through the berm.
2.
The internal failure occurred progressively within the grout piles reinforced berm will cause the berm to lose its ability to support the excavation wall. However, such a phenomenon was not easy to be reflected in the parameters input to numerical analysis. Without taking this phenomenon into account, numerical analysis will overestimate the stability of the berm.
3.
The composite undrained shear strength of grout piles reinforced clay varies with loading conditions. When subjected to lateral unloading, the mobilized shear strength can be much less than that subjected to lateral and/or
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
axial loading. So, a smaller strength reduction factor α should be used to obtain the composite shear strength from Eq. 1. 4.
The unconfined compressive strength determined from the cored specimens of grout piles will overestimate the strength actually mobilized in the field. So, it should be treated only as the upper bound compressive strength of grout pile.
REFERENCES 1.
Duncan, J. M. and Chang, C. Y., 1970, Nonlinear analysis of stress and strain in soils, Journal of the Soil Mechanics and Foundations Division, ASCE, 96(5), 637-659.
2.
Ghaboussi, J. and Pecknold, D. A., 1984, Incremental finite element analysis of geometrically altered structure, International Journal for Numerical Methods in Engineering, 20, 2051-2064.
3.
Hsieh, P. G. and Ou, C. Y., 1997, Use of the modified hyperbolic model in excavation analysis under undrained condition, Geotechnical Engineering, SEAGS, 28(2), 123-150.
4.
Hsieh, P. G. and Ou, C. Y., 1998, Shape of ground surface settlement profiles caused by excavation, Canadian Geotechnical Journal, 35, 1004-1017.
5.
Potts, D. M., 2002, Numerical analysis – A virtual dream or practical reality, 42nd Rankine Lecture, Imperial College, London..
6.
Zienkiewicz, O. C., 1977, The Finite Element Method, 3rd edition, McGraw-Hill, London.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
A NEW SLOPE STABILITY APPROACH USING CALCULUS OF VARIATIONS, AND SAFETY AND SENSITIVITY ANALYSIS Enrique Castillo 1 and Roberto Mínguez 2 Departament of Applied Mathematics and Computational Sciences, University of Cantabria Avda. de Los Castros s/n, 39005 Santander, Spain
[email protected] Departament of Applied Mathematics, University of Castilla-La Mancha Avda. Camilo José Cela nº 2, 13071 Ciudad Real, Spain
[email protected]
ABSTRACT The paper presents a new method for slope stability analysis that incorporates calculus of variations and safety and sensitivity analysis. First, the calculus of variations technique is used to determine the worst sliding line using a discretization method that is stated as an optimization primal problem. Then, different design strategies are described: (a) the classical, based on safety factors, (b) the modern, based on failure probabilities, and (c) a new combined strategy, in which the reliability index is determined and combined with the safety factor method in such a way that the most stringent condition prevails. Finally, a marginal sensitivity analysis is performed by using the dual variables of the dual problem associated with an artificially created primal optimization problem. PROLOGUE This paper has been written on the occasion of the 70th birthday of Prof. Raymond J. Krizek, to whom the authors of this paper are deeply indebted. E. Castillo, who was fortunate in having Prof. Krizek as his Ph. D. thesis advisor, an enriching event that transformed his life, wants to express, with this paper, his recognition and gratitude to Prof. Krizek, who will survive through all his many disciples, around the world, of the first and successive generations. Roberto Mínguez is one example of this. He does not know personally Prof. Krizek, but he is aware that the work he is doing, including this paper, would not had been possible without Prof. Krizek's dedication and work. Special mention and recognition must also be given to Claudia, her wife, Raymond's loyal companion, in the good and in the hard times. 1. INTRODUCTION There are two main different ways of dealing with safety: (a) the classical approach, based on safety factors, and (b) the modern approach, that uses probabilities of failure. The classical approach is questioned because is does not give a clear idea of how far we are from failure, and a design based on failure probabilities is demanded (see Ditlevsen and Madsen (1996), Wirsching and Wu (1987), and Wu et al. (1989)). However, the modern approach is also criticized because of its sensitivity to the probability assumptions, specially to tail behavior (see Galambos (1987), Castillo (1988), and Castillo et al. (1996a,1996b)). In this paper we show how to use both methods using calculus of variations, and we introduce a new combined method called the safety-factor-failure-probability method. A standard classical stability analysis consists of determining the safety factor associated with a collection of sliding lines selected by the engineer, and choosing the one leading to the minimum safety factor, that can be taken as the safety factor of the slope (see Schuster and Krizek (1978)). The set of tentative sliding lines is normally a parametric family.
Department of Civil and Environmental Engineering Infrastructure Technology Institute Northwestern University, Evanston, Illinois, 60208
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
However, since this trial procedure does not cover the whole set of possible sliding lines, this process is not guaranteed to lead to the minimum value. This is not a problem if the resulting value for the safety factor is close enough to the actual minimum, but can be a problem if it is far from it. Alternatively, a non-parametric approach can be used. The most important and powerful non-parametric approach is that based on calculus of variations, that has been applied to soil mechanics in the past (see, for example, Chen and Giger (1971), Chen and Snitbhan (1975), Baker and Garber (1977a, 1977b, 1978), Revilla and Castillo (1977), and Castillo and Luceño (1982, 1983)). This technique is a generalization of the problem of maxima and minima, in which the maximum or minimum of functionals, instead of functions, is looked for. A functional is an application of the set of functions y(x) on the set of real numbers. In the slope stability problem the function y(x) represents the sliding line, and the real number is the corresponding safety factor. So, the functional here associates to each sliding line its safety factor. In most approaches, the functional can be written as the quotient of two integrals, that represents the ratio of stabilizing to sliding forces or moments. As pointed out by Cornell (1971), a system's reliability is that of all potential slip surfaces, and the failure probability of a system will be larger than the failure probability for a single slip surface. The difference between these two probabilities will depend on the correlation between the failure probabilities of the different potential slip surfaces. The calculus of variations approach has clear advantages with respect to other methodologies, because: (a) in a classical approach, it is guaranteed that the worst sliding line is obtained, i.e., no other line leads to a smaller safety factor, and (b) in a modern approach, the system reliability is obtained, since the worst sliding line for each combination of parameters, is dealt with. Finally, once a slope has been designed, it is interesting to perform a sensitivity analysis to know, how and how much, the different model parameters influence the safety. In this paper we propose a method for performing such an analysis. The paper deals with a variational approach of the slope stability problem, and discusses how to deal with it in three different ways, the classical, the modern and the mixed approaches. In Section 2 the calculus of variations method for slope stability is discussed. In Section 3 and 4 it is described how a classical and a modern engineer can solve the slope stability design problems using calculus of variations. In Section 5 the mixed safety-factor-failure-probability method is presented. Section 6 deals with sensitivity analysis. Finally, Section 7 gives some conclusions. 2. CALCULUS OF VARIATIONS APPROACH Revilla and Castillo (1977), Castillo and Revilla (1977), or Castillo and Luceño (1978, 1980, 1981), based on the Janbu method (see Janbu (1954)), proposed the following functional:
x n +1
∫
F=
x1
c 2 + ( y ( x) − y ( x)) tan φ (1 + y ' ( x)) γ dx y ' ( x ) tan φ 1+ F
[1]
x n +1
∫ ( y( x) − y( x)) y' ( x)dx
x1
where F is the safety factor, y (x) is the slope profile (ordinate at point x ), y(x) is the ordinate of the sliding line at point x, c is the cohesion of the soil, φis the angle of internal friction of the soil, γ is the unit weight of the soil, H is the slope height, and x1 and xn+1 are the coordinates of the sliding line end points. The engineer must be careful when selecting the functionals for slope stability. In fact, Castillo and A. Luceño (1983) showed that some functionals that have been proposed in the past are unbounded, and then not valid for design. In particular, they proved that other functionals, as (1), are bounded, and consequently, valid for slope design.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Note that (1), for a given slope profile y (x) , relates five important variables φ, c, γ, H and F, i.e., we are dealing with a 5dimensional space. Of course, we can work in this space, but this complicates things unnecessarily and hides the deep structure of the slope stability problem. Dimensional analysis, by means of the π-theorem, reveals us that expression [1] can be written in terms of the three nondimensional variables in the set
{F , N =
c ,ψ = tan φ} γH
[2]
as u n +1
∫
F=
[ N + ( z(u) − z(u))ψ ] (1 + z' 1+
u1
2
(u ))
z ' (u )ψ F
du
[3]
z n +1
∫ ( z (u ) − z(u)) z ' (u)du
z1
where u=
y (uH ) y (uH ) x ; z (u ) = ; z (u ) = . H H H
[4,5,6]
are the non-dimensional x coordinate, the non-dimensional slope profile and sliding line, respectively. This allows us to consider a simpler 3-dimensional space, and reveals that, as long as the non-dimensional values N and ψ in [2] remain constant, the safety of the slope, and the non-dimensional slope profile and sliding line remain the same. For example, the safety factors associated with two slopes with Ψ=0.6, c=20 kN/m2, γ=21kN/m3 and H=10 m, and ψ=0.6, c=2 kN/m2, γ =21kN/m3 and H=1 m are identical, though the slope profile and the sliding line have been applied the same scale change as the H
F = H ( N ,ψ ).
[7]
Expression (3) shows that given z (u ) and z(u), there is a relation between F, N and ψ. However, we can go even further, if we realize that in expression [3], N, ψ and F can all be multiplied by any constant k, and the same relation still holds, i.e., kF = kH ( N ,ψ ) = H ( kN , kψ )
[8]
in other words, it satisfies a functional equation, known as the functional equation of the homogeneous functions, which general solution can be written as1 (see Aczél (1966), Castillo and Ruiz (1992)) F
ψ
1
= h(
N
ψ
) ⇔ F * = h( N * )
It is assumed that H(.,.) function really depends on both variables N and ψ.
[9]
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
This reveals that our problem can be stated in terms of only two non-dimensional variables F*=F/ψ and N*=Nψ. In fact, (3) can be written as
u n +1
∫
1=
[N
*
]
+ z (u ) − z (u ) (1 + z ' 2 (u )) du
F * + z ' (u )
u1
[10]
z n +1
∫ ( z(u) − z(u)) z' (u)du
z1
Of course, in our presentation of the slope stability problem we could have started with [10] instead of [1], but this decision would have hidden our readers the beauty and deep concepts below non-dimensional analysis, on one side, and functional equations, on the other. The final important conclusion is that our problem is a 2-dimensional problem in nature. Then, it can be stated as: Minimize F* with respect to u1, un-1 and z(u) subject to [10] and N*=N0. Solving this problem for different values of N0, a plot relating F* and N* can be drawn, and then the safety coefficient F can be obtained from the plot for any values of the data variables φ, c, γ and H. The calculus of variations implementation of the slope stability problem consists of minimizing F* with respect to u1, un-1 and z(u). This problem can be solved analytically. However, it leads to complicated differential equations and transversality conditions (see Luceño (1979)). The obvious alternative consists of a numerical solution, that is the option selected in this paper.
vn-1
zn
vn
zn+1=zn+1
zn-1
vn z3 z v2 z1=z1 v1 2 v1 z2
zn-1
v2
vn-1
z3
∆u u1
zn
u2
u3
un-1
un
un+1
Figure 1: Illustration of the slope stability problem, showing the sliding line and associated discretized polygonal. 2.1-. DISCRETIZATION OF THE PROBLEM. Using the following formulas (see Figure 1):
∆u =
un +1 − u1 z +z ; ui = u1 + (i − 1) ∆u; vi = i +1 i ; 2 n z −z v 'i = i +1 i ; vi = zi (ui + ∆u / 2); ∆u
[11,12,13] [14,15]
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
the equation (10) can be easily discretized as n
1=
∑
[N
*
]
+ v i − v i (1 + v' i 2 ) ∆u F * + v' i
i =1
[16]
n
∑ (v
i
− v i ) v ' i ∆u
i =1
Then, the slope stability analysis reduces to Minimize F *
[17]
u0 ,u n +1 , zi ;i = 2,...,n
subject to n
∑ i =1
(v i − v i )v ' i −
[N
*
]
+ (v i − v i ) (1 − v' i 2 ) = F * + v' i zj = N
*
0
[18]
z j ; j = 1, n + 1
[19]
N 0*
=
= c /(γHψ )
[20]
The important thing of the calculus of variations method is that for each combination of soil parameters φ, c, γ, and slope H, it gives the corresponding critical sliding line and safety factor F i.e., it informs us about whether (F <1) or not (F >1) the failure will occur, and through which sliding line. In other words, it informs us about the behavior of the system and not of particular sliding lines. The reader must realize that the study must pass through the non-dimensional parameters F+=1/ψ and N*. z 0.5 z (u)
arctan(π u) π
-1
1
u
1
-0.5
Figure 2: Selected slope profile for the illustrative example. 3. THE CLASSICAL APPROACH In this section we solve the slope stability problem based on safety factors, with no consideration of the statistical properties of the parameters involved, i.e., we assume that all parameters are deterministic.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
3.1.-DERIVATION OF THE STABILITY RELATION BETWEEN N* AND F*. As one example, in this section we assume a slope with profile (see Figure 2)
y ( x) = H arctan(πx / H ) / π ; −∞ < x < ∞ that implies
z (u ) = arctan(πu ) / π ; −∞ < u < ∞ Solving the stability problem [17]-[20], for all possible values of the non-dimensional number N*, a plot that allows determining the safety coefficient for any value of N* can be obtained. This plot is shown in Figure 3, where the optimal values of F* as a function of N* (F* =h(N*)) are shown. The corresponding sliding lines for different values of N* are also given in Figure 4. The graph in Figure 3 allows solving any slope stability problem by the classical approach. 3.2.- CLASSICAL SOLUTION OF THE STABILITY PROBLEMS. The problem of slope design can arise in several different forms, as: Classical design problem 1: Obtain the maximum allowable N* for given ψ and safety factor F. In this case we calculate F*=F/ ψ, enter the graph in Figure 3 and determine N*. Once N* is known, we can calculate the value of H in terms of the soil parameters c, γ, ψ, as H=c/(N*γψ), or the value of c, in terms of H, γ and ψ, as c=N*γψH. Example 1: Assume that the desired safety factor is F=2 and φ=10º, so 1/ ψ =1/tan(φ)=5.67. In this case we calculate F*=2x5.67=11.34, enter the graph in Figure 3 and obtain N*=1.27. If c=17 kN/m2 and γ =22.8 kN/m3, then the height of the slope must be H=17x5.67/(1.27x22.8) =3.33 m, or, if H=5 m, γ=22.8 kN/m3 and ψ=1/5.67=0.176, then, the design value becomes c=1.27 x 22.8 x 5 x 0.176=25.48 kN/m2. Classical design problem 2: Obtain the required ψ for given height H, γ, c and safety factor F. In this case we need to use a graphical method . We know that tan(β)=F*/N*=F/N, so we trace a line from the origin with slope β. The intersecting point with the curve F*=h(N*) is the solution of the problem (see Figure 3). To avoid this method, the graphic in Figure 3 can be transformed into the equivalent one in Figure 5, where now the single relation between F* and N* is transformed into many relations between ψ, N and F.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
To solve the problem we enter with N=c/(γH) and the given F in Figure 5, and determine the resulting value of ψ.
