@prefix vivo: . @prefix edm: . @prefix ns0: . @prefix dcterms: . @prefix dc: . @prefix skos: . vivo:departmentOrSchool "Applied Science, Faculty of"@en, "Civil Engineering, Department of"@en ; edm:dataProvider "DSpace"@en ; ns0:degreeCampus "UBCV"@en ; dcterms:creator "Wu, Guoxi"@en ; dcterms:issued "2008-09-11T19:27:40Z"@en, "1992"@en ; vivo:relatedDegree "Master of Applied Science - MASc"@en ; ns0:degreeGrantor "University of British Columbia"@en ; dcterms:description "Previous investigations of the seismic stability of Sardis Dam indicated that the entire central silt core and a weak clayey silt layer in the dam foundation are susceptible to liquefaction which could result in large losses of shear strengths in these liquefied soils. The post-liquefaction behaviour of Sardis Dam was evaluated using new flow deformation analysis technique which was developed by Finn and Yogendrakumar (1989). The analysis showed the potential for large displacements including a great loss of freeboard during the design earthquake. A strategy of designing remedial measures to limit deformations to a tolerable amount was adopted over the conventional factor of safety approach. Various levels of remediations were investigated using TARA-3FL. The remediation procedure adopted for field trials was anchoring the upstream slope to the foundation using rows of rectangular prestressed reinforced concrete piles. Estimating post-liquefaction deformations for this remediation scheme posed challenging problems in analysis."@en ; edm:aggregatedCHO "https://circle.library.ubc.ca/rest/handle/2429/1859?expand=metadata"@en ; dcterms:extent "4516448 bytes"@en ; dc:format "application/pdf"@en ; skos:note "SEISMIC INDUCED FLOW DEFORMATION ANDREMEDIATION STUDY OF SARDIS DAMbyGuoxi WUB.Eng. Nanjing Institute of Architechtural Engineering, 1984M.A.Sc. Tongji University, Shanghai, PRC, 1987A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFMASTER OF APPLIED SCIENCEinTHE FACULTY OF GRADUATE STUDIESCIVIL ENGINEERINGWe accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIAApril 1992© Guoxi WU, 1992In presenting this thesis in partial fulfilment of the requirements for an advanced degree atthe University of British Columbia, I agree that the Library shall make it freely availablefor reference and study. I further agree that permission for extensive copying of thisthesis for scholarly purposes may be granted by the head of my department or by hisor her representatives. It is understood that copying or publication of this thesis forfinancial gain shall not be allowed without my written permission.CIVIL ENGINEERINGThe University of British Columbia2324 Main MallVancouver, CanadaV6T 1Z4Date: ABSTRACTPrevious investigations of the seismic stability of Sardis Dam indicated that the entirecentral silt core and a weak clayey silt layer in the dam foundation are susceptible toliquefaction which could result in large losses of shear strengths in these liquefied soils.The post-liquefaction behaviour of Sardis Dam was evaluated using new flow deformationanalysis technique which was developed by Finn and Yogendrakumar(1989). The analysisshowed the potential for large displacements including a great loss of freeboard during thedesign earthquake. A strategy of designing remedial measures to limit deformations to atolerable amount was adopted over the conventional factor of safety approach. Variouslevels of remediations were investigated using TARA-3FL. The remediation procedureadopted for field trials was anchoring the upstream slope to the foundation using rowsof rectangular prestressed reinforced concrete piles. Estimating post-liquefaction defor-mations for this remediation scheme posed challenging problems in analysis.iiTable of ContentsABSTRACT^ iiList of Tables ivList of Figures^ vAcknowledgement vi1 INTRODUCTION^ 12 DESCRIPTION OF SARDIS DAM^ 63 PREVIOUS SEISMIC STUDIES 93.1 Previous Investigation ^93.2 Potential for Liquefaction in Fine Grained Soils ^113.3 Conclusions Drawn from Previous Seismic Study ^ 144 THEORY OF POST LIQUEFACTION ANALYSIS 184.1 Conventional Static Equilibrium Method ^ 184.2 The Deformation Analysis Theory 185 RESIDUAL SHEAR STRENGTH^ 245.1 Introduction ^245.2 Determination of Residual Strengths ^ 265.2.1 Steady state concept ^26iii65.2.2^A correction between residual strengths and SPT blowcounts . . .5.3^Evaluation of Residual Strength on Sardis Dam ^LIQUEFACTION DEFORMATION ANALYSIS3233366.1 Initial Stress Conditions in Dam and Foundation ^ 366.1.1^Finite element mesh ^ 366.1.2^Soil material properties 396.1.3^Pool water level and water force ^ 426.2 Description of Liquefaction Analysis 426.2.1^Liquefiable materials ^ 426.2.2^Soil properties after liquefaction ^ 436.2.3^Results of liquefaction analyses 467 GENERAL REMEDIATION STUDY OF SARDIS DAM 537.1 General Remediation Schemes ^ 537.2 Effect of the Pore Water Pressure of the Non-Liquefiable Sand Shell 577.2.1^Loss of freeboard ^ 587.2.2^Horizontal pressures against the plug ^ 587.2.3^Ratio of shear stress to strength of the plug ^ 617.2.4^Horizontal displacement of the plug 617.3 Effect of Plug Width ^ 637.3.1^Loss of freeboard 637.3.2^Horizontal pressures against the plug ^ 667.3.3^Ratio of shear stress to strength of the plug ^ 667.3.4^Horizontal displacement of the plug 697.4 Effect of Plug Strength ^ 717.4.1^Loss of freeboard 71iv7.4.2^Horizontal pressures against the plug ^717.4.3^Ratio of shear stress to strength of the plug ^747.4.4^Horizontal displacement of the plug ^767.5 Summary on General Remediation Studies ^ 768 REMEDIATION STUDY OF PILE-REINFORCED SECTION^848.1 Equivalent Composite Material Properties of the Pile Reinforced Section^848.2 Comparison of Results Between the Pile-Reinforced Section And the Plug 998.2.1^Loss of freeboard ^998.2.2^Horizontal pressures against the remediated zone ^ 1008.2.3^Ratio of shear stress to strength of the remediated zone ^ 1008.2.4 Horizontal movement of the remediated zone ^ 1039 SUMMARY AND CONCLUSIONS^ 105Bibliography^ 108vList of Tables6.1 Parameters of Strength and Stiffness Used in the Construction Analysis . 416.2 Variations of Residual Strengths in the Weak Clayey Silt (psf) ^ 446.3 Parameters of Strength and Stiffness after Liquefaction ^ 456.4 Summary Results of Liquefaction Analysis (WT277 model) ^ 507.1 Different Plug Remediation Schemes ^ 567.2 Summary Results of Plug Remediation Studies ^ 788.1 Composite Strengths and Moduli of the Pile-Reinforced Section ^ 99viList of Figures2.1 Typical Section of Sardis Dam (Finn et al., 1990a) ^73.1 Longitudinal Section Showing Liquefiable Zones in Shell and FoundationSands 139 m(450 ft) Upstream from Centerline (Finn et al., 1990a) . . .^103.2 Plan View of Liquefiable Weak Clayey Silt Zones (Finn et al., 1990a) . .^153.3 Longitudinal Section Showing Liquefiable Weak Clayey Silt Zones 139m(450 ft) Upstream from Centerline (Finn et al., 1990a) ^164.1 Adjusting Stress-Strain State to Post-Liquefaction Conditions (after Finnet al., 1990b) ^215.1 Types of Stress Strain Curve for Consolidated Undrained Triaxial Testson Clean Sand ^ 255.2 Changes in Driving Stresses and Undrained Shear Strength During anEarthquake ^ 255.3 Steady State Line ( void ratio vs effective confining pressure) ( after Pouloset al, 1985) ^285.4 Steady State Line (void ratio vs steady state strength) (after Poulos et al, 1985) ^285.5 Poulos Procedure for Determining Steady State Strength ( after Poulos etal., 1985) ^315.6 Cross Section of Lower San Fernando Dam Showing Liquefaction Zone(Seed et al., 1988) ^31vii5.7 Tentative Relationship Between Residual Strength and SPT N Values forSands (after Seed et al., 1988) ^ 345.8 Residual Strength in Topstratum Clay — Sardis Dam (Finn et al., 1990a) 346.1 Finite Element Mesh Showing Element Distribution ^ 376.2 Finite Element Mesh Showing Node Distribution 386.3 Distribution of Soil Material Zones ^ 406.4 Variation of Post Liquefaction Configurations with Minimum ResidualStrength — WT277 model ^ 476.5 Variation of Loss of Freeboard with Minimum Residual Strength ^ 486.6 Variation of Horizontal Displacement with Minimum Residual Strength . 516.7 Maximum Ratio of Shear Stress to Strength in the Weak Clayey Silt VersusMinimum Residual Strength ^ 527.1 Finite Element Mesh Showing Location of Remedial Pile Plug ^ 547.2 Detailed Distribution of Remedial Elements ^ 547.3 Variation of Loss of Freeboard with Residual Strength (PWP effect) . . . 597.4 Pressure Distribution on Downstream Face of Remediated Section (PWPeffect) ^ 607.5 Pressure Distribution on Upstream Face of Remediated Section (PWP effect) 607.6 Distribution of Ratio of Shear Stress to Strength at Downstream Face ofRemediated Section (PWP effect) ^ 627.7 Maximum Ratio of Shear Stress to Strength in Remediated Section versusResidual Strength (PWP effect) ^ 627.8 Distribution of Downstream Horizontal Movement of Remediated Section(PWP effect) ^ 64viii7.9 Horizontal Displacement at Downstream Edge with Residual Strength(PWP effect) ^647.10 Variation of Loss of Freeboard with Residual Strength (Plug width effect) 657.11 Pressure Distribution on Downstream Face of Remediated Section (Plugwidth effect) ^677.12 Pressure Distribution on Upstream Face of Remediated Section (Plugwidth effect) ^677.13 Distribution of Ratio of Shear Stress to Strength at Downstream Face ofRemediated Section (Plug width effect) ^687.14 Maximum Ratio of Shear Stress to Strength in Remediated Section versusResidual Strength (Plug width effect) ^687.15 Distribution of Downstream Horizontal Movement of Remediated Section(Plug width effect) ^707.16 Horizontal Displacement at Downstream Edge with Residual Strength(Plug width effect) ^707.17 Variation of Loss of Freeboard with Residual Strength (Plug strength effect) 727.18 Pressure Distribution on Downstream Face of Remediated Section (Plugstrength effect) ^737.19 Pressure Distribution on Upstream Face of Remediated Section (Plugstrength effect) ^737.20 Distribution of Ratio of Shear Stress to Strength at Downstream Face ofRemediated Section (Plug strength effect) ^757.21 Maximum Ratio of Shear Stress to Strength in Remediated Section versusResidual Strength (Plug strength effect) ^757.22 Distribution of Downstream Horizontal Movement of Remediated Section(Plug strength effect) ^77ix7.23 Horizontal Displacement at Downstream Edge with Residual Strength(Plug strength effect) ^777.24 Loss of Freeboard versus Plug Strength ^797.25 Maximum Ratio of Shear Stress to Strength in Plug versus Plug Strength 807.26 Maximum Horizontal Displacement in Plug Versus Plug Strength . . . 817.27 Variation of Typical Post liquefaction Configurations after Remediation —Plug Strength 3000 psf, Plug Width 120 ft) ^838.1 Cross Section of Sardis Dam Showing Remediation Piles ^858.2 Plane View of Layout of Remediation Piles ^858.3 Finite Element Model of Single Pile-Soil System ^878.4 Boundary Conditions and Loading Distribution of Single Pile-Soil System 888.5 Shear Stress - Strain Relationship for the Pile-Reinforced Section, Eleva-tion 173 to 185 ft ^8.6 Shear Stress - Strain Relationship for the Pile-Reinforced Section, Eleva-tion 185 to 191 ft ^8.7 Shear Stress - Strain Relationship for the Pile-Reinforced Section, Eleva-tion 191 to 200 ft ^8.8 Shear Stress - Strain Relationship for the Pile-Reinforced Section, Eleva-tion 200 to 205 ft ^8.9 Shear Stress - Strain Relationship for the Pile-Reinforced Section, Eleva-tion 205 to 215 ft ^8.10 Shear Stress - Strain Relationship for the Pile-Reinforced Section, Eleva-tion 215 to 220 ft ^8.11 Shear Stress - Strain Relationship for the Pile-Reinforced Section, Eleva-tion 220 to 230 ft89909192939495x8.12 Shear Stress - Strain Relationship for the Pile-Reinforced Section, Eleva-tion 230 to 240 ft ^968.13 Shear Stress - Strain Relationship for the Pile-Reinforced Section, Eleva-tion 240 to 250 ft ^978.14 Variation of Shear Strength in Pile - Reinforced Section Versus Elevation 988.15 Pressure Distribution on Downstream Face of Pile - Reinforced Section . 1018.16 Distribution of Ratio of Shear Stress to Strength at Downstream Face ofPile - Reinforced Section ^ 1028.17 Distribution of Downstream Horizontal Movement of Pile - ReinforcedSection ^ 104xiAcknowledgementI would like to thank my research supervisor, Professor W.D. Liam Finn, for his helpfulguidance, suggestions and encouragement throughout my research.)diChapter 1INTRODUCTIONSardis Dam is a hydraulic fill, flood control structure located in Northwestern Missis-sippi. It has a length of 4600 m (15,000 ft) and a maximum height of 36 m (117 ft).From previous seismic investigations, it has been determined that the central silt core ofthe dam and a weak clayey silt layer in the dam foundation are susceptible to liquefac-tion under the design earthquake. Liquefaction could result in large movements of theupstream slope and a substantial loss of freeboard. The crucial problem of Sardis damis how to evaluate the post-liquefaction behaviour of the dam and how to select effectiveremediation measures.In general, there are two major problems confronting the soil engineers dealing with asituation where soil liquefaction may occur. The first problem is determining the condi-tions required to trigger soil liquefaction. This process is usually termed soil liquefactionpotential assessment. For sands, soil liquefaction potential may be evaluated by usingSeed's liquefaction assessment chart based on the in-situ Standard Penetration Tests(SPT). For soils with plastic fines, soil liquefaction potential can be assessed by usingthe criteria which was developed by Wang (1979). Evaluation of soil liquefaction poten-tial has been described by Finn (1985), which is based on the level of the dynamic orexcess pore water pressure generated during an earthquake loading. The assessment ofsoil liquefaction potential made during the previous seismic studies of Sardis Dam willbe presented in chapter 3.1Chapter 1. INTRODUCTION^ 2The second problem dealing with soil liquefaction is determining the performance ofthe soil structures after soil liquefaction. This process includes the determination of theresidual strengths of the liquefied soils and the assessment of overall stability of the soilstructures and the level of permanent deformations.The residual strength of the