Missouri University of Science and Technology Scholars' Mine

International Conference on Case Histories in (1993) - Third International Conference on Case Histories in Geotechnical Engineering

03 Jun 1993, 10:30 am - 12:30 pm

Prediction of Structural Slurry Wall Behavior

Mark Tamaro Thornton-Tomasetti/Engineers, New York, New York

Pablo Lopez Mueser Rutledge Consulting Engineers, New York, New York

Sibel Pamukcu Lehigh University, Bethlehem, Pennsylvania

Follow this and additional works at: https://scholarsmine.mst.edu/icchge

Part of the Geotechnical Engineering Commons

Recommended Citation Tamaro, Mark; Lopez, Pablo; and Pamukcu, Sibel, "Prediction of Structural Slurry Wall Behavior" (1993). International Conference on Case Histories in Geotechnical Engineering. 23. https://scholarsmine.mst.edu/icchge/3icchge/3icchge-session05/23

This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.

This Article - Conference proceedings is brought to you for free and open access by Scholars' Mine. It has been accepted for inclusion in International Conference on Case Histories in Geotechnical Engineering by an authorized administrator of Scholars' Mine. This work is protected by U. S. Copyright Law. Unauthorized use including reproduction for redistribution requires the permission of the copyright holder. For more information, please contact [email protected]. 1!1!!1111!1! Proceedings: Third International Conference on Case Histories in Geotechnical Engineering, St. Louis, Missouri, ~ June 1-4, 1993, Paper No. 5.06

Prediction of Structural Slurry Wall Behavior MarkTamaro Sibel Pamukcu Engineer, Thornton-Tomasetti/Engineers, New York, New York Assistant Professor, Department of , Lehigh University, Bethlehem, Pennsylvania Pablo Lopez Engineer, Mueser Rutledge Consulting Engineers, New York, New York

SYNOPSIS The following study was undertaken with the intent of improving the ability to accurately predict the behavior of structural slurry walls. An existing wall, employed during the construction of the Washington D.C. subway system, was examined using four different analysis methods. The actual stresses and displacements of this wall were measured, providing a basis for investigating the accuracy of the different analysis techniques. The results obtained warrant the use of one particular approach, referred to in this study as the "Beam on Elastic Method". This method provided the most useful simulation of the /structure interaction that occurred during construction of the subway, in terms of accuracy and amount of work required.

INTRODUCTION . An example of the typical stages of construction for a structural slurry wall are Structural slurry walls have become an presented in figure 1. increasingly popular method of supporting deep excavations in sites where large scale underpinning of adjacent structures would normally be required. They can be utilized as permanent structural components or as temporary retaining systems. The versatility of applications as as efficiency of construction, often make these walls the most suitable retaining method during the construction of deep foundations, and cover , and deep vertical shafts.

HISTORICAL DEVELOPMENT During the early 1900s mud slurry (sodium montmorillonite ) began to be used for the construction of petroleum , panel being excavated and serving a dual purpose of flushing drilling filled with slurry tailings to the surface, and providing circumferential support to the well walls. By the 1950s bentonite slurry was introduced as a means of supporting the excavation of deep . In this application, the Figure ~- S1urry wall construction hydrostatic head and density of the slurry is maintained at a level which provides sufficient sequence lateral support to prevent the side walls of the excavation from caving in. Deep trenches constructed in this manner, are Structural slurry walls were first used often used as cut-off walls and structural during the late 1940s in Milan Italy, where foundation walls. Cut-off walls refer to they were incorporated into the construction of barriers constructed for the purpose subway tunnels and deep building foundations of controlling ground water flow or pollution (Kyle, 1967). migration, where the slurry filled is Presently, slurry wall technology has backfilled with low permeability material. developed to the point where it competes with This type of wall is well suited for containing more conventional retaining methods, such as contaminated ground water, encapsulating land sheet piling or soldier beam and lagging fills, and repair of earth dams. systems (Kapp, 1969). In many cases, Structural slurry walls are generally used as structural slurry walls have proven to be the retaining systems during the construction of most economical method of construction, and subgrade structures. These walls are occasionally, the only feasible option. constructed by installing reinforcing steel cages into slurry filled trenches followed by the displacement of slurry through placement of