20 z (u) 18
16
N
ψ
arctan (π u) π c γH
tan (φ )
F ψ
-1
u
1
1
-0.5
14
F*
z 0.5
12 10 4
8
3
6
2 4 1 2
0
β
0
0.5
0.2
0.1
1
1.5
N*
2
2.5
N ψ
Figure 3: The optimal values of F* as a function of N* and enlarged region close to origin. Example 2: Assume that H=10 m, γ =22.8 kN/m3, c=17 kN/m2 and F=1.5. If we enter the graph in Figure 5, with N=17/(22.8x10)=0.15 and F=1.5, we obtain ψ =0.17. If we draw the line with slope tan(β)=F/N =1.5/0.15=10 we obtain, in Figure 3, F*=8.8 so ψ = F/F*= 1.5/8.8= 0.17. Classical design problem 3: Obtain the safety factor F for given H, γ, c and ψ. In this case we calculate N*=c/(γ H ψ), enter the graph in Figure 3, determine F* and calculate the value of F= F* ψ. 0.5 N * = 0.003 0
3
-0.5
z arctan (π u) π N
z( u)
-1
N*
ψ
-1.5
ψ tan (φ )
-2 -1
0
1
u
Figure 4: Sliding lines for different values of N*.
2
3
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Example 3: If H =10 m, γ =22.8 kN/m3, c=17 kN/m2 and ψ =0.7673, we calculate N*= 17/(10x22.8x0.7673)= 0.097, enter the graph in Figure 3, obtain F*=2.8 and finally get the value of F=2.8x0.7673=2.15. 4-. MODERN APPROACH In this section we solve the stability problem using failure probabilities (Alonso(1976)). 2.5 z (u) N
2
ψ
arctan ( π u) π c
z 0.5
γH
tan (φ )
-1
u
1
F= 8
1
-0.5 F= 7
Failure region 1.5
ψ
F = 1.50 1
F= 6
F = 1.75 F= 5 F= 2
0.5
F = 2.50
F= 4 F= 3
F= 1 F = 1:25 0 0.1
0.2
0.3
0.4
0.5
N
Figure 5: The optimal values of F as a function of N. Contrary to the assumption in the previous section, we assume here that all the variables involved are random variables. Since the non-dimensional failure probability problem depends on only two variables F+=1/ψ and N*, in fact we can work only with these two variables. Note that, since the failure condition is F=1, F+ is not F*. Perhaps one is temptated to work with the random variables γ, c, φ and H, and make assumptions about the joint density of this set of variables. However, as it has been shown, the slope problem depends only on the joint density of F+ and N*. Thus, no matter which joint density be selected for γ, c, φ and H, the final result will be dependent only on the resulting joint density of (F+,N*). In other words, without loss of generality, the joint density of (F+,N*) can be assumed. 4.1.- CALCULATING THE RELIABILITY INDEX OF A GIVEN DESIGN. To simplify our treatment we assume that F+ and N* are normals, but dependent, because the ψ parameter appears in both F+ and N*. More precisely we assume that: F + ≈ N ( µ F + , σ F2 + ) ; N * | F + ≈ N ( µ N * + β 0 ( F + − µ F + ), σ N2 * |F + ) ;
where µF+ and µN* are the means of F+ and N*, respectively, σF+2 is the variance of F+, N*|F+ is N* conditioned on F+, σN*|F+2 is the variance of N* given F+, and β0 is the regression coefficient of N* on F+. In the following we assume that µF+=F+0 and µN*= N*0, i.e., they are equal to the classical design values F+0 and N*0, respectively.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
As it is customary, to calculate the probability of failure we proceed to transform the initial set of random variables into a set of independent standard normals using the transformation (see Rosenblatt (1952), Ditlevsen and Madsen (1996) and Madsen et al.(1986)): ZF +
=
F + − µF +
=
F + − µF +
σF+ vF + µ F + N − µ N * − β0 (F + − µF + ) N * − µ N * − β0 (F + − µ F + ) = σ N * |F + vN * | F + µ N * *
ZN*
=
[21]
where vF+ is the coefficient of variation of F+, and vN*|F+ is the coefficient of variation of N*, given F+. Then, calculating the reliability index β reduces to:
MinimizingZ
F+
,Z
N*
β = ZF + 2 + ZN* 2
[22]
subject to
ZF +
=
F + − µF +
=
F + − µF +
vF + µ F + σF+ N − µ N * − β0 ( F + − µF + ) N * − µ N * − β0 ( F + − µF + ) = σ N * |F + vN * | F + µ N * *
ZN*
=
F+
= h( N * )
[23]
It is important to mention that we are solving the problem in non-dimensional form. This means that an infinite set of problems are solved at once. To conclude in a real case, we need to use the resulting non-dimensional values obtained after this process and recover the dimensional parameters for our particular case. In the following we present two different methods: the approaches 1 and 2 below. 4.1.1.- APPROACH 1. Unfortunately, no close expression (F+=h(N*)) for F+ in terms of N* is available. Thus, one way to overcome this difficulty consists of obtaining (using least squares, least absolute values, the maximin method, etc.) an approximate analytical expression F*=h’(N*), for example, a polynomial of degree m. The following polynomial of degree 8 gives a good approximation F+
= 1.35 + 18.19 N * − 46.6 N *2 + 111.26 N *3 − 155.56 N *4 + 127.52 N *5 − 60.23N *6 + 15.14 N *7 − 1.565 N *8
4.1.2.- APPROACH 2. Since the failure region, that is determined by solving the optimization problem [17] –[20], has not a closed analytical expression, the problem has to be solved using an iterative scheme, as the one given in the following algorithm, based on the bisection method: Algorithm 1: Calculating the reliability index β Input: The statistical parameters of the problem µF+, µN*, vF+, vN*|F+ and β0, and an error bound ε to control the convergence of the process. Output: The non-dimensional design values N*, F+ and the reliability index β
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Step 1: Choose two initial values, one very small N*1=0.001 and one large N*2=5, and calculate the associated failure values F+1 and F+2, respectively, by solving the optimization problem [17] – [20] once for each. This is equivalent to entering the graph in Figure 3. Step 2: Calculate the values of β1 and β2 associated with the sets ( N*1, F+1) and ( N*2, F+2), respectively, using Expressions [21] and [22]. Step 3: If |β1- β2|<ε, go to Step 7. Otherwise, let N3=0.51N1+0.49N2 and N4=0.49N1+0.51N2. Step 4: Calculate the associated failure values F+3 and F+4, respectively, by solving the optimization problem [17] – [20] once for each, or entering the graph in Figure 3. Step 5: Calculate the values of β3 and β4 associated with the sets ( N*3,,F+3) and ( N*4,, F+4), respectively. Step 6: If β4< β3 let N*1=N3; F+1= F+3 and β1= β3. Otherwise, let N*2=N4; F+2= F+4 and β2= β4, and go to Step 3. Step 7: Return N*=( N*1+ N*2)/2, F+=(F+1+ F+2)/2 and β=( β1+ β2)/2. 4.2.- MODERN SOLUTION OF THE STABILITY PROBLEMS. Since in the modern approach the designed slope is required to have a reliability index larger than a given value β0, and not necessarily the resulting β value will satisfy this constraint, the process need to be iterated until the desired value of β be obtained. The problem of slope design under the modern point of view can also arise in different forms, as: Modern design problem 1: Obtain the allowable µ N* for given statistical parameters µF+, vF+, vN*|F+, β0, and reliability lower bound β 0. This problem can be stated as:
MinimizingZ
F+
,Z
N*
,µ
N*
β = ZF + 2 + ZN * 2
[24]
subject to
ZF +
=
F + − µF +
=
F + − µF +
σ F+ vF + µ F + N − µ N * − β0 (F + − µ F + ) N * − µ N * − β0 (F + − µ F + ) = σ N * |F + vN * | F + µ N * *
ZN*
=
F+
= h( N * )
β
> β0
[25]
and solved using the approximation h’(N*), for h(N*). If an approximation is not available, we can proceed as follows: we start with a tentative value of µN*, solve the problem [22]-[23], and obtain β. If the resulting value of β is smaller than the required bound β0, we increase the value of µN* and repeat the process until we get β = β0. As applications of this case we can determine H in terms of µN*, γ, ψ and c, or c in terms of µN *, γ, ψ and H. Example 4: Assume that we have µF+=5.67, vF+=0.1, vN*|F+=0.1, β0=1, and a reliability lower bound β0=2. Then, the solution of [24]-[25] leads to
F + = 4.584; µ N * = 1.464; N * = 0.294; β = 2.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
If, alternatively, we use the above procedure, we start with a tentative µN*=2 solve the problem [22]-[23], and obtain β=2.931. Since the resulting value of β is larger than the required bound β0=2, we decrease the value of µN* using the updating formula:
µ N( k*+1) = µ N( k*) + ρ ( β 0 − β ( k ) ) = 2 + 0.5(2 − 2.931) = 1.534. Next, we repeat the process until we get β=β 0, as it is illustrated in Table 1. (k ) Iteration K µ N *
1 2 3 4 5 6
2.000 1.534 1.471 1.465 1.465 1.465
β (k ) 2.931 2.126 2.012 2.001 2.000 2.000
Table 1: Illustration of the iterative process for Example 4. Modern design problem 2: Obtain the allowable µF+ for given reliability lower bound β0 and statistical parameters µN , vF+, vN*|F+, β0. This problem can be stated as:
MinimizingZ
F+
,Z
N*
,µ
F+
β = ZF
2 +
+ Z N*
2
[26]
subject to
ZF +
=
F + − µF +
=
F + − µF +
vF + µ F + σF+ N − µ N * − β0 (F + − µF + ) N * − µ N * − β0 (F + − µ F + ) = σ N * |F + vN * | F + µ N * *
ZN*
=
F+
= h( N * ) = µ N µF +
µN* β
[27]
> β0
and solved using the approximation h’(N*), for h(N*). If an approximation is not available, we can proceed as follows: we start with a given value of tentative µF+, we calculate µN*= µN µF+, solve the problem [22]-[23], and obtain β. If the resulting value of β is smaller than the required bound β0, we decrease the value of µF+ and repeat the process until we get β = β 0. The resulting value of µF+ is the desired value. Example 5: Assume that we have µN=0.28, vF+=0.1, vN*|F+=0.1, β0, and a reliability lower bound β0=2. Then, the solution of [26]-[27] leads to
F + = 7.141; N * = 0.643; µ N * = 2.459; β = 2. Modern design problem 3: Obtain the reliability index β for given statistical parameters µN*, µF+, vF+, vN*|F+, β0 and βN. In this case we simply solve the problem [22]-[23], and obtain β Example 6: Assume that we have µN*=2, µF+=5.67, vF+=0.1, vN*|F+=0.1, and β0=1, then, solving the problem [22]-[23], we obtain β=2.931.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
5-. THE SAFETY-FACTOR-FAILURE-PROBABILITY METHOD In this section we present the safety-factor-failure-probability method that combines the classical and the modern approaches. The idea is to design a slope (choose the value of µN*) such that the actual safety factor F and the actual reliability index β satisfy F> F0 and β > β0, i.e.:
MinimizingZ
F+
,Z
N*
,µ
N*
,F
β = ZF + 2 + ZN* 2
[28]
subject to
ZF +
=
F + − µF +
=
F + − µF +
σ F+ vF + µ F + N − µ N * − β0 (F + − µ F + ) N * − µ N * − β0 (F + − µ F + ) = σ N * |F + vN * | F + µ N * *
ZN*
=
[29]
F + = h( N * ); Fµ F + = h( µ N * );
β > β 0;
F > F 0;
The design can be done using the following algorithm: Algorithm 2 (Safety-factor-failure-probability method) Input: The statistical parameters µF+=µ1/ψ, vF+, vN*|F+ and β0, the safety and reliability lower bounds F0 and β0, a ρ value for changing the tentative safety factor, and an error ε to control convergence of the process. Output: The actual safety factor Factual and reliability index β values, and the design values for the classical F* and µN*, and the modern F+ and N*, approaches. Step 1: Let Factual=F0 and Fprevious=F0+2ε. Step 2: If |Factual-Fprevious|< ε go to Step 6. Otherwise, let Fprevious=Factual, F*= Factual µF+. Step 3: Solve the problem [17] –[18], or enter the graph in Figure 3 with F*, and obtain µN*. Step 4: Solve the optimization problem [22]-[23], and obtain the design values F+ and N*, and β. Step 5: If β > β 0, then go to Step 2. Otherwise, let Factual=Factual+ ρ (β0 – β) and go to Step 2. Step 6: Return Factual, β, F*, µN*, F+ and N*. 6-. SENSITIVITY ANALYSIS To obtain the sensitivities of β with respect to µF+, vF+, vN*|F+ and β0 we consider them as artificial variables in Step 2, and add the corresponding constraints. In other words, we transform the optimization problem in Step 2 into the equivalent one:
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Minimize with respect to ZF+, ZN* , µF+, µN*, vF+, v+N*|F+, and β0.