liquefied soil plays a key role in the assessment of thepost-liquefaction behaviour (Seed,1987). Generally, two methods are used to evaluate theresidual strengths of liquefied soils. Poulos (1985) proposed that the residual strength bedetermined from undrained triaxial tests on undisturbed samples with appropriate cor-rections for differences between field and test void ratios. Seed (1988) recommended thatthe residual strength be determined using a correlation between the residual strengthsand SPT blowcounts developed by analysis of case histories. The determination of resid-ual strengths by the two methods is discussed in detail in Chapter 5.The stability assessment of soil structures includes two approaches, static equilibriumanalysis and the estimation of permanent deformations. The static equilibrium approachis based on the concept of an acceptable factor of safety. The deformation approach isbased on the concept of acceptable deformations.The conventional method in assessing the stability of an earth structure involving soilliquefaction is the static equilibrium method. In this method, the factor of safety againsta shear failure along a specified potential failure surface is determined by using residualstrengths in the liquefied soils as the mobilized shear strengths. The minimum value ofthe factors of safety for all possible potential failure surfaces is used to define the stabilityChapter 1. INTRODUCTION^ 3of the soil structures under the earthquake loading. The purpose of the static equilib-rium method is determining whether a shear failure will occur after soil liquefaction. Inearth structures zones of liquefied soils or zones of degraded strengths may lead to ac-ceptable deformations although the factor of safety based on the original geometry of thestructure is less than what is normally considered acceptable and in some cases less thanunity. The use of deformation criteria can lead to substantial savings in remediation costs.Therefore, for these reasons and in keeping with the concept of designing dams foracceptable deformations proposed by Newmark (1965), there is a tendency to move awayfrom the factor of safety concept and to evaluate the extent of necessary remedial mea-sures on the basis of a tolerable amount of deformation for the low probability eventspecified by the design earthquake. This deformation approach requires a reliable methodof estimating post-liquefaction deformations.A nonlinear finite element method for analyzing the post liquefaction response forsoil structures has been developed by Finn and Yogendrakumar (1989) which is incorpo-rated in the computer program TARA-3FL. The program has the capability of computingpotential flow deformations. The basic methodology of this flow analysis method is tosimulate the sequence of shear strengths in liquefied soils. In the proposed flow analysis,the soil structure is analyzed by using initial or pre-earthquake soil strengths and modulibefore applying a seismic loading. The stress-strain field prior to the seismic loading isdetermined. When the seismic loading is applied, rises of dynamic porewater pressureslead to reductions in shear strengths and shear moduli of saturated soils which causeunbalanced shear stresses in the soil structure. As the unbalanced shear stresses are re-distributed throughout the soil structure, a new stress-strain field is established and flowdeformations are obtained for this level of shear strengths. The process continues untilChapter 1. INTRODUCTION^ 4all strengths reach either residual or minimum values for the level of shaking. For designpurposes it is more convenient to evaluate flow deformations using static analysis undergravitational loads. If deformations under gravitational loading are acceptable then theeffects of seismic loading are evaluated.The deformation analysis method described above is used for the post-earthquakedeformation analysis of Sardis Dam. The analysis showed that large deformations inthe upstream slope of the dam could occur under an earthquake loading for the normalpool elevation. Remediation measures are proposed and their effectiveness in controllingdeformations and the loss of freeboard are evaluated by using TARA-3FL.OUTLINE OF THESISChapter 2 provides a brief description on Sardis dam .Chapter 3 presents the previous seismic investigations on Sardis dam. Potential liq-uefaction zones and critical zones are illustrated.Chapter 4 presents the basic theory of static post-liquefaction deformation analysis.A brief review of the development of earthquake induced deformation analysis and itsapplications are described.Chapter 5 describes the procedures for determining residual strengths based on lab-oratory undrained triaxial tests or in-situ Standard Penetration Tests. The advantagesand disadvantages of these procedures are discussed.Chapter I. INTRODUCTION^ 5Chapter 6 presents post-liquefaction deformation analyses of Sardis dam. Finite el-ement modelling, determination of strength and stiffness parameters and the results ofpost liquefaction deformation analysis are given in the chapter.Chapter 7 presents the results of remediation studies. The performances of Sardisdam after soil liquefaction for different levels of remediation are compared.The preferred remedial method for Sardis dam is anchoring the dam to the foundationby driving large piles through the upstream slope. This procedure is discussed in chapter8. A preliminary assessment of the effectiveness of this procedure is evaluated by usingTARA-3FL analysis.Chapter 9 presents conclusions drawn from the studies in previous chapters and makessome suggestions for further studies.Chapter 2DESCRIPTION OF SARDIS DAMSardis Dam, a hydraulic fill , flood control structure, is located in northwestern Missis-sippi, approximately ten miles southeast of the town of Sardis on the little TallahatchieRiver, a tributary of the Yazoo River. Sardis along with three other dams (Arkabutla,Enid and Grend) are the principal features of the Yazoo Basin Headwater project. Thepurpose of the associated reservoirs is flood control, however, they also provide opportu-nities for recreation and enhance local navigation on the Yazoo River.The total length of Sardis Dam is approximately 4600 m ( 15,000 ft), with a maxi-mum height of 36m ( 117 ft). The central portion of the dam, located in the floodplainof the Little Tallahatchie River, ranges in height from 28 m (90 ft) to 36 m (117 ft), andis approximately 2620 m (8500 ft) long. This central portion of the dam was constructedby hydraulic filling, and consists of a predominantly silt core surrounded by a sand shell,shown in Fig. 2.1.The dam foundation consists of a 3 m (10ft) to 6 m (20 ft) thick zone of natural siltyclay, designated as the topstratum clay, Fig. 2.1, which extends approximately 370 m( 1200 ft) upstream of the dam centerline. In the areas of the original streambed, thetop stratum clay was missing and a 3 m (10 ft) thick silty clay rolled fill was placed inthose areas. The topstratum clay is underlain by pervious alluvial sands (substratumsands) which are approximately 12 m (40ft) thick and are underlain by Tertiary silts and6SPILLWAY CRESTEL. 281.4 lirTOP OF DAM 14EL 311.4EL 282.4,371.5PRESSURE RELIEFWELLS NOT SHOWN25EL 249.4^I BERM=DUMPED RIPRAP ON----GRAVEL BLANKET='CORE --1- --`MOULDER^ TOE DRAIN SYSTEMTOPSTRATUM CLAYE22=1111111\"I- ------LEGEND CLAY OR COMPACTED FILLHYDRAULICALLY PLACED SANDSILTSALLUVIAL DEPOSITChapter 2. DESCRIPTION OF SARDIS DAM^ 7Figure 2.1: Typical Section of Sardis Dam (Finn et al., 1990a)clays. During dam construction , the topstratum clay was removed from beneath thedownstream portion of the dam to help control under seepage. The project was built inthe late 1930's.Hydraulic studies indicate that the probable maximum flood (PMF) would resultin a reservoir level 3 m (10 ft) below the embankment crest. The limited dischargecapacity of the outlet works prevents maintaining a specified reservoir level during peri-ods of even moderate rainfall. The difficulty in maintaining a constant reservoir level hasa significant impact on the feasibility of performing remedial work on the upstream slope.Proximity of Sardis Dam , a hydraulic fill structure , to the New Madrid area, a re-gion of significant historical seismicity , led to concern about the possibility of seismicallyinduced liquefaction of portions of the dam and foundation and the stability of the damChapter 2. DESCRIPTION OF SARDIS DAM^ 8under seismic loading. The U. S. Army Corps of Engineers, Vicksburg District, under-took several studies to evaluate the probable behaviour of the dam during and after anearthquake . The results of these investigations indicated that some remedial measureswere necessary to improve the stability of Sardis Dam during seismic loading.Chapter 3PREVIOUS SEISMIC STUDIES3.1 Previous InvestigationUsing the results of field and laboratory testing, conventional seismic assessment proce-dures (described in next chapter) were followed to predict the effects of the maximumcredible earthquake on the structure. In Sardis dam, many borings were drilled and stan-dard penetration tests(SPT) were performed both in the dam and the foundation soils.Laboratory testing of undisturbed samples taken from the borings included classificationtests and static and cyclic triaxial tests.From the previous investigations, it was concluded that the downstream stability ofthe dam is adequate during and after the earthquake. However, previous investigatorsfound zones with the potential for liquefaction or significant strength loss which couldthreaten the upstream stability of the dam. These zones include the hydraulically placedsilt core, and a discontinuous layer of weak clayey silt located in the foundation beneaththe upstream slope of a 310 m ( 1000 ft) long portion of the dam. Preliminary field ex-ploration also indicated the possible existence of discontinuous layers of weak clayey siltin other areas of the dam foundation. The upper 3 m (10 ft ) to 9 m (30 ft) of sand shellalong the lower portion of the upstream slope was also identified as having a potentialfor liquefaction; however, loss of strength in this zone has a relatively small effect on thestability of the dam.9Chapter 3. PREVIOUS SEISMIC STUDIES^10MI uouguat woos Ism•: renew ma totropuoemaaz sum cum ANTON rt. asE21smos Are sure sues=3 CLAY • SLTSFigure 3.1: Longitudinal Section Showing Liquefiable Zones in Shell and FoundationSands 139 m(450 ft) Upstream from Centerline (Finn et al., 1990a)Fig. 3.1 shows zones of potential liquefaction within the shell and substratum sandswhere the predicted factor of safety against liquefaction is less than unity. This is a sec-tion about 139 m (450 ft) upstream. There was some concern that unacceptable excessporewater pressure might be generated in areas where the predicted factor of safety isbetween 1.0 and 1.25, and these zones are shown with hatched vertical lines. The studyshowed that liquefaction might occur in the upper portions of the upstream shell begin-ning between 31 m ( 100 ft) and 77 m (250 ft) upstream of the centerline and extendingat least. 139 m ( 450 ft) upstream.The factor of safety with respect to upstream stability of the dam would still be ade-quate except in areas where the weak clayey silt layer occurs beneath the upstream slopeChapter 3. PREVIOUS SEISMIC STUDIES^ 11within 77 m (250 ft) of the centerline, even though the silt core might liquefy along theentire length of dam.3.2 Potential for Liquefaction in Fine Grained SoilsField data was evaluated and additional exploration and testing were conducted to locateany other zones of weak clayey silt. Discontinuous layers of the weak clayey silt weresubsequently identified in two areas outside the original 310 m (1000 ft) long section.Liquefaction potential of the weak clayey silt in the original investigation was deter-mined using the Chinese criteria developed by Wang (1979). These criteria are :• per cent finer than 0.005 mm < 20%• liquid limit, LL < 35%• natural water content > 0.9 LL• liquidity index,/,„ < 0.75Liquefaction or significant loss of shear strength will occur for soils which satisfy all fourcriteria. In addition, any fine grained soils for which the standard penetration resistanceN < 4 were also assumed to liquefy or suffer significant strength loss whether they sat-isfied the Chinese criteria or not. The Chinese criteria were applied strictly with noaccount taken of uncertainties in the measurement of the parameters in the criteria. Inthe later investigations these uncertainties are taken into account.Chapter 3. PREVIOUS SEISMIC STUDIES^ 12In the original investigation, the residual strength of the weak clayey silt in the foun-dation was established on the basis of judgement, the Seed's (1987) criterion, and lab-oratory vane tests. Sample disturbance and subsequent reconsolidation prior to testingresulted in higher values for the residual strengths determined by the laboratory vanetests than would be expected in situ. A more elaborate procedure was adopted for thelater investigations involving use of cone penetration testing and field vane test data.In the later investigation, Woodward Clyde consultants (1989) suggested that al-lowances should be made for uncertainties in the measured values of the parameters inthe criteria. They recommended ignoring the liquidity index and making the followingchanges in the measured soil properties before applying the criteria:• decrease per cent fines by 15%• decrease LL by 5%• increase water content by 3%These changes increased significantly the extent of the soils vulnerable to liquefac-tion and strength loss so that almost the entire length of the dam required remediation.Therefore, the Vicksburg District engineers reviewed reports on the scatter in measuredindex properties in U.S. Corps of Engineers' laboratories over the last 30 years to de-termine the likely range of variation in test data. In addition they conducted tests onsamples of standard soils of low to medium plasticity in their own laboratory to establishthe scatter in their data. These standard soils are used to check comparability of testingprocedures between different Corps of Engineers' laboratories and private laboratories.As a result of these