Third International Conference on Case Histories in Geotechnical Engineering 695 Missouri University of Science and Technology http://ICCHGE1984-2013.mst.edu METHOD OF CONSTRUCTION designed to act as a vertical load bearing element. structural slurry walls follow a general construction scheme as outlined below. CURRENT DESIGN METHODS 1.) Slot trenches are dug along the perimeter of the proposed excavation. During excavation, Presently in the u.s., there are no specific bentonite slurry is continually pumped into the codes or standards in existence that directly open trench, maintaining a hydrostatic head at regulate the design of structural slurry least several feet above the ground water walls. Design engineers currently use table. geotechnical design guides originally developec 2.) When the trench is excavated to its final for flexible retaining systems to determine thE depth, reinforcement cages, consisting of loading and boundary conditions for a given reinforcing bars andjor steel beams, are wall. Structural codes (ACI and AISC) are ther lowered into place, forming panels. When used to satisfy the strength requirements of vertical steel beams are not used, stop- the proposed walls. Currently used methods arE end devices are implemented to form the joints either analogous to classical that separate panels. Concrete is then design methods, or are modified versions of thE tremmied into each panel, displacing the slurry classical approach. However, two methods upwards where it is pumped out and cleaned for studied in this paper differ from the standard re-use. design procedures, as they attempt to simulate 3.) After the wall is completed, excavation the interaction between the structural slurry begins within the perimeter of the wall. The wall and the soil (soil/structure interaction). excavation is performed in stages, where each Four analysis methods, ranging from simple stage is completed by placing tie-back or strut empirical techniques to highly sophisticated supports at levels specified in the design. models were compared in this paper. Descriptions of these four methods follow.

BENEFITS OF STRUCTURAL SLURRY WALLS Terzaghi-Peck Method (Bowles, 1988) Slurry walls have dramatic benefits over conventional retaining methods when applied in Thi7 is the most elementary method of analysis several specific cases. For example, slurry ava1lable. It was developed for the design of walls are commonly used in areas where ground bracing where the soil conditions were uniform water tables are high. The advantage of using or clay and the water table was below an impermeable concrete wall versus soldier subgrade. The analysis sequence is as follows. beams and timber lagging is often exhibited in First, the active on the the reduced cost of de-watering the site, and wall is determined using Rankine or coulomb the protection of adjacent structures. A site theory (Bowles, 1988). The maximum ordinate of enclosed by a concrete slurry wall requires the lateral pressure is then multiplied by an little dewatering and the surrounding ground appropriate factor. This value of pressure is distributed as a uniform or trapezoidal water table is usually maintained at its normal loading, depending on whether the material level. In contrast, an excavation supported by being supported is sand or clay. A typical a soldier pile and timber lagging system may wall configuration and associated loading for encounter significant dewatering problems as granular backfill is shown in figure 2. A well as possible settlement of surrounding simple support is assumed at the level of structures due to draw-down of the natural subgrade, where the rotational restraints are water table. Slurry walls are also well suited released, leaving the wall to be analyzed as a for projects that require deep excavations. The depth of a slurry wall is usually series of simple-beams, or as a continuous controlled by the limitations of the trenching indeterminate beam. equipment, therefore walls can be constructed Although this method is commonly used to to depths well beyond 100 feet, provided the determine the maximum pending moment in a proper equipment is used. Slurry walls also proposed wall, it was originally intended to be prevail in situations where construction noise used for estimating maximum strut loads. and vibration must not exceed a certain level (e,g. in urban locations where adjacent buildings are occupied). In these cases, the Net Pressure Method with Support noise and vibration generated from driving Settlements (Tamaro & Kerr; 1990) sheeting or soldier beams can not be tolerated. In this approach, the designer determines the The use of structural slurry walls also Rankine active pressure along the entire length eliminates the need for underpinning of adjacent structures. Often, when sheet pile of the wall, as well as the passive pressure retaining systems are used, elaborate beginni~g at subgrade for each stage of underpinning schemes must be developed in order excavat1on. The net pressure resulting from to avoid vertical settlements resulting from the superposition of the active and passive the horizontal deformations of flexible steel pressures then represents the lateral loading sheeting. ~n.t~e wall. The point where the net pressure 1n1t1ally reaches zero, is taken to be the La~tly, slurry walls can be designed to rema1n as the perimeter walls of substructures point of zero moment (assuming there is (e.g. basement walls of buildings). In these suffi~ient passive pressure developed). The cases the temporary struts can be replaced by wall 1s truncated beyond this point and a subgrade floor beams, and the wall can be simple support is assumed. As the excavation