β = ZF + 2 + ZN * 2 subject to
ZF +
=
F + − µ F+ +
ZN*
=
N * − µ N+ * − β 0 ( F + − µ F+ + )
F+ Fµ F +
= h( N * ) = h( µ N * )
vF + µ F +
vN+ * | F + µ N+ *
[30]
β > β 0 ; F > F 0 ; µ F+ + = µ F + ; vF+ + = vF + ; vN+ * | F + = vN * | F + ; β 0+ = β 0 ;
Then, the values of the corresponding dual variables become the desired sensitivities. The sensitivities of F* with respect to N*0, can be directly obtained from the dual of the problem [17] – [18], because N*0 appears in the last constraint. The above algorithms have been implemented in GAMS (see Castillo et al. (2001)) and used to solve the following example.
∂β 0
∂F 1.734
∂β ∂µ F +
∂β ∂v N * | F +
∂β ∂v F +
∂β ∂β 0
0.104
-1.675
-18.72 -2.136
Table 2: Sensitivities of the reliability index to the model parameters. Example 7: In this example we illustrate the approaches 1 and 2. For the approach 1 we use the data F 0 = 2.2;
β 0 = 2.0;
vF + = 0.1; vN * |F +
µ F + = µ1 / ψ = 5.67; = 0.1; β 0 = 1.
Solving the problem (28) - (29) we obtain:
F + = 4.562; N * = 0.292; β = 2.04; F = 2.2; µ N * = 1.487. and the sensitivities in Table 2. And for the approach 2, we assume F 0 = 1.2; β 0 = 2.0; ρ = 0.5; µ F + = µ1 / ψ = 5.67; vF + = 0.1; vN * | F + = 0.1; β 0 = 1.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Then, applying Algorithm 2, we get:
F * = 12.34; µ N * = 1.465; F + = 4.58; N * = 0.294; β actual = 2.0; Factual = 2.18. The process is illustrated in Table3, that shows the quick convergence of the process. Note that the safety factor constraint F> F0=1.2 is not active, while the reliability index constraint β > β0=2 is active. Finally, the sensitivities of the reliability index β to the statistical parameters µN*, µF+, vN*|F+, vF+ and β0 have been calculated using the method described in Section 5, and the results are shown in Table 4, where one can see that the reliability index increases 1.82 units per unit increase of µN*, and decreases -0.562,-1.6,-18.4 and -2.1, per unit of increase of µF+, vN*|F+, vF+ and β0, respectively . Classic Iteration Safety Factor F µ N * 1 2 3 4 5 6 7
1.200 2.040 2.159 2.175 2.177 2.177 2.177
0.594 1.335 1.448 1.463 1.465 1.465 1.465
Design Point
µ F*
N*
6.806 11.568 12.246 12.334 12.345 12.347 12.347
0.411 0.310 0.296 0.295 0.294 0.294 0.294
F+ 5.491 4.707 4.599 4.585 4.584 4.583 4.583
β 0.32 1.761 1.969 1.996 1.999 2.000 2.000
Table 3: Illustration of the iterative process.
∂β
∂β
∂β ∂v N * | F +
∂β ∂v F +
∂β ∂β 0
∂µ N *
∂µ F +
1.82
-0.562
-1.6
-18.4
-2.1
Table 4: Sensitivities of the reliability index to the model parameters. 7-. CONCLUSIONS The methodology presented in this paper provides a rational and systematic procedure for slope stability analysis. Using the proposed combined safety-factor-failure-probability method, the engineer is capable of simultaneously giving bounds for the safety factor and failure probability, so that the most stringent conditions prevail. In addition, a sensitivity analysis can be easily performed by transforming the input parameters into artificial variables, that are constrained to take their constant values. The provided example illustrates how this methodology can be applied and proves that the proposed method is really practical and useful. Some extra advantages of the proposed methods are: 1. They take full advantage of the optimization packages. 2. The non-dimensional treatment of the problem allows identifying the minimum set of non-dimensional parameters that really play a role in the slope problem. In addition, it allows solving an infinite set of problems at once. 3. While the classical approach allows solving the problem by hand, i.e., entering a graph, the modern and mixed approaches require an iterative procedure and a computer.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
4. Sensitivity values are given, by free, since they are the values of the dual problem of a primal problem in which the data are converted into artificial variables, and constraints, locking these variables to their actual values, are added. 5. The calculus of variations method associates to each occurrence or instantiation of the soil parameter values, the worst sliding line, that is dependent on the instantiation being considered, Thus, not particular sliding lines are considered but the whole system of possible sliding lines. Consequently, the safety of the system is directly considered. In other words, not the probability of particular sliding lines is dealt with, but the failure probability of the whole system. 8-. ACKNOWLEDGMENTS We thank Iberdrola, the University of Castilla-La Mancha and the Dirección General de Investigación Científica y Técnica (DGICYT) (project PB98-0421), for partial support of this work. REFERENCES 1.
Aczél, J., 1966, Lectures on functional equations and their applications. Vol. 19, Mathematics in Science and Engineering. Academic Press.
2.
Alonso, E., (1976), Risk analysis of slopes and its application to slopes in sensitive Canadian clays. Geotechnique, London, 25(3), 453-472.
3.
Baker, R. and Garber, M., 1977, Variational approach to slope stability. Proceedings of the 9th International Conference on Soil Mechanics and Foundations Engineering, Tokio,Vol 2, 9-12.
4.
Baker, R. and Garber, M., 1978, Theoretical analysis of the stability of slopes. Geotechnique, nº. 4, 395-411.
5.
Castillo, E., 1988, Extreme Value Theory in Engineering. Academic Press, New York.
6.
Castillo, E., Conejo, A., Pedregal, P., García, R. and Alguacil, N., (2001), Building and Solving Mathematical Programming Models in Engineering and Science. Pure and Applied Mathematics: A Wiley-Interscience Series of Texts, Monographs, and Tracts, New York.
7.
Castillo, E. and Luceño, A., 1982, A critical analysis of some variational methods in slope stability analysis. International Journal for Numerical and Analytical Methods in Geomechanics, 6:195-209.
8.
Castillo, E. and Luceño, A., 1978, One application of the calculus of variations to bearing capacity of foundations. In Second International Conference on Applied Numerical Modeling, pages 1-11, Madrid.
9.
Castillo, E. and Luceño, A., 1980, Application of the calculus of variations to the vertical cut off in cohesive frictionless soil (discussion). Geotechnique, 30(1):1-16.
10.
Castillo, E. and Luceño, A., 1981, Application of the calculus of variations to the vertical cut-off in cohesive frictionless soil. (discussion by J. de Jong). Geotechnique, 30:295-296.
11.
Castillo, E. and Luceño, A., 1983, Variational methods and the upper bound theorem. Journal of Engineering Mechanics (ASCE), 109(5):1157-1174.
12.
Castillo, E., and Revilla, J., 1977, The calculus of variations and the stability of slopes. In Proceedings of the 9th International Conference on Soil Mechanics and Foundations Engineering, volume~2, pages 25-30, Tokyo.
13.
Castillo, E. and Ruiz, R., 1992, Functional Equations in Science and Engineering. Marcel Dekker.
14.
Castillo, E., Solares, C., and Gómez, P., 1996a, Tail sensitivity analysis in bayesian networks. In Proceedings of the Twelfth Conference on Uncertainty in Artificial Intelligence (UAI'96), Portland (Oregon), Morgan Kaufmann Publishers, San Francisco, California, 133-140.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
15.
Castillo, E., Solares, C., and Gómez, P., 1996b, Estimating Extreme Probabilities Using Tail Simulated Data. International Journal of Approximate Reasoning.
16.
Castillo, E., Solares, C., and Gómez, P., 1997, High Probability One-sided Confidence Intervals in Reliability Models. Nuclear Science and Engineering, Vol. 126, 158-167.
17.
Chen, W. F., and Giger, M. W., 1971, Limit analysis of stability of slopes, Journal of the Soil Mechanics and Foundations Division, ASCE.
18.
Chen, W. F., and Snitbhan, N., 1975, On slip surface and slope stability analysis. Soils and Foundations, 15, 41-49.
19. 20.
Cornell, C. A., 1971, First-order uncertainty analysis of soil deformation and stability. Proceedings of the 1st Internation Conference on Application of Statistics and Probability to Soil and Structural Engineering, 129-1440, Hong Kong.
21.
Ditlevsen, O. and Madsen, H. O., 1996, Structural reliability methods. Wiley, Chichester, New York.
22.
Freudenthal, A. N., 1956, Safety and the probability of structural failure. Transactions, ASCE 121, 1337-1397.
23.
Galambos, J., 1987, The Asymptotic Theory of Extreme Order Statistics. Robert E. Krieger Publishing Company. Malabar, Florida.
24.
Garber, M., and Baker, R., 1977, Bearing capacity by variational method. Journal of the Geotechnical Engineering Division, ASCE, 103, 1209-1225.
25.
Janbu, N., 1954, Application of composite slip surfaces for stability analysis. Proc. European Conf. Stability of Earth Slopes, Stockholm 3, 43-49.
26.
Luceño, A., 1979, Análisis de los métodos variacionales aplicados a los problemas de estabilidad enmecánica del suelo. Utilización del teorema de la cota superior. Ph. D. Thesis. University of Cantabria. Santander, Spain.
27.
Madsen, H. D., Krenk, S., and Lind, N. C., 1986, Methods of structural safety. Prentice Hall, Inc., Englewood Cliffs, N. J..
28.
Revilla, J., and Castillo, E., 1977, The calculus of variations applied to stability of slopes. Geotechnique, 27(1):1-11.
29.
Rosenblatt, M., 1952, Remarks on a multivariate transformation. Ann. Math. Stat. 23(3),470-472.
30.
Schuster, R, and Krizek, R., (1978), Landslides analysis and control. Special Report 176. Transportation Research Board. National Academy of Sciences. USA.
31.
Wirsching, P. H., and Wu, Y. T., 1987, Advanced reliability methods for structural evaluation. J. Engineering Mechanics Division, ASCE 109(2), 19-23.
32.
Wu, Y. T., Burnside, O. H., and Cruse, T. A., 1989, Probabilistic methods for structural response analysis. In W. K. Lam and T. Belytschko (eds.) Computational mechanics of Probabilistic and Reliability Analysis. Elmepress International, Annandale, Va, 181-196.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
UNCERTAINTIES IN SIMPLIFIED LIQUEFACTION ANALYSIS OF SOIL DEPOSITS Rafael Blázquez, Susana López, Vicente Navarro, Jesús Sánchez and Emilio González Geotechnical and Geological Engineering Division, Department of Civil Engineering Universidad de Castilla – La Mancha Ciudad Real, Spain
[email protected] [email protected] [email protected] [email protected] [email protected]
ABSTRACT The evaluation of the liquefaction potential of a soil layer subjected to seismic or sea waves require accounting for the uncertainties in the dynamic loads and in the properties of the soil. Since the origin of the two sources of uncertainty is obviously different, the question of ranking its relative importance in the overall accuracy of the predictions deserves some attention. Rigorously speaking, such a question can only be answered in probabilistic terms, within the framework of a reliability model convolving a hazard definition of the load with a geomechanical formulation of the phenomenon (e.g., an effective stress or a cumulative damage liquefaction model). At the present stage this kind of approach is too complicated for professional use and becomes biased by the statistical aspects of the input, which obscure the geotechnical contribution to the global uncertainty. To clarify this matter and get some guidance in practical applications, a simplified deterministic model of liquefaction has been explored. The model, of analytical form, can be applied to situations of seismic and seafloor liquefaction. In both cases the safety factor at any point within the layer, under undrained or drained conditions, can be computed. In this paper, using the above approach, a sensitivity study on the parameters governing liquefaction has been conducted and the relative influence of those parameters in the observed variations of the safety factor has been investigated. 1.- Introduction Liquefaction of saturated cohesionless soil is the transformation of the material from a solid state to a liquefied state as a consequence of increased pore pressure and reduced effective stress. During liquefaction the soil behaves as a viscous liquid: the skeleton transfers temporarily the intergranular stresses to the pore water, thereby reducing to practically zero the bearing capacity of the soil. Liquefaction of a soil deposit can result from the application of either static or dynamic loading. In the latter case, the mechanics of the phenomenon (1-D formulation) is depicted in Fig 1-a for situations of storm waves and seismic motions acting, respectively, at the top and bottom boundaries of the layer. In both phenomena, since the total pressures within the soil mass do not vary with time, at the onset of liquefaction, the effective stress principle requires (Fig. 1-b): σ’v0 = udyn
Department of Civil and Environmental Engineering Infrastructure Technology Institute Northwestern University, Evanston, Illinois, 60208
(1)
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
which means that, at the time (tl) and depth (zl) of liquefaction, the dynamical pore pressure (udyn) should be equal to the corresponding initial vertical effective stress (liquefaction criterion). 2.- Liquefaction model for undrained level ground For the purposes of this study, a simplified theoretical model of the liquefaction susceptibility of a uniform soil layer has been developed. The basic steps of the model are as follows: 1.
The excess pore pressure at any depth within the soil, ug(t), increases linearly with time: ug Neq t = = tl σ'v0 Nl
(2)
(ug = dynamically generated pore pressure; σ’v0 = initial effective stress) (Neq = no. of cycles of equivalent harmonic loading; Nl = no. of cycles to liquefaction) 2.
3.
The liquefaction cyclic strength curve, τl(Nl), is of fatigue type: τl = a Nl-b σ'v0 Dr (a, b = constants; Dr = relative density of soil)
(3)
The average shear stress induced in the soil by either sea or seismic waves can be computed by elasticity theory: γw H K z -K z Sea waves: τav(z) = e (4-1) 2 cosh(K d) τav(z) = 0.65
Seismic waves:
σv0 amax r g d σ'v0
(4-2)
2π is the wave number (L = wave length), L and d is the water depth (Fig 1-a). In eq. (4-2), amax is the peak seismic acceleration at the surface of the ground and rd is the deformation factor (Fig. 2-a), that has been fitted with the exponential formula: In eq. (4-1) γw is the unit weight of water, H is the wave height, and K =
rd = e-r z 1
(r1 = 0.0165 for z ≤ 25 m.)
(5)
which closely approximates the linear expression (Fig. 2-b): rd = 1 – 0.015 z
(6)
frequently used in engineering practice (Tokimatsu and Yoshimi, 1983). With these premises, assuming undrained conditions throughout the layer, the time to liquefaction can be obtained combining properly equations (4), (3) and (2). The safety factor against liquefaction at time t and depth z is then calculated by means of the equation: Fu =
tl Nl 1 a Dr γb1/b 1/b = = Q t Neq Neq e-pz
(7)
where, for earthquake excitation: p = r1 and for ocean wave loading:
Q=
g 0.65 amax γsat
(7-1)
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
p=K
Q=
2 cosh (K d) HK
(7-2)
In both cases, γb and γsat are the buoyant and the saturated unit weight of the soil, respectively.