studies the following changes in measured properties were adoptedChapter 3. PREVIOUS SEISMIC STUDIES^ 13before applying the Chinese criteria ( again ignoring the liquidity index) :• decrease the fines content by 5%• decrease the liquid limit by 2%• increase the water content by 2%This change reduced the length requiring remediation to about 1700 m (5500 ft).The impact of the Chinese criteria on the extent of remediation necessary for stabilityappeared to be so critical that an investigation of Chinese procedures was undertakenby Koester (1990). The Chinese determine the liquid limit using a fall cone rather thanthe Casagrande device generally used in North America. Using a standard Chinese fallcone and following Chinese standard SD 128-007-84, Koester (1990) showed that thefall cone gives a liquid limit about 3% or 4% greater than the Casagrande device. TheKoester study is not complete and findings relative to the liquid limit should be viewedas tentative.On the basis of all the above studies the following changes in measured index prop:ertied were finally adopted to account for uncertainty before application of the Chinesecriteria:• decrease the fines content by 5%• increase the liquid limit by 1%• increase the water content by 2%Chapter 3. PREVIOUS SEISMIC STUDIES^ 14These changes reduced the length requiring remediation to about 926 m (3000 ft).The zones in the topstratum clay vulnerable to strength loss based on the N<4 Ruleand the Chinese criteria and taking account of uncertainty are shown in plan, Fig. 3.2,and in longitudinal section at a location 76 m ( 250 ft) upstream from the centerline ofthe dam, Fig. 3.3. Those zones are designated as weak clayey silt layers found outsidethe 1988 remediation berms.3.3 Conclusions Drawn from Previous Seismic Study1.) The silt core of the dam may liquefy along the entire length of the dam.2.) The post-liquefation factor of safety with respect to the upstream stability of thedam would still be adequate except in areas where the weak clayey silt layer occurringbeneath the upstream slope.3.) The Chinese criteria for evaluating the potential for liquefaction or significantstrength loss in clayey soils, based on liquid limit, water content and per cent fines <0.005 mm, can have a major impact on the extent of remedial measures necessary toachieve stability in earth structures with potentially liquefiable fine grained materials.4.) Before applying the Chinese criteria the uncertainties in the measured soil prop-erties should be taken into account in a reasonably conservative manner. This may bedone by adjusting the measured water content, liquid limit, the fines content before ap-plying the criteria. The amount of these adjustments should be based on the estimated------..„.,,•i,i ,- w ...... , .-„,.. -;-----,^) x^Ii. f .....e'...^' - ..,,, ..)_02)^' -,',--' -^,^,,...I4,1dA'4; I I 11111 I 1 Ol)tIlullllu^1111144* „I^4„^1114 [117. •',.. WOVEWAVE17TTTITTTT 11111T 11^i 111 rilei i 11111 I I i^I 1 1111111111h I I 11 .1 '1111H 11%1 1^ii211 [111111 11 1B111111111111L1111111 111^111111111111111111111111j)1 11 ) 111 11N:2-77; 71\"(114g1--- ^ •AN Al 30dtommeilmsr4^Nosanommsvaroo •■•■•• •■••••■•••1SOFIN f Y.M. NO 110DEMON 104•111•NINom NO IISAROIS EARTHQUAKE S ILIUMsCHIRESE CRITERIA - PLANte • ••■•• .“1.11118, 0.111.01COMOIIFigure 3.2: Plan View of Liquefiable Weak Clayey Silt Zones (Finn et al., 1990a)A 1 Av ..,.• ,,,44404._,111111 WE IS OINESE CRITERIAsuoss wry samemi. CLAY • SOSFigure 3.3: Longitudinal Section Showing Liquefiable Weak Clayey Silt Zones 139 m(450ft) Upstream from Centerline (Finn et al., 1990a)Chapter 3. PREVIOUS SEISMIC STUDIES^ 17variability in data appropriate for the laboratory conducting the tests. In the absence ofsuch specific information, the adjustments noted above of increasing the liquid limit by1%, the water content by 2%, and decreasing the fines content by 5% may be considered.These adjustments reflect conservative estimates of the variability to be expected fromvery experienced personnel operating under high standards of quality control. Due noteshould be taken of the possible sensitivity of the liquefaction assessment to minor changesin the measured parameters.5.) The zones of weak clayey silt vulnerable to strength loss are determined based onthe N<4 Rule and the Chinese criteria and taking account of uncertainty. Those zonesare found outside the 1988 remediation berms and need to be remediated.Chapter 4THEORY OF POST LIQUEFACTION ANALYSIS4.1 Conventional Static Equilibrium MethodThe static equilibrium method is widely used to assess the stability of an earth dam afterliquefaction in the dam or foundation. In this method, the factors of safety against shearfailures along all potential failure surfaces are determined based on the residual strengthsof the liquefied soils.In the case of dams some deformations are acceptable although such deformationssmay indicate temporary factors of safety less than unity. Therefore, in accordance withthe concept of designing dams for acceptable deformations proposed by Newmark (1965),there is a tendency to move away from the factor of safety concept and to evaluate theextent of necessary remedial measures on the basis of a tolerable amount of deformationfor the low probability event specified by the design earthquake. This approach requiresa reliable method of estimating post-liquefaction deformations.4.2 The Deformation Analysis TheoryPost-liquefaction deformations for soil structures can be computed by using a nonlinearfinite element code developed by Finn and Yogendrakumar (1989). This finite element18Chapter 4. THEORY OF POST LIQUEFACTION ANALYSIS^ 19code is incorporated in the computer program TARA-3FL. The program has the capa-bility of computing potential flow deformations. This flow deformation analysis methodmake it possible to assess the seismic behaviour of soil structures incorporating zones ofliquefiable soils.When a saturated or partially saturated soil structure undergoes an earthquake load-ing, the dynamic pore water pressures in the saturated soils increase, shear strengthsof the saturated soil decrease. As the dynamic porewater pressures increase more andmore, liquefaction may occur, and the shear strengths of the liquefied soils may reduceto steady state strengths or residual strengths.In the present preliminary study, it is assumed that the residual strengths will betriggered in all soil elements that will liquefy according to the criteria developed by Seedet al. (1985), and the analysis is concentrated on the post-liquefaction behaviour only.In the deformation analysis involving soil liquefation, the first requirement is a triggeringcriterion to switch the strength of any liquefiable soil in the dam to the residual strengthat the proper time during the dynamic analysis. Two criteria are available , the peakstrain criterion of Castro et al. (1989) and the stress ratio criterion of Vaid and Chern(1985). These criteria are not used in the present analysis. This analysis simply focuseson the fact that the residual strengths are reached but ignores when the residual strengthswill be reached.In the proposed deformation analysis, the soil structure is analyzed using initial orpre-earthquake soil strengths and moduli before applying a seismic loading. A stress-strain field prior to the earthquake loading is thus determined. When the seismic loadingChapter 4. THEORY OF POST LIQUEFACTION ANALYSIS^20is applied, increases in dynamic porewater pressures in saturated soils lead to reduc-tions in shear strengths and shear moduli of the saturated soils which cause unbalancedshear stresses in the soil structure. As the unbalanced shear stresses are redistributedthroughout the soil structure, a new stress-strain field is reached, and flow deformationsare obtained for this level of shear strengths.In any particular element in a dam, the shear stress-strain state which reflects pre-earthquake conditions is specified by a point P on the stress-strain curve as shown in Fig.4.1. When liquefaction is triggered, the strength will drop to the steady-state value. Thepost-liquefaction stress-strain curve cannot now sustain the pre-earthquake stress-straincondition, and the unbalanced shear stresses are redistributed throughout the dam. Inthe liquefied elements, the stresses are adjusted according to the following equation,Of ,^Of ,= ^ao-i + — cry^ (4.1)(9cr'„^(9-ywhere T = f (c ,-y). This process leads to progressive deformation of dam until equi-librium is reached at the state represented by P2.For static preliminary studies of flow deformations the basic idea is simulating thereduction sequence of the shear strengths in liquefied soils or non-liquefied soils by con-ducting a series of static finite element analyses. For each subsequent step of the defor-mation analysis, the previous shear strengths are reduced by a small percentage such as5%. A final flow deformation configuration is determined as the final post-earthquakeshear strengths ( residual shear strengths ) are reached in all liquefied elements. Since thedeformation may become large, it is necessary to update progressively the finite element2000af dc,..a an;^mP I2 YBe foreliquefactionT. f(crr;,,r )o f _,— • u yay5^10^15SHEAR STRAIN,AfterliquefactionChapter 4. THEORY OF POST LIQUEFACTION ANALYSIS^21Figure 4.1: Adjusting Stress-Strain State to Post-Liquefaction Conditions (after Finn etal., 1990b)Chapter 4. THEORY OF POST LIQUEFACTION ANALYSIS^ 22mesh. Each calculation of incremental deformation is based on the current shape of thedam, not the initial shape as in a conventional finite element analysis. If the resultingflow deformations are not acceptable, the design must be not used. If they are acceptable,then the dynamic form of the analysis is used to check the design.To model the nonlinear stress-strain behaviour of the soil material, an incrementalelastic approach is used where the soil is assumed to be isotropic and elastic during theload increment. The two-dimensional stress-strain relationship is determined by a pairof elastic stiffness constants, tangent bulk modulus Bt and tangent shear modulus G t .The bulk modulus can be expressed by:crBt = Kb • Pa\" (m (4.2)whereKb = bulk modulus constantPa = atmospheric pressurecorn = effective mean normal stressn = bulk modulus exponentThe constants Kb and n are determined by triaxial tests (Duncan and Chang, 1970).The tangent shear modulus G t is determined by using a hyperbolic shear stress-strainmodel based on the maximum shear modulus , the shear strength, and the shear strain(Finn and Yogendrakumar, 1989). The maximum shear modulus Gmax may be inputdirectly if known or calculated by the program using the following equations (Seed andIdriss, 1970):Chapter 4. THEORY OF POST LIQUEFACTION ANALYSIS^23for sands,Gmax = K2max • Pa • ( Cri4 ) 115^(4.3)for clays,Gmax = K, • Su^(4.4)where Su is the undrained shear strength and K2max and K, are parameters to beestimated or determined. The maximum shear modulus may be determined based onlaboratory tests or on the shear wave velocity from conventional seismic crosshole tests.If direct data are not available, the maximum shear modulus of soil can be estimatedbased on the relative density of sands or on the level of shear strain for clays (Seed andIdriss, 1970).Chapter 5RESIDUAL SHEAR STRENGTH5.1 IntroductionLiquefaction involves large unidirectional shear deformations. When the soil is strainedbeyond the peak strength, the undrained strength drops to a value that is maintainedconstant over a large range in strain, as illustrated by curve 1 in Fig. 5.1. The steady stateof deformation for any mass of particles is that state in which the mass is continuouslydeforming at constant volume, constant normal effective stress, constant shear stress,and constant rate of shear strain. The steady state is achieved only after the structureis completely remolded and all particle orientation effects have reached a steady statecondition and after all particle breakage, if any, is complete. The steady state existsonly during deformation. The shear strength under steady state condition is called theundrained steady state strength or residual strength.If the driving shear stresses due to gravity on a potential slip surface through lique-fied materials in an embankment are greater than the undrained steady state strength,deformations will occur until the driving stresses are reduced to values compatible withstatic equilibrium (Fig. 5.2). The more the driving stresses exceed the steady statestrength the greater the deformations to achieve equilibrium. The driving shear stressesto be used in analyzing liquefaction are not the shear stresses resulting from placementor consolidation of the soil, but rather are the minimum shear stresses that are necessary24Chapter 5. RESIDUAL SHEAR STRENGTHAXIAL STRAIN (%)Figure 5.1 Types of Stress Strain Curve for Consolidated UndrainedTriaxial Tests on Clean Sand25UNITAS.STRENGTHICGCD10IINMAL coDRIVINGSTRESSslide^slide',movement^movementss^starts^stops\\ ,s‘ssfinal driving stressconsistent with post-slideconfigurationstart ofearthquakeshakingIr^steady-state orresidual strengthSHEAR STRAINFigure 5.2 Changes in Driving Stresses and Undrained Shear StrengthDuring an EarthquakeChapter 5. RESIDUAL SHEAR STRENGTH^ 26to maintain equilibrium of soil mass under external or body forces. The driving shearstresses that should be compared to the steady state shear strength are shear stresseswhich will continue to be applied to soil as deformation occur.The residual strength is the crucial element controlling post- liquefaction deforma-tions. Therefore available procedures in determining the residual strength are criticallydiscussed below.5.2 Determination of Residual Strengths5.2.1 Steady state conceptCasagrande (1936) in his classic paper discussed the significance of the critical void ratiosof cohesionless soils. In his paper, Casagrande pointed out that a sand with a void ratiogreater than the critical void ratio tends to contract upon monotonic shearing, whereasa sand with a void ratio less than the critical value tends to expand. Thus, when a satu-rated sand is sheared in an undrained state, positive pore pressures are developed if thevoid ratio is greater than this critical value (loose