22

ilt

bul,FL, U

e dlowitsed sanlemstre

termed L-8 Storage Reservoirlls witsite isMarshArea 1ee Rivt of thea Fedeon of frary pu

and provide protection against overtopping during hurricane events,

Engineering Geology 151 (2012) 8999

Contents lists available at SciVerse ScienceDirect

Engineering

e lssupply to West Palm Beach.The reservoir is owned and operated by the South Florida Water

Management District (SFWMD) and was previously mined by PalmBeach Aggregates, Inc. (PBA) as a commercial limestone quarry. The res-ervoir, shown in Fig. 1, was constructed during the period 20032008,consisting of seven interconnected cells that had beenmined to approx-imately elevation 4.5 m North American Vertical Datum, NAVD, (allelevations in this paper are NAVD-1988, unless otherwise noted),

posits have relatively low permeabilities, which allow utilization ofdry mining techniques for the quarry operation. These characteristicsare also uniquely suited for water storage and withdrawal with min-imal interaction with the surrounding ground water system.According to current geologic literature, the shallow geologic stratainclude variably lithied members referred to as the Bermont Forma-tion (Petuch and Roberts, 2007). The upper-most stratum at the site isidentied as the Fort Thompson Formation. Overall, the ages of thethen the reservoir was further deepened b13 m to provide additional storage capa(Bromwell and Ying, 2010). Perimeter e

Corresponding author. Tel.: +90 216 677 1630/197E-mail addresses: [email protected] (A.T. zer)

(L.G. Bromwell).1 Tel.: +1 863 6672345; fax: +1 863 6672662.

0013-7952/$ see front matter 2012 Elsevier B.V. Allhttp://dx.doi.org/10.1016/j.enggeo.2012.09.011rpose is to help sustainals in order to facilitaterovide additional water

Bend geological formation, which underlies the entire L8SR, is anarea of unique geological character. The shallow cemented sandyshell of the carbonate-rich zone and the underlying shelly sand de-desired ows and levels within adjacent canrestoration of the Loxahatchee River and pThe Loxahatchee Reservoir, also(L8SR), consists of 7 interconnected ceity of approximately 56106 m3. TheBeach County, Florida, north of the Artuge (also termed Water Conservationoriginal headwaters of the Loxahatch2009). The reservoir is an integral parEverglades Restoration Plan (CERP),prove water quality and the distributiida (Christmann et al., 2009). Its primh a water storage capac-located in western Palmall National Wildlife Ref-or WCA-1), within the

er (Bromwell and Ying,overall Comprehensiveral/State project to im-esh water in south Flor-

as well as additional water storage (zer et al., 2009).

2. Site geology

Geologic conditions at the L8SR site include a surcial layer of peatand organic sands underlain by sand intermixed with lenses of silty toclayey sands, overlying a relatively thick deposit of well-cementedsand and shell of marine origin (Land et al., 1973). The 20-Mile1. Introduction elevation +7 m surround the reservoir. The embankments wereconstructed from overburden excavated during the mining process,Stability assessment of an earth dam on s

A. Tolga zer a,, Leslie G. Bromwell b,1

a Civil Engineering Department, Okan University, Tuzla Campus, 34959 Akrat/Tuzla, Istanb AMEC-BCI Engineers & Scientists Inc., 2000 E. Edgewood Drive Ste.215, 33803 Lakeland,

a b s t r a c ta r t i c l e i n f o

Article history:Received 11 August 2011Received in revised form 12 September 2012Accepted 16 September 2012Available online 4 October 2012

Keywords:Rapid drawdownTransient seepageSlope stabilityCoupled analysisSedimented silt/clay tailings

This case study describes thconsisting of low strength,within the construction limCounty, Florida. Fully couplperformed for operationalmethods (LEM) and nite etions utilizing fully coupledity during drawdown.

j ourna l homepage: www.y dredging to elevationcity up to 56106 m3

arth embankments to

8; fax: +90 2166771486., [email protected]

rights reserved./clay tailings foundation: A case study

TurkeySA

esign, and construction of an earth dam built on recently sedimented tailingsplasticity silt/clay, and variably silty sand. The tailings were encountered

of Cell 7 at the 4 km2 Loxahatchee (L-8) Reservoir in western Palm Beachteady state and transient seepage models with slope stability analyses wered rapid drawdown conditions, respectively, using both limit equilibriument stress-based modeling (FESBM) techniques. In addition, stability calcula-ss/pore pressure analysis were performed to analyze rapid drawdown stabil-

2012 Elsevier B.V. All rights reserved.

Geology

ev ie r .com/ locate /enggeoformations present at the PBA facility range from approximatelytwo million years to less than 10,000 years.

3. Field exploration program

This paper concerns the design of the southern perimeter em-bankment of Cell 7 using both limit equilibrium methods (LEM) andnite element stress-based modeling (FESBM) methods to provide

below the ground surface to the top of the rock formation. Thesesoils ranged from 2.5 to 7 m in total thickness, depending on theamount of overburden at the boring locations and the elevation ofthe underlying rock formation. The surcial organic soils were re-moved to facilitate mining operations, except at sporadic locationswhere organic soils were covered by overburden. Limestone and/orlimestone formations were encountered within the borings underly-ing the surcial sands. These are the primary mineable units encoun-tered at the site, and varied between 0.5 and 6 m in total thickness.

Rock coring was performed to determine the quality of the rockand to obtain representative samples. Recorded values for the RockQuality Designation (RQD) ranged between 0 and 48 and recoveryratio ranged between 17 and 83%. Blow counts (N-values) recordedwithin these strata ranged from 24 blows per foot to practical refusal(N-value>50 blows per 15 cm). Intermittent layers of silty and clay-

Cells 1&2

Cell 3 Cell 4

Cell 5

Cell 7 Cell 6

90 A.T. zer, L.G. Bromwell / Engineering Geology 151 (2012) 8999adequate slope stability during normal operation as well as rapiddrawdown conditions. Cell 7, as shown in Fig. 2, is encompassed bythree exterior levees referred to as Cell 7 West, South, and East. Cell7 encompasses a storage area of nearly 0.5 km2. It is bordered onthe south side by a 230 kV Florida Power & Light (FP&L) transmissionline easement, the L-8 canal along the east side, and the process plantalong the west side. A conveyance structure in the form of an openchannel was dredged to connect Cell 7 to Cell 6 through FP&L's trans-

NORTH

Fig. 1. Aerial view of Loxahatchee Reservoir.mission line easement area. Cell 7 originally served as a processwater/tailings disposal pond for limestone quarry operations, whichpreceded reservoir construction at the site.