L
a max
H
WATER
z
d po
b
z b
; Dr, cv
D
a)
W.T. Initial vertical effective stress 'vo
DEPTH, z
SOIL
; Dr, cv
Time = t
zl
Time = t l
PORE PRESSURE INCREMENT, u
b)
Fig. 1.- Mechanics of liquefaction in oceanic and seismic environments a) Scheme of the problems (free – field level ground) b) Onset of liquefaction
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
rd 0
0.2
0.4
0.6
0.8
1
1.2
0 20
Dpeth (ft)
40 60 Average values
80 100 120
a)
rd 0.6
0.65
0.7
0.75
0.8
0.85
0
z (m)
5 10 15 20
rd = 1- 0.015 z rd = e-0.0165z
25
b) Fig. 2.- a) Computed values of the deformation factor, rd b) Proposed expressions to fit the deformation factor data, rd
0.9
0.95
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
3.- Effect of drainage on the liquefaction potential of the soil The preceding derivation is strictly correct for infinitely deep deposits with a low hydrodynamic factor (Whitman, 1970). In such a case the duration of the load is very short compared to the consolidation time and there is practically no flow of water outside the seabed. However, in the cases of highly permeable soils or slowly acting load (such as sea wave load), the hypothesis of no drainage is no longer valid. In those cases the distribution of pore water pressures in the layer, at time t, is governed by the equation: cv
∂ 2u ∂ u ∂ ug = – ∂t ∂t ∂z2
(8)
in which u represents the total excess pore pressure over the hydrostatic (generated minus dissipated) and cv is the coefficient of consolidation of the soil. In the following, this parameter is assumed to be constant within the layer and proportional to the coefficient of permeability of the soil. Although such an assumption is hardly justifiable in liquefaction processes it simplifies greatly the calculations without significant loss of accuracy. To derive mathematically the solution of the partial differential equation (8), the term ∂ug/∂t must be computed first. Taking into account eq. (2): ∂ ug ∂ ug ∂ N σ'v0 Neq = = Nl td ∂t ∂N ∂t
(9)
and substituting Nl from eq. (7), yields: ∂ ug = C z e-pz/b ∂t
(10)
where for seismic load : C=
0.65 amax γsat1/b 1 (γb)b T g a Dr
(10-1)
C=
1 HK 1/b (γ )b T b a Dr cosh(K d)
(10-2)
1
and for ocean wave load: 1
1 td is the “predominant period” of the input motion, that, in the In the above equations: td = load duration; b1=1- and T = Neq b case of sea waves, becomes (Wiegel, 1964): T=
2πL g tanh (K d)
(11)
Next, inserting eq. (10) into eq. (8), the balance of the simultaneous generation and dissipation of pore pressures that continuously occurs inside the porous sediment can be obtained for both types of dynamic actions. It is required at this point to introduce adequately the thickness of the deposit, D, and the boundary and initial conditions for the solution, namely (Fig. 1-a):
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Boundary conditions: Pervious top:
[u(z,t)]z=0 = 0
(12-1)
Impervious bottom:
∂ u (z,t) = 0 ∂z z=D
(12-2)
Pervious bottom:
[u(z,t)]z=D = 0
(12-3)
Initial condition: [u(z,t)]t=0 = 0
(12-4)
Following this path, if the excess pore-pressure curves are expressed in the following form (Layas, 1982; Rahman and Jaber, 1986; Blázquez and Martínez, 1988): u(z,t) = w(z) + v(z,t)
(13)
and the function v(z,t) is decomposed into a product of two independent functions of time and space: v(z,t) = T*(t) Z*(z)
(14)
eq. (8) can be integrated analytically. The final results are listed below. Case 1: layer with impervious bottom: w(z) = -
C 2- r (r D -1) z e r D + (z r - 2) e r z cv r3
v(z,t) = S an e
-(n1 π / D)2 cv t
(n=0 → ∞)
(15-1)
n1 p z D
sin
(15-2)
where :
an =
r=-
p b
(15-3)
n1 =
2n + 1 2
(15-4)
- 2 C (D3 r3) -2 D n1 π r + er D (D2r2 (-1+rD) + n12 π2 (1+r D)) sin (n1 π) cv r3 (n13 π 3 + D2 n1 π r2)2
(
)
(15-5)
Case 2: layer with pervious bottom:
(
-C 2 (-1+ er D) z + D (2 - er D r z + (-2 + r z) er z) w(z) = cv r 3 D n p z -(n π/D)2c t v D e
v(z,t) = S an sin (n=0 → ∞)
1
)
(16-1)
(16-2)
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
where :
an =
(
)
2 C 2 D4 n π r4 + D3 er D n π r3 (n2 π2+ r D (-2 + r D))cos (n π) cv r3 (n3 π 3 + D2 n π r2)2
(16-3)
Then, the factor of safety against liquefaction in the case of drained soil can be written as: Fd = 4.
σ'v0 γb z = u(z,t) w(z) + v(z,t)
(17)
Numerical example
To check the validity of the simplified model described above, a sample problem has been solved. The basic parameters of the model are listed in Table I. TABLE I.- Model parameters EARTHQUAKE LOAD
SEA WAVE LOAD
BOTH LOADS γb = 1 T/m3 a = 0.5 b = 0.25
Dr = 45 % td = 6 hr H = 12 m T = 15 sec L = 200 m d = 40 m D = 100 m
Dr = 60 % td = 6 sec amax = 0.15 g T = 0.5 sec D = 50 m
The evolution with time of the dynamic isochrones, in the case of seismic loading, is compared in Fig. 3 for an undrained (Fig. 3-a) and a drained (Fig. 3-b) soil layer. In the first case the depth of the layer does not affect the shape of the isochrones, whereas, for a 50 m thick drained stratum, theses curves are independent of the permeability at the base of the deposit. s'v0; u (T/m2) 0
5
10
15
20
25
2
s'v0; u (T/m ) 30
35
40
0
0
D=∞
15
20
25
30
35
2
40
D = 50 m cv = 10 m2/s
5
cv = 0 m /sec
10
10
15
15
20
s'v0
zl = 9 m
30
20 z (m)
z (m)
10
0
5
25
5
25
s'v0
zl = 10 m
30
35
1 sec
35
1 sec
40
2 sec
40
2 sec
45
4 sec
45
4 sec
50
tl = 5.2 sec
50
tl = 7 sec
a) Undrained layer
b) Drained layer
Fig. 3.- Seismic liquefaction development in soil stratum with undrained (a) and drained (b) conditions
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Similarly, Figs. 4-a and 4-b show the results of the analysis for a seafloor soil subjected to a wave storm. In this case, the geometrical and wave parameters of the problem must meet the condition (Horikama, 1978): H 1 2 π d < tanh L 7 L
(18)
which simply means that the wave steepness, H/L, can not increase beyond the critical value corresponding to the breaking of the wave. s'v0; u (T/m2)
2
s'v0; u (T/m ) 0
5
10
15
20
25
30
35
0
40
0 10
D=∞
20
cv = 0 m2/sec
0.2
0.3
0.4
D = 100 m
0.5
cv = 10 m2/sec
s'v0
20
I.B. 5 min I.B. 10 min
s'v0
30
I.B. 1 hour
40
40
zl = 8 m
50
z (m)
z (m)
0.1
0
P.B. 5 min
60
P.B. 10 min
60
P.B. 1 hour
80
70 80
5 min
90
12 min
100
20 min
100
I.B.: Impervious bottom P.B.: Pervious bottom
120
No liquefaction
a) Undrained layer
b) Drained layer
Fig. 4.- Sea wave liquefaction of soil deposit with different permeability properties From the results presented in Fig. 4 it is concluded that drainage plays an important role in the onset of liquefaction of ocean beds, due to the retardation effect over the mechanism of the pore pressure generation. This effect is more pronounced near the base of the deposit if the bedrock is permeable, leading to greater safety factors than for the impervious boundary. As concerns the pore pressure generation rates at a given point within the soil profile, ∂u/∂t, they are constant and exponentially – decreasing, in the undrained case (eq. 2) and the drained case (eqs. 15-2 and 16-2), respectively. This implies that the safety factors decrease at all depths as: Fu ÷
1 t
Fd ÷ e-αt
(see eq. 7)
(see eq. 17)
(19)
Fig. 5 exemplifies this assertion for the different loading and permeability conditions considered before.
z=5m
10.00 9.00 8.00 7.00 6.00 F 5.00 4.00 3.00 2.00 1.00 0.00 0.00
z=5m
cv = 10 m2/sec
2
cv = 10 m /sec 60.00
Fd Fu
40.00
Fd
F
Fu 20.00
2.00
4.00
6.00
8.00
10.00
12.00
t (sec)
a) Earthquake loading
0.00 0.00
2.00
4.00
6.00
t (hr)
b) Ocean wave loading
Fig. 5.- Time variation of the safety factor against liquefaction at a given location in the soil deposit
8.00
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
5.
Parametric study
In order to check the sensitivity of the model to the various factors involved, a parametric study of the factor of safety against liquefaction has been carried out. The ranges of variation of the variables about the reference values shown in Table I are: soil parameters: (homogeneous soil)
40 % ≤ Dr ≤ 80 % 0.8 T/m3 ≤ γb ≤ 1.2 T/m3 0.4 ≤ a ≤ 0.6 0.2 ≤ b ≤ 0.3 10-2 m2/sec ≤ cv ≤ 10 m2/sec
seismic loading :
0.1 g ≤ amax ≤ 0.2 g 0.2 sec ≤ T ≤ 0.8 sec td = 6 sec
sea wave loading:
10 m ≤ H ≤ 14 m 150 m ≤ L ≤ 250 m 10 sec ≤ T ≤ 20 sec td = 12 hrs.
As can be inferred form the proposed model, in general, for given values of space and time, the safety factor varies with the type of load, the drainage of the soil and the permeability conditions at the base of the deposit. In the following, these aspects of the problem are investigated separately, for earthquake and wave - induced liquefaction phenomena, in the light of the parametric study just proposed. 5.1.- Earthquake loading Figures 6 and 7 summarize the results of the parametric liquefaction analysis of an infinitely deep soil layer at depths of 5 m., 10 m. and 15 m. As would be expected, the liquefaction potential is very sensitive to changes in the relative density of the soil and / or the amplitude of the input acceleration, whereas variations in the unit weight and/or the cyclic resistance of the material are comparatively less important. Furthermore, the susceptibility to liquefaction decreases (other parameters being the same) as we go deeper into the layer, due to the correlative increment of the initial effective pressure, which prevents the occurrence of liquefaction (eq. 1). With regard to the influence of drainage, a similar analysis has been conducted with a 50 m – deep saturated sand layer with a free - draining top boundary and either an impervious or a pervious bottom boundary. The computed results (Fig. 8) indicate that, in both cases, there is a threshold value of the coefficient of consolidation (and therefore of the permeability of the soil) below which the layer behaves essentially as undrained and no differences between Fu and Fd are observed. In the example considered herein this critical value is of the order of 1 m2/sec. According to these findings, a value of cv = 10 m2/sec has been selected to elucidate the effect of drainage in liquefaction processes. Regardless of the boundary condition at the bottom of the layer, an improvement of the factor of safety is consistently found when the dynamic pore pressures are allowed to dissipate to some degree at the time they are being generated. This conclusion is graphically displayed in Figs. 9 and 10, which show again the strong dependence of the drained safety factor on the relative density of the soil and the peak ground acceleration, as well as the beneficial effect of drainage in preventing liquefaction (Fd > Fu in all cases). It is also remarkable the sensitivity of the drained model to the number of cycles of equivalent harmonic load applied to the layer. For a given duration of the earthquake, the safety factor Fd increases linearly with the “predominant period” of the input, irrespectively of the permeability conditions at the bottom of the deposit (Fig. 11). This conclusion comes from the fact that liquefaction is viewed in the model as a low – cycle fatigue type of process: as more load (⇒ shear stress) reversals are u 1 applied to the soil more damage is caused to the material, and the liquefaction degree = increases with the F σ'v0 d frequency of the excitation (1/T).
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Dr
0.40 0.00
0.50
0.60
0.70
0.80
0.50 1.00 1.50
z=5 m z=10 m
Fu 2.00
z=15 m
2.50 3.00 3.50 4.00
a max (g)
0.10 0.00
0.12
0.14
0.16
0.18
0.20
0.50 1.00 1.50 Fu 2.00
2.50 3.00 3.50 4.00
Fig. 6.- Parametric study of earthquake liquefaction. Undrained soil layer
z=5 m z=10 m z=15 m
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
a 0.40 0.00
0.45
0.50
0.55
0.60
0.50 1.00 1.50 Fu
2.00 2.50 3.00 3.50 4.00
b 0.20 0.00
0.22
0.24
0.26
0.28
0.30
0.50 1.00 1.50 Fu
2.00 2.50 3.00 3.50 4.00
γ b (T/m 3) 0.80
0.90
1.00
1.10
1.20
0.00 0.50 1.00 Fu
1.50
z=5 m
2.00
z=10 m
2.50
z=15 m
3.00 3.50 4.00
Fig. 7.- Parametric study of earthquake liquefaction. Undrained soil layer
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
cv (m2/sec) 1.0E-02 0.00
1.0E-01
1.0E+00
1.0E+01
0.50 1.00 1.50
z=5 m
Fd 2.00
z=10 m
2.50
z=15 m
3.00 3.50 4.00
Fig. 8.- Influence of the coefficient of consolidation on the seismic liquefaction pattern. Drained soil layer (D = 50 m) with pervious or impervious bottom boundary
Dr 0.40 0.00 0.50
0.50
0.60
a max (g) 0.70
0.80
0.10 0.00 z=5m
F 2.00 2.50 3.00 3.50 4.00
0.14
0.50
1.00 1.50
0.12
0.16
0.18
0.20 z=5m
1.00 Fu Fd
1.50 F 2.00 2.50 3.00 3.50 4.00
Undrained deposit : D = ∞ ; cv = 0 m2/sec. Drained deposit: D= 50 m.; cv = 10 m2/sec.