sands). The development of positivepore pressures leads to a reduction in the effective normal stress and consequently toa reduction in shearing strength. It was found that the critical void ratio varied withthe effective confining pressure. The line describing such a relation on a critical voidratio versus effective confining pressure plot is unique for a given sand and is called thesteady state line. The undrained steady state shearing strength is represented by thestrength corresponding to the effective confining pressure on the steady state line andthe associated void ratio.Chapter 5. RESIDUAL SHEAR STRENGTH^ 27Since it is the steady state shear strength that is needed for a liquefaction analysis ,it is convenient to plot the results of the undrained triaxial tests in terms of void ratioversus undrained steady state strength on the failure plane, as shown in Fig. 5.4, ratherthan in terms of u38 , as shown in Fig.5.3. The corresponding steady state, undrainedstrength of the soil, Sus , can be determined from the Mohr diagram and expressed in thefollowing form:sine' cos'S„„ = (5.1)1 — sin01 cr 3swhere 0 1 is the effective friction angle , and u31 , the effective minor principle stress,corresponding to the steady state condition.The steady state strength of liquefied sands,Sus , generally can not be determined di-rectly by undrained shear tests on undisturbed samples from the field . Such contractivesoils are very difficult to sample. They are likely to densify during sampling, transporta-tion and the process of setting up the tests. Therefore the tests cannot be conducted atthe field void ratio. A procedure for dealing with this problem has been proposed byPoulos et al. (1985). In the procedure the steady state strength of a good qualityundisturbed sample is determined at the laboratory void ratio after reconsolidation in thelaboratory. It is then assumed (1) that there is a unique relationship ( the steady stateline) between steady state strength and void ratio ; (2) that the slope of the steady stateline is the same for reconsitituted samples of that sand as for undisturbed samples; and(3) that the slope of the steady state line in independent of the method by which samplesare reconstituted in the laboratory. Thus by performing tests on reconstituted samples,end of consolidationsteady state line• idilative^ contractive/steady state strength linecompacted spedmens•• •/ •undisturbed samples28Chapter 5. RES/DUAL SHEAR STRENGTHeffective confining pressureFigure 5.3 Steady State Line (void ratio vs effective confining pressure)(after Poulos et al., 1985)a)0ill32oa)0111:0undrained steady state strength SsuFigure 5.4 Steady State Line (void ratio vs steady state strength)(after Poulos et al., 1985)Chapter 5. RESIDUAL SHEAR STRENGTH^ 29the slope of the steady state line for these samples can be established and used to pre-dict the steady state strength of the undisturbed sample at the void ratio correspondingto its in-situ condition. The procedure for the accomplishing this is illustrated in Fig. 5.5.The steady state line in e - log Ssu space is obtained by tests on reconstituted samplesat different void ratios. Then S.,, is measured in undrained compression tests on goodquality undisturbed samples from the field. A representative value of this laboratorysteady state strength, (S.„)L is plotted in Fig. 5.5 at the void ratio at failure in thelaboratory, eL . A line is drawn through the point ((S.,,)L ,eL ) parallel to the line forreconstituted samples and the strength corresponding to the field void ratio, e f , givenby this line, is taken as the steady state strength in the field, (58„)fThere are some difficulties with the use of this procedure. First the differences be-tween the steady state lines of the field samples can be very wide. It is clear that selectionof a representative steady state strength poses significant problem for a designer or an-alyst. Secondly there is controversy over whether the undrained steady state strength isstress path dependent or not.Castro et al. (1985) hold the view that steady state strength is independent of thestress path. Vaid et al. (1990) conducted an extensive test program on Ottawa sandand a tailings sand to investigate the effects of stress path on steady state strength usingextension and compression tests. They found that the steady state strength was greatestin compression. In the case of loose Ottawa sand, the ratio of steady state strength incompression to the strength in extension was 10:1; for the loose tailings sand, the ratiowas about 6:1. Furthermore the range in void ratios exhibiting contractive behaviourin extension is much larger than in compression. These tests were conducted on waterChapter 5. RESIDUAL SHEAR STRENGTH^ 30pluviated sands to simulate the depositional process in nature. Such sands are inherentlyanisotropic and their response to loading depends on the orientation, 3, of the majorprincipal stress relative to the plane of deposition. The data strongly suggest that thesteady state or residual strength of soils in the field is a function of p (Vaid et al., 1990).The dependence of S8 on has important practical implications. The angle variesalong the curved failure surface of an embankment from = 0 (compression loading) nearthe crest to /3= 90 degree (extension loading) at the toe (Fig. 5.6). Therefore steadystate strength based on data from compression tests would appear to be applicable onlynear the upper part of the failure surface. The strength should decrease and reach itslowest value in extension near the toe. Thus the average steady state strength in theliquefied zone of the lower San Fernando dam may be much less than that value mea-sured in laboratory compression tests. Seed et al. (1988) have shown from back analysisof the San Fernando dam that the average steady state strength in situ was substantiallyless than the average strength measured in compression. tests.Discrepancies between steady state strength in compression measured on undisturbedsamples and the average steady state strength computed by back analysis of the San Fer-nando slide are credited almost entirely to densification. The quantitative effect of stresspath in steady state strength suggested by the work of Vaid et al. (1990) may accountfor a substantial part of the difference noticed in the San Fernando dam studies. Thiseffect is also crucial to a reliable stability analysis.There are clearly sharp differences between recent research finding and current prac-tice in the determination of steady state strength from laboratory tests. The key assump-tion underlying current practice that the steady state strength is a function of void ratioSteady - state Line forRe-constituted Samples1(SUS)f^(Sus)LChapter 5. RESIDUAL SHEAR STRENGTH^ 31Steady - state Strength, Sus (Log scale)Figure 5.5 Poulos Procedure for Determining Steady State Strength(after Poulos et al., 1985)1160 -,960 -Figure 5.6 Cross Section of Lower San Fernando Darn Showing UquefactionZone (after Seed et al., 1988)Chapter 5. RESIDUAL SHEAR STRENGTH^ 32only needs further investigation. More studies on the effects of stress path are needed toestablish a generally acceptable position on this very important problem.5.2.2 A correction between residual strengths and SPT blowcountsSeed (1987) developed an alternative approach to determining steady state strength. Heanalyzed the stability of a number of field cases in which large deformations occurredafter liquefaction and developed a correlation between in situ steady state strength andrepresentative (N1 )60 values. This correlation was updated in 1987 by refining some ofthe previous analyses and incorporating new data from embankment failures during theChilean earthquake of 1985 (De Alba, et al., 1987). The latest version of the correlationis shown in Fig. 5.7.There are two challenges for the designer in using this correlation. The first is relatedto the range in strength at a given penetration resistance. At the low penetration resis-tances associated with very contractive materials the range in strength is many times theminimum value. This makes it very difficult to decide on an appropriate value for residualstrength at low penetration resistances. The second challenge relates to the selection ofa representative (N 1 )60 . The dispersion of (N 1 ) 60 values can be very wide in the field.This dispersion can be particularly troubling when analyzing case histories to determinesteady state strengths. An excellent discussion of all the difficulties associated with theanalysis of case histories of flow deformation to determine steady state strength may befound in Seed et al. (1988).Chapter 5. RESIDUAL SHEAR STRENGTH^ 335.3 Evaluation of Residual Strength on Sardis DamThe residual shear strength in the liquefied silt core in Sardis Dam was estimated tobe 5 kPa (100 psf) based on Seed's correlation between corrected standard penetrationresistance (N1 )60 and residual strength shown in Fig. 5.7; see Seed et al, 1988.In the original investigation , the residual strength of the weak clayey silt in thefoundation was established on the basis of judgement, the Seed (1987) criterion , andlaboratory vane tests. Sample disturbance and subsequent reconsolidation prior to test-ing resulting in higher values for the residual strengths determined by the laboratory vanetests than would be expected in situ. In the later investigation, the peak and residualin situ strength of the weak clayey silt were determined by field vane tests in the top-stratum clay. These residual strength (Fig. 5.8) , however, were not used in the stabilityassessment prior to 1988 because cone penetration tests in adjacent locations appearedto indicate the presence of silt and sand lenses that might have allowed some drainageduring the vane tests. Such drainage would result in higher strength values. Woodwardclyde Consultants (1989) reviewed the results of classification tests conducted on samplesof the soils tested by the field vane and concluded that lenses of sand or silt sufficient tocause significant internal drainage were not present. Furthermore, they concluded thateven if lenses small enough not to have been detected in the classification were present,they were unlikely to have allowed significant drainage at the typical rates for conductingthe vane tests (about 6 degree per minute in the post-peak phase of the test).The field vane tests were used to estimate the ratios of peak undrained strength, S u ,and residual strength, S,,,, to the effective overburden pressure cry' 0 and the sensitivitySu /5'8u . The peak undrained strength was also estimated from cone penetration test dataChapter 5. RES/DUAL SHEAR STRENGTH^ 341200ci)CL800ca92:0 400a)cc01620^4^8^12Equivalent Clean Sand (N 1)60Figure 5.7 Tentative Relationship Between Residual Strength andSPT N Values for Sands (after Seed et al., 1988)20200-^o16^C,-^0e^0 ^0^0^12^o noo o o-o^Do^o^0oO^ooo oOS^n n^0-0O0^o^o (bO^o%°°^0^o04^o^9, I\"^o00^0 0-^0 00itap 0 0^_ 0_r, von9C1D,^ct0O^096e^400^SOO^1200^1600RESIDUAL STRENGTH, psiFIELD VANE SHEARFigure 5.8 Residual Strength in Topstratum Clay - Sardis Dam(after Finn et al.,1990a)211-zz0Chapter 5. RESIDUAL SHEAR STRENGTH^ 35using the equation developed by Robertson and Campanella (1983),S,, = qc cv015(5.2)where qc = cone tip resistance and o-,,, = total overburden pressure. In addition,estimates were obtained of the undrained strength required to give factor of safety in therange of 1.5 to 2.0. This range in factor of safety was considered appropriate for the pre-earthquake condition of the dam. A conservative assessment of all these results suggestedan average Su /u„„ ratio for the weak clays and silts of about 0.2 to 0.3 . Based on theseresults for Sedcr„,,, the work of Pyles (1981) on the shearing resistance of cohesive soils atlarge strains, the S../avo ratios and sensitivities from the field vane tests, a conservativevalue of the residual strength for the weak clayey silt was estimated to be S./o-vi c, =0.075.The residual shear strength of the non-liquefiable top stratum clay was estimated at100 kPa (2000 psf) on the basis of undrained triaxial tests.Chapter 6LIQUEFACTION DEFORMATION ANALYSIS6.1 Initial Stress Conditions in Dam and FoundationThe mechanical behaviour of Sardis dam is simulated by modelling the performance ofa typical cross section . The actual three dimensional problem is simplified as a twodimensional plane strain problem. Both initial stresses after the construction and the de-formations after liquefaction were computed by using the computer program TARA-3FL( Finn and Yogendrakumar, 1989) described earlier. The basic theory has been reportedby Finn (1985,1990) and Finn et al. (1990).6.1.1 Finite element meshFor the purposes of simulating the initial stress conditions and the subsequent liquefac-tion behaviour of the dam , a finite element mesh is constructed as shown in Fig. 6.1and Fig. 6.2, which show the element and node numbers of the mesh. This finite elementmesh consists of 576 nodes and 502 quadrilateral elements.The boundary conditions for analyzing the construction of dam and foundation aredefined along all boundaries. The surfaces of the dam and the free field are free to movehorizontally and vertically.36403.1.1101:2:41111111111t• t.1:11:1:111 410 41141114111::: 1:111111t ..364 365 3E41 367 3E0 389 370 311 312 373 7 315 376 377^980 36 405 387 0. 390 391 392 393 394 395MI 307 303 304 305 306 300 309 310 11312 313301 314 320315 316 317 310 319 321 322 323 324 325 126 321 3281 31131S71-333027Q367993^318 339040 341 342 343 344 345 146 347 148 349 350 351 152 353 354 355 356 157 158 359 360241 242 243 244 245 240 247 248 24 2104 252 253 254 255 256 257 250 259 200 261 262 203 204 2e! 298 267 200 71 I7' 270 2791236 201 202 203 204 295 256 267 260 209 290 291 292 293 294 295 290 297 296 299 300181 162 161 164 165 1E0 107 108 108 190 9 112 193 194 195 196 197 195 199 200 201 202 203 204 292 206 207 206 911 II^!I! 214 1 916154 56II\" 216 219E21 221 222 223 224 225 226 227 220 229 230 231 212 215 254 215 238 217 230 219 243121 122 121 124 125 126 127 120 129 130 3 132 153 134 135 135 137 136 139 143 141 142 141 144 147.146 147 140 51 5 5! 