A eld exploration program was performed to dene the general-ized site-specic soil and rock stratigraphy in order to evaluate theexisting foundation conditions for design of the reservoir embank-ments. A combination of mechanical auger and standard penetrationtest (SPT) borings was completed to depths ranging between 3.5 and20 m below the existing grade using tracked and pneumatic drill rigs.

In the vicinity of the southern embankment of Cell 7, seven soilstrata generally characterize the subsurface geotechnical prole, asshown in Fig. 3. Typically, the surcial soils consist of a combinationof interbedded layers of variably silty and clayey sands extending

incremental loading (IL) consolidation tests were performed on

L-8 CAN

PROCESS WATER/TAILINGS DISPOSAL POND (CELL 7)

PROCESS PLANT NORTH

CELL 7 SOUTH CELL 7 WEST

CELL 7 EAST

Fig. 2. Aerial vieundisturbed silt/clay tailings extracted from the southwestern bermof Cell 7 to provide soil properties for settlement analyses. The resultsof these tests are presented in Table 4.

Based on the soil types encountered, the southwestern Cell 7 typ-ical cross section was divided into layers as follows: embankment ll,sandy ll material, very loose sand, silt/clay tailings, silty sands, andunderlying sands with variable quantities of shell and shell fragments(Figure 3). In-situ and laboratory permeability tests were performed

AL

CONVEYANCE STRUCTURE

CELL 6 ey sands were encountered within the rock. In addition, within thefootprint of the southwestern embankment of Cell 7, low plasticityclay, silt, and variably silty sand tailings, which originate from theplant washing and sizing operations, were encountered. Relativelylow N-values were recorded within the low plasticity clay and silttailings, ranging from weight-of-rod to 10 blows per foot. N-valuesbetween 1 and 50 blows per foot were recorded within the variablysilty sand tailings.

4. In-situ and laboratory testing

Both in-situ and laboratory tests were performed to determine hy-draulic conductivity and strength properties of the materials to use inLEM and FESBMmodels. Laboratory strength tests were conducted onselected samples of embankment ll and silt/clay tailings. Triaxialshear tests (TX) were performed on representative grab samples ofoverburden used to construct the perimeter embankment (BCI,2007; zer et al., 2009). The soil samples were compacted in the lab-oratory to 95% of Modied Proctor maximum density, which was theminimum compaction criteria for the as-built embankments. The re-sults of the TX testing are presented in Table 1.

In addition, three TX tests, one direct shear (DS) test, and one di-rect simple shear (DSS) test were performed on selected undisturbedsilt/clay tailing samples extracted from the southwestern berm of Cell7. The results of these tests are presented in Tables 2 and 3. Also, fourw of Cell 7.

tions for different types of soils built into the software (Geo-Slope

1 Embankment Fill 4 Silt/Clay Tailings2 Sandy Fill Material 5 Silty Sand3 Very Loose Sand 6 Sand with Shell

9

0

-9

-18

12

3 45

6

Foundation Sand with Shell I

Foundation Sand with Shell II

Bench El. -4.5 m 10H:1V

4H:1V

Crest El. 7 m Crest Width: Min. 4 m, Max 8 m.

Normal Operating Pool Level El. 4.5

nce

tion

(mete

rs, N

AVD)

15 m

of

91A.T. zer, L.G. Bromwell / Engineering Geology 151 (2012) 8999for seepage modeling. Laboratory permeability tests were conductedon representative soils used to construct the southern perimeter em-bankment of Cell 7. The soil samples were compacted in the laborato-ry to 95% of Modied Proctor maximum density. In addition, in-situpump well tests were conducted to estimate the permeabilities ofthe subsurface soils (sandy ll material, very loose sand, silty sand,sand with shell, and foundation sands with shell) during the eld in-vestigations. The horizontal permeability and ratio of vertical to hor-izontal permeabilities were calculated based on data obtained fromin-situ and laboratory tests. The soil layer permeability values areshown in Table 5. These values were used in the steady state andtransient seepage analyses.

5. Seepage and stability analyses methods used

According to Lambe and Silva (2003), the components involved indetermining an accurate factor of safety for earth structures are situ-ation (events), geometry (composition), stress and stress path, porepressure, strength, level of stability and probability of failure. Withthe introduction of advanced software products along with increasedcomputing power, powerful techniques such as nite element com-puted stresses have gained popularity in geotechnical engineeringdesign. In this design study for the southwestern embankment ofCell 7, coupled steady-state and transient seepage and deformationanalyses using both LEM and FESBM methods were performed inorder to design the earth dam on silt/clay tailings, taking into accountthe components listed by Lambe and Silva (2003).

5.1. Seepage analyses

Steady-state seepage analyses and transient seepage analyseswere made in order to predict the location of the phreatic surface

-27 6030

Dista

Elev

a

0

Fig. 3. Original design cross-sectionand pore water pressures for steady state and drawdown conditions,respectively. The typical cross-section of the Cell 7 southwestern em-bankment was analyzed, using the nite element program SEEP/W(Geo-Slope International Ltd., 2008a) to determine the ow throughthe embankment based on the embankment geometry, boundarywater pressures, soil prole, and vertical and horizontal soil perme-abilities (hydraulic conductivities).

Table 1Multi-stage triaxial (TX) test results.

Grab sample ID Effective stress envelope

Cohesionc '(kPa)

Friction angle '(degrees)

White sand 0 36Light gray sand 0 38International Ltd., 2008a), and two closed form equations based oncurve t parameters (Van Genuchten, 1980 and Fredlund and Xing,1994). Empirical relations based on the grain size distribution forsandy soils presented by Yang et al. (2004) were used in this studyto estimate Fredlund and Xing (1994) curve tting parameters to de-ne the water characteristics curve for sandy soils within the prole.

The Fredlund and Xing (1994) curve tting parameters for sandysoils are summarized in Table 6, and water retention curves areshown in Fig. 4. The retention curve parameters for foundation sandlayers are irrelevant for the drawdown models since these layers re-main always saturated.

Since the permeability of the silt/clay tailings are very low(Table 5) unsaturated phenomena play a signicant role in their be-havior (Zandarin et al., 2009). The Aubertin et al. (2003) methodwas used to determine the water characteristic curve for silt/clay tail-ings, since it can be applied for a variety of materials, including tail-ings (Geo-Slope International Ltd., 2008a). The resulting curve isalso shown in Fig. 4.