Fig. 9.- Effect of drainage on earthquake liquefaction potential
Fu Fd
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
γb (T/m3) 0.80 0.00
0.90
b
1.00
1.10
z = 5m
0.50
0.50
1.00
1.00
1.50 F
0.20 0.00
1.20
F
Fd
2.50
0.24
0.26
0.28
Fu
2.00
Fd
2.50
3.00
3.00
3.50
3.50
4.00
4.00
a 0.40 0.00
0.45
0.50
0.55
0.60 z = 5m
0.50 1.00 1.50
Fu
F 2.00
Fd
2.50 3.00 3.50 4.00
Fig. 10.- Effect of drainage on earthquake liquefaction potential
T (sec) 0.20 0.00
0.30
0.40
0.50
0.60
0.70
0.80
0.50 1.00 1.50
Fd
2.00 2.50
0.30 z = 5m
1.50
Fu
2.00
0.22
z=5 m z=10 m z=15 m
3.00 3.50 4.00
Fig. 11.- Variation of the safety factor with the period characteristics of the seismic motion. Drained layer
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
5.2.- Sea wave loading Ocean beds are usually loose to medium cohesionless sediments prone to liquefaction as a result of the progressive build up of pore water pressures under severe storm conditions. In order to investigate sea floor liquefaction, a parallel analysis to the one described in the previous section has been implemented for a soil with a relative density varying between Dr = 30 % and Dr = 60 % (average value = 45 %) in either an undrained or a drained state. Figs. 12 and 13 show the results for an undrained layer. It can be realized that, in all cases, for the ranges of variation of the parameters of the model, the soil liquefies completely to a depth greater than 15 m. below the ocean floor. Obviously (Fig. 12) the greater the relative density and the unit weight of the soil, the greater the resistance against liquefaction , whereas the fatigue parameters, a and b, appear to be also important. The wave depth and the wave height clearly govern the liquefaction potential, although they operate in opposite directions (Fig. 13). On the contrary, the wave length seems to have little effect on the occurrence of the phenomenon. The effect of drainage has been investigated for a finite deposit (50 m - thick), with the same material and load characteristics as above and different values of the coefficient of consolidation. The influence of this parameter is represented in Fig. 14, which clearly demonstrates how, beyond the threshold value cv ≥ 0.5 m2/sec, a drastic increase in the factor of safety against liquefaction is found (Fig. 14). For the purposes of comparison, Figs. 15, 16 and 17 show, respectively, the sensitivity of the drained model to the soil, load and geometrical parameters for two different coefficients of consolidation (1 m2/sec and 10 m2/sec) and two different bottom boundaries of the layer (pervious or impervious). Although the comparison is made only at the shallower control point (z = 5 m.), it can be seen that the behavioral patterns are qualitatively the same as the ones derived in the undrained case. However, the factors of safety increase very significantly, and no liquefaction occurs in the examples analyzed. It is also observed that the effect of drainage at the bottom of the layer is to accelerate the dissipation of the dynamic pore pressures, and, consequently, to reduce the liquefaction potential near the surface of the ground. Besides, this effects are more pronounced for high permeable soils (cv = 10 m2/sec) and shallow layers (D ≤ 100 m), as it is clearly shown in Fig. 17. 6. Conclusions A simple unified approach to the formulation of the safety factor against liquefaction of a saturated sand deposit subjected to harmonic seismic or ocean waves has been developed. The model computes analytically the dynamic pore pressure increase at different depths in the deposit, taking into account the internal and external drainage conditions of the soil. Based on the results obtained from a parametric study of the factors that affect the factor or safety, the following conclusions are drawn: 1. 2. 3. 4. 5. 6.
Liquefaction is a fatigue type of failure, and, as such, is very sensitive to the duration of the load. Safety factors decrease drastically with time in seismic environments, while in ocean environments they reach stable (time – independent) values after many cycles of wave loading. In both undrained and drained situations, the liquefaction potential is very sensitive to changes in the relative density of the soil, the depth within the layer and the amplitude of the input excitation, whereas variations in the unit weight and/or the cyclic resistance of the material are comparatively less important. There is a threshold value of the coefficient of consolidation (both for seismic and ocean wave excitation) below which the internal drainage of the soil does make little difference in the computed liquefaction behavior. For high cv values, the dissipation rate of the dynamic pore pressures counteracts the generation rate, leading to high values of the drained safety factors for both types of loading. In the case of ocean waves, the boundary conditions at the base of the deposit are important for points located near the surface of shallow sediments (thickness less than 100 m). In such a case the factor of safety is greater for deposits with a pervious base than for those with an impervious one. The wave depth plays a stabilizing role in ocean floors: the greater the wave depth the smaller the liquefaction susceptibility, regardless of the internal and external drainage conditions of the soil.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
gb (T/m3)
Dr 0.30 0.00
0.40
0.50
0.60
0.80 0.00
0.10
1.00
1.10
1.20
0.10
0.20 Fu
0.90
0.20
0.30
z=5 m
0.40
z=10 m
Fu 0.40
z=10 m
0.50
z=15 m
0.50
z=15 m
z=5 m
0.30
0.60
0.60
0.70
0.70
0.80
0.80
a 0.40 0.00
0.45
0.50
b 0.55
0.60
0.20 0.00
0.10
0.24
0.26
0.28
0.30
0.10
0.20 0.30
0.22
0.20 z=5 m
Fu 0.40
z=10 m
0.50
z=15 m
Fu
0.30
z=5 m
0.40
z=10 m
0.50
z=15 m
0.60
0.60
0.70
0.70
0.80
0.80
Fig. 12.- Sensitivity of the undrained model of sea wave liquefaction to the soil parameters
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
H (m) 10.00 0.00 0.10 0.20 0.30 Fu 0.40 0.50 0.60 0.70 0.80
11.00
12.00
d (m) 13.00
14.00
z=5 m z=10 m z=15 m
20.00 0.00 0.10 0.20 0.30 Fu 0.40 0.50 0.60 0.70 0.80
30.00
40.00
50.00
60.00
z=5 m z=10 m z=15 m
L (m) 150.00 0.00 0.10
200.00
250.00
0.20 0.30
z=5 m
Fu 0.40
z=10 m
0.50 0.60
z=15 m
0.70 0.80
Fig. 13.- Sensitivity of the undrained model of sea wave liquefaction to the wave depth and wave parameters
cv (m2/s)
1.0E-02 0.00
1.0E-01
1.0E+00
1.0E+01
10.00 20.00 30.00 40.00 Fd 50.00
60.00
z=5 m z=10 m z=15 m
70.00 80.00 90.00 100.00 Fig. 14.- Effect of consolidation coefficient on the safety factor. Drained deposit (pervious or impervious bottom boundary) subjected to sea wave loading
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Dr 0.30 0.00
0.40
0.50
0.60
gb (T/m2)
0.80 0.00
z=5m
20.00
0.90
1.00
1.10
1.20 z=5m
z=5m
20.00
Fd 40.00
cv [m2/sec]
60.00
40.00 Fd
cv [m2/sec]
60.00
80.00
impervious Cv = 10 pervious Cv = 10
100.00
impervious Cv = 1
impervious Cv = 10
80.00
pervious Cv = 10
100.00
impervious Cv = 1 pervious Cv = 1
pervious Cv = 1
a 0.40 0.00 20.00 Fd
40.00
0.45
0.50
b 0.55
0.60 z=5m
0.25
0.30
z=5m
20.00 2
cv [m /sec]
60.00 80.00 100.00
0.20 0.00
Fd
40.00
cv [m2/sec]
60.00 80.00
impervious Cv = 10 pervious Cv = 10
100.00
impervious Cv = 10 pervious Cv = 10
impervious Cv = 1
impervious Cv = 1
pervious Cv = 1
pervious Cv = 1
Fig. 15.- Sensitivity of the drained model of sea wave liquefaction to the soil parameters
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
10.00
10.50
11.00
11.50
12.00
z=5m
z=5m
H (m)
12.50
13.00
13.50
L (m)
14.00
150.00
0.00
0.00
10.00
10.00
20.00
20.00
30.00
30.00
40.00 Fd 50.00
40.00 Fd 50.00
60.00
60.00
70.00
70.00
80.00
80.00
90.00
90.00
100.00
100.00
170.00
190.00
210.00
230.00
250.00
T (sec) 10.00
12.00
14.00
0.00
16.00
18.00
20.00
z=5m
6
10.00 20.00 30.00 40.00
Fd 50.00 60.00 2
cv [m /sec]
70.00 80.00
impervious Cv = 10
90.00
pervious Cv = 10 impervious Cv = 1
100.00
pervious Cv = 1
Fig. 16.- Sensitivity of the drained model of sea wave liquefaction to the load parameters
D (m)
d (m) 20.00 30.00 40.00 50.00 60.00 0.00 10.00 20.00 30.00 40.00 Fd 50.00 60.00 70.00 80.00 90.00 100.00
0.00
z=5m
2
cv [m /sec]
impervious Cv = 10 pervious Cv = 10 impervious Cv = 1 pervious Cv = 1
200.00
400.00
600.00
800.00 1,000.00 1,200.00
0.00 10.00 20.00 30.00 40.00 Fd 50.00 60.00 70.00 80.00 90.00 100.00
z=5m
cv [m2/sec] impervious Cv = 10 pervious Cv = 10 impervious Cv = 1 pervious Cv = 1
Fig. 17.- Sensitivity of the drained model of sea wave liquefaction to the geometrical parameters
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
7. References Blázquez, R. and Martínez, F. (1988), Probabilistic analysis of seafloor liquefaction, XXI World Conference on Coastal Engineering, Torremolinos, Málaga, Spain, Vol. 2, pp. 1352 – 1367. Layas, F.M. (1982), Response and stability of ocean floor soils under random waves, PhD thesis, North Carolina State University, Raleigh, N.C., U.S.A. Rahman, M.S. and Jaber, W.Y. (1986), A simplified drained analysis for wave – induced liquefaction in ocean floor sands, Soil and Foundations, Vol. 26, Nov. 3, pp.37-68. Tokimatsu, K. and Yoshimi, Y. (1983), Empirical correlation of soil liquefaction based on SPT N-value and fines content. Soils and Foundations, Vol. 23, No. 4, pp. 56-74. Whitman, R.V. (1970), The response of soil to dynamic loadings (Final Report). Contract Report No. 3-26, U.S. Army Engineer Waterways Experiment Station, Vicksburg, Miss. U.S.A. Wiegel, R.L. (1964), Oceanographical engineering, Prentice – Hall Series in Fluid Mechanics. Richard Skalak Editor, U.S.A., 532 pp.
