influence of matric suction on geotextile reinforcement-marginal soil interface strength

lable at ScienceDirect

Geotextiles and Geomembranes 42 (2014) 139e153

Contents lists avai

Geotextiles and Geomembranes

journal homepage: www.elsevier .com/locate/geotexmem

Influence of matric suction on geotextile reinforcement-marginal soilinterface strength

Danial Esmaili 1, Kianoosh Hatami*, Gerald A. Miller 2

School of Civil Engineering and Environmental Science, University of Oklahoma, 202 W. Boyd Street, Room 334, Norman, OK 73019, USA

a r t i c l e i n f o

Article history:Received 8 June 2013Received in revised form11 January 2014Accepted 27 January 2014Available online 28 February 2014

Keywords:GeotextilesMoisture reduction factorMarginal soilsSoil matric suctionInterface strengthPullout resistance

* Corresponding author. Tel.: þ1 (405) 325 3674; faE-mail addresses: [email protected] (D.

(K. Hatami), [email protected] (G.A. Miller).1 Tel.: þ1 (405) 325 5911; fax: þ1 (405) 325 4217.2 Tel.: þ1 (405) 325 4253; fax: þ1 (405) 325 4217.

0266-1144/$ e see front matter � 2014 Elsevier Ltd.http://dx.doi.org/10.1016/j.geotexmem.2014.01.005

a b s t r a c t

This paper presents descriptions and results of multi-scale pullout and interface shear tests on a wovenpolypropylene (PP) geotextile reinforcement material in a marginal quality soil. A main objective of thesetests was to develop a moisture reduction factor (MRF) for the pullout resistance equation in thecurrently available design guidelines. The tests were carried out at different overburden pressure andgravimetric water content (GWC) values. The differences in the soil-geotextile interface strength amongthe cases with different GWC values were used to determine the corresponding MRF values. Results ofthe study indicate that the reinforcement interface strength and pullout resistance could decreasesignificantly as a result of the loss in the matric suction (e.g. by 42% between the cases of 2% dry and 2%wet of the soil optimum moisture content). It is concluded that wetting of the soil-geotextile interfaceduring construction or service life of a reinforced soil structure can measurably reduce the interfacestrength and pullout resistance of the geotextile reinforcement which needs to be accounted for indesign. Results of the study will be also useful to estimate the difference in the pullout capacity ofgeotextile reinforcement in a marginal soil when placed at different GWC values during construction. Themethodology described in the paper could be used to expand the database of MRF results to include awider range of soil types and geotextile reinforcement for practical applications.

� 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Transportation agencies worldwide are faced with the persis-tent problem of landslides and slope failures along highways, rail-roads and other transportation-related infrastructure. Repairs andmaintenance work due to these failures are extremely costly. InOklahoma, many of these failures occur in the eastern and centralparts of the state due to higher topography and poor soil type(Hatami et al., 2010a,b, 2011). A recent example of these failures is alandslide on the US Route 62 in Chickasha, Oklahoma (Fig. 1).

An ideal solution for the construction or repair of highwayslopes and embankments is to use coarse-grained, free-drainingsoils to stabilize these structures as recommended by designguidelines and specifications for Reinforced Soil Slopes (RSS) andMechanically Stabilized Earth (MSE) structures in North America(e.g. Elias et al., 2001; Berg et al., 2009). However, coarse-grainedsoils are not commonly available in many parts of the world.

x: þ1 (405) 325 4217.Esmaili), [email protected]

All rights reserved.

Consequently, the fill material and transportation costs can beprohibitive depending on the location of the borrow source for thehigh-quality soil.

One possible solution in such cases is to use locally availablesoils as construction materials because they would require signif-icantly less material transportation, fuel consumption and gener-ated pollution as compared to using high-quality offsite soils. It hasbeen estimated that fuel costs could constitute as much as 20% ofthe total transportation costs of high-quality soils (Ou et al., 1982).