5.2. Stability analysis

Critical loading conditions that may occur over the operational lifeof the structure were analyzed in order to determine the minimumcalculated factor of safety (FOS). These loading conditions and thecorresponding analyses are commonly designated as the steady-Partially and fully saturated conditions can be modeled usingSEEP/W. Volumetric water content functions for the drainage of theupstream slope during drawdown were determined for the unsatu-rated zone above the phreatic surface. In order to dene volumetricwater content functions, SEEP/W offers four methods: one based ongrain size (Aubertin et al., 2003), one based on a sample set of func-

120 15090 180

(meters)

Cell 7 Southwestern Embankment.state seepage and rapid drawdown cases. Several different analy-ses are available to determine the FOS of the embankments, includingLEM (method of slices), boundary element methods (Jiang, 1990),and nite element computed stresses (Matsui and San, 1992;Dawson et al., 1999; Grifts and Lane, 1999; Belirgen, 2007; Krahn,2007; Alkasawneh et al., 2008, and Huang and Jia, 2009). In thisstudy both LEM and FESBM were used for the design of the

Table 2Multi-stage triaxial (TX) and DS test results for southwestern embankment of Cell 7.

Soil Effective stress envelope

Cohesion c ' (kPa) Friction angle ' (degrees)

Silt/clay tailings (TX test#1) 15 40.3Variably silty sand tailings (TX) 17 39.3Silt/clay tailings (TX test#2) 9 36.5Silt/clay tailings (DS) 6 32.0

alban.kuriqiHighlight

alban.kuriqiHighlight

southwestern embankment of Cell 7 for both steady state and draw-down conditions.

LEM is based on the method of slices, and the FOS is dened as theratio between available and mobilized shear strength along the criti-cal failure plane. LEM, which is based only on the principles of statics,has long been used in geotechnical practice for its simplicity com-pared to other methods. However, Krahn (2003) indicated thatsince the LEM does not deal with strain and displacements it doesnot satisfy displacement compatibility. Krahn (2003) recommendeda stressstrain based approach using nite element computed stress-es (FESBM) inside a limit equilibrium framework. In this method, -nite element computed stresses are rst calculated then they areimported into conventional limit equilibrium analysis.

On the other hand, Duncan (1996) indicated that accurate LEMmethods of slices give the same values of FOS as obtained from frictioncircle analyses, log spiral analyses, and nite element analyses. Duncan(1996) also reported that themaximum difference between FOS valuescalculated by methods that satisfy limiting equilibrium conditions is

properties and the cost of nite-element analyses are high (Duncan,1996).

In this design study, slope stability analyses were performed forthe southwestern embankment of Cell 7 using the computer programSLOPE/W (Geo-Slope International Ltd., 2008b). SLOPE/W imple-ments various LEM (including General Limiting Equilibrium Method,Ordinary or Fellenius Method, Bishop's simplied Method, Janbu'ssimplied Method, Spencer Method, MorgensenPrice Method,Corps of Engineers Method, LoweKaraath Method, Sarma Method,and Janbu's generalized Method). For LEM practice, Krahn (2003)recommended using Morgenstern and Price (1967) or Spencer(1967) methods since these methods satisfy both force and momentequilibrium. Also, the Spencer (1967) method was recommended byWright (1969) after examining eight different LEM. In this study, allthe reported FOS values using LEM are calculated using the Spencermethod. LEM steady-state slope stability was investigated using Ef-fective Stress Analyses (ESA). ESA uses predicted pore pressures and

Table 3DSS test results for southwestern embankment of Cell 7.

Soil Parameter Test 1 Test 2

Silt/clay tailings Vertical consolidation stress (kPa) 190 380Shear strain at the peak shear stress (%) 9.9 16.6Corrected peak shear stress (kPa) 46 97Su 0vc

0.24 0.25

dex

Horizontal (kh) Vertical (kv)

Table 6Fredlund and Xing (1994) tting parameters for sandy soils.

Soil a(kPa)

n m

Embankment ll (compacted sand) 1.74 4.47 0.62Sandy ll material 1.77 5.04 0.67Very loose sand layer 1.65 4.71 0.71Silty sand 2.59 4.39 0.59Sand with shell 1.67 5.37 0.66

92 A.T. zer, L.G. Bromwell / Engineering Geology 151 (2012) 8999about 12%, and concluded that an accuracy of about 6% for calculatedFOS values is close enough for practical purposes.

Performing stability analyses using FESBM based on nite elementcomputed stresses inside a limit equilibrium frame work allowsconsideration of soil-structure interaction effects (Krahn, 2003). Inaddition, by calculating stresses using nite element techniques, thereis no need to make assumptions about the interslice forces (Krahn,2003). Using nite elementmodeling it is possible tomodel many com-plex conditions, including non-linear stressstrain behavior, non-homogeneous conditions, and changes in the geometry; however,both the efforts involved in developing the appropriate engineering

Table 4IL consolidation test results.

Soil Void ratioeo

Compression inCc

Silt/clay tailings 0.785 0.159Silt/clay tailings 0.532 0.116Silt/clay tailings 0.492 0.081Silt/clay tailings 1.050 0.238

Table 5Soil permeability values used in seepage analyses.

Soil Elevation (meters)Embankment ll (compacted sand)a Crest (El +7 m) to El +3 mSandy ll materialb El. +3 m to +1.5 mVery loose sand layerb El. +1.5 m to 0.5 mSilt/clay tailingsc El. 0.5 m to 1.5 mSilty sandb El. 1.5 m to 3 mSand with shellb El. 3 m to 4.5 mFoundation sand with shell 1b El. 4.5 m to 15 mFoundation sand with shell 2b El. 15 m to 26 ma Permeability values were from laboratory triaxial test results.b Permeability values were based on in-situ pump well test results.c Permeability values were back calculated from IL consolidation test data.effective stress strength parameters, ' and c '.In order to perform FESBM, the stress-deformation computer pro-

gram SIGMA/W (Geo-Slope International Ltd., 2008c) was utilized inconjunction with SLOPE/W. In order to calculate FOS of ne-grainedsoils during rapid drawdown, consideration of stress induced porepressures and their dissipation, as well as seepage induced pore pres-sures are recommended (Belirgen, 2007). Fully coupled nite ele-ment computed stress/pore-pressure analysis using SIGMA/W alongwith SEEP/W and SLOPE/W provided FOS values during rapid draw-down by taking stress induced pore-pressures into account. In addi-tion, seepage induced rapid drawdown models were also run.