Teaching & Management Early 1970’S Engineering Education at Northwestern M. H. Farzin, EES Corporation, USA Teaching Research – The Successful Legacy of RJK R. D. Holtz, University of Washington, USA Role of University Educators in Developing Future Leaders of Engineering Enterprise S. Abdelhamid and T. B. Edil, CH2M Hill, USA, and University of Wisconsin-Madison, USA
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
EARLY 1970'S ENGINEERING EDUCATION AT NORTHWESTERN M. Hassan Farzin EES Corporation 1350 West Fifth Avenue, Columbus, Ohio 43212
[email protected]
ABSTRACT Most Employers expect their employees with higher educational degrees to carry at least some "management" functions, in addition to their specialized technical responsibilities. The term management of a program is meant as a generic word that is applied to directing, controlling, or handling of an operation. An operation consists of: (a) a variety of work programs and associated administrative and financial functions, requiring management, and (b) a work force involving different labor classes, requiring leadership. Management and leadership are two faces of the same coin. The former term is applied for management of programs and processes, while the latter term is used to indicate leadership of individuals. In other words, people are not managed; they should have leaders capable of leading them through complex work arrangements, with the ultimate purpose of completion of a program and achievement of a defined goal. Management of programs requires certain tools, such as systems to track the status of costs and schedules, budgeting and finance, and contract administration. Leadership of people also requires tools, such as leaders’ communications skills, to influence and motivate people with relevant and appropriate characteristics to work together and achieve a defined goal. INTRODUCTION In the early 1970's, most graduates with higher educational degrees from U.S. universities hired by most U.S. employers were expected to be involved in "management" of technical and technology related programs, in addition to the specialized technical responsibilities. I was not exempted from this pattern of work assignment, and shortly after joining my employer (at the time, a very large U.S. East Coast A/E and constructor organization) I was assigned responsibility for management of a technical program. The responsibility consisted of a small segment of a multi-billion dollar program for design and construction of a Nuclear Power Generating Station. It had taken me over twenty years of classroom, field, and laboratory education and research work, culminating in a Ph.D. degree in engineering, to reach where I was. During my university education years, I had attended three different universities, and studied more than fifty courses in a variety of topics and disciplines. Once in my assignment with my employer, I discovered that the only classical knowledge I and, as far as I knew, most graduates from engineering programs, lacked were subjects of "management," and leadership. Management and leadership qualifications, therefore, had to be learned on an ad hoc basis, with hit-and-miss results, on the job, and over time. My on-the-job-training as a manager progressed well and took time, but I survived the process, and learned many tools that since those days have been dramatically improved, and have found even wider use for management and successful control of large (and small) scale programs. Today, given appropriate technical knowledge and use of relevant management tools, Department of Civil and Environmental Engineering Infrastructure Technology Institute Northwestern University, Evanston, Illinois, 60208
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
budgets, schedule of delivery of material and parts, and completion times, can all be forecasted with sufficient accuracy and long enough lead times that could permit corrective action initiation, where needed, to prevent cost and or schedule slippage of a program. Leadership education and training, I discovered, was entirely a different matter, as I will discuss later. From this personal experience, it became obvious to me that a proper education and training in "management," and leadership, in addition to the intensive training in many technical subjects we received at Northwestern (and other universities I had attended previously) would have been extremely helpful, if not essential. Education in "management" and leadership would help technical people who must lead, influence, and energize individuals, and deal with the realities of funding, budgets, and schedules of deliveries. In other words, graduates should be prepared to enter the world of business, in a general sense of the word business. It was a delight when the news arrived that such a program had finally been initiated by Professor Raymond Krizek at Northwestern, under his personal guidance and leadership. The Master of Project Management Program should place the Northwestern Engineering education at the top of the list for anyone who is headed for the world of reality, i.e., the world of business and management of technological operations, or any type of operations. BACKGROUND After graduation from Northwestern in the Summer of 1973, I got a job with the then a premier East Coast A/E and Construction Management firm, specialized in complex facility designs and implementations, including engineering, design and construction of Nuclear Power Generating Stations. This all was happening at a time when nuclear power generation was starting to take hold as a valid, economically superior, technically challenging but quite possible source of electricity for the remainder of the twentieth century and beyond. As events turned, within a few months of the beginning my new job, the war of 1973 broke out in the Middle East. Doubling of gasoline prices, as well as long lines at gas stations, contributed to the image of nuclear power generation as an important, if not inevitable, source of energy for the country; perhaps for the rest for the world. This meant that my employer's business quickly increased, and appeared to be very stable for the long haul. At the time I was assigned as the Project Specialist for the Millstone Nuclear Power Station, in Connecticut. This meant that I was second (among many seconds) under the Project Engineer for this multi-billion dollar project. My job on that project, and many projects to come during the following years, was to set up engineering and design specifications for geotechnical investigations, coordinate it with various other engineering disciplines, place the work out for bid, select a desirable contractor to carry out the actual field work, assure quality of field investigations while they were being conducted, interpret the results such that they were useful for the rest of the engineering disciplines, write (and do many rewrites of) the report of the results of the field work in a form that it could be placed in PSAR (Preliminary Safety Analysis Report) so that it could be submitted to the regulatory bodies, including AEC, and later to the NRC, and appear before the regulatory bodies, as well as the client, respond to questions regarding relevant subjects, and defend the relevance, accuracy, and appropriateness of the work and the results. Although my responsibilities could be described in further detail, the point I am trying to make is that in addition to technical specialty needed to conduct this work, these activities also required planning, estimating, budgeting, contracting and purchasing, and financing. Furthermore, the work also required interaction with other professionals, some of them people who would be working for you and/or with you, some of them your superiors, as well as personnel of subcontractors in different labor classes. In other words, the management portion of the work under my responsibilities required management of programs, and leadership of professionals. MANAGEMENT AND LEADERSHIP The reality was, and still is, that a university graduate with a higher degree in technical disciplines, employed by many employers is expected to carry at least some "management" functions, in addition to their specialized technical responsibilities. The term management of a program used herein is meant as a generic word that is applied to directing, controlling, or handling of an operation. An operation consists of: (a) a variety of work programs and associated administrative and financial functions, requiring management, and (b) a work force involving different labor classes, requiring leadership.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Here is the important practical issue in the management of technical programs. Management and leadership are two faces of the same coin. The former term is applied for management of programs and processes, or inanimate things, while the latter term is used to influence and lead individuals. In other words, people are not managed; they should have leaders capable of leading them through the complex work arrangements, with the ultimate goal of completion of a defined program. Management of programs requires certain tools, such as systems to track the status of costs and schedules, budgeting and finance, and contract administration. Most, if not all of these tools currently exist, or are available in commercial form; one needs only to learn their use and varied applications. Although learning such a variety of systems and their applications would require time, there is a reasonable chance that anyone with sufficient attention to detail can learn these systems. In the process of managing the work, all new recruits learn most of these management tools. Use of management tools has become widespread and necessary for successful control and completion of any scale project involving a variety of technical disciplines, including design and implementation efforts. Leadership of people also requires tools, such as leaders’ communications skills, to influence and motivate people with relevant, appropriate, but varied characteristics to work toward, and achieve the defined goal. In practical terms, student’s education and training for the leadership side of the management coin has not been systematically addressed. It is not totally clear how one would teach, or train students for leadership. A recent book by John Maxwell1 has defined leadership as influence. This appears to be a very acceptable definition of leadership. Accepting John Maxwell’s definition, the issue remains unchanged: how to teach and train students involved in a classical technical education to acquire business leadership qualities that most will need during their working life period? LEADERSHIP CHARACTERISTICS John Maxwell, in his book (Ref.1) has a list of characteristics for a leader, which include the following: The leader coaches his workers, The leader depends upon the good will of his workers, and The leader inspires enthusiasm. To achieve these and additional characteristics desirable for a leader, a leadership education and training program should, at a minimum, include a number of topics. These include: 1. Team Selection and Work Assignment.-- A leader does not necessarily do the work, or do the work alone. A leader makes things happen by appropriate assignment of the personnel as they are matched with the work to be done. Hence, a leader’s training should include education for selection of an important resource: Personnel. Discussion on this topic should include occasions when a team is provided to a leader and manager along with the job assignment. This is specially true for managers at the low and middle segment of the leadership structure, who generally do not have the luxury of selecting their team members. A leader’s training should cover handling of such difficulties: given time, a leader should be able to have all members of his team heading in the same common direction. 2.
Communications.-- This means that leadership education should include communications training, and emphasize that a leader must be able to continuously and clearly communicate with the members of the group. Communications of the type discussed here is not as simple as it may sound. Normally, in the process of work, managers become so preoccupied that they easily overlook communication with the group members, assuming that members are close enough to each other to be informed of the changes occurring almost continuously. The net result of such a breakdown in communications is that slowly the group cohesion may begin to disappear. Soon, the entire program will start showing signs of fatigue. A leader makes sure that everyone has the satisfaction of contributing to the progress of the work, so that everyone feels responsible for, and good about, the success of the program and quality of the finished product.
3.
Consultation and Involvement.-- As a result of a good communications arrangement, effective leadership requires that members of a group working together be provided an opportunity to discuss their ideas and get involved in the workings of the group. A successful leadership education program should provide training as to how to set up operations such that members of the group are given meaningful opportunity to get involved in the work assignment, and provide opinions and suggestions for work improvements. The more members are consulted, even in those items that may appear too
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
trivial to bother with, the more likely is a program’s success. Some of the ideas and solutions to existing problems presented by members may not make sense. However, listening to those ideas, and letting the members of a team to listen to the ideas and discuss them for final resolution, is an important facet of the leadership. This practice also keeps the leader from becoming the bad guy for rejecting a loose idea: he could leave that unpleasant task to other members of the team. 4.
Knowledge of Subject Matter.-- In theory, one does not need to be an expert in rocket science to be able to lead a rocket design group, provided that you know management principals, and you are a leader in other respects. In practice, however, it helps if you know rocket science. In a sense, since the subject of discussion is leadership in a technical program, it becomes almost imperative to have a leader who also is knowledgeable in the technical subject matter. The greater his expertise, the better his communication skills relative to the technical subject matter, and the better his chances for being accepted as an effective leader.
SUMMARY It is of paramount importance for anyone who is headed for the world of business after graduation from the university environment to learn and be familiar with the subject of "management," and learn and practice elements of leadership. Most, if not all, of the mid- and higher level technical positions in today’s economy will have a large (and growing) element of management and leadership components to them. In the early 1970's at Northwestern, education and training for the skills of program management and leadership were not widely available. Almost twenty years later, such a program was finally initiated at Northwestern by Professor Raymond Krizek, under his guidance and leadership. It has been a source of success by itself, as well as for the engineering education at Northwestern. REFERENCES 1.
Maxwell, J. C., Developing the Leader Within You, Publisher: Thomas Nelson; ISBN: 0785266666; 2nd Rev edition (December 19, 2000)
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
TEACHING RESEARCH--THE SUCCESSFUL LEGACY OF RJK Robert D. Holtz University of Washington Department of Civil & Environmental Engineering, Box 352700, Seattle, WA 98195
[email protected]
INTRODUCTION Ray Krizek has had a profound and positive influence on my academic career. At the time I was his student, however, I was not sufficiently mature to realize that his approach to doctoral student supervision would be so beneficial. Working as part of “Team Krizek” was a big challenge, and yet it provided many opportunities that later proved to be very valuable to me, both as a post-doctoral researcher and as a professor. Dr. Krizek taught us how to write proposals, plan and carry out the work, and publish the results—in short, to do all the things necessary for a successful career in research and academia. We didn't know it at the time, and we certainly didn't call it mentoring, but that’s what it was. Dr. Krizek was a very effective mentor for his doctoral students. While the Krizek method may not work for all students and all professors, it appears to be more successful than the other common approaches to graduate student supervision and mentoring—especially for producing successful researchers and professors. In order to provide some perspective on the RJK method, other approaches to dissertation supervision are described below along with their advantages and disadvantages. Many times, Ray Krizek would tell us students what it takes to succeed as an academic; his ideas were often illustrated by examples of current and past students as well as his own experiences. Five pieces of advice that I particularly remember are given in the paper. Finally, I close with a few remarks about the legacy of RJK. DIFFERENT APPROACHES TO THE SUPERVISION OF DOCTORAL STUDENTS There are a number of ways professors supervise their doctoral students. Although I suspect that most of us would deny that this applies to us, the approach usually taken seems to be more a reflection of the professor’s personality and style than any conscious attempt to find a procedure that is most appropriate for the individual student. If the student can adapt to the professor’s style, then the PhD student will probably complete a dissertation successfully. At the risk of overgeneralization or glaring omissions, the approaches to PhD supervision that I have observed in my academic career seem to fall into four categories: hand-holding, sink-or-swim, impossible dream, and the RJK method. 1. The “hand-holding” approach This approach is common in big multi-year projects involving several student researchers. Because the work needs to be well organized and well coordinated, it may even involve post-doctoral researchers or other permanent departmental and research staff. In this case, the student is given a fairly detailed plan for her dissertation research. The original proposal was written by the professor or other former students and staff, and there rarely is any involvement of the student in establishing Department of Civil and Environmental Engineering Infrastructure Technology Institute Northwestern University, Evanston, Illinois, 60208
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
the scope of work or the technical approach. The student only has to be moderately successful at doing what was proposed, write it up, and voila! A PhD. In many cases, there are frequent, often weekly, meetings with the professor just to be sure things are on track. The meetings may involve others on the project team and may include members of the student’s PhD advisory committee. Occasionally even the sponsor’s technical representative is involved in these meetings. Proponents of “hand holding”: Well, you probably know some… Upside: if the student stays on track and does not have any serious technical difficulties, then she gets the degree with a minimum of pain and uncertainty. Research students learn to be good team players. Downside: the fresh PhD does not directly learn about proposal writing, research planning, or how to work independently. 2. The “sink or swim, hands-off, you are on your own” approach In this approach, the professor meets with the student and they both agree on a topic that should make a suitable thesis. Then the student attempts to do the research with minimum input from or interaction with the professor. Many times, the student gets frustrated with a lack of obvious progress and little positive feedback, and he may become “in absentia” and go get a real job. Now, in some cases, absentia students do actually manage to complete a dissertation. They may be motivated by pressure from their spouses, employment contracts, university imposed time limits, etc. Then after years of little to no communication, the student suddenly appears with a thesis in hand and he expects/ hopes/demands that it be approved as is. What happens if it is not acceptable to the professor or to some of the committee members? What if the thesis needs substantial revisions, or worse additional research? What if the time limit is exceeded? Proponents of “sink-or-swim”: Famous Professors Upside: for Famous Professors who are often very busy, there is a minimum of effort and time involved. For the student, IF he succeeds in writing and successfully defending a thesis, then he is really able to work on his own, generate his own ideas, and will likely become a successful academic. But no thanks to Famous Professor… Downside: Especially with absentia students, the quality of the research is often poor, and graduation rates appear to be very low. The geotechnical world used to be littered with students of Famous Professors, students who were very bright and capable but who never quite finished and often were bitter about their experience with Famous Professor. 3. The impossible dream It is tempting for both a professor and a student to try to solve one of the remaining big geotechnical problems, the solution of which would bring them both fame and fortune. These big problems require a lifetime of work devoted to the topic, and thus are unrealistic for a PhD dissertation. When the dissertation topic is so broad in scope, the student finds that after a few years’ work, it is going to be impossible to ever solve the problem and complete an acceptable thesis. I have not seen many “impossible dream” topics recently, although they were quite common in the early days of soil mechanics. (I think my own original thesis topic to relate shear strength to quantified clay fabric was an “impossible dream”, even though it would have never brought us fame or fortune!) I have seen a problem that is a corollary of the impossible dream. The thesis topic suggested by the professor and agreed to by the student is relevant, important, and considered reasonable and doable by both. But as the work progresses, major differences develop between the expectations of the professor and the ability of the student researcher to make “satisfactory progress” (defined by the professor, of course) on the topic. This disparity leads to conflicts between the professor and student researcher. The professor is disappointed at the inability of the student and the apparent lack of progress, and the student is continually frustrated by not being able to produce acceptable results.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