Third International Conference on Case Histories in Geotechnical Engineering 696 Missouri University of Science and Technology http://ICCHGE1984-2013.mst.edu uniformly street level distibuted street level Rankine active pressure l-----1~jj1:)::==:::;:::====~ cr' = 0.65k cr' a a v 1---~~~~~ rigid si:nple support:: 1---o-~:::: ;.:~ Rankine l------1~:~::~=====1 active ;:;' pressure •'• ~---+1···· Structural •'• o' = k o' •'•;•;• Slurry tvall a a v ,.....+---~::;. ,. .. ,. H,.+-----tt.-.·~======1 ...... ,. 1--...... 1----o-1·:-: :~:: Rankine passive r.~~:=:=:::=::==t pressure 1--~---.J·~~~; o'p = k p o'v f---+-t------... .·,·•'•' :~:· building •' surcharge final dspth of hydr_ostatic exacavation pressure Net pressure Figure 2. Terzaghi-Peck Method Figure 3. Net Pressure with Support Settlements Method stages progress and braces are installed, the wall is then analyzed as a continuous beam. The passive pressure is recalculated for each excavation stage prior to placement of the next brace, requiring the net pressure to be updated as well. The critical moment is then street level determined from the analysis of each stage, and the wall is designed accordingly. The effect of wall movements are also addressed in this method. During the first stage, the wall displacement is determined at supports the depth where the future brace is to be placed. In the next stage the brace is installed with this displacement introduced as i support settlement. This process is Rankine active =ontinued for each of the remaining stages as pressure lllustrated in figure 3. By incorporating o'a k a o'v :hese initial displacements into the model, a = nore accurate prediction of the distribution of noments along the depth of the wall is >btained. pressure leam on Elastic Foundation Method (Haliburton, k cr' .979) p v

~he Beam on Elastic Foundation approach is tnique in that it utilizes springs to simulate 1oth the elastic and inelastic behavior of the 1ubgrade . Using this technique, the wall .s loaded by Rankine or Coulomb active pressure 1n the unexcavated side while the subgrade Rankine spring stiffness ·eaction on the excavated side is simulated by pressure .t-rest pressure and a series of springs whose equals the subgrade o' = k o' .tiffness is equivalent to the subgrade 0 0 v modulus .odulus. A typical wall subjected to these oading conditions is shown in figure 4. Figure 4. Beam on Elastic Foundation Similar to the previous method, the braces or ie-backs are modeled as elastic springs, and Method