However, locally available soils for the construction of reinforcedslopes in many parts of the world are of marginal quality (e.g. soilswith more than 15% fines). Geosynthetic reinforcement is a well-established and cost effective technology for the construction andrepair of slopes and embankments (e.g. Berg et al., 2009). Forinstance, it has been reported that reinforcing marginal soils couldhelp reduce the cost of fill material by as much as 60% (Keller, 1995).However, proper drainage and adequate soil-reinforcement interfacestrength are essential elements for reinforced soil structures builtwith marginal soils in order to provide safe and satisfactory perfor-mance during their service life. Mechanical response of marginalsoils and that of their interface with geosynthetic reinforcement arecomplex and may include strain softening, excessive deformationand loss of strength as a result of wetting (Zheng et al., 2013). Loss of

mailto:[email protected]



http://crossmark.crossref.org/dialog/?doi=10.1016/j.geotexmem.2014.01.005&domain=pdf

www.sciencedirect.com/science/journal/02661144

http://www.elsevier.com/locate/geotexmem

http://dx.doi.org/10.1016/j.geotexmem.2014.01.005



Fig. 1. Failed slope of a highway embankment in Chickasha, OK. Note: the height andinclination angle of the slope are z12 m and 17�, respectively.

D. Esmaili et al. / Geotextiles and Geomembranes 42 (2014) 139e153140

strength due to increase in the GWC can be especially crucial at thesoil-reinforcement interface because depending on the type of geo-synthetic employed (e.g. geotextiles) this interface can act as a po-tential slip plane. This can be understood by noting the fact thatinterface friction coefficient values are typically less than unity formany types of geotextiles (e.g. Koerner, 2005).

Moreover, failure of reinforced earthen structures that are builtwith marginal soils and lack a proper drainage system may simplyoccur in the form of significant wetting-induced deformations asopposed to complete collapse. As a result, there have been docu-mented cases of serviceability problems and failures of thesestructures related to the use of marginal soils without adequatecare in their design and/or construction (e.g. Zornberg andMitchell,1994; Mitchell and Zornberg, 1995; Christopher et al., 1998;Koerner et al., 2005; Sandri, 2005).

Hamid and Miller (2009) studied the shearing behavior of anunsaturated low-plasticity fine-grained soil using a modified directshear test apparatus in which the matric suction of the soil spec-imen was controlled. Their results showed that the matric suctioncontributed to the peak shear strength of unsaturated interfacesbut did not significantly influence their post-peak shear strength.However, variations of the net normal stress affected both the peakand post-peak shear strength values.

Liu et al. (2009) carried out a series of direct shear tests to studythe interface shear strength of geogrids and geotextile embedded insand and gravel. The test results showed that the shear strength ofthe soil-geotextile interface was 0.7 and 0.85 of the soil shearstrength for Ottawa sand and gravel, respectively. Results alsoindicated that the shear strength of soil-geogrid interfaces wasgenerally higher than that of soil-geotextile interfaces. The soil-geogrid interface shear strength was found to vary between 0.89and 1.01 of the soil shear strength for the types of geogrids tested.

Anubhav (2010) conducted a series of direct shear tests toexamine the shear stress-displacement behavior of sand-geotextileinterfaces. The results indicated that the peak shear strength of theinterface between the sand and a coarse-textured geotextile wassignificantly higher (i.e. up to 35%) than that for the interface be-tween the sand and a fine-textured geotextile. The results alsoshowed that the shear displacement at peak shear strengthincreased with overburden pressure.

In this study, it is postulated that adequate internal drainagecapacity exists in the reinforced soil structure to prevent thedevelopment of positive pore water pressure in the soil. However,seasonal variations of the moisture content due to precipitation orsubsurface water infiltration could still result in significant changesin matric suction during the service life of the structure in theabsence of a proper global drainage system in the structure, or if forinstance, the existing drainage system is compromised as a result ofexcessive clogging. A conservative design approach for slopes and

embankments is to assume that the embankment soil is fullysaturated. However, this is not an ideal design approach for thefollowing reasons: the soil properties from tests carried out on fullysaturated soil samples (i.e. dry unit weight, cohesion and frictionangle) do not realistically represent actual field conditions. This isbecause the soil is never placed and compacted in a fully saturatedcondition during construction. In addition, since the magnitudes ofhydraulic conductivity in fine-grained unsaturated soils areextremely low, it is usually unlikely that a significant portion of theslope would ever become saturated even under extreme rainfallconditions. However, in addition to hydraulic conductivity, thedegree of saturation in unsaturated soils also depends on the hy-draulic gradient which could be significant due to matric suction.Furthermore, saturated soil samples in the laboratory cannot becompacted to the specified relative compaction (e.g. 95% ofmaximum dry unit weight) to represent field conditions. Therefore,their measured properties would underestimate the correspondingfield values. As a result, an embankment design using saturated soilproperties will be neither optimized nor realistic.