4. The RJK method The following comments are based on the experiences of me and my classmates during 1966-1970. I have no idea if Ray Krizek has continued to use this method with all his students, but in those days the RJK approach to doctoral supervision seemed be almost custom-designed for each student. Our individuality was recognized and respected, as were our different backgrounds, abilities, and technical interests. Students often were able to select their own dissertation research topics, although when a funded project was involved, more direction was obviously required. We often ended up in groups or teams working on related aspects of a research topic. For example, Wally Baker’s Anisotropic Mohr-Coulomb study turned into work on clay fabric and shear strength by Don Sheeran and myself, and then to the theses by Tunch Edil, Salah Abdelhamid, and others. Dr. Krizek provided us lots of opportunities to write proposals and to review and critique published papers. We were encouraged to write up any studies of a topic that might become a section or chapter in our thesis…or it might not; it didn’t matter. Ray typically critiqued and edited these short papers, and then generously (in the pre-word processing era) had them typed up and duplicated for distribution to the other graduate students. In a few cases, if he felt a paper was potentially publishable, he would urge us to do the necessary additional work, suggest a “home” for the paper, and even help us with the submission process. His encouragement to work on something interesting and to write up our preliminary results was a good learning experience for those of us aiming for an academic career. (In my own case, my Northwestern-RJK experience was excellent preparation for a post-doctoral stint at the Swedish Geotechnical Institute and an assistant professorship at Purdue.) One of the current buzzwords in engineering education is teamwork. Well, that is nothing new for the graduate students in soils at Northwestern, at least when I was a student. We considered ourselves to be part of the entire NU soil mechanics team—faculty, graduate students, and staff, such as Dominic and Hugo. As students, we were encouraged to participate along with the faculty in seminars (ours and departmental), Chicago Soil Mechanics Lecture Series, and Chicago ASCE soils group meetings. We all worked together, for example, on a big open house for the Midwestern geotechnical graduate programs in 1970. I always appreciated the fact that all the soils faculty—Professors Osterberg, Krizek, and Franklin—were faithful attendees at these events. I learned from them that, as long as I am town and no matter how busy I might be, always go to seminars, lectures, and even student presentations. ADVICE FROM PROFESSOR KRIZEK TO ASPIRING ACADEMICS When I came to Northwestern in the fall of 1966 to work on a PhD, Ray was still an Assistant Professor. Although he had only been on the faculty a couple of years, he already had well-developed ideas about research and academic life (perhaps nurtured during his teaching time at Maryland?) Discussions with him about technical problems often turned into conversations about what was necessary to have a successful academic career. He often used living examples—current students or recent graduates—to illustrate his points. (I often wondered what he told the other students about me!) While I’m sorry I don’t remember all of his advice, here are a few gems for those of us intending to be academics: 1. It is “publish or perish”, and be sure to also bring in research money! Deans and Department chairs like to see lots of publications and a decent amount of outside funding from their young faculty. Even in the late 1960s, this was one of the “rules” of the academic game. Ray told us: “If you don’t like the rules, then don’t play the game.” There are plenty of other satisfying career paths available to bright, hard-working PhDs, and most even in teaching don’t require publishing and chasing research money to be successful. But if you do take an academic position at a research university, don’t complain later about the pressure to publish and bring in research funds. 2. Another piece of advice—difficult for me, an experimentalist weak in theory—was to stay out of the lab and avoid doing experimental research, especially early in your academic career. Labs are expensive, and department chairs don’t like it because you are always asking them for more money for your lab, but yet you don’t have much to show (e.g., publications) for all your hard work and their money. Because it takes valuable tenure time to build a lab and develop a laboratory research program, it is difficult to show published results during those early critical years. 3. Ray also told us, only half jokingly, how to treat experimental data so it would look better than it really was. Besides plotting the data on 10-cycle semi-log paper, these “tricks” included using larger-than-normal symbols around the data points
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
and then omitting the data points themselves after positioning these symbols close to the curve-fit line. I’m sure there were others that I’ve forgotten. 4. “There is always a home for that paper.” Even if it has been already rejected by one or more of our traditional journals, Ray was a master at finding a place to publish a paper. My favorite (and perhaps Ray’s, too?) was the Journal of the Franklin Institute. Another journal I discovered while working in Sweden was Archiwum Hydrotekniki (published in Poland and always looking for papers in English). My Swedish favorite is Acta Polytechnica Scandinavica. 5. When you write a proposal, have some of the research work already done. Show “preliminary results” to demonstrate the likely viability of your approach. Otherwise, reviewers will always complain: “…but how do we know the proposed approach/technique/method/procedure… will really work? (“Well, that’s why we want to do the research!”) 6. Finally, Ray occasionally talked about the different graduate programs in the USA, describing their strengths and weaknesses. As I recall, he favored programs with the faculty having different technical interests so that graduate students would be exposed to traditional soil mechanics and foundation engineering, as well as the then newer areas of physicalchemical properties and rock mechanics. He assumed that groundwater and seepage, soil dynamics, engineering mechanics, applied mathematics, and geology would be a part of the traditional geotechnical engineering education. In smaller programs, a balance of faculty interests was considered to be especially important. THE LEGACY OF RJK The program of this Symposium, as well as Ray’s publications, awards and prizes, and his election to the National Academy of Engineering all attest to his amazing record of technical and professional accomplishments. It is truly remarkable for anyone to make important contributions in so many different geotechnical areas. I for one am glad Ray did not follow Prof. Ray Yong’s advice, offered in 1967 or 68, and paraphrased here: Ray Krizek works in too many different technical areas! If he really wants to make a name for himself in soil mechanics, then he should concentrate his efforts and those of his students on one or maybe two specialty areas, no more! As you know, Ray was the second President of the Geo-Institute, and as a founding member of the G-I Board of Governors, he played a key role in its formation and early development. His leadership, vision, and attention to detail were crucial to whatever continuing success the G-I has had, and he set the performance standards for all future G-I presidents. I wrote the following (slightly edited) comments in a letter to Ray 10 yr ago on the occasion of his 60th birthday. I hope you will agree that they are an appropriate closing. Ray, you have a very loyal group of students who have benefited greatly from having you as their major professor. You provided a marvelous atmosphere at Northwestern for personal and professional development. I learned how to think and work independently, and yet at the same time interact constructively with fellow students and other faculty. You also taught me how to write proposals and papers, do twenty-seven things at once, and keep a large number of graduate students reasonably happy and inspired. I always felt you put the students’ interests first, and I have tried to follow your model with my own students. You also taught us how to work very hard and get things done, and you showed us how to outwit bureaucracy and silly rules. At this time, I know you’re going to look back on your many technical and professional accomplishments, especially your contribution to the entire Department of Civil Engineering at Northwestern in the leadership role you have played as chair of the department. Many of those accomplishments are very tangible and can be easily counted. But I also hope you will take pride in the accomplishments of your students, which often are less tangible and more a matter of inspiration and style. Thanks for being a good role model and a good friend. Congratulations, Ray, on your 10th, 40th, 60th, and 70th anniversaries, and many happy returns!
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
ROLE OF UNIVERSITY EDUCATORS IN DEVELOPING FUTURE LEADERS OF ENGINEERING ENTERPRISE 1
Salah Abdelhamid, 2Tuncer B. Edil
CH2M HILL 1700 Market Street, Suite 1600, Philadelphia, PA 19103-3916
[email protected] 2
Civil & Environmental Engineering and Geological Engineering, University of Wisconsin-Madison 1415 Engineering Drive, Madison, WI 53706
[email protected]
ABSTRACT This paper highlights the role of university educators in developing future leaders of engineering enterprise through the example set by Professor Raymond J. Krizek of Northwestern University as experienced by two of his doctoral students who pursued entirely different career tracks, one in academe and the other in engineering services industry. The authors identify six areas of development: technical, professional and applied engineering, business management, communications, teamwork and people/social skills. INTRODUCTION In the summer of 1969, a few weeks before humankind first stepped on the moon, Tunch and I arrived in America, and it was in Evanston to be specific. We came from two different countries on opposite sides of the Mediterranean: Tunch from Turkey and I Egypt. We have left much we loved behind for the promise of a high caliber education, better life, and some relief of the turbulent political conditions in our home countries. The Cold War was at a peak at that time. The Soviets have just crushed the last inkling of freedom in Checkoslovakia, hundreds of nuclear warheads were pointed at Turkey, and Egypt was practically fully under Soviet domination. We both landed in America with only a few names in our pockets; at the very top of my very short list was Ray Krizek’s. Over the following 4 years, we both, as well as many of our colleagues, came to the realization of how crucially important the influence and contributions of the “Northwestern Community” of people were on our lives and future. Very quickly, we all became part of that family-like community, and names like Ray, Claudia, Jorj, Ruth, Holtz, Sheeran and Dominic became as common in our lives as those we left behind thousands of miles to the east. Tunch and I graduated from Northwestern in early 1970s, and took two completely different career tracks. Tunch stayed in academe and wound up spending the next quarter century in Madison Wisconsin, within driving distance of Ray and the beloved campus of NU. I went straight to the engineering services industry, moved to the East Coast and had very little interaction with either. But when I started working with Tunch on this paper couple of month ago, the memories and feelings were as fresh as could be. Our reflections and assessments of what Ray Krizek and Northwestern have contributed to our lives, characters and careers are remarkably consistent. Today, Tunch and I will attempt to capture what those years meant to us in terms of shaping our professional and technical capabilities, our work ethics, as well as, and most importantly perhaps, our human and social characters as citizens of this our chosen country. We will try to do that in the context of what we now understand, hopefully a lot clearer than then, to be the fundamental role of a university educator, primarily with Dr. Krizek as the model. It’s meant to be a lighthearted chat with many old friends and associates, a memory refresher perhaps for many of you. We both hope that the laughs, if they do happen, will be mostly on us two, and that we don’t bore you too much with historic details.
Department of Civil and Environmental Engineering Infrastructure Technology Institute Northwestern University, Evanston, Illinois, 60208
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
The making of a well-rounded university graduate, particularly in the special but important case of a foreign graduate student, involves six major areas of development. These are: technical, professional and applied engineering, business management, communications, teamwork and people/social skills. Tunch and I will alternate sharing with you what Dr. Krizek, Dr. Osterberg, and others at Northwestern have done for us in each of those areas. Tunch will start with the technical skills. Tunch…! TECHNICAL I came to Northwestern with great zeal for academic achievement and with an objective of becoming a university professor. I came to Evanston several weeks after the summer school started, which provided attendance in courses offered by guest instructors to college professors through NSF funding that Ray had secured. I took very unique courses offered by experts through these summer programs, which later provided me with sound grounding in some fundamentals, e.g., clay mineralogy, physico-chemical properties of soils, finite elements, consolidation theories, groundwater flow. Later courses offered in the regular Northwestern program also helped a lot. In particular, Ray’s Soil Rheology was unique and helped me much as an academician, though I have to admit that I would get an embarrassingly low score in the first exam if it was not to extraordinary tutoring effort of Salah the night before the exam. He had an idea about where the questions would come from and a strategy, which I did not. Engineering Properties of Soils was a course that encouraged teamwork as we worked in groups of two and spent many Saturdays in the laboratory trying to figure out some soil test procedure on our own. Of course, Jorj’s three advanced soil mechanics courses were the place where I learned geotechnical engineering (which was not called that yet in those days) and that helped me a lot in my consulting activities later in life. Ray also sent Salah and me to the University of Arizona to attend a summer school and learn about soil fabric techniques from the experts, which turned out to be crucial for progress in our thesis research. In summary, I had a series of good courses that provided both fundamentals for an academic life and also practical for engineering life. Ray’s efforts to attract those NSF summer courses and also make them available to the graduate students were a unique contribution to my graduate education. Overall we knew, however, that being a geotechnical engineer (in academia or in practice) meant being realistic, having a feel for soil behavior, and a keen interest in disseminating your experiences. We knew the importance of balance between theory, laboratory, and most significantly field investigation. Now, let me turn this lectern to Salah to tell us how the practical side of things worked out. Salah….! WORLD-CLASS PRACTICE OF ENGINEERING To become a world-class engineering practitioner, one must have a first-hand exposure to a wide variety of field applications, professional ethics and social contexts. Accordingly, a University Educator must have a world-view of the profession, its related fields of business, as well as national, economic, socio- political and geographic influences thereto. He should also be able to bring captains of the engineering industry, as well as its leading technical authorities, in touch with the student body, as an integral part of their learning experience. This all happened to us at NU, thanks to the Krizek-Osterberg “Tag Team”. Let me just elaborate a bit about this aspect. I chose NU over Caltech and others because of its unique Work-Study Program, which Ray recommended in his first letter to me. At least 50% of the Soils class at the time was involved in some Work Study program; that was a definite plus. In my case, the 7 months, which I spent, working for Clyde Baker and Co. at STS, were my absolute first and purest experience with practical geotechnical engineering, and they confirmed for me the desire to be a practitioner rather than an academician. It is there where many of us developed a real appreciation for the challenges of dealing with real soils, particularly Chicago’s elusive Clay, and its “made-famous-worldwide-by-Osterberg” Hard Pan. At STS, I ran my first consolidation test, drilled my first hand-augured hole, tested my first undrained shear sample. But what was so unique about this whole Work-Study Program is that it almost directly segued into, and supported most of what Ray Krizek, Jorj Osterberg, and Gus Franklin talked to us about during the very first year at Northwestern. As such, it was uniquely useful, and thoroughly enjoyable. After 7 months at STS, Ray and Jorj secured a teaching assistantship for me. Soon thereafter, my original hunch was firmly confirmed: NU’s learning environment, and the Krizek - Osterberg Team as sponsors/coaches/teachers and advisors were perfect for what I needed; a development program that is well balanced between theory and practice on one side, and between laboratory, field and design on the other. Moreover, the “Mentorship Model” that they established continues to be the most
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
exemplary to-date; it’s a benchmark of commitment that we often fail to achieve in today’s professional world, particularly in industry. Today, I believe more than ever that an excellent graduate program in any engineering branch must show diversity on many fronts, in order to become a “world-class” program. And “program diversity” is what I believe was Professor Krizek’s main and crucial contribution to our development at Northwestern. During those early years of ours, such diversity encompassed 4 main areas: technical content, student background, opportunities for practicing engineering and exposure to sources of learning outside of Northwestern’s own. Let us touch briefly on each of these. •
Technical areas of coursework and research, which in our case here ranged from Ray’s insistence that my electives include Parmelee’s “Pre-stressed Concrete Design” course, all the way to Ray Young’s “Physico - chemical Behavior of Soils”, which I still can’t fathom more than 20% of it! Simultaneously under Dr. Krizek’s purview, Scott Jin was downstairs playing chemist with 4000 dredge spoil leachate columns, while Peter Krugman was wrestling with one of the largest finite elements models for highway culvert design. Meanwhile, Tunch and I were too busy learning white clay ceramics from Masters Holtz and Sheeran.
•
National and academic background of the student body. This was truly the most valuable and fun part of the Krizek legacy. Count them please and let me know if I left out someone or some country here, but we had 24 nationalities, and more than 20 universities represented in Krizek’s student body between 1970 & 74. We had people from Guatemala, Spain, Libya, Hungary, Portugal, Saudi Arabia, Germany, Sudan, Thailand, Latvia, Japan, Palestine, Australia, Belgium, China, Turkey, Egypt, Columbia, Switzerland, India, Checkoslovakia, Iran, Singapore, Trinidad, South Africa, and even….. Baltimore!
•
Opportunities for hands-on practical engineering assignments. Professors Krizek and Osterberg, both accomplished consultants in their own right, provided numerous such opportunities to their graduate students, including the WorkStudy programs at Chicago area firms. They both also excelled in regularly engaging their graduate students in meaningful tasks associated with their consulting work. There were also the field research programs, e.g., the Dredge Spoils Leaching Study Project, which Dr. Krizek directed for the Toledo District Corps of Engineers. This was an extensive multi-disciplinary field sampling and monitoring programs, particularly by university standards. Working on this study, we developed a great deal of first hand understanding of soil chemistry’s role in a real engineering problem,e.g. the environmental acceptability and structural stability of dredged spoils, which, has recently become a major challenge area for today’s marine terminal expansion projects nation-wide. Not only did this project expanded and enhanced our technical side, but it was also a complete exercise in project management and integrated data management systems.