Third International Conference on Case Histories in Geotechnical Engineering 697 Missouri University of Science and Technology http://ICCHGE1984-2013.mst.edu the displacements that develop in each stage construction. are superimposed on the subsequent stages. Stress-strain characteristics, as well as, This method also incorporates the concept of strength and bulk modulus values had to be limit analysis in that the springs that form approximated by using previously documented the subgrade reaction can be limi~ed in values of samples with similar soil capacity (the magnitude of the spring reactions classifications, due to the lack of triaxial is set as an upper bound equal to the subgrade test data (Duncan, 1980). passive pressure), thereby acting in a manner similar to the plastic behavior of the subgrade soil. For example, during the course of a Wall Loading solution iteration, if the reaction force in a spring exceeds the passive limit (yielded The loading scheme applied on the unexcavate spring), the model is revised by removing the side of the wall included three components. yielded spring and applying its passive limit These were the lateral earth pressure, the as a load. This logic is continued along the hydrostatic pressure, and the length of the wall until the springs no longer lateral component of the bearing pressure of yield. The repetitive nature of this analysis adjacent building foundation. makes the use of an iterative computer program The lateral earth pressure was determined a necessity for obtaining a quick solution. using Rankine theory (Bowles, 1988), where t In addition to the complexity of the effective vertical stress is multiplied by a analysis, another potential problem arises when coefficient representing either the active o using this method. The designer must use a passive lateral pressure. Wall was realistic value for the subgrade modulus considered during the computation of the (spring stiffness). If the value selected or Rankine coefficients of lateral earth pressu computed is mis-representative of the actual This resulted in lateral earth pressures bei soil behavior, the accuracy of the solution dependent on the angle of internal friction will be compromised. This is especially true for each soil stratum. when determining the deflected shape of the Typical hydrostatic pressures were include wall. in the analyses at depths which remained constant on the unexcavated side of the wall and varied on the excavated side for the Finite Element Method (Filz, 1990) different construction stages. Lastly, the surcharge loading created by a Use of a two dimensional finite element mesh to adjacent building foundation was taken from model a plane slicing through a section of both original project specifications provided by the structural wall and the soil surrounding it Washington Metropolitan Area Transit Authori provides two improvements over the previously This loading was trapezoidal in shape, mentioned methods. Most importantly, it allows beginning at the base of the building the soil to be modeled using properties that foundation (twenty feet below the ground are determined from tests of actual soil surface) at a magnitude of 1.2 kips per squa samples from the site. This eliminates the foot over a depth of sixty feet. From a dep need for making assumptions such as equating of sixty feet to eighty four feet the load the passive resistance to spring reactions or decreased linearly from 1.2 kips per square setting a fictitious subgrade point reaction. foot to 0 kips per square foot. These press It also provides the user with resulting diagrams were combined by superposition to f vertical displacements of the soil surface one loading diagram, applied per horizontal adjacent to the wall. This feature is unique foot of wall. to the Finite Element Method, as it is the only Active loads calculated in this manner wer method which includes interaction of the used for the analyses of the Net Pressure wi surrounding soil. Support Settlements and the Beam on Elastic Foundation Methods. For the Terzaghi-Peck Method the magnitude of the active soil DETERMINATION OF BEHAVIOR OF EXISTING WALL pressure calculated at the base of the wall reduced by a factor of 0.65 as suggested by It is necessary to determine several parameters Bowles, (1988). This aajusted soil pressure before analyzing a wall using the methods was distributed evenly along the length of t: discussed in the previous section. The wall, while the building surcharge and existing soil parameters, lateral loading on hydrostatic pressures were superimposed as the wall, flexural stiffness of the wall, and additional loads. the stiffness of the support members must all For the Net Pressure with Support SettlemeJ be estimated. Approaches taken, and and the Beam on Elastic Foundation solutions assumptions made during the calculation of the the subgrade passive pressure was determined above parameters associated with the wall for each excavation stage. Rankine theory w; analyzed in this comparative study, are used to estimate the passive resistance on tl presented below. excavated side of the wall. During the last two excavation stages the high shear strengtl of the over-consolidated clay layers were Soil Parameters included in the calculation of the passive pressure because they added substantially to Values for the angle of internal friction, the passive resistance of the subgrade soils. cohesion, and unit weight of the soil at the It was also.assumed that during construction, site under investigation, were estimated from the excavat1on was de-watered to a level two the original soil samples prior to feet below subgrade at each stage.

Third International Conference on Case Histories in Geotechnical Engineering Missouri University of Science and Technology http://ICCHGE1984-2013.mst.edu 698 Surry Wall Street Level El. 0 o- N " 4 th Street S.W. Fill ;: H""'Vf" 41-29 ... slurry walls Silty Clay 41 =26 • Silty Sand ~-34 I Terminal I! ! 1\ Refrigeration " c Gravelly Sand ~ =36 l?uilding I~ _g I ---}V l:i; panel Silty Sand 41 = 34 numbers I r,,.~ .. .. I Overconsolidated Su = 2.5 ksf Clay 41 = 25 IStreet "D" I -··' Overconsolidated Su = 4.0 ksf 1 I 78+00 I Clay 41 = 25 ,, Plan View Overconsolidated Su = 5.0 ksf -~EI. -84 It Clay <1>=34

Figure 6. Soil Profile Terminal Refrigeration Building compacted fill stresses. If the increased stress exceeded the pre-consolidation stress in the lower clay completed strata, consolidation and settlement would occur. The occurrence of differential X - Segtion settlement in areas adjacent to the site would have had profound effects on the existing Figure 5. Slurry wall site concrete and masonry buildings including severe cracking and possible structural failure.