Compression index(eld adjusted)Cc

Recompression indexCr

0.172 0.0090.124 0.00530.089 0.00280.270 0.0244

Layer no Permeability (m/s)1 7.0105 4.0105

2 3.0105 1.5105

3 3.0105 1.5105

4 1.2108 5.8109

5 2.4105 6.7106

6 1.4105 9.1106

7 1.8105 8.8106

8 1.1105 4.8106

alban.kuriqiHighlight

The linear-elastic model uses the general Hooke's Law, and thusrelates strain increments to stress increments, and it affords a system-atic method for relating modulus values to stresses (Duncan, 1996).The main drawbacks of this model are that it cannot consistently dis-tinguish between loading and unloading phases (Schanz et al., 1999),and also doesn't model plastic deformations in a fully logical way0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.1 1 10 100 1000

Vol

umet

ric W

ater

Con

tent

Suction (kPa)

Embankment FillSandy FillVery Loose SandSilt/Clay TailingsSilty SandSand with Shell

Fig. 4. Soil-water characteristics curves.

Table 8Modied cam-clay parameters for silt/clay tailings.

Soil eo

Silt/clay tailings 1.050 0.270 0.0244

93A.T. zer, L.G. Bromwell / Engineering Geology 151 (2012) 8999In order to perform FESBM steady-state slope stability analyses,the stressstrain relationships of the materials in the model need tobe dened. SIGMA/W includes six different soil constitutive modelsincluding the linear-elastic model, anisotropic elastic model, non-linear elastic hyperbolic model, elasticplastic model, and modiedCamClay model (Geo-Slope International Ltd., 2008c).

The linear-elastic model is the simplest soil constitutive model,and only requires Elastic Modulus (E) and Poisson's ratio ().Duncan (1996) pointed out that whenever the simplest representa-tion of soil behavior would result in acceptable accuracy, there is noneed to use a very complex constitutive model to analyze the prob-lem. In addition, the linear-elastic model always ensures a solutionsince there are no convergence difculties as with non-linear consti-tutive models (Krahn, 2003).

The linear-elastic model was used for all the sandymaterials in thecross-section, since the strains for the sandy materials are expected tobe very small. In addition, it was considered that a reasonable stressdistribution can be obtained using simple linear-elastic model param-eters for sandy materials since there are no zones of hard materials ora soft core in the southwestern embankment of Cell 7.

In order to perform steady-state FESBM slope stability calculationsfor tailings consisting of low strength, low plasticity silt/clay, and var-iably silty sand, two other constitutive models, namely the non-linearelastic hyperbolic and modied CamClay models, were used in addi-tion to the linear-elastic model, and the results were compared.SIGMA/W uses the formulation developed by Duncan and Chang(1970) for the non-linear hyperbolic model. The advantage of thismodel is that the required model parameters can be obtained directlyfrom triaxial strength tests, which are commonly performed to deter-mine engineering properties.Table 7Material properties for both LEM and FEM analysis.

Material Unitweight, (kN/m3)

Internalfriction angle, ' (degrees)

Cohesion,c ' (kPa)

Poissonratio,

Elasticmodulus,E (MPa)

Embankment(compacted ll)a

18 36 0 0.35 56

Sandy llmaterialb

17 28 0 0.30 20

Very loose sandb 15 25 0 0.30 15Silt/clay tailingsa 14 36 9 0.40 15Silty sandb 17 29 0 0.35 20Sand with shellb 18 33 0 0.35 50Foundation sandwith shell 1b

20 35 0 0.35 80

Foundation sandwith shell 2b

20 38 0 0.35 100

a Based on laboratory triaxial test results.b Based on SPT correlations.(Duncan, 1996). After considering the low compression ratio, (0.06to 0.13), relatively low void ratio (0.5 to 1.0), and also fairly lowand narrow range for the plasticity index (5 to 18), the Duncan andChang hyperbolic model (Duncan and Chang, 1970) was selected asone of the constitutive models for the FESBM stability calculationsof silt/clay tailings.

The third constitutive model selected for silt/clay tailings was themodied CamClay model, since relatively low SPT blow count valueswere recorded within these tailings, ranging between weight-of-rodand 10 blows. The modied CamClay model's formulation inSIGMA/W is based on Britto and Gunn (1987). It is an effective stressbased model and undrained analysis cannot be performed using thismodel, which is considered a drawback of the model.

When performing rapid drawdown stability analyses based onFESBM, the rst step was establishing the long-term seepage condi-tions using SEEP/W. To establish stress/pore pressure analysis,in-situ stress conditions have to be established using SIGMA/W. Thisphase of the model uses pore pressure conditions from parentlong-term seepage analyses, and the reservoir pressure was appliedas a uid-pressure type of boundary condition along the upstreamface of the reservoir. After establishing long-term seepage andin-situ stress conditions, the SIGMA/W transient model was complet-ed by simulating the removal of the pool water by applying twoboundary conditions on the upstream face of the reservoir. They arethe stressstrain uid pressure boundary function and the hydraulicpressure boundary function. Then, nite element computed stressesduring drawdown stage are brought into limit equilibrium frame-work (SLOPE/W) to complete the analyses.

6. Seepage and stability analyses results

6.1. Steady state seepage and long-term stability analysis results

Earth dams reach equilibrium after prolonged water storage undernormal operational pool levels. It can take many years to establishsteady state seepage through an earth dam. Steady state stabilityanalysis does not take into account the time required to reach equilib-rium. Steady state seepage analyses (long-term) were made using thenite element program SEEP/W in order to predict the equilibrium lo-cation of the phreatic surface within the embankment. Reservoirlevels of 4.5 m (normal operating pool), 5 m (maximum operating

Table 9Upstream static slope stability analysis with steady state seepage.

Method Calculated FOS

Normaloperating pool

Maximumoperating pool

Maximumsurcharge poolLEM 2.91 3.05 3.26FESBMa 2.81 2.94 3.16FESBMb 2.80 2.93 3.15FESBMc 2.89 3.04 3.28

a With linear-elastic model for silt/clay tailings.b With non-linear hyperbolic model for silt/clay tailings.c With modied cam-clay model for silt/clay tailings.

alban.kuriqiHighlight

less than the variability in soil engineering properties. Thus, LEM

1

34

2

5 6

Foundation Sand with Shell I

Foundation Sand with Shell II

1 Embankment Fill 4 Silt/Clay Tailings2 Sandy Fill Material 5 Silty Sand3 Very Loose Sand 6 Sand with Shell

Fig. 5. Steady-state critical failure plane determined by LEM for operational pool (FOS=2.9).