5' 158 159116(161 162 161 164 105 I68 167 tee 109 170 171 172 175 174 175 176 17 7 170 179 100el 62 ee 54 03 ee 67 68 01 70 71 72 73 74 71 70 71 76 79 00 01 02 03 04 05 06 07 06 91 9391 94 II 96 97 90 99 IOC 101 102 103 104 101 106 107 108 109 110 111 112 115 114 113 116 117 118 119 120I 1 4 5 0 7 0 9 16 11^12 13 14 13 16 17 10 111 20 21 22 21 24 25 26 27 20 2^31 32 51 34 56 17 SO 53 40^41 42 41 44 41 48 47 40 49 50 51 52 53 54 55 36 57 50 59 60Figure 6.1: Finite Element Mesh Showing Element Distributionrn363 3364110•;..4401111lik.AI!: **top., .475.4/111.1110111111kk1641427 7 437ra alelligINNIMOk 11111111101$4437^,..fmtiONIVIIIIIIIIIIIIPIrg^II, I 1111111111v.6...mullimmemprum,,. mit up ^• ' 'IIMPOIMPIRIPAIIINIMINMPIPIMPIVI ^1^ i. !Li' T.. Ali n,ItIVIIIIIIRP9 111 ,1 ,1MIRENTE9W 14.,5W11111W4ik11R'IM'WAN;: NI gii iii;INITIMII;i1111111M1R4F11WANIIRFORMil. 7473WINIIIIIPIIMPIIM IIIIINIMMOVIII WIPP* 1911011111111414/11111PIPPITMIRRIMM•721111144111111014 1/11INIMMINFIBIIII 1•., 0.11111111.1111,1411111MINWRIPPII.7.11111141114111111 111 1 111111101MUltAmi 1,..^1101111110111111110111111111111111Millit 70Figure 6.2: Finite Element Mesh Showing Node DistributionChapter 6. LIQUEFACTION DEFORMATION ANALYSIS^ 39Nodes on the base are fixed in both directions. At the two side boundaries the nodesare free to move vertically but are fixed horizontally. Therefore , the two side boundarieshave to be far enough so that the building of the dam could not cause significant hori-zontal movementE at the two side boundaries.The rest of the nodes in the mesh are set free to move in both X and Y directions.6.1.2 Soil material propertiesParameters defining material properties are assigned to each element to reflect realisticstrengths and moduli. Generally the elements are grouped to represent zones of differentmaterials. In each zone, the same material parameters are used. Fig. 6.3 shows thedistribution of these material zones.The dam foundation consists of a 10 to 20-foot thick zone of natural silty clay , des-ignated as top stratum clay and modelled as material #2 and #5 . The top stratum clayis underlain by pervious alluvial sands , designated as substratum sands and modelledas material #1, approximately 40-foot thick. The substratum sands are underlain bytertiary silts.Sardis Dam consists of a central silt core, constructed by hydraulic filling. The satu-rated (liquefiable) and drained (above water table) parts of the silt core were modelledas material #8 and #9 , respectively. The silt core is surrounded by a sand shell. Thedrained part of the sand shell was modelled as material #10 . The saturated part ofthe sand shell was modelled as material #6 ( non-liquefiable) and #7 ( liquefiable). Thecrest of the dam consists of a compacted clay fill which was modelled as material #11.Elevation (ft)silt clay #5weak silt clay #2gu'10MF4P5t'l54^305 ^tri1'1=.-00'70 btx.105 P:1ti304.-4=.K1-C# I IFoundation Sand #1Fig. 6.3 Distribution of Soil Material ZonesChapter 6. LIQUEFACTION DEFORMATION ANALYSIS^ 41Table 6.1: Parameters of Strength and Stiffness Used in the Construction AnalysisMaterial# C (psf) Kb n Kc K2ma -y (pcf)1 35 0 6182 0 61 1252 0 2000 13230 0 1400 1205 0 2000 13230 0 1400 1206 35 0 6182 0 61 1257 35 0 6182 0 61 1258 20 300 4054 0 40 1209 20 300 4054 0 40 12010 35 0 5067 0 50 12511 0 750 5150 0 1453 115Materials:#1 - foundation sand#2 - weak clayey silt#5 - clayey silt#6 - sand shell (non-liquefiable)#7 - sand shell ( liquefiable )#8 - silt core (liquefiable)#9 - silt core ( above water table)#10 - sand shell (above water table)#11 - rounded clay capIn the program TARA-3FL, soil materials have been classified into two types : sandsand clays. For the construction sequence, drained strength and modulus parameters areused for sands, and undrained strength and modulus parameters are used for clays. Af-ter liquefaction, undrained strengths and moduli are used for all materials. Table 6.1shows the soil parameters used in the construction sequence, which are obtained fromthe previous study. The parameters used for the liquefaction analysis will be presentedin section where post-liquefaction analysis is considered.Chapter 6. LIQUEFACTION DEFORMATION ANALYSIS^ 426.1.3 Pool water level and water forceThe conservation level of the reservoir is at elevation 277 ft. In this thesis, the perfor-mance of the dam after liquefaction was examined at this reservoir elevation. The damcrest is at elevation 312 ft, for an initial freeboard of 35 ft. If the material is submergedbelow the phreatic line ( Fig. 6.3), the buoyant unit weight is used in analyses.The water forces acting on left side of the relatively impermeable silt core should beincluded in the initial stress analysis of the construction. These forces shown in Fig. 6.3are perpendicular to the core face. The distributed water forces are replaced by equivalentconcentrated normal forces at the nodes and then resolved in the vertical and horizontaldirections in an approximately representation of reservoir effects during liquefaction.The construction of the foundation and the dam was modelled by a 13-layer construc-tion sequence, where incremental stresses, strains and deformations were computed afterthe placement of each new layer. Final results of static analysis during construction stagewere printed out by the program. In this way , the initial stress conditions in the dambefore liquefaction were determined.6.2 Description of Liquefaction Analysis6.2.1 Liquefiable materialsPrevious investigators had determined the zones with potential liquefaction or significantstrength loss. These are :Chapter 6. LIQUEFACTION DEFORMATION ANALYSIS^ 43• Hydraulically placed silt core (saturated) - material #8;• Upper 10 to 30-foot of sand shell along the upstream slope - material #7;• Discontinuous layer of weak clayey silt 5 ft thick located approximately 7 ft intotop stratum clay - material #2.The residual strength for the liquefied silt core is 100 psf, for the liquefied sand shell is400 psf, and for the weak clayey silt is 0.075cr i,'„, where a is the effective vertical stress.Table 6.2 shows the variation of residual strengths in the weak clayey silt with locationsin the dam foundation, a minimum value of 100 psf of the residual strengths was used inthe analysis.6.2.2 Soil properties after liquefactionThe maximum shear modulus and bulk modulus for sands can be calculated by followingequations:Gmax cri 0 5= 21 . 7K 2,,,x Pa (^) (6.1)criB = KbPa(---; )n^(6.2)where Pa is atmospheric pressure, 2117 psf; um is the mean normal stress. The otherconstants are shown in Table 6.1.Using the appropriate mean normal stress from the final results of the constructionanalysis, we can calculate the maximum shear modulus Gm,„ and bulk modulus B at theend of construction for the drained parameters. When drained strengths and moduli areshifted into undrained parameters, undrained strengths and moduli are approximated byChapter 6. LIQUEFACTION DEFORMATION ANALYSIS^ 44Table 6.2: Variations of Residual Strengths in the Weak Clayey Silt (psf)Element WT277 model241-255 100256 100257 109258 121259 133260 147261 160262 173263 186264 211265 246266 287267 342268 398269 449270 504271 549272 597273 621274 671275 728276 766277 742278 687279 610280 551281 515Chapter 6. LIQUEFACTION DEFORMATION ANALYSIS^ 45Table 6.3: Parameters of Strength and Stiffness after LiquefactionMaterial# 0° C (psf) Kb n K° K2max 7(Pef )1 35 0 6182 0 61 1252 0 0.0750L 13230 0 1400 1205 0 2000 13230 0 1400 1206 35 0 6182 0 61 1257 0 400 6182 0 3825 1258 0 100 4054 0 1384 1209 20 300 4054 0 40 12010 35 0 5067 0 50 12511 0 750 5150 0 1453 115maintaining the same values of these parameters. For the liquefaction or undrained con-dition, the bulk modulus exponent n is usually set to be zero and Kb adjusted accordingly.The moduli expressions for undrained parameters are :Gmax — Kc • SuB = Kb • P.The undrained equivalent shear modulus constant IC c and bulk modulus constant Kbcould easily be evaluated by using equations (6.3) and (6.4) given the initial undrainedvalues of Gmc,„ B, and undrained strength.There are three potential zones of liquefaction which are grouped as materials #2,#7 and #8. The soil parameters after liquefaction are shown in Table 6.3.For Sardis Dam, the reduction of shear strengths and moduli of the liquefied weakclayey silt in the top stratum clay (material #2) mainly controls the final deformations(6.3)(6.4)Chapter 6. LIQUEFACTION DEFORMATION ANALYSIS^ 46of the dam. The loss of strength and stiffness in the upstream sand shell after its lique-faction has a relatively minor effect. The strengths of the liquefied soils were reduced by5% for each step in the sequence of strength reduction from initial undrained strength toresidual strength. This led to a corresponding decrease in the undrained shear modulus.When the residual strength in any zone is reached, the residual strength is used as theshear strength in this zone for the subsequent steps of the analyses. For each new run,the calculated finite element mesh and stresses from the previous run were taken as theinput data for the next run.6.2.3 Results of liquefaction analysesWhen the shear strength of the weak clayey silt decreases from 2000 psf to residualstrengths, large deformations are induced in the dam. The water level of the dam isassumed to be at the elevation of 277 ft , and it is termed as the WT277 model. Fig. 6.4shows the developement of the deformed shape of the dam as the shear strengths in theweak clayey silt are reduced.The variation of the loss of freeboard is illustrated in Fig. 6.5 for various levels ofminimum residual strength. The residual strengths of the weak clayey silt vary dependingon the initial effective stresses. The minimum value of these residual strengths in thislayer for the current liquefaction step is designated as the minimum residual strength.The increase in the loss of freeboard is gradual with the decrease in the minimum residualstrength. For the WT277 model , the loss of freeboard begins to increase very rapidlyafter the minimum residual strength drops to 400 psf. When the minimum residualstrength reaches 100 psf, a crest vertical displacement of 45.5 ft is predicted from theanalysis.a■ww...■■■Imm■ww ^.IMEMMO=IMEMIMEMOMBIN••=1 M•1=•■••■MOMNIIMMOOMMINN■OIMNIIMIER=1•=11MINIMMMMMOMMIMIMIMINMMINE ^MM=MIN.IDW 110111...I •rowMinimum Residual Strenght 500 psfChapter 6. LIQUEFACTION DEFORMATION ANALYSIS^ 474111,-—At sq111■._■■•■••••■••••■■•■■■■=■=MN•IMM• 1••■•■■^ ••••••••••• 111•10^ /MOMMOMMNIMINIMMEMM•^ IMMEP=PENMINMIN.MMI^■IMIIIMMIINIMM .Mea=/MdIMNMMM^ MI= ^ =111•1•1•MMINMMINNIMMNIMMinimum Residual Strength 2000 psfMinimum Residual Strength 200 psf ^■■•■■••••• 111■1■11=1Mo..1■■=4•1•.' MMENII•■•■•■■•■•■•■•■=1MOMMIN^Men....MOOMM.M.MEMINEMMMe.....^INI■011•4■111•11■111■1•••••■•MNIMIIMMONOIMMIONMOM IMMO.= IMIIMWM 8MIIM^0111BMIHNINEM=AMMIMMOM■IIMMIMMENNIMIMONI.■MIMI■NaMMMIIMN■Mta■MIENSIMINM■MMIMINNI IMMO/a.vertical scale two times horizontal scaleMinimum Residual Strength 100 psfhorizontal scale (ft) 0^ 400^800408 BOOFigure 6.4: Variation of Post Liquefaction Configurations with Minimum ResidualStrength - WT277 model48Chapter 6. LIQUEFACTION DEFORMATION ANALYSIS30 ^25 —/if weak - ayey siftfoundation sand1001,200^1,000^800^600^400^200Minimum Residual Strength (psf)Fig. 6.5 Variation of Loss of Freeboardwith Minumum Residual strength0Chapter 6. LIQUEFACTION DEFORMATION ANALYSIS^ 49Based on the original configuration of the dam for the WT277 model, the water levelof the dam is at elevation 277 ft and the dam crest at elevation 312 ft, for an initial free-board of 35 ft. Again from Fig. 6.5, the dam is predicted to overtop at minimum residualstrength 120 psf of the weak clayey silt. Hence, remediation measures are required if anadequate freeboard is to be maintained.The variations of horizontal displacements at the midpoint of the upstream slope(point A) are shown in Fig. 6.6 for various levels of minimum residual strengths. Thehorizontal displacements increase dramatically when the minimum residual strength isless than 400 psf. When the minimum residual strength drops to 100 psf, the horizontaldisplacement of 100 ft is predicted.Fig. 6.7 illustrates the variations of maximum ratios of the shear stress to the shearstrength in the weak clayey silt. This ratio increases constantly with the reduction ofthe minimum residual strength and reaches the steady state value of 1.0 at the minimumresidual strength of 615 psf. This implies that the weak clayey silt fails before the mini-mum residual strengths 100 psf are reached.Tables 6.4 summarizes the final results from the liquefaction analysis. The overallperformance of the dam is very poor after liquefaction with its original configuration.The loss of freeboard is 45.5 ft after liquefaction. The maximum horizontal displacementof the dam is 100 ft. The dam will fail along the weak clayey silt layer under the upstreamslope. Therefore, remediation measures are required to maintain an adequate freeboardof the dam. A zone of improved soil must be created in the dam to resist the upstreammovements after liquefaction. Certain requirements of strength and stiffness of this zoneChapter 6. LIQUEFACTION DEFORMATION ANALYSIS^50Table 6.4: Summary Results of Liquefaction Analysis (WT277 model)Residual Strength (psf) Loss in Freeboard(ft) Horiz. Disp. (ft) Shear Stress / Su1000 0.015 0.001 0.574800 0.114 0.005 0.76615 0.463 0.052 1.00500 0.911 0.229 1.00408 1.56 0.717 1.00300 3.72 3.500 1.00200 9.15 21.80 1.00154 26.2 50.90 1.00120 38.7 81.28 1.00100 45.5 100.8 1.00have to be met to keep an adequate freeboard of the dam and to prevent a shear failurealong the weak layer.weak clayey siltfoundation sandth^th^o^1^1Chapter 6. LIQUEFACTION DEFORMATION ANALYSIS^512520W 200weak clayey silt ....