The focus of this paper is on pullout capacity of geotextilereinforcement in marginal soils which is an important designconsideration in internal stability of reinforced soil structures.Based on the above discussion, since RSS are typically constructedwith the soil compacted at or the vicinity of the optimum (gravi-metric) moisture content (OMC ¼ GWCopt), the mechanical prop-erties of the soil for the design of RSS need to be determined in thelaboratory at the corresponding GWC values. Therefore, a primaryobjective of the study is to quantify the magnitude of reduction inthe pullout capacity of geotextile reinforcement as a result of loss ofmatric suction in the unsaturatedmarginal soil due to wetting. Thiscan lead to excessive deformations and even failure of the rein-forced soil structure. However, the influence of the soil GWC on thereinforcement pullout capacity and the resulting factors of safetyagainst failure is not explicitly accounted for in the current designguidelines and provisions. In this study, a moisture reduction factor(MRF), denoted by m(u), is proposed to account for the pulloutresistance of geotextile reinforcement in the design of reinforcedsoil structures with marginal soils. The MRF value is a function ofthe soil GWC value (and hence of the soil suction), whichmakes thepredicated value of the pullout resistance more accurate and reli-able for design purposes.

It should be noted that due to the very low hydraulic conductivityof unsaturated marginal soils, measuring the change in the pulloutresistance of a significant size soil-geotextile specimen in drainedconditions (e.g. conforming to the ASTM D6706 test protocol) isextremely time consuming. Therefore, in this study themarginal soilin the pullout tests described in this paperwere placed at prescribedGWC values ranging from OMC-2% to OMCþ2% to determine thecorresponding MRF values. Consequently, the MRF values in thisstudy do not exactly represent the reduction in the reinforcementpullout capacity as a result of wetting of a soil mass compacted at aninitial GWC and unit weight. However, they provide some quanti-tative data that could be used to estimate the magnitude of suchreduction for design purposes. Furthermore, these results are moredirectly applicable to determine the expected pullout capacity of thereinforcement in themarginal soil if placed at any GWC valuewithinthe range between OMC-2% and OMCþ2%.

2. Theory

2.1. Reinforcement pullout capacity in reinforced soil structures

For internal stability, the pullout resistance per unit width (Pr) ofthe reinforcement in reinforced soil structures is determined usingEq. (1). It is defined as the maximum tensile load required to

80

100

)

D. Esmaili et al. / Geotextiles and Geomembranes 42 (2014) 139e153 141

generate outward sliding of the reinforcement through the rein-forced soil mass (Elias et al., 2001; Berg et al., 2009):

Pr ¼ F*as0vLeC (1)

where:

Le: Embedment or adherence length in the resisting zone behindthe failure surfaceC: Reinforcement effective unit perimeter; e.g., C ¼ 2 for strips,grids, and sheetsLeC: Total surface area per unit weight of reinforcement in theresistive zone behind the failure surfaceF*¼ tandpeak: Pullout resistance (or friction-bearing-interaction)factordpeak: Equivalent peak friction angle of the soil-geosyntheticinterfacea: A scale effect correction factor to account for a nonlinearstress reduction over the embedded length of highly extensiblereinforcements0v: Effective vertical stress at the soil-reinforcement interface

Pullout tests are typically used to obtain the parameters a and F*

for different combinations of soil and reinforcement materials.Tests are typically performed on samples with a minimumembedded length of 600 mm as recommended in related guide-lines (e.g. ASTM D6706). The correction factor a depends on theextensibility and the length of the reinforcement. For extensiblesheets (i.e., geotextiles), the recommended value of a is 0.6 (Berget al., 2009). The parameter F* (especially in reinforcement typessuch as geogrids and welded wire mesh) includes both passive andfrictional resistance components (e.g., Palmeira, 2004; Abu-Farsakhet al., 2005; Berg et al., 2009).