13

42

5 6

Foundation Sand with Shell I

1 Embankment Fill 4 Silt/Clay Tailings2 Sandy Fill Material 5 Silty Sand3 Very Loose Sand 6 Sand with Shell

raw

94 A.T. zer, L.G. Bromwell / Engineering Geology 151 (2012) 8999pool), and 5.5 m (maximum surcharge pool) were used in theanalyses.

The input parameters for the ESA (steady state) analyses are sum-marized in Table 7. Shear strength values and elastic modulus for theembankmentll and silt/clay tailing layerswere calculated based on tri-axial tests (Tables 1 and 2). For the rest of thematerials, conservative es-timates of the material properties were made based on SPT data.

The input parameters for the modied CamClay model derivedfrom IL consolidation test results were used for silt/clay tailings, assummarized in Table 8. SIGMA/W allows the user to dene the mod-ulus of elasticity versus stress curves for non-linear hyperbolicmodels, thus the modulus versus stress curve from TX test resultsshown in Table 2 are used for non-linear models of silt/clay tailings.

The upstream slope stability with steady state seepage using theow properties from Table 5 is shown in Table 9 for reservoir levelsof 4.5 m (normal operating pool), 5 m (maximum operating pool),and 5.5 m (maximum surcharge pool). Fig. 5 shows the critical failureplane for the case of normal operating pool calculated using LEM. Rel-atively close agreement was found between the critical failure planes

Fig. 6. Phreatic line after 10 days of dobtained from both LEM and the three FESBM methods for thesteady-state analysis.

Based on the results shown on Table 9, the upstream embankmentslope for all reservoir levels exceeded the minimum required FOS of1.5 for static slope stability under steady-state seepage conditionsestablished by the United States Army Corps of Engineers (USACE,2003). Regardless of the selected soil behavior (linear-elastic,

1

1.1

1.2

1.3

1.4

1.5

0 2 4 6 8 10 12 14 16 18Fac

tor o

f Saf

ety

(Base

d on F

ESBM

)

Elapsed Time (Days)

Drawdown from 4.5 mDrawdown from 5.0 mDrawdown from 5.5 m

Fig. 7. Results of rapid drawdown analysis.analysis using a method that satises both force and moment equilib-rium (i.e., Spencer's Method) performed equally as well as FESBMbased models for this study. Among the FESBM models, thelinear-elastic model can be used for low plasticity silt/clay tailingssince it provided reasonable stress conditions and did not requiregreat effort to obtain the input parameters relative to other morecomplex constitutive models.

6.2. Rapid drawdown stability analysis results

Rapid drawdown is dened as the sudden lowering of the reser-voir water level, and the rate of acceptable drawdown needs to be de-ned for embankment dams. If the water within the embankmentnon-linear hyperbolic, and modied CamClay) for the FESBMmodeling of silt/clay tailings material, reasonably close agreementwas obtained with the FOS value using Spencer's LEM method. Themaximum difference between calculated FOS methods using LEMand FESBM are within approximately +4 and 1%, which is much

Foundation Sand with Shell II

down from normal operating pool.cannot drain fast enough as the reservoir is lowered, internal porepressure will build within the dam and may cause the upstreamslope to fail. There are numerous reported drawdown failures in theliterature including Pilarcitos Dam, Walter Bouldin Dam, numerousriverbank slopes along Rio Montario (Duncan et al., 1990), and bankslopes of Licking River (Londono, 2004).

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Tota

l Hea

d (m

eters)

Elapsed Time (Days)

Transient Seepage ModelStress/Pore Pressure Coupled ModelReservoir Pool

Fig. 8. Pore water pressure response.

In order to perform drawdown slope stability analyses, transientseepage analyses were rst performed utilizing SEEP/W. In order toperform transient seepage analyses for the rapid drawdown condi-tion, steady state seepage analyses (long-term) were rst performedutilizing SEEP/W in order to predict the equilibrium location of thephreatic surface within the embankment. Once the steady-state anal-ysis was completed, transient seepage modeling was performed byapplying a drawdown rate of 0.3 m/day along the upstream face ofthe embankment as a moving head boundary condition.

Based on the results of the transient seepage analyses, ow netswere generated for the southwestern embankment of Cell 7. Thenthe ow nets were imported into conventional limit equilibrium anal-ysis (SLOPE/W) to perform stability analyses according to the USACE(2003) method. Fig. 6 shows the phreatic line after 10 days of draw-down from normal operating pool level. A dot within the middle ofthe silt/clay tailings layer in Fig. 6 is a reference point, which will bediscussed further regarding the pore-pressure response.

zer et al. (2009) reported the results of coupled transient seep-age models with rapid drawdown slope stability analyses using LEMin three stages in accordance with the procedures described inUSACE (2003). The method was developed by Lowe and Karaath(1960), and later modied by Wright and Duncan (1987), andDuncan et al. (1990). zer et al. (2009) concluded that using the tri-axial shear test results and the USACE (2003) procedure, the mini-mum required factors of safety were satised (varied between 1.4and 1.6). However, analyses using undrained strength values fromDSS testing (Table 3) gave a calculated minimum factor of safetyvalues on the order of 1.0 (varied between 1.0 and 1.03), which is un-acceptably low for rapid drawdown.

In this study, coupled stress/pore pressure modeling results fromrapid drawdown analyses were used in FESBM modeling in order toinclude the effect of stress conditions and soil-structure interactionduring the drawdown. The rapid drawdown transient seepage analy-ses were performed for reservoir levels of 4.5 m (normal operatingpool), 5 m (maximum operating pool), and 5.5 m (maximum sur-charge pool). For the L8SR, the design drawdown rate is 0.3 m/day.