^------------------------------------remediated zone180•weak ieyey siltwidthplug strength 3000 psfplug width 120 ft0.05^0.1^0.15^0.2^0.25Downstream Horizontal Displacement(ft) -- SIDE A160 0 0.3 0.3550% PWP in NL sand shellNo PWP in NL sand shellAremediated zoneweak clayey siltwidthplug strength 3000 psfplug width 120 ftChapter 7. GENERAL REMEDIATION STUDY OF SARDIS DAM^ 64Figure 7.8 Distribution of Downstream Horizontal Movement of RemediatedSection (PWP effect)0.25.S\" 0.2C000.15CEcoa.0 0.1;30N'000.0501,200^1,000^800^600^400^200^0Minimum Residual Strength (psf)Figure 7.9 Horizontal Displacement at Downstream Edge with Residual Strength(PWP effect)•weak Clayey siltPlug Width 120 ft,3000psfPlug Width 80 ft,3000psfPlug Width 120 ft,2000psf0Plug Width 80 ft,2000psfB^remedlated zonewidth I50% PWP In NL sand shall4Fco 3Oa)u_O 2totoO101,20051,000^800^600^400Minimum Residual Strength (psf)200 0Chapter 7. GENERAL REMEDIATION STUDY OF SARDIS DAM^65Figure 7.10 Variation of Loss of Freeboard with Residual Strength(Plug width effect)Chapter 7. GENERAL REMEDIATION STUDY OF SARDIS DAM^667.3.2 Horizontal pressures against the plugFig. 7.11 and Fig. 7.12 shows the horizontal stresses on the remediation zone for betweenthe two different plug widths. The horizontal stresses against the downstream side (sideA) greatly increase after soil liquefaction, by a maximum increment of 9240 psf. Thehorizontal stresses reach the maximum value around the weak clayey silt. It should benoted that 80% or more of the total driving forces are carried by the plug section withinand close to the weak clayey silt zone. The very low strength and stiffness in the weakzone causes a stress concentration in this area. For the plugs with same strengths, thehorizontal stresses against the downstream face are essentially the same regardless of theplug width because the total loadings coming from the downstream of the dam is thesame as long as the plug is wide enough to prevent failure or large deformation.7.3.3 Ratio of shear stress to strength of the plugThe plug width has little influence on the shear stresses in the plugs with same strengths.Fig. 7.13 illustrates the ratio of the shear stress to the shear strength at the downstreamface of the plug after liquefaction. Because of the stress concentration in the weak clayeysilt zone, the shear stresses in this area increase very much. The ratios of the shear stressto the shear strength increase rapidly with the decrease of elevation and reach their max-imum value of 0.68 at the weak clayey silt layer.The maximum ratio of the shear stress to the shear strength in the plug increasesas the residual strengths decrease from 800 psf to 300 psf, as shown in Fig. 7.14. Asthe minimum residual strength in the weak clayey silt layer is greater than 800 psf, thedriving force is carried mainly by the soil. When the minimum residual strength of thePlug Width 120 ft,3000psf Plug Width 80 ft ,3000psf after constructionO 0Plug Width 120 ft,2000psf Plug Width 80 ft,2000psf^Chapter 7. GENERAL REMEDIATION STUDY OF SARDIS DAM^ 67240 —^0 A remedisted zone0220 —weak clayey silt20018050% PWP In NL sand shell0^2,000^4,000^6,000^8,000^10,000^12,000^14,000Downstream Horizontal Stress (psf) -- SIDE AFigure 7.11 Pressure Distribution on Downstream Face of Remediated Section(Plug width effect)widthweak lam silt160d tweak clayey siltFr, 200o^ t +50% PWP In NL sand shellPlug Width 120 ft,3000psf Plug Width 80 ft ,3000psf after construction^ -o^Plug Width 120 ft, 2000psf Plug Width 80 ft, 2000psf160^I^ I^ I^1 0 1,000 2,000 3,000 4,000^5,000Upstream Horizontal Stress (psf) — SIDE BFigure 7.12 Pressure Distribution on Upstream Face of Remediated Section(Plug Width effect)180220240/// weak clayey sillwidthremeoleted zoneowe•leysy siltfomented zonewidth50% PWP In NL sand shellPlug Width 120 ft,3000psf Plug Width 80 ft ,3000pst—a--Plug Width 80 ft, 2000psf Plug Width 120 ft,2000psfPlug Width 120 ft,3000psf--- -8----Plug Width 80 ft,3000psfaPlug Width 120 ft,2000psf0Plug Width 80 ft, 2000pstB plug j■ remediated zone/// weak !clayey siltwidth•50% PWP In NL sand shellChapter 7. GENERAL REMEDIATION STUDY OF SARDIS DAM250682402302202 210to>812001901801700^0.2^0.4^0.6^0.8^1Ratio of Shear Stress to Strength --- SIDE AFigure 7.13 Distribution of Ratio of Shear Stress to Strength at DownstreamFace of Remediated Section (Plug width effect)1vlS 0.8.ca)i-C7)0 0.6(13V3P.N.c 0.4U)OOTo'Er 0.2201,200^1,000^800^600^400^200^0Minimum Residual Strength (psf)Figure 7.14 Maximum Ratio of Shear Stress to Strength in RemediatedSection versus Residual Strength (Plug width effect)Chapter 7. GENERAL REMEDIATION STUDY OF SARDIS DAM^69weak clayey silt is less than 300 psf, almost all of the driving forces are carried by theremedial plug section, and the curve goes flat.7.3.4 Horizontal displacement of the plugThe maximum horizontal displacement occurs around the weak clayey silt layer. Fig.7.15 illustrates the response of horizontal displacements on the downstream side of theplug (side A). It is interesting to note that the horizontal movement of the plug alongthe weak clayey silt is larger than the horizontal movements on the upper half of the plug.Although a plug section with a large width could reduce the horizontal movementby some extent, the shear strength of the plug has a stronger control on the horizontalmovement. A plug with a shear strength of 3000 psf and a width of 80 ft would providea much stronger resistance to the horizontal forces than a plug with a shear strength of2000 psf and a width of 120 ft does. For plugs with the shear strength of 3000 psf, themaximum horizontal displacements are 0.26ft and 0.24 ft for the plug width of 80 ft and120 ft, respectively. For plugs with the shear strength of 2000 psf, the maximum horizon-tal displacements are 0.52 ft and 0.48 ft for the plug width of 80 ft and 120 ft, respectively.Fig.7.16 shows the development of the horizontal movement at the downstream edgeof the plug with the variation of the minimum residual strength. The horizontal displace-ment increases at a constant rate after the minimum residual strength is less than 600 psf.260•0E 220C00IL 200240remedlated zone180 1 1widthweak clayey silt50% PWP in NL sand shell0.1^0.2^0.3^0.4^0.5Downstream Horizontal Displacement(ft) — SIDE AFigure 7.15 Distribution of Downstream Horizontal Movement of RemediatedSection (Plug width effect)160 0 060.500.4C00.— 0.3\"roC0N8 0.1Plug Width 120 ft,3000psfPlug Width 80 ft, 3000psta^Plug Width 120 ft, 2000psfPlug Width 80 ft, 2000psfPlug Width 120 ft,3000psfPlug Width 80 ft, 3000psfPlug Width 120ft,2000psfPlug Width 80ft, 2000psf...remedlated zonePlug• 7//7 weak Iclayey siltwidth50% PWP in NL sand shell•Chapter 7. GENERAL REMEDIATION STUDY OF SARDIS DAM^ 701,200^1,000^800^600^400^200^0Minimum Residual Strength (psf)Figure 7.16 Horizontal Displacement at Downstream Edge with Residual Strength(Plug width effect)Chapter 7. GENERAL REMEDIATION STUDY OF SARDIS DAM^717.4 Effect of Plug StrengthA parametric study is made to examine the effect of the strength of the plug on theperformance of Sardis Dam after the remediation. The plug width of 120 ft and the 50%PWP in NL sand shell are kept unchanged during this comparison study. The strengthsof the plug are 1000 psf, 2000 psf, 3000 psf, 4000 psf and 8000 psf.7.4.1 Loss of freeboardThe variation of the loss of freeboard with the minimum residual strength in the weakclayey silt is illustrated in Fig. 7.17 for the five plug strengths chosen. The strengthmodels of 2000 psf , 3000 psf , 4000 psf and 8000 psf have a very close response in theloss of freeboard. However, the stronger shear strength in the plug the smaller its dropof the dam crest. The difference in the loss of freeboard is not significant because theplugs of 120 ft width with strength greater than 2000 psf provide sufficient resistanceagainst driving forces from the downstream of the dam. The loss of freeboard is around4.5 ft for these cases as the minimum residual strength drops to 100 psf. But when thestrength of the plug drops to 1000 psf, the loss of freeboard of the dam is 5.7 ft. Theaverage strength of 1000 psf in the remediation plug was not considered acceptable.7.4.2 Horizontal pressures against the plugFig. 7.18 and Fig. 7.19 show the horizontal stress responses against the downstreamside ( side A) and the upstream side (side B). The horizontal stresses increase after soilliquefaction compared to those before soil liquefaction at the elevations above the weakclayey silt layer and decrease at the elevations below the weak clayey silt layer. The64co0.0°.,)). 3\"60 21plug strength 1000 psfplug strength 2000 psfplug strength 3000 psf0plug strength 4000 psfplug strength 8000 psgrernedisted zonesysy siltwidthplug width 120 ft50% PWP in NL sand shellChapter 7. GENERAL REMEDIATION STUDY OF SARDIS DAM^720^ 1^I^I^1 1,200^1,000^800 600 400 200^0Minimum Residual Strength (psf)Figure 7.17 Variation of Loss of Freeboard with Residual Strength( Plug strength effect)240 Aafter constructionplug strength 4000 psf plug strength 8000 pig- - - - -^--•-- -remedIsted zoneweak clayey siftwidthplug width 120 ft50% PWP in NL sand shellplug strength 1000 psf plug strength 2000 psi plug strength 3000 psf-o180i^ t^ 1 1,000 2,000 3,000Upstream Horizontal Stress (psf) — SIDE B4,0001600220E\"0g 200Chapter 7. GENERAL REMEDIATION STUDY OF SARDIS DAM^ 73• remedl•ted zoneAZT3 lam antwidth............... ............... ..........................■•• •^plug width 120 ft50% PWP in NL sand shellweak clayey siltplug strength 1000 psf plug strength 2000 psf plug strength 3000 psfplug strength 4000 psf plug strength 8000 psg^after construction1^I^I^I^I 2,000^4,000^6,000^8,000^10,000Downstream Horizontal Stress (psf) — SIDE A12,000^14,0002402200id 200w1801600Figure 7.18 Pressure Distribution on Downstream Face of Remediated Section(Plug strength effect)Figure 7.19 Pressure Distribution on Upstream Face of Remediated Section(Plug strength effect)Chapter 7. GENERAL REMEDIATION STUDY OF SARDIS DAM^74horizontal stresses show a triangular distribution along the downstream face of the plugafter liquefaction. The maximum horizontal stresses against the downstream faces of theplugs occur just above the weak clayey silt layer. These maximum horizontal stresses are9160, 10866, 11710, 12100, 12405 psf for the plugs with the shear strengths of 1000 psf,2000 psf, 3000 psf, 4000 psf and 8000 psf, respectively.The plug with a higher shear strength needs to penetrate deeper to provide sufficientmoment resistance. The pressures against the downstream face of the stronger plugs arehigher than those on the weaker plugs. The higher the plug strength is, the more the hori-zontal stresses against the downstream face of the plug increase after liquefaction. On theother hand, the pressures against the upstream face of the strong plug are less than thoseof the weak plugs. Therefore, the overturning moments about the point at which theplug intersects with the foundation sands are high for the plug with a high shear strength.7.4.3 Ratio of shear stress to strength of the plugA comparison of the ratios of the shear stress to the shear strength in the plugs for thefive different plug strengths is given in Fig. 7.20. The maximum ratios of the shear stressto the shear strength in the plug decrease as the strengths of the plugs increase. Fig. 7.21shows the relationship between these maximum ratios and the reduction of the minimumresidual shear strength of the weak clayey silt. These maximum ratios have the valuesof 0.1 to 0.4 before liquefaction occurs, and they increase after liquefaction. When theminimum residual strength is less than 300 psf , the curve goes flat and most of the shearforce is carried by the plug. The final maximum ratios of the shear stress to the shearstrength are 0.99 (failure) , 0.85, 0.68 ,0.55 and 0.29 for the plugs with strengths of 1000psf, 2000 psf, 3000 psf ,4000 psf and 8000 psf, respectively.plug strength 1000 psf^plug strength 2000 psf^plug strength 3000 psiplug strength 4000 psf^plug strength 8000 psgAremedlated zoneweak clayey siltwidthplug width 120 ft‘,17^ 50% PWP in NL sand shellweak clayey s^....................................................................ilt.................................................................cr)\"a 0.4rofStill 0.22E., 0.8in0in 0.6(4a)0)0.Oc1.201,200 1,000^800^600^400Minimum Residual Strength (psf)200 000............. .. .... ...plug width 120 ft50% PWP in NL sand shell0 -0-00plug strength 1000 psfplug strength 2000 psfplug strength 3000 psfaplug strength 4000 psfplug strength 8000 psfChapter 7. GENERAL REMEDIATION STUDY OF SARDIS DAM^750^0.2^0.4^0.6^0.8^1Ratio of Shear Stress to Strength - Side AFigure 7.20 Distribution of Ratio of Shear Stress to Strength at DownstreamFace of Remediated Section (Plug strength effect)2402200200LTJ180160Figure 7.21 Maximum Ratio of Shear Stress to Strength in RemediatedSection versus Residual Strength (Plug strength effect)Chapter 7. GENERAL REMEDIATION STUDY OF SARDIS DAM^767.4.4 Horizontal displacement of the plugFig. 7.22 illustrates the distribution of the horizontal displacements in the plug withelevation. Above the weak clayey silt which is located at the elevations between 200 ftand 205 ft, the horizontal movements are essentially constant for the plugs with highstrength such as 4000 psf or 8000 psf. The plug with strength of 2000 psf deflects muchmore than those plugs with the strengths of 3000 psf, 4000 psf and 8000 psf.The increasing rate of horizontal movement becomes high when the plug strength isless than 2000 psf; see Fig. 7.23. The maximum horizontal movements are 3.29ft, 0.48ft,0.24ft, 0.16ft and 0.07 ft for the plugs with strengths of 1000 psf, 2000 psf, 3000 psf, 4000psf and 8000 psf, respectively.The large horizontal displaceMent and the rapid increase of the horizontal displac-ment were observed for the plug with strength of 1000 psf.7.5 Summary on General Remediation StudiesThe results on the general remediation study are summarized in Table 7.2. Plug strength1000 psf is too weak to provide enough resistance after liquefaction. This plug strengthresults in a large horizontal displacement and a low factor of safety against a shear failureof the plug. Plug strength 3000 psf may restrain the maximum horizontal displacementof 0.24 ft and the maximum ratio of shear stress to strength of 0.68. Plug strength 3000psf would be an appropriate strength of the remedial plug. On the other hand, plug220240o.\"t 200weak clayey siltI^ i^ I^ i 0.2 0.4 0.6 0.8Downstream Horizontal Displacement (ft) • Side A18oI-11600Aremedleted zonelam slitwidthplug width 120 ft50% PWP in NL sand shellplug strength 1000 psf^plug strength 2000 psi^plug strength 3000 psfplug strength 4000 psf plug strength 8000 pig0.2widthplug width 120 tt50% PWP in NL sand shellr01,200 2001,000^800^600^400Minimum Residual Strength (psf)1.4plug strength 1000 psfplug strength 2000 psfplug strength 3000 psfplug strength 4000 psfplug strength 8000 psfAremedIsted zoneplus•weak clayey silt^0-^.... - ....................... ...........