Subsurface Site Conditions Wall Construction The site conditions of the Federal center Two parallel slurry walls were installed project were representative of the typical seventy feet apart, extending to a maximum conditions for which structural slurry walls depth of eighty four feet and are approximately are best suited. The specific features were a eleven hundred feet long in plan. Occupied high water table, a sub-stratum layer of low buildings were located near the walls along permeability high soils, and a both sides of the excavation, leaving little site surrounded by structures highly sensitive room for construction equipment and requiring to settlement. strict control of ground settlements associated Boring samples taken at the site described with horizontal wall movements. These the geological profile as a thin layer of fill, buildings imposed surcharge loads on the walls followed by several layers of compact, medium in addition to typical horizontal earth and to coarse sand, and two layers of dense over­ hydrostatic pressures. consolidated clay located approximately sixty Both walls were constructed as a series of to ninety feet below street level as shown in seven foot long panels separated by steel figure 6. soldier beams. A trench was excavated for each The two clay layers had sufficient bearing panel using a special thirty two inch wide, capacity to support the vertical loads from the nine ton clamshell bucket followed by the walls, while their relatively high shear placement of W30 x 211 soldier beams, connected strengths provided significant passive by a reinforcing cage, at each end of the resistance at the base of the wall. The clay trench. Concrete was then placed into the layers also acted as an impermeable layer trench by methods, forming a series of essential for limiting groundwater from flowing separate panels. After completion of the wall under the base of the wall and into the panels, the excavation was advanced in five excavated area. stages. The natural water table, was located twenty Initially, the subgrade level was lowered to two feet below street level. Variations in the a depth two feet below the position of the water level, particularly draw-down associated first set of struts to be installed. At this with dewatering the excavation, could have level, the struts were placed and excavation caused an increase in the effective vertical continued to a depth two feet below the next

Third International Conference on Case Histories in Geotechnical Engineering 699 Missouri University of Science and Technology http://ICCHGE1984-2013.mst.edu experience with similar projects and empirical movements for the last two stages of relationships, or limited by the serviceability excavation. The deflections exhibited in th of the wall. stages leading up to this point, do however, appear to provide an adequate approximation the actual measured values. The discrepanci Beam on Elastic Foundation Method that occur in the late stages may be explain by considering the accumulating effect of The Beam on Elastic Foundation Method, produced under-estimated support settlements on the conservative moment values. It resulted in over-all deflected shape. When the initial critical moments that were of the same computed deflections are lower than the actu magnitude, or larger than the measured moments deflections, the introduction of these valu' in the wall for both positive and negative wall as support settlements over a series of stag, bending. Use of this method for determining can compound the errors. To minimize the maximum moments should yield a safe design. effect of this problem, the model could be r• However, obtaining a solution requires the analyzed with increased support settlements, ability to perform a sophisticated computer thus assuring that the predicted deflections analysis. would be conservative.

Finite Element Method Beam on Elastic Foundation Method The moments produced by the Finite Element The Beam on Elastic Foundation Method produc4 analysis were more conservative than the Beam conservative deflection values for all stage! on Elastic Foundation Method. The increase in In addition, this method addresses the the moment values is mainly attributed to an possibility of deflections below subgrade. underestimation of the passive soil strength. This not only contributes to the conservativ4 Use of these values would also produce a safe nature of the solution, but also identifies design. However, these results would produce an instability conditions at the base of the wa: "over designed" wall leading to higher by displaying excessive base movements. The construction costs. results obtained by using this method can be The moments from this method also exemplify a improved by varying the flexural stiffness major shortcoming of this analysis technique. along the depth of the wall. This allows th4 The solution proved to be highly sensitive to designer to use the "gross" moment of inerti< inaccuracies of the soil element properties. along segments of the wall that are subjectec Large amounts of time spent on achieving a to low stress levels, while using the "crack4 solution may not be warranted due to this moment of inertia along wall segments that sensitivity. undergo stresses sufficient to produce tensi4 cracking in the concrete. However, it is important to note that such refinements may 1 Geotechnical Conclusions be justified until more accurate estimates Ol subgrade moduli can be developed. The geotechnical engineer is mainly concerned with the settlement of structures adjacent to the excavation site. These settlements are Finite Element Method caused by disturbances in the soil beneath neighboring structures, such as a reduction of As previously mentioned, the Finite Element the groundwater table, or lateral soil solution was adversely affected by inaccurat4 movements. Structural slurry walls have proven soil properties. In terms of lateral to be successful in maintaining the level of deflections this resulted in a translational the local water table, but excessive wall movement at the base of the wall of one inch, movements must be controlled to assure the as shown in figure 9. This horizontal safety of surrounding structures. With this in translation magnified the deflections due to mind, the analysis methods were judged on their bending over the remainder of the wall. ability to accurately predict the wall Because the measured deflection of the actuaJ movements throughout the excavation process. wall showed no movement at the base, it can l concluded that the Finite Element results arE dubious. Terzaghi-Peck Method Although this solution proved to be unreliable, the time spent on the Finite The Terzaghi-Peck Method is unsuitable for Element Method may be justified by the fact determining deflections of slurry walls that that it does produce the values of vertical are constructed with more than one level of displacements adjacent to the wall. In this bracing. This method does not consider the study settlement under the adjacent building fact that walls can accumulate displacements amounted to one inch, approximately uniform prior to placement of braces, or that rigid along the base of the building. This body movements can occur. information is valuable but dependent on accurate soil information. Net Pressure with Support Settlements Method CONCLUDING REMARKS The results obtained from this study indicate that the Net Pressure with Support Settlements This study was undertaken in order to improve Me.thod under-estimates the lateral wall the understanding of how a structural slurry