Because linear-elastic analysis seemed adequate to obtain reasonablestress conditions for silt/clay tailings during steady state stability analy-ses, the linear-elastic soil properties were also used for the FESBM anal-yses of rapid drawdown. As recommended by zer et al. (2009)undrained DSS test results (Table 3) were used for slope stability calcu-lations for the silt/clay tailings. As pointed out by Ladd and DeGroot(2003), DSS tests better represent typical stress conditions imposed ona soft foundation layer by an overlying earth embankment. During DSStesting, the shear stresses are applied horizontally, which is most likelythe case in the eld, rather than the approximately 45 shear planethat occurs in triaxial tests. In addition, theDSS test representsmobilizedshear strength that would likely be the failure mechanism underundrained shear conditions for embankments on soft clays. In addition,DSS test results will generally show a lower strength value for soft cohe-sive soils thanwill triaxial compression tests (Ladd and DeGroot, 2003).

The results of transient stability models based on FESBM areshown in Fig. 7. Analyses using the FESBM modeling indicated mini-mum factor of safety values on the order of 1.0 (varied between1.03 and 1.06) which is in close agreement with the LEM valuesusing Spencer's method (factor of safety values varied between 1.0and 1.03). For the rapid drawdown case, the maximum difference be-tween calculated FOS methods using LEM and FESBM are within

4

4.0

4.5

7

ng Reservoir Pool

FO

mid st

lace

oun

ion

(a)

b) Ti

95A.T. zer, L.G. Bromwell / Engineering Geology 151 (2012) 89990.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

0 1 2 3 4 5 6

Res

ervo

ir Po

ol E

leva

tion

duri

Dra

wdo

wn

(m)

This line corresponds to therapid drawdown for pumpe

6 m. wide toe buttress

9 m. wide over excavation and rep

1 Embankment Fill 4 Silt/Clay Tailings2 Sandy Fill Material 5 Silty Sand3 Very Loose Sand 6 Sand with Shell

F

Foundat

(ElapsedFig. 9. (a) Results of rapid drawdown analysis when toe buttress and over-excava1

1.5

2

2.5

3

3.5

8 9 10 11 12 13 14 15

S

nimum required FOS = 1.3 against orage dams (USACE, 2003)

ment area1

34

2

56

dation Sand with Shell I

Sand with Shell II

Fact

or o

f Saf

ety

me (Days)tion are implemented. (b) Design cross-section and the critical failure plane.

approximately 3%, which is, much less than the variability in soil en-gineering properties. This indicates that the normal stress distributionalong the slip surfaces are in good agreement in limit equilibrium andnite element based models.

The generated excess pore water pressure behavior during rapiddrawdown using transient seepage analysis and the fully coupledstress/pore pressure model are compared within the silt/clay layerat the location shown in Fig. 6. As shown in Fig. 8 both the transientseepage and stress/pore pressure models produced nearly identicalresults. Consequently, the coupled stress/pore pressure model usinglinear-elastic material properties resulted in reasonable values ofstress distribution during rapid drawdown. Thus, LEM analysisusing a method that satises both force and moment equilibrium

(i.e., Spencer's Method) performed equally as well as FESBM basedmodels for this study.

Regardless of the model (FEM or FESBM) selected to analyze therapid drawdown stability of the reservoir, the factor of safety valuesagainst rapid drawdown were on the order of 1.0, which is unaccept-ably low for the rapid drawdown condition. Therefore, the design ofthe Cell 7 southwestern embankment called for the over-excavationof very loose or soft deposits of mine tailings encountered beneath theembankment footprint within 6 to 9 m of the inside toe. In addition, a6-meter wide toe buttress was added from the mine bottom at4.5 mup to an elevation of 0.5 m to provide additional resistance againstslope instability in the event of rapid drawdown of the reservoir andalso to create a bench for a possible future roller compacted concrete

(a)

(b)

(c)

Shot Rock Placement at elevation -4.5 m

Embankment Construction

Natural Ground at elevation +6 m

Tailings Removal

Tailings Removal

96 A.T. zer, L.G. Bromwell / Engineering Geology 151 (2012) 8999Fig. 10. (a) Embankment ll placement; (b) tailings removal; and (c) Cell 7 southern embankment in operation.

7ed T

e med

oun

ion

ace

(a

(b

97A.T. zer, L.G. Bromwell / Engineering Geology 151 (2012) 89990.0

0.51.0

1.52.0

2.53.0

3.54.0

4.5

0 1 2 3 4 5 6

Res

ervo

ir Po

ol E

leva

tion

durin

g D

raw

dow

n (m

)

Elaps

This line corresponds to thrapid drawdown for pump

1 Embankment Fill 4 Silt/Clay Tailings2 Sandy Fill Material 5 Silty Sand3 Very Loose Sand 6 Sand with Shell

F

Foundat

6 m. wide toe buttress

9 m. wide over excavation and replslope protection system. These stabilization measures resulted in a dra-matic increase in the calculated factor of safety against rapid drawdownfrom approximately 1.0 to 2.8 (Figure 9).

7. Construction details and post-construction performance

The southwestern embankment design called for the over-excavation of very loose or soft deposits of mine tailings encounteredbeneath the embankment footprint within 6 to 9 m of the inside toe.The extent of over-excavation depended on the location of tailingsalong the alignment. Unsuitable tailings were replaced with suitableembankment ll material to increase the factor of safety of the slopeand to reduce potential post-construction settlement. In addition, atoe buttress consisting of well-graded limestone rockwas added duringconstruction to provide additional resistance against slope instability inthe event of rapid drawdown of the reservoir.

Vertical construction of the southern Cell 7 embankment extend-ed from the pit bottom at Elevation4.5 m to the crest at Elevation+7 m (Figure 10a). Prior to the start of construction, the water levelin Cell 7 was lowered to Elevation5 m in order to excavate tailingsmaterial and prepare the foundation for the embankment. Thesouthern alignment of Cell 7 was constructed from the pit bottom(Elevation 4.5 m) because of constraints imposed as a conse-quence of original mining operations and the location of the trans-mission line easement relative to the proposed construction.

Cell 7, as shown in Fig. 2, is bordered on the south side by a 230 kVFP&L transmission easement, and a conveyance structure in the formof an open channel was dredged to connect Cell 6 to Cell 7 throughthe transmission line easement area. Seepage into the south side ofCell 7 from the adjacent Cell 6, which was at a higher water level, re-quired installation of a steel sheetpile wall across the conveyance

Fig. 11. (a) Results of rapid drawdown analysis for the as-b1

1.5

2

2.5

3

3.5

8 9 10 11 12 13 14 15

Fact

or o

f Saf

ety

ime (Days)

Reservoir PoolFOS

inimum required FOS = 1.3 against storage dams (USACE, 2003)

1

46

4

4

5

dation Sand with Shell I

Sand with Shell II

6

ment area

)

)channel, which had been backlled in order to allow lowering of thewater in Cell 7. The sheetpile wall reduced the seepage and allowedexcavation of the tailings material.