^. .. . .... - -•- • - -^-•-•-•.. . .01.20C.rot 0.8Chapter 7. GENERAL REMEDIATION STUDY OF SARDIS DAM^ 77Figure 7.22 Distribution of Downstream Horizontal Movement of RemediatedSection (Plug strength effect)Figure 7.23 Horizontal Displacement at Downstream Edge with Residual Strength(plug strength effec)Chapter 7. GENERAL REMEDIATION STUDY OF SARDIS DAM^78Table 7.2: Summary Results of Plug Remediation Studiesplugstrength(psf)loss ofFreeboard(ft)Maximum Lr-8.ratio in plug1 Maximum Hori.Displacement(ft)width 120ft 1000 5.67 0.99 3.292000 4.56 0.85 0.4850% PWP 3000 4.50 0.68 0.244000 4.49 0.55 0.168000 4.40 0.29 0.07width 80ft 2000 4.56 0.86 0.5250% PWP 3000 4.54 0.69 0.26width 120ftNo PWP 3000 2.36 0.74 0.32strength higher than 3000 psf is not necessary for the remediation of this dam.From the variation of the loss of freeboard with plug strength, Fig. 7.24, plug strength2000 psf is appropriate to control the loss of freeboard of the dam. Plug strength 2000psf may cause the maximum ratio of the shear stress to strength of the plug as high as0.85; see Fig. 7.25. Hence it is reasonable to select plug strength 3000 psf to limit thisratio to a range of 0.70 which may be adequate for an engineering design. Fig. 7.26shows the maximum horizontal displacement in the plug versus plug strength. Whenthe plug strength is less than 2000 psf, the maximum horizontal displacement in theplug increases dramatically with the decrease of the plug strength. Again plug strengthof 3000 psf would be adequate to meet design requirements on the displacement of theplug. Plug width 120 ft is reliable to meet all design purposes.Therefore plug strength of 3000 psf (with modulus 900,000 psf) and plug width of120 ft are selected to be the general requirements of the remediated zone of Sardis Damto provide a satisfactory control on both the loss of freeboard and the shear failure alongplug width 120 ft50% PWP in NL sand shellplug width 80 ft50% PWP in NL sand shellAplug width 120 ftNo PWP in NL sand Shell00Chapter 7. GENERAL REMEDIATION STUDY OF SARDIS DAM^79876-4-2co0.1Da)e) 50U)cnJ43210,000^8,000^6,000^4,000^2,000^0Plug Strength (psf)Figure 7.24: Loss of Freeboard versus Plug Strength1.2 plug width 120 ft50% PWP in NL sand shellplug width 80 ft50% PWP in NL sand shellplug width 120 ftNo PWP in NL sand Shell12,0008,0000.210,000 0a)c.75 0.80U)U)a)U)cts.c/S 0.600cc*)0.46,000^4,000Plug Strength (psf)Chapter 7. GENERAL REMEDIATION STUDY OF SARDIS DAM^80Figure 7.25: Maximum Ratio of Shear Stress to Strength in Plug versus Plug Strengthcr)_2 1.5acE.Ea)URSaU)5 1Tut0N0IEE*;<-`z 502 .010,000^8,000^6,000^4,000Plug Strength (psf)2,000 0plug width 120 ft50% PWP in NL sand shell___e__.plug width 80 ft50% PWP in NL sand shell,,plug width 120 ftNo PWP in NL sand Shell0Chapter 7. GENERAL REMEDIATION STUDY OF SARDIS DAM^81Figure 7.26: Maximum Horizontal Displacement in Plug Versus Plug StrengthChapter 7. GENERAL REMEDIATION STUDY OF SARDIS DAM^82the weak layer.The plug with strength of 3000 psf and width of 120 ft was adopted to remediateSardis Dam. The overall improvement on the performance of the dam after liquefactionis evident; see Fig. 7.27. After remediation, the loss of freeboard is 4.50 ft, and the max-imum horizontal displacement of the dam is 0.24 ft. Although the overall configurationof the dam does not change very much, significant distortions in the liquefied silt coreare observed.Chapter 7. GENERAL REMEDIATION STUDY OF SARDIS DAM^83__.....uummuiran11•1111•11111■1111■._^ re:/%1F...awwwwww.....wwwrelmwmA,■•■••■•■■^^ IMAW■11111=4WWWWINOMMOWINEWAI WAAZ1, =''''f==IMMIW=WWM=A/WWEINWWAIIMMANWAMW•Alwri■^ONN^=WM= 1•011111MMIMIMMINEMMIMOINIMIMMISMIMANNINIIIMMI.pI.1 MINNIE ■••■■•IMIN MIN •■■•■■■••■•=11•1■111•■■”IIMMEMMEI....m.s.Minimum Residual Strength 2000 psf. -4 latillho^■■■•■•••■■■■^oamow.s.m.Minimum Residual Strenght SOO psfMinimum Residual Strength 200 psf.W........■,■===‘,WWW•WWWAN.....14MM OII.WWWW11■11 .1•WWWWW6■WWWWWWAIW===^=NMI=Mb ^ IMINIMMINVIMIIIMI =EMI= OMNIONINIMMIIMENNOMINIMINIMMIIIMMINI.^IM•Mow■M•11■10111MI^MIIMIMMIMIMIMI”M=111•11•••••■11•M••••■•■■■■•••■•■•■■•■•■=11■MMinimum Residual Strength 100 psf^vertical scale two times horizontal scale800horizontal scale (ft)^ 400Figure 7.27: Variation of Typical Post liquefaction Configurations after Remediation -Plug Strength 3000 psf, Plug Width 120 ft)Chapter 8REMEDIATION STUDY OF PILE-REINFORCED SECTION8.1 Equivalent Composite Material Properties of the Pile Reinforced SectionIn practice, the remediation requirements of Sardis Dam can be met by driving pilesthrough the section to be remediated. The proposed layout of the remediation piles isshown in Fig. 8.1 and Fig. 8.2. The width of the pile-reinforced zone is 120 feet. 24-inchsquare prestressed concrete piles are arranged in the remediation zone with a center-to-center spacing of 12 feet in two cross horizontal directions. The remediation piles needto penetrate 15 feet into the foundation sand from the bottom of the weak clayey silt layer.Since the reinforced pile group distributes loads in both horizontal and vertical di-rections, the problem involves a three dimensional pile-soil interaction system. 3-D flowanalysis is not available yet, so 2-D plane strain analysis was employed to simulate thepile-soil system. Again finite element code TARA was used, which contains 2-D planestrain bilinear isoparametric elements. For this purpose equivalent composite materialproperties are required for the pile-reinforced section.Woodward-Clyde consultants (1991) conducted a 3-D analysis on a single pile, whichprovided the composite stress strain curves for the pile-reinforced zone. In their study,the 3-D nonlinearity finite element code, NONSAP, developed by Bath et al. (1974) was841 02'10.SOIL200^400 f tGeo Scale01111^1111I IFigure 8.2: Plane View of Layout of Remediation PilesI2Figure 8.1: Cross Section of Sardis Dam Showing Remediation Piles, 2' ,^10'^2 :^10'^2',Chapter 8. REMEDIATION STUDY OF PILE-REINFORCED SECTION^85Chapter 8. REMEDIATION STUDY OF PILE-REINFORCED SECTION^86used. Since the program does not have a beam element, the 3-D 8-node isoparametric el-ement was chosen to model the pile and its surrounding soils. The single pile-soil systemconsists of a 24-inch-square prestressed concrete pile embedded about 15 feet into thedense foundation sand with a pile center-to-center spacing of 12 feet; see Fig. 8.3. Thissingle pile-soil system was loaded laterally. The lateral loads were assumed to distributetriangularly along elevation; see Fig. 8.4. In their study the baseline load was defined tobe the shear force that can cause shear stresses of 230 psf in the weak clayey silt layer,i.e, 33 kips for the single pile case ( 12 by 12 feet area ). Load levels were increased bymultiplying the baseline load by different integer numbers. The compression strength ofprestressed concrete was assumed to be 6000 psi in their analysis. After performing 3-Danalysis on the single pile-soil system with various load levels, they developed the shearstress-strain relationships shown by the solid lines in Fig. 8.5 through Fig. 8.13 for usein the finite element analysis of the global deformation of the dam.The composite shear stress-strain curves obtained by Woodward-Clyde consultants(1991) are used to represent the composite shear stress-strain characteristics of the pile-reinforced zone of remediated soil. Since the hyperbolic stress-strain relationship is usedin TARA, those composite shear stress-strain curves were approximated by hyperboliccurves. These hyperbolic curves are shown by the dashed lines in Fig. 8.5 through Fig.8.13.The shear strength and shear modulus (initial modulus) obtained from the hyperbolicmodel are designated as the composite shear strength and the composite shear modulus.Those composite strengths and moduli for different elevations of the reinforced zone areshown in Table 8.1. Fig. 8.14 presents the variation of the composite shear strengths inthe pile-reinforced section versus elevations. The composite shear strengths vary fromLinePile211208205202.5200197Chapter 8. REMEDIATION STUDY OF PILE-REINFORCED SECTION^87Figure 8.3: Finite Element Model of Single Pile-Soil SystemChapter 8. REMEDIATION STUDY OF PILE-REINFORCED SECTION^88LinesLines C)Sand ShellStiff ClayWeak Clayey SlitSubstratum Sand714=77 7M-7Figure 8.4: Boundary Conditions and Loading Distribution of Single Pile-Soil SystemChapter 8. REMEDIATION STUDY OF PILE-REINFORCED SECTION^895,0004,000Composite Stress - Strain CurveHyperbolic Modelling Curve... ..............coco 3,000coyrCl)a)a)C)E3. 2,000Hyperbolic Modelling:Shear StrengthShear Modulus1,000State Reached After Soli Liquefaction4100 PSF443 TSF0.01^0.02^0.03^0.04Average Shear StrainFigure 8.5: Shear Stress - Strain Relationship for the Pile-Reinforced Section, Elevation173 to 185 ft1,000Chapter 8. REMEDIATION STUDY OF PILE-REINFORCED SECTION^904,000Composite Stress - Strain CurveHyperbolic Modelling CurveAHyperbolic Modelling:Shear Strength 3600 psfShear Modulus 540 tsfState Reached After Soil LiquefactionI^I 0.01^0.02^0.03Average Shear Strain10.04^0.05Figure 8.6: Shear Stress - Strain Relationship for the Pile-Reinforced Section, Elevation185 to 191 ft3,000Chapter 8. REMEDIATION STUDY OF PILE-REINFORCED SECTION^913,5003,000......^...............Composite Stress - Strain CurveHyperboll-a -Modetling Curve^Hyperbolic Modelling:Shear Strength 3100 psfShear Modulus 1240 tsfStateReached-Ater-Soil-Liquefaction^1,0005002,500U)ag 2,000c73a)cn 1,500Ea.a)>0.01^0.02Average Shear StrainFigure 8.7: Shear Stress - Strain Relationship for the Pile-Reinforced Section, Elevation191 to 200 ftChapter 8. REMEDIAT1ON STUDY OF PILE-REINFORCED SECTION^923,5003,000A2,500Composite Stress - Strain19—CurveHyperbolic lvtodelling Curve ^9c-oca.co2cn '2 000inifsa)_ccna)^ ,1 500co a)><1,0005000State Reached After Soil LiquefactionHyperbolic Modelling:Shear StrengthShear Modulus3600 psf180 tsf0^0.02^0.04^0.06^0.08Average Shear StrainFigure 8.8: Shear Stress - Strain Relationship for the Pile-Reinforced Section, Elevation200 to 205 ftCurveComposite Stress - Straine__Hyperbolic Modelling:Shear Strength 3300 psfShear Modulus 396 tsfState Reached After Soil LiquefactionHyperbok Modelling CurveChapter 8. REMEDIATION STUDY OF PILE-REINFORCED SECTION^933,5003,0002,500acn,g2) 2,000a)ci)a) 1,500cu1,0005000.02^0.04Average Shear StrainFigure 8.9: Shear Stress - Strain Relationship for the Pile-Reinforced Section, Elevation205 to 215 ftComposite Stress - Strain CurveHyperbolic Modelling CurveChapter 8. REMEDIATION STUDY OF PILE-REINFORCED SECTION^94=.... ............. ...........Hyperbolic Modelling:Shear Strength 2000 psfShear Modulus 240 tsfState Reached After Soil Liquefaction2,5002,000C.rn 1,500U);12Encaa)cr)5 1,00050000^0.01^0.02^0.03^0.04^0.05Average Shear StrainFigure 8.10: Shear Stress - Strain Relationship for the Pile-Reinforced Section, Elevation215 to 220 ft1,000U)a)ei5• 800a)ct2• 600Chapter 8. REMEDIATION STUDY OF PILE-REINFORCED SECTION^95Compoitte Stress - Strain CurveHyperbolic Modelling CurveA.... ........Hyperbolic Modelling:Shear Strength 1600 psfShear Modulus 176 tsfState Reached After Soil Liquefaction0.01^0.02^0.03Average Shear Strain0.04 0.05Figure 8.11: Shear Stress - Strain Relationship for the Pile-Reinforced Section, Elevation220 to 230 ftComposite Stress - Strain Curve—fa-i--Hyperbolic Modelling Curve ....Hyperbolic Modelling:Shear Strength 700 psfShear Modulus 168 tsfState Reached After Soil LiquefactionIChapter 8. REMEDIATION STUDY OF PILE-REINFORCED SECTION^96700600500='sC13a.U)ir) 400ir)ida)..cci)a)Esa)><2001000.01^0.02^0.03Average Shear StrainFigure 8.12: Shear Stress - Strain Relationship for the Pile-Reinforced Section, Elevation230 to 240 ftComposite Stress - Strain Curve-El-Hyperbolic Modelling Curve&.p.... ... ........... 2s,..,..—,h; .......... °Hyperbolic Modelling:Shear Strength 230 psfShear Modulus 80 tsfState Reached After Soil Liquefaction250200=-(r)au) 150u)a)ZEa)..acoa)eis 100a)><500.01^0.02Average Shear StrainChapter 8. REMEDIATION STUDY OF PILE-REINFORCED SECTION^97Figure 8.13: Shear Stress - Strain Relationship for the Pile-Reinforced Section, Elevation240 to 250 ftChapter 8. REMEDIATION STUDY OF PILE-REINFORCED SECTION^98240220C?a 200W1801600^1,000^2,000^3,000^4,000^5,000Composite Shear Strength in Pile-Reinforced Section (psf)Figure 8.14: Variation of Shear Strength in Pile - Reinforced Section Versus ElevationChapter 8. REMEDIATION STUDY OF PILE-REINFORCED SECTION^99Table 8.1: Composite Strengths and Moduli of the Pile-Reinforced SectionElevation (ft) Composite Strength (psf) Composite Modulus (psf)240-250 230 80,000230-240 700 168,000220-230 1600 176,000215-220 2000 240,000205-215 3300 396,000200-205 3600 180,000191-200 3100 1240,000185-191 3600 540,000173-185 4100 443,000230 psf at elevation 245 ft to 4100 psf at elevation 180ft, with an average value of 2900psf.8.2 Comparison of Results Between the Pile-Reinforced Section And thePlugAs the composite shear strengths and moduli of the pile-reinforced section have beendetermined, the post-liquefaction behaviour of Sardis Dam was reassessed with the pile-reinforced remediation in place in the upstream slope as shown in Fig. 8.1. In theanalysis, 50% of pore water pressure is assumed to be generated in the non-liqufiablesand shell.8.2.1 Loss of freeboardThe results of the pile-reinforced section were compared with the results of the plugmodel which has a constant shear strength of 3000 psf and a width of 120ft. When min-imum residual strength in the weak clayey silt layer drops to its minimum value of 100Chapter 8. REMEDIATION STUDY OF PILE-REINFORCED SECTION^100psf, the loss of freeboard is 4.50 ft for the plug, and it is 4.74 ft for the pile-reinforcedsection. Since the composite shear strengths of the upper part of the pile-reinforcedsection , between the elevations of 215 ft and 240 feet, are less than those of the plugmodel. Therefore, the pile-reinforced section has a little less resistance than plug modelto displacement in the upper section of the zone.8.2.2 Horizontal pressures against the remediated zoneFig. 