Third International Conference on Case Histories in Geotechnical Engineering Missouri University of Science and Technology 700 http://ICCHGE1984-2013.mst.edu 20 REFERENCES Beer, Ferdinand P., Johnston, Russel E. Jr., street level Mechanics of Materials, New York, McGraw­ Hill Inc., 1981. o-1---...-m • Backflgured deflections u.-- El. -4' Bowles, Joseph E., Foundation Analysis and . ~, --- Net Pressure Design 4th Ed., New York, McGraw-Hill Inc., ~ 1988. g ' ~' - Beam on Elastic Found. II) -20 '~ -ll- Rnite Element Chapra, Steven c., Canale, Raymond P., 14-_._- El. -26'' ~ Numerical Methods For Engineers, New York, ~ ~ :::J \ McGraw-Hill Inc., 1988. en \ .,._struts -g ~ \ Duncan, J.M., Byrne, Peter, Wong, Kai S., 5.... -40 ·• ~ Mabry, Phillip, strength Stress-strain and 01 El. -44' '~ Bulk Modulus Parameters for Finite Element I E I I Analysis, Virginia Tech, Blacksburg Virginia, I ,g I 4 1980. .c I /lo a -so EPA, "Slurry Trench Construction for Pollution ii.l 0 . Migration Control", Report No. 540/2-84-001, ... Cincinnati, Ohio, 1984 . . \ Filz, George, Clough, Wayne g., Duncan, J. M., -80 Subgrade Draft User's Maual for Program Soilstruct El.-62' (Isotropic) Plain-strain with Beam Element, Virginia Tech, Blacksburg Virginia, 1990. Haliburton, T. Allan, Soil-Structural -100 +---.--.--~-r--...-----.-~--. Interaction, Technical Publications No. 14, . 1 0 2 3 4 Oklahoma State University, 1979.

Wall Displacement (ln.) !COS Corporation, company brochure. Skempton, A. W., "The Bearing Capacity of Figure 9. Deflection Comparison Clays", Building Research congress, Division I, Part 3, London, 1951. Tamara, George J., "Perimeter Wall for New wall behaves when constructed in a manner World Center, Hong Kong", ASCE Journal of similar to the procedures followed in this Construction Division, 1981. example. For the wall under investigation, it is apparent that the major differences between Tamara, George J., Kerr, William c., "Diaphragm the analyzed behavior and the actual behavior Walls- Update on Design and Performance", of the wall can be attributed to an incorrect Cornell ASCE Soils & Foundations Special prediction of the inward displacements of the Publication, 1990. wall prior to the placement of brace levels. This most likely has caused inaccuracies Terzaghi, Karl, "Evaluation of Coefficients of resulting in under-conservative, or critical Subgrade Reaction", Geotechnigue, Vol. 5, predictions in the Terzaghi-Peck and Net December, 1955. Pressure models. The Beam on Elastic Foundation and Finite Element Methods both Ware, Kenneth R., Underpinning Methods and allowed the base of the wall to displace Related Movements, RE~T Proceedings, Vol. 2, laterally, and in most stages produced 1974. conservative deflection and moments in the wall. However, the Finite Element Method's sensitivity to soil parameters precludes its use unless reliable triaxial test results for the appropriate load ranges are available. Therefore, it can be concluded that the use of the Beam on Elastic Foundation Method is the most reliable method of analysis among the four approaches under investigation for designing structural slurry walls. However, in the absence of a computer program capable of determining a solution to this model, the Net Pressure with Support Settlements Method could be used to attain a safe design, provided that a range of initial support settlements are analyzed.

Third International Conference on Case Histories in Geotechnical Engineering 701 Missouri University of Science and Technology http://ICCHGE1984-2013.mst.edu