After the pit bottom was exposed and dewatered by rim ditching,placement of embankment ll began. In some areas along the em-bankment, wet conditions were encountered in excavations aftermine tailings had been removed. These excavations were backlledwith well-graded compacted limestone rock to provide a stable work-ing platform for the placement and compaction of embankment ll(Figure 10a).

All unsuitable mine tailings were removed to the extent practicaland replacedwith clean compactedll after the pit bottomwas exposedin order to increase the factor of safety of the slope and to reduce poten-tial post-construction settlement (Figure 10b). Embankment ll wasthen placed by large earth-moving equipment and spread by bulldozersin 30-cm thick lifts before compaction.Waterwas added as necessary toachieve a moisture content near optimum, and compaction wasperformed by large vibratory roller compactors. The compacted surfaceof each lift was scaried before the overlying layer was placed. The res-ervoir was further deepened by dredging to elevation 13 m to pro-vide additional storage capacity. After the completion of embankmentconstruction, sod was placed over the embankment (Figure 10c).

Six SPT conrmation borings were performed on August 16 and 17,2007 along the Cell 7 South embankment crest as a part of thepost-construction conrmation process. The borings generally encoun-tered layers of compacted sand and variably silty sand. However, pocketsof tailings were encountered between Elevations +2.3 and +4 m, 1and +1.5 m, and 2 to 3 m during the performance of one of theseborings. Recorded SPT blow counts in the tailings ranged from 2 to 13blows per foot, indicating a very soft to stiff consistency for the soil unit.This worst case scenario was modeled at the permitted drawdown rate

uilt worst case scenario. (b) The critical failure plane.

2007

2007

2008

2008

2008

2009

2009

te

e re

98 A.T. zer, L.G. Bromwell / Engineering Geology 151 (2012) 8999and the calculated factor of safety against rapid drawdown was slightlyreduced from 2.8 (Figure 9) to 2.5 (Figure 11) which still satised theminimum required FOS against rapid drawdown (1.3).

Cell 7 embankment construction was completed on August 2007.As shown in Fig. 12, the reservoir has been through 4 drawdowncycles since completion. The average drawdown rate was approxi-mately 12 cm/day which was less than the design drawdown rateof 30 cm/day. Maximum daily drawdown during the last 4 yearswas 25 cm/day which is relatively close to the design drawdownrate. Frequent visual inspection of the physical condition of thesouthern embankment of Cell 7 has been performed since the initiallling and during subsequent drawdowns. No indications of struc-tural instability have occurred indicating that the toe buttress and

07/0

6/

05/1

0/

02/0

2/

01/0

6/

29/0

9/

27/0

1/

27/0

5/

Fig. 12. Post-construction performanc-1

0

1

2

3

4

Res

ervo

ir Po

ol E

leva

tion

(m, N

AVD)

Daremoval of the silt/clay tailings materials 6 to 9 m inside the toehas provided structural integrity. In addition to the visual observa-tions, post construction behavior of the embankment is alsomodeledunder the historical pool levels (Figure 12). Post construction analy-ses and observations indicate that themeasures taken to increase thestability against rapid drawdown loading were successful.

8. Conclusions

This paper describes the design of Cell 7 of the L8SR in westernPalm Beach County, Florida, part of which was built on recentlysedimented tailings consisting of low strength, low plasticity silt/clay, and variably silty sand. Fully coupled steady state and transientseepage models with slope stability analyses were performed for op-erational and rapid drawdown conditions, using both LEM and FESBMtechniques. The results of the analyses showed that a LEM analysisusing a method that satises both force and moment equilibrium(i.e., Spencer's Method) agreed closely with more complex FESBMbased models.

Additionally, results using a linear elastic soil behavior model forFESBM analysis were compared with more complex constitutivemodels for the low plasticity silt/clay and variably silty sand tailings.The linear elastic soil behavior model gave comparable and reason-able values of stress distribution during both steady state and rapiddrawdown analyses. However, for the modeling of higher plasticitytailings materials, more complex material models should be studied.Excess pore pressure response during rapid drawdown was com-pared between transient seepage and stress/pore pressure models,and it was found that both approaches produced nearly identical re-sults. Consequently, the coupled stress/pore pressure model usinglinear-elastic material properties resulted in reasonable values ofstress distribution during rapid drawdown. It was found that rapiddrawdown governs the design, and regardless of the type of stabilityanalysis (either LEM or FESBM), undrained DSS strength test resultsfor the low plasticity tailings should be considered for the designagainst rapid drawdown failure. It would be prudent for a design en-gineer to perform the analyses using the results from both triaxial andDSS tests to dene the shear strength behavior of undrained soils, anduse the lower, more critical, factor of safety value for the design.2.4

2.45

2.5

2.55

2.6

2.65

2.7

2.75

2.8

2.85

2.9

24/0

9/20

09

22/0

1/20

10

22/0

5/20

10

19/0

9/20

10

17/0

1/20

11

17/0

5/20

11

14/0

9/20

11

Fact

or o

f Saf

ety

Reservoir elevationFactor of Safety

garding factor of safety calculations.References

Alkasawneh, W., Malkavwi, A.I.H., Nusairat, J.H., Albataineh, N.A., 2008. Comparativestudy of various commercially available programs in slope stability analysis. Com-puters and Geotechnics 35, 428435.

Aubertin, M., Mbonimpa, M., Bussiere, B., Chapuis, R.P., 2003. A model to predict thewater retention curve from basic geotechnical properties. Canadian GeotechnicalJournal 40, 11041122.

BCI Engineers and Scientists, 2007. Design Basis Memorandum: L8 Storage ReservoirPalm Beach County, FL. Prepared for Palm Beach Aggregates, Inc.

Belirgen, M.M., 2007. Investigation of stability of slopes under drawdown conditions.Computers and Geotechnics 34, 8191.

Britto, A.M., Gunn, M.J., 1987. Critical state soil mechanics via nite elements. JohnWiley and Sons, Inc., New York.

Bromwell, L.G., Ying, K., 2009. Predicted vs. measured slope erosion at L-8 reservoirduring hurricane Jeanne. Association of State Dam Safety Ofcials (ASDSO) DamSafety'09, Hollywood, FL.