8.15 presents the post-liquefaction pressure distribution against the downstreamface of the remediated section (side A). The maximum horizontal pressure against thepile-reinforced section is about 10000 psf, compared with a horizontal pressure of 12000psf against the uniform plug. But apart from the peak value near the weak layer, thepressure distributions are very similar.8.2.3 Ratio of shear stress to strength of the remediated zoneThe ratio of the shear stress to the composite shear strength of the pile-reinforced sectionis used to examine the shear resistance of the pile-reinforced section. Fig. 8.16 shows theratios of the shear stress to the shear strength in the remediated section versus elevationswhen the minimum residual strength in the weak clayey silt is 100 psf. For the pile-reinforced section, these ratios are high at elevations between 215 ft and 240 ft becauseof the relatively low shear strengths there, and they are low at elevations below 215 ftbecause of relatively high shear strengths. The maximum ratio of the shear stress to theshear strength of the remediated zone occurs at the level of the weak clayey silt layerwith a value of 0.49. This relatively low ratio indicates a relatively high factor of safetyplug strength 3000 psfplug width 120 ftPile - Reinforced Section- G.after constructionChapter 8. REMEDIATION STUDY OF PILE-REINFORCED SECTION^101Plug240 —220 —0^ ....................................weak clayey silt180160I^1^I^I^1^I 2,000 4,000 6,000 8,000 10,000 12,000 14,000Downstream Horizontal Stress (psf) --- SIDE A200(Do-0Figure 8.15: Pressure Distribution on Downstream Face of Pile - Reinforced Section.....................weak clayey siltPlugplug strength 3000 psfplug width 120 ft....................^rChapter 8. REMEDIATION STUDY OF PILE-REINFORCED SECTION^102260Pile - Reinforced Section240180220 —0ctsLu200 ^—160 ^0 0.1^0.2^0.3^0.4^0.5^0.6^0.7Ratio of Shear Stress to Shear StrengthFigure 8.16: Distribution of Ratio of Shear Stress to Strength at Downstream Face ofPile - Reinforced SectionChapter 8. REMEDIATION STUDY OF PILE-REINFORCED SECTION^103against a shear failure of the remediated zone.8.2.4 Horizontal movement of the remediated zoneFig. 8.17 illustrates the downstream horizontal movements of the remediated zone afterliquefaction. The pile-reinforced section shows a larger displacement at the top of theremediated section because of the lower composite shear strengths and moduli in the up-per regions. The maximum horizontal movement of the pile-reinforced section is about0.49 ft, compared with the plug model of 0.24 ft. These values are within tolerable limitsIt is clear that 24-inch square prestressed concrete piles at 12 ft centres arranged inthe remediated zone can provide sufficient strength and stiffness to control the deforma-tions of Sardis Dam within tolerable limits provided they can be designed structurely tocarry individually the required moments and shears. This is the next stage in the designprocess and will be carried out by structural engineers.220FCO-...7,co>a)uJ2001801600Chapter 8. REMEDIATION STUDY OF PILE-REINFORCED SECTION^104260 ^PlugPile - Reinforced Sectionplug strength 3000 psf_ plug width 120 ft240 —/weak clayey siltI^1^i0.2^0.4 0.6^0.8Downstream Horizontal Displacement(ft) -- SIDE AFigure 8.17: Distribution of Downstream Horizontal Movement of Pile - Reinforced Sec-tionChapter 9SUMMARY AND CONCLUSIONSSardis Dam is expected to suffer a great loss of shear strengths in the liquefied soils duringthe design earthquake shaking. Studies have been made to examine the post-liquefactionperformance of the dam. First, post-liquefaction deformation analyses were conductedon the initial configuration of the dam without remediation. Results of this analysisshow that a significant loss of freeboard and large horizontal displacements of the damwill occur after liquefaction. Furthermore, necessary remedial measures were proposed,and studies have been made to determine the general requirements of both the extent ofthe remediation and the average properties of the proposed remediated section. Finally,the studies show that remediation requirements can be met by driving 24-inch squarepiles into the remediated section. The research performed for Sardis Dam leads to theconclusions below:1. For Sardis Dam without remediation, the overall performance of the dam is verypoor after liquefaction. The loss of freeboard of the dam is 45.5 ft after soil liq-uefaction. The dam will fail along the weak clayey silt layer under the upstreamslope.105Chapter 9. SUMMARY AND CONCLUSIONS^ 1062. The dam is to be stabilized by a 120ft wide remediated zone that crosses the weakfoundation layer and provides an adequate resistance to sliding or bending.3. Parametric studies show that an average shear strength of 3000 psf and a shearmodulus of 900,000 psf are needed to supply the necessary resistance to shearingand bending.4. The remediation zone is to be located with its downstream face at the slope breakin the upstream slope near the conservation level of the pool.5. Porewater pressures up to 50% of effective overburden pressures can be developedin the sand shell without deformations exceeding tolerable levels.6. The preferred method of meeting the strength and stiffness requirements of theremediated section is to drive prestressed concrete piles across the weak clayey siltlayer. Preliminary studies show that 24-inch square prestressed concrete piles at12 ft spacing would be adequate.The above studies have indicated that the remediation requirements for the strengthand stiffness can be met by driving piles into the dam. Research piles have been drivenin a section in order to test the feasibility of driving piles into the foundation sands.The studies have indicated that it is feasible to drive the 24-inch piles. In-situ testswere performed to evaluate densification of the sand shell by the pile driving. Increasingdensification increases the ability of load transfer between the piles. These data are nowbeing analyzed. Studies have begun to determine the shear and moment both staticallyand dynamically for which the individual piles must be designed structurely. The reliableChapter 9. SUMMARY AND CONCLUSIONS^ 107determination of the shear and moment is a. difficult problem in analysis, and variousmethods for analyzing the dynamic response of the piles are being evaluated to determinewhich method of analysis may be best.Bibliography[1] Bathe,K-J,Wilson,E.L.,Iding,R.H.,(1974), \" NONSAP, A Structural Analsis Pro-gram for Static and Dynamic Response of Nonlinear Systems\", Report No. VCSESM 74-3, College of Engineering, University of California, Berkeley, California,February.[2] Casagrande, A. 1936. \"Characteristics of Cohesionless Soils Affecting the Stabil-ity of Slopes and Earth Fills, \" Jounal of the Boston Society of Civil Engineers,reprinted in: Contribution to Soil Mechanics, 1925 to 1940, Boston Society of CivilEngineers, 1940, pp257-276.[3] Castro, G., T.O. Keller ans S.S. Boynton.(1989). Re-evaluation of the Lower SanFernando Dam. Report No. 1, USACE, Waterways Expermental Station, Vicks-burg, Mississippi, September.[4] Castro, G., Poulos, S. J. , and Leather F.D. 1985. \" A Re-examination of the Slideof the Lower San Fernando Dam,\" Journal of Geotechnical Engineering Division,ASCE, Vol. 111, GT9.[5] De Alba, P., H.B., Seed, E. Retamal, and R.B., Seed. 1987. Residual Strengthof Sand from Dam Failures in the Chilean Earthquake of March 3, 1985. Earth-quake Engineering Research Center, Report No. UCB/EERC-87-11, University ofCalifornia , Berkeley, September.[6] Duncan, J.M.,and Chang, C.Y., 1970,\"Non-linear Analysis of Stress and Strain inSoils\",Proceedings,ASCE,Vol. 96, No. SM5, pp.1629-1653.[7] Finn, W.D. Liam, 1981. Liquefaction Potential Development since 1976. Interna-tional Conference on Recent Advances in Geotechnical Engineering and Soil Dy-namics, St. Louis, MO, Vol. 2, pp. 655-681.[8] Finn, W.D.L.,Ledbetter, R.H., Fleming JR,R.L., Templeton,E.T., Forrest,T.W.,Stacy,S.T., (1990a). \" Dam on Liquefiable Foundation : Safety Assessment andRemediation\"[9] Finn, W.D. Liam, and Byrne, P.M., (1976), Estimating Settlements in Dry SandsDuring Earthquakes, Canadian Geotechnical Journal, Vol. 13, No. 4, pp. 355-363.108Bibliography^ 109[10] Finn, W.D. Liam, M. Yogendrakumar, N. Yoshida, and H. Yoshida. (1986). TARA-3: A Program for Nonlinear Static and Dynamic Effective Stress Analysis, SoilDynamic Group, University of British Columbia, Vancouver, B.C., Canada.[11] Finn, W.D. Liam and Yogendrakumar (1989). TARA-3FL - Program for Analysisof Liquefaction Induced Flow Deformations, Dept. of Civil Engineering, Universityof British Columbia, Vancouver, B.C., Canada.[12] Finn, W.D. Liam, M. Yogendrakumar, R. C. Lo, and R.H. Ledbetter. (1990b).Seismic Response of Tailings Dams, State of the Art Paper, Proc., Int. Symposiumon Safety and Rehabilitation of Tailings Dams, Internation Commission on LargeDams, Sydney, Australia, May.[13] Finn, W.D. Liam. (1985) Dynamic Effective Stress Response of Soil Structures;Theory and Centrifugal Model Studies, Proc. 5th Int. Conf. on Num. Methods inGeomechanics, Nogoya, Japan, Vol. 1, 35-46.[14] Idriss, I.M., J. Lysmer, R.N. Hwang, and H.B. Seed. (1973) QUAD4: A Com-puter Program for Evaluating the Seismic Response of Soil Structure By VariableDamping Finite Element Procedures, Rept. No. UCB/ERRC-73-16, University ofCalifornia, Berkeley.[15] Klohn, E.J., Maartman, C.H., Lo, R.C.Y., Finn, W.D.L., 1978. Simplified SeismicAnalysis for Tailings Dams, Proc. ASCE Geotech. Engng. Div., Specialty Conf. onEarthquake Engineering and Soil Dynamics, Pasadena, California, June, Vol. 1,pp. 540-556.[16] Koester, J.P. (1990). Letter Report to Vicksburg District, U.S. Army Corps ofEngineers.[17] Lee,K.W., and Finn, W.D.L., 1978, DESRA-2: Dynamic Effective Stress ResponseAnalysis of Soil Deposits with Energy Transmitting Boundary Including Assess-ment of Liquefaction Potential, Soil Mechanics Series Report No. 38, Dept. of CivilEngineering, University of British Columbia, Vancouver, B.C.[18] Lysmer, J., Udaka,T., Tsai, C.F., and Seed, H.B., (1975). FLUSH: A ComputerProgram for Approximate 3-D Analysis of Soil Structure Interaction Problems,Report No. EERC75-30, Earthquake Engineering Research Center, University ofCalifornia, Berkeley.[19] Martin, G.R., Finn, W.D.Liam, and Seed, H.B., 1975. Fundementals of Lique-faction Under Cyclic Loading, Journal of the Geotechnical Engineering Division,ASCE, Vol. 101, GT5, May,423-438.Bibliography^ 110[20] Newmark, N.M. (1965). Effects of Earthquakes on Dams and Embankments, 5thRankine Lecture, Geotechnique, Vol. 15, No. 2, June, pp139-160.[21] Poulos, S. J., Castro, G., and France, J. W. 1985. \"liquefaction Evaluation Proce-dure,\" Journal of the Geotechnical Engineering Division, ASCE, pp772- 792.[22] Pyles, M. (1981). The Undrained Shearing Resistance of Cohensive Soils at LargeDeformation, Ph.D. Dissertation, University of California, Berkeley.[23] Robertson, P.K. and R.G., Campanella, 1983. \"Interpretation of Cone PenetrationTests,\" Part II (clay), Canadian Geotechnical Journal, Vol. 20, No. 4.[24] Seed, H.B., and Idriss, I.M., 1970. Soil Moduli and Damping Factors for DynamicResponse Analysis, Report No. EERC 70-10, Earthquake Engineering ResearchCenter, University of California, Berkeley.[25] Seed, H. B., R.B., Seed, L.F. Harder and H.-L. Jong, 1988. Re-Evaluation of theSlide in the Lower San Fernando Dam in the Earthquake of February 9, 1971.Report No. UCB/ERRC-88/04, University of California, Berkeley, April.[26] Seed, H.B., Tokimatsu, K., Harder, L.F. and Chung, R.M. 1985 \"Influence of SPTProocedures in Soil Liquefaction Resistance Evaluvations,\" Journal of GeotechnicalEngineering, ASCE, Vol. 3, No. 12.[27] Seed, H. B. 1987, Design Problem in Soil Liquefaction, Journal of GeotechnicalEngineering, ASCE, Vol. 113, No. 7, pp. 827-845.[28] Seed, H.B., and Martin, G.R., The Seismic Coefficients in Earth Dam Design, Jnl.Soil Mech. Fdn. Div. ASCE Vol. 92, No. SM3, May, 1966, pp. 25-28.[29] Seed, R.B., and L.F., Harder Jr. (1990). SPT-Based Analysis of Cyclic Pore Pres-sure Generation and Undrained Residual Strength. Proceedings, H. Bolton SeedMemorial Symposium, J. Michael Duncan (ed.), Vol. 2, May, pp. 351-376.[30] Seed, H.B. (1979). Consideration in the Earthquake Resistance Design of Earthand Rockfill Dams, 19th Rankine Lecture, Geotechnique 29, No. 3, pp. 215-263.[31] Spencer, E., (1973). Thrust Line Criterion in Embankment Stability Analysis,Geotechnique, Vol. 23,No. 1, pp. 85-167.[32] USACE (1989). User's Guide: UTEXAS2 Slope Stabiltiy Package, Vol. II. Theoryby Task Group on Slope Stability , Instruction Report GL-87-1, Final Report, U.S.Army Corps of Engineers, Washington, D.C.Bibliography^ 111[33] Vaid, Y.P., and J.C. Chem. (1985). Cyclic and Monotonic Undrained Responseof Saturated Sands. ASCE National Convention, Session Advances in the Art ofTesting Soils Under Cyclic Loading, Detroit, October 21-25, pp. 12-0 147.[34] Vaid, Y.P., and Finn, W.D.Liam, 1979. Effect of Static Shear on Liquefaction Po-tential, Journal of the Geotechnical Engineering Division, ASCE, Vol. 105, GT10,OCT., pp. 1233-1246.^•[35] Vaid, Y.P., E.K.F. Chung, and R.H. Kuerbis, 1990. \"Stress Path and Steady State,\"Canadian Geotechnical Journal, Vol. 27, No. 1.[36] Wang, W. 1979, \"Some Findings in Soil Liquefaction,\" Water Conservancy andHydroelectric Power Research Institute, Beijing, China, August.[37] Woodward Clyde Consultants 1989, Private Communication[38] Woodward Clyde Consultants 1991, Private Communication"@en ; edm:hasType "Thesis/Dissertation"@en ; vivo:dateIssued "1992-05"@en ; edm:isShownAt "10.14288/1.0050511"@en ; dcterms:language "eng"@en ; ns0:degreeDiscipline "Civil Engineering"@en ; edm:provider "Vancouver : University of British Columbia Library"@en ; dcterms:publisher "University of British Columbia"@en ; dcterms:rights "For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use."@en ; ns0:scholarLevel "Graduate"@en ; dcterms:title "Seismic induced flow deformation and remediation study of Sardis Dam"@en ; dcterms:type "Text"@en ; ns0:identifierURI "http://hdl.handle.net/2429/1859"@en .