•
A sustained connection between theory and practice through a visiting lecturer program of international experts, leading educators, and renowned practicing engineers. I am sure most of you remember how remarkable was the Civil Engineering visiting lecturer program in the 1970’s. We had at least a dozen such speakers come from all over the world to talk every school year. Of the international names, I particularly remember Professor Salazaar of Portugal, Janbu of Norway and Witke of Karlsrue. The American roster ranged from Mitchell and Duncan of Berkeley, to Wissa of MIT, Schmertman of Florida, and Saada of Case Western. What was particularly helpful regarding listening to these giants speak about their experience, is the way it was all tied into what we were doing at that time, as well as to applied engineering in general. Let me close this point with another Krizek nugget. After a guest lecture by Professor Adel Saada, in which he presented a study of shear strength of sands, Ray took over the podium to close. He thanked Dr. Saada for speaking, paused reflectively before adjourning the audience, turned to us and said: “And by the way, next time you’re computing bearing capacity for shallow foundations, you probably won’t need this paper. You still have to use the SPT blow count, and assume 28 degrees for loose sand and 34 for dense.” That said it all! I remember, and do often quote him for saying that “…for a bunch of Soil Engineers, this and other theoretically-oriented courses we offer you at NU, are aimed at helping you to pose the right problem correctly, much less so for learning how to solve the equations”. Ladies and gentleman, no statements can sum up what I took out of NU than those few Krizek words.
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
Let me now move another step away from academia, to talk about “Business management”. BUSINESS MANAGEMENT It is evident today that successful engineering enterprises need seasoned managers at all levels, to navigate them through the overwhelming challenges of highly globalized and volatile marketplace. Accordingly, the term “well-rounded engineering executive” no longer means, necessarily, one with a career spanning design, consulting and construction experiences. Rather, it always additionally implies a knack for salesmanship, a keen sense of business and a sound financial judgement. MBA’s, for better or for worst, have replaced engineering Masters Degrees for many engineers, as the desirable second degree. Although it was still the seventies, somehow you’d think Krizek knew this is about to happen! Nothing is more unique to Ray Krizek’s fulfillment of his role as an Educator than what he demonstrated to us about managing a business: Northwestern’s Soil Mechanics R&D business that is. Directly or by example, I learned from him how to set priorities and allocate resources, to accomplish program objectives on time and within budgets. Like any project delivery organization, our little geotechnical group had a bunch of projects to perform, a finite mission for each, which was mostly R&D oriented, a limited set of resources to do it, with a highly structured template for its disbursement (student labor, CE Department’s machine shop and NU’ funding) and somehow deceptively flexible schedule. How Ray managed to accomplish all that is frankly beyond me, particularly when assessed according to prevailing business management practices and efficiencies. I think he not only managed to do it, but also actually relished that part of his mission. In so doing, Ray Krizek developed a pioneering formula for how business management skills and principles can be taught as part of an engineering program. Professor Krizek had a clear vision of the values and priorities to uphold as he conducted his business. He shared it often, and acted on it all the time; he deeply believed that we’re there to learn from, as well as teach, one another, he included. In today’s business world, they finally figured out what to call this. They refer to it as the “Learning Organization”, one of the most sophisticated growth models for highly leveraged business teams, to compete effectively within an evolving marketplace. Dr. Krizek dedication to the welfare of his students was all encompassing and beyond doubt. He demanded reciprocity for his hard work and commitment, and we gave it back happily. Business ethics were crucial and reflected on the quality of the work done, the integrity with which deviations from plans and schedule slippages were reported (I probably held the record for that one!) and for the mutual trust among the team, and which flew directly from Ray himself. There was also Ray’s ability to mix and match the different talents and temperaments within the group to get the work done, and most effectively. There were those that knew how to manage the dollars (Max Geiger), those that knew how to direct research teams (Don Sheeran), those that built great heavy equipment (Dominic), and those that tinkered with high-precision instrumentation (Schmidt). There were the mathematicians (Manolo and Roberto), computer jocks (Krugman), the lab wizards (Salem & Jin), and the teaching-types (Neil Kay, who single handedly taught me how to run a killer consolidation test!). It was the ultimate in business team orchestration, and Ray conducted seamlessly and almost effortlessly, for he was a natural Maestro! The music was good. This multi-tasking and highly leveraged management culture tremendously expanded the skill set with which most of us left Northwestern, and also enhanced my orientation for a career with the industry. Each of us learned and applied more skills than the average graduate student usually acquires. To be a graduate student under Ray, meant that you plan a technical approach, define tasks accordingly, schedule budgets and write research proposals. You developed technical and management progress reports for all those NSF and other research projects, you taught classes and lab sessions), mentored other students, participated in field investigation campaigns, and worked in the CE shop with Dominic & Co. It was all in a day’s work under Ray and which is, in my humble opinion, a world-class program in business management. Tunch now will describe the importance of communication skills and teamwork. COMMUNICATION More than ever before in its history, engineering in today’s global economy and multi-cultural marketplaces, demands excellence in inter-personal as well as inter-organizational communication. Whether it is writing an organizational intraoffice memorandum, drafting the executive summary of a proposal, or giving a speech at the retirement party of one’s
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
revered thesis adviser, the need for effective communication dominate 50 to 75% of an engineering leaders business and professional life. At Northwestern, it certainly was not all about acquiring new knowledge that marked our graduate education, particularly Ray’s advisees. There were important skills to be learned that were invaluable in our future professional lives. For instance, we learned creative ways of dealing with academic bureaucracy since Ray took great pride and pleasure in telling us how he beat the system. We were organized as a group of four students working on clay fabric, who met weekly, reviewed progress, divided work, and proceeded. We would also meet with Ray, though less frequently, to hear his thoughts on the overall direction of the research and seek his help in solving problems and getting money for equipment and supplies. This semiautonomous management style was the hallmark of my training as a future academician. We also gained experience, for instance, in recruitment of new graduate students, such as Jack Rosenfarb, Chawla and Kutay. We learned self-management, i.e., setting up goals, pursuing them without significant external impetus, work as a team and deal with its shortcomings in exchange of its long-term benefits. We would not have developed the methods of quantitative determination of clay particle orientation and apply this common basis to a variety of problems in our respective theses, if we had not worked together as a team. Ray provided us with a group of undergraduates to help us in our work. Some of the brightest undergrads worked for me, and managing them was a great way to train as an academic advisor. They were a diverse group of people and required different approaches, and in the process, they too learned much. We also received significant training in written communication. In the beginning we would draft letters to suppliers and manufacturers. Ray would edit them and I would watch in amazement how he, with a few strokes of very small handwriting and elaborate arrows, would change some barely legible, definitely foreign sounding text into clear and precise English. As a young professor when I was on my own and literally without any senior colleagues, I knew how to write proposals and papers, and yes effective memos to the administration. Ray has perfected the art of recommendation letter writing. What he says and what he does not say all communicate a meaning. I am not sure all the readers can decipher all the nuances and messages conveyed. To conclude, watching Ray in “editorial action” established for me the significance of precise and elegant text, and the discipline and hard work that go into the process. Let me now move from this to a very important topic “Teamwork.” TEAMWORK Ray created an environment conducive to teamwork when we were at Northwestern, a time that very well may be the peak for the size of his research group. As an example of how his students were encouraged to work as a team, I will describe the backroom. “Backroom” was dedicated to basic material behavior of clays and consisted of a controlled-temperature-room, a general soil fabric laboratory and an office. The characters who occupied or worked in this room and how they interacted with each other is telling of how teamwork was the operating theme and how knowledge and training passed from generation to generation. I trace the back roomers to Wally Baker, Jr. I never met him then but knew all about him. I believe Bob Holtz and Don Sheeran were the next people. They had built the slurry consolidometers and were working on clay fabric methods such as the optical microscopy when Salah and I were recruited by them. I had a choice between developing and applying a computer code to a boundary-value problem or conduct experimental work on clay fabric. Bob and Don were surprised why I would go to this seemingly open-ended, unstructured and perhaps difficult area but they were pleased nonetheless. Salah and I spent the next year or two in training with Bob and Don. We learned the fine art of plumbing, designing temperature-control systems, machining, soil testing, electronic measurement, and yes clay fabric methods (optical microscopy and x-ray diffraction). Eventually Bob graduated and Don went to McGill but Salah and I had new recruits, Chawla, Kutay, and later Jack Rosenfarb. In 1970, Ray sent Salah and me to the University of Arizona to learn transmission electron microscopy and develop a deeper appreciation of clay fabric from the experts. The main objective was to perfect our techniques for preparing and identifying the fabric of kaolinitic clays. This central task was accomplished eventually with the collective effort of the group consisting of Chawla and Kutay in addition to Salah, Jack Rosenfarb and I, each tackling and leading a certain aspect of the problem. I remember showing the first series of definitive scanning electron micrographs in a seminar in Spring 1972. At that time, I had no idea how my thesis would turn out, I had difficulty seeing any light at the end of the tunnel, and felt that I may never get it done. Immediately after the seminar, Ray came and indicated that this was a breakthrough. He was right though I did
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
not really understand what he was saying at the time. Within 6 months my thesis was essentially complete with clay fabric, mechanical testing, and an analytical model adapted from composite materials with the help of Prof. Toshio Mura. We all eventually graduated and Ray had another generation or two of clay fabric students. This story is very exemplary of the spirit of teamwork, which we enjoyed working under Dr. Krizek, and which was so essential in reaching our academic goals. There are I am sure other stories such as the participants in the dredgings and concrete culvert projects can tell. I did not participate in those but I heard how Enrique Socias handled so effectively food catering for the field crew and how Chawla finally found a concrete pipe section to take a nap in the middle of a hot afternoon of hard work. Working with fellow graduate students, working with other professors, working with other departments and universities all went into my graduate education in Ray’s program. Many elements of what I learned from Ray at Northwestern are manifest in my program at Wisconsin. Our soils group is known for its teamwork and extensive collaboration in the department and seen as a model for other groups. We take care of our own as I remember Ray used to take care of his own e.g., spending many hours trying to solve a student’s visa problem or securing employment for a spouse (like my wife Berrin or Salah’s wife Zahera, who both worked for Ray). Let me now turn this back once more to Salah to conclude with one of his favorite topics: People and Social Skills. Salah…! PEOPLE AND SOCIAL SKILLS I am about to share something for the first time, so get your kleenex tissues out! Ray Krizek is one of only a handful of people that I, for better or for worse, credit for influencing the direction of my life in a major way. He was the very first American to be my boss, mentor, teacher, and friend, and thus from and through him I came to learn so much. Although this part of Krizek’s role as educator will probably be rather personal, I believe it is as applicable, and more needed, today than ever before. This part also brings in focus the foreign graduate student vantage point more than any other in today’s presentation. The first time I met Dr. Krizek was an absolute riot. It was day # 4 for me in America, a typical hot and humid August day in Evanston. I arrived in Ray’s office in my blue blazer and red silk tie, soaked wet in my sweat after the 15-minute walk up Sheridan Avenue from the bus station. I walked in the room to shake hands with my almighty professor, who received me very warmly in his checkered yellow and green Bermuda shorts, and half-sleeve shirt. For an hour or more, he talked very enthusiastically, in elaborate details sometimes, but most of it was over my head. A few years later, I confessed to him that between his Check-Bohemian-Baltimore- Midwestern accent and my limited Quasi-British-Arabic English at the time, I practically didn’t even grasp what the context of the conversation was. But he later told me that he was impressed with my presentation anyway!! Kind of scary, isn’t it? Humor aside, within a month, Dr. Krizek sponsored my very first trip to a ballpark ever; it was the Cubs playing the White Sox at Wrigley field. Ray persuasively argued that no student of his could possibly succeed unless he understood what baseball is all about? But during the game, while explaining to me the totally incomprehensible rules of the game, he reminded me more than once that it might actually take me years to really understand what’s going on the field. Ray: I am still reminded of that by my Son Omar! Then came the first Thanksgiving dinner, the first Christmas Holiday Dinner, the first suburban backyard picnic, complete with hot dogs and- what else- non-stop baseball. Bob Krizek, who was about 6 then, managed to swing the bat, but Kevin at 2, was attached to Claudia’s hip, literally. Without much fanfare or fuss, we were gradually and sincerely drawn into the Krizek Team, almost as family members. Not that everything was rosy all around! Problems and conflicts did arise, and that’s when Ray set many more examples as to how to handle people. That’s when our social skills were pulled, pushed, and honed. Although more comfortable with strategic-level management, Ray was a taskmaster who didn’t shy from details, whenever he had to deal with problems. And he was as good, as tough and blunt as any manager I ever worked for. He didn’t rely on memos or deputies to tell me I was screwing up. He just showed up in the “back room”, of course after 6:00PM, stretched on the couch, and fired away. In about an hour or so, I would know exactly what I needed to do. To this day, I am still struggling to get to that level of effective candor in directing my associates. Let me close with this one, which is very dear to me. There was a time when I too felt that perhaps Dr. Krizek didn’t think I could finish my thesis; he later told me that he never doubted that I would. He did complain about my lack of focus on my
Geotechnical Materials: Measurement and Analysis: R.J. Krizek Commemorative Symposium: 3 Aug 2002
research, but said he knew I could and should finish. Coming from a man whom I saw bluntly advising others not to pursue a Ph.D. program, that meant much to me. But it also conveyed to me another strong principle. We must know our associates and subordinates well enough, and develop sufficient feel for their capabilities, to actually be able to take ownership of their careers, and help sustain their self-confidence when they doubt themselves. Today’s world of business seriously lacks this particular leadership component. It is a costly shortcoming, for it undermines employee loyalty and limits intraorganizational synergy. There is no easy way to conclude this chat. Tunch and I appreciate the opportunity to share with you these cherished memories and sincerely felt reflections about Ray Krizek. But above all, we are grateful for the opportunity to finally tell him to his face, and with so many of his colleagues and associates present, what he meant to many of us as an educator, sponsor and mentor. We are indebted to him and to Northwestern for such a rich learning experience, and unique growing process. Thank you Professor Krizek.