Bromwell, L.G., Ying, K., 2010. Correlation of predicted and measured slope erosion.ISCE-5, 5th International Conference on Scour and Erosion, San Francisco, CA.

Christmann, C.W., Bromwell, L.G., Schwartz, M., 2009. Soil-bentonite slurry wall forseepage control at L-8 reservoir. Association of State Dam Safety Ofcials(ASDSO) Dam Safety'09, Hollywood, FL.

Dawson, E.M., Roth, W.H., Drescher, A., 1999. Slope stability analysis by strength reduc-tion. Geotechnique 49 (6), 835840.

Duncan, J.M., 1996. State of art: limit equilibrium and nite element analysis of slopes.Journal of the Geotechnical Engineering Division, ASCE 122 (7), 577596.

Duncan, J.M., Chang, C.Y., 1970. Nonlinear analysis of stress and strain in soils. Journalof the Soil Mechanics and Foundation Engineering Division, ASCE 96 (SM5),16291653.

Duncan, J.M., Wright, S.G., Wong, K.S., 1990. Slope stability during drawdown. (May1990) Proceedings of the H. Bolton Seed Memorial Symposium, vol. 2, pp. 253272.

Fredlund, D.G., Xing, A., 1994. Equations for the soilwater characteristics curve. Cana-dian Geotechnical Journal 31, 521532.

Geo-Slope International Ltd., 2008a. Seepage Modeling with SEEP/W 2007, An Engi-neering Methodology, 3rd edition. Geo-Slope International Ltd., Calgary: Alberta,Canada.

Geo-Slope International Ltd., 2008b. Stability Modeling with SLOPE/W 2007, An Engi-neering Methodology, 4th edition. Geo-Slope International Ltd., Calgary: Alberta,Canada.

Geo-Slope International Ltd., 2008c. Stress-deformation Modeling with SIGMA/W2007, An Engineering Methodology, 3rd edition. (Calgary, Alberta, Canada).

Grifts, D.V., Lane, P.A., 1999. Slope stability analysis by nite elements. Geotechnique49 (3), 387403.

Huang, M., Jia, C.-Q., 2009. Strength reduction FEM in stability analysis of soil slopessubjected to transient unsaturated seepage. Computers and Geotechnics 36, 93101.

Jiang, Y.S., 1990. Slope Analysis Using Boundary Elements. Springer-Verlag Publishers,New York.

Krahn, J., 2003. The 2001 R.M. Hardy Lecture: the limits of limit equilibrium analyses.Canadian Geotechnical Journal 40, 643660.

Krahn, J., 2007. Limit equilibrium, strength summation and strength reductionmethods for assessing stability. Proceedings of 1st Canada-U.S. Rock MechanicsSymposium, Vancouver, B.C., Canada, pp. 2830.

Ladd, C.C., DeGroot, D.J., 2003. Recommended practice for soft ground site characterization:Arthur Casagrande Lecture. 12th Panamerican Conference on Soil Mechanics and Geo-technical Engineering, MIT, June 2225.

Lambe, T.W., Silva, F., 2003. Evaluating the stability of an earth structure. Proceedingsof 12th Pan-American Conference on Soil Mechanics and Foundation Engineering,Cambridge, MA, pp. 27852790.

Land, L.F., Rodis, H.G., Schneider, J.J., 1973. Appraisal of the Water Resources of EasternPalm Beach County Florida. U.S. Geologic Survey.

Londono, A.C., 2004. Bank instability resulting from rapid ood recession along theLicking River, Kentucky. Master of Science Thesis, University of Cincinnati.

Lowe, J., Karaath, L., 1960. Stability of earth dams upon drawdown. Proceedings of theFirst PanAmerican Conference on Soil Mechanics and Foundation Engineering,Mexico D.F, pp. 537552.

Matsui, T., San, K., 1992. Finite element slope stability analyses by shear strength re-duction technique. Soils and Foundations 32, 5970.

Morgenstern, N.R., Price, V.E., 1967. The analysis of the stability of embankments as-suming parallel interslice forces. Geotechnique 17, 1126.

zer, A.T., Faulk, W., Bromwell, L.G., 2009. Rapid drawdown stability of dam on silt/claytailings foundation. Association of State Dam Safety Ofcials (ASDSO) Dam Safe-ty'09, Hollywood, Florida, USA.

Petuch, E.J., Roberts, C.E., 2007. The Geology of the Everglades and Adjacent Areas. CRCPress, Boca Raton, Florida.

Schanz, T., Vermeer, P.A., Bonnier, P.G., 1999. The hardening soil model: formulationand verication. Beyond 2000 in Computational Geotechnics 10 Years of PLAXIS.Balkema, pp. 281296.

Spencer, E., 1967. A method of analysis of embankments assuming parallel inter sliceforces. Geotechnique 17 (1), 1126.

Unites States Army Corps of Engineers (USACE) Engineering and Design, 2003. SlopeStability Manual, No. EM 1110-2-1902.

Van Genuchten, M., 1980. A closed-form equation for predicting the hydraulic conduc-tivity of unsaturated soils. Soil Science Society of America Journal 44, 892898.

Wright, S.G., 1969. A study of slope stability and the undrained shear strength of clayshales. Ph.D. Dissertation, University of California, Berkley, CA.

Wright, S.G., Duncan, J.M., 1987. An examination of slope stability computation proce-dures for sudden drawdown. Miscellaneous Paper GL-87-25. U.S. Army EngineerWaterways Experiment Station, Vicksburg, MS.

Yang, H., Rahardjo, H., Leong, E.-C., Fredlund, D.G., 2004. Factors affecting drying andwetting soilwater characteristic curves of sandy soils. Canadian GeotechnicalJournal 41, 908920.

Zandarin, M.T., Oldecop, L.A., Rodriguez, R., Zabala, F., 2009. The role of capillary waterin the stability of tailing dams. Engineering Geology 105, 108118.

99A.T. zer, L.G. Bromwell / Engineering Geology 151 (2012) 8999

Stability assessment of an earth dam on silt/clay tailings foundation: A case study1. Introduction2. Site geology3. Field exploration program4. In-situ and laboratory testing5. Seepage and stability analyses methods used5.1. Seepage analyses5.2. Stability analysis

6. Seepage and stability analyses results6.1. Steady state seepage and long-term stability analysis results6.2. Rapid drawdown stability analysis results

7. Construction details and post-construction performance8. ConclusionsReferences