It isworth noting that in the case of extensible reinforcement (e.g.geotextiles), a wetter soil results in smaller amount of reinforcementextension before pullout as compared to an otherwise identical soilin a drier condition. In Eq. (1), smaller reinforcement extensibilityresults in a larger value for a which could point toward a larger Prvalue for the case of a wetter soil, which is erroneous. However, thecombined term aF* is expected to always decreasewith the soil GWCvalue. Nevertheless, in this paper Eq. (1) is modified in the form:

Pr ¼ F*as0vLeCmðuÞ (2)

by introducing a moisture reduction factor (MRF), m(u), to explicitlyaccount for the influence of the soil GWC value on the soil-reinforcement pullout capacity as described in more detail in thispaper.

0

20

40

60

0.00010.0010.010.1110

Per

cent

fine

r by

wei

ght (

%

Diameter (mm)

Fig. 2. Gradation curve of Chickasha soil from sieve analysis and hydrometer tests. Thevertical broken line shows the location of the #200 sieve.

2.2. Extended Mohr-Coulomb failure envelope

According to a theory proposed by Fredlund et al. (1978) theshear strength of an unsaturated soil can be expressed using thefollowing two stress variables: the net normal stress, which is thedifference between the total stress and the pore air pressure(sn � ua) and the soil matric suction, which is the difference be-tween the pore air and porewater pressures (ua � uw). Based on theFredlund et al. (1978) approach, Miller and Hamid (2005) proposedthe following equation to determine the shear strength of unsat-urated soil-structure interfaces:

ss ¼ C0a þ ðsn � uaÞtan d0 þ ðua � uwÞtan db (3)

Where:

C0a: Adhesion intercept

sn: Normal stress on the interfaceua: Pore air pressured0: Angle of friction between the soil and the structure coun-

terface with respect to (sn � ua)uw: Pore water pressuredb: Angle of friction between the soil and the structure coun-terface with respect to matric suction (ua � uw)

In unsaturated soils, Mohr’s circles representing failure condi-tions correspond to a three-dimensional (3D) failure surface, wherethe shear stress (s) is the ordinate and the two stress variables(sn � ua) and (ua � uw) are the abscissas. The planar surface formedby these two stress variables is referred to as the extended Mohr-Coulomb failure envelope (Khoury et al., 2011; Hatami et al., 2013).

3. Laboratory tests

3.1. Materials

3.1.1. Soil classification and propertiesThe soil used for the pullout and interface shear tests in this

study was a lean clay found on US Route 62 in Chickasha, OK. Thegradation and fines content of the soil were determined using theASTM D422 (ASTM, 2007) and D1140 (ASTM, 2006) test methods.The test results are given in Fig. 2 and Table 1, which indicate thatthe soil is classified as CL and A-6, based on the USCS and AASHTOsoil classification systems, respectively.

The maximum dry unit weight and OMC value of the Chickashasoil from modified Proctor tests were determined asgdmax ¼ 17.3 kN/m3 and OMC ¼ 18%, respectively. A series of directshear tests (ASTM D3080, 2011) was carried out on the soil at threedifferent GWC values (i.e. OMC-2%, OMC and OMCþ2%) and at arate of 0.06 mm/min to determine its shear strength parameters(i.e. c0 and B0) with the results as shown in Table 1.

3.1.2. Geosynthetic reinforcementAwoven polypropylene (PP) geotextile (Mirafi HP370) was used

for the pullout and interface shear tests in this study. The me-chanical response of the geotextile was found as per the ASTM

Table 1Summary of Chickasha soil properties.

Value

Lean clayLiquid limit 38.0Plastic limit 20.0Plasticity index 18.0Specific gravity 2.75Gravel (%) 0.0Sand (%) 10.6Silt (%) 49.4Clay (%) 40.0Maximum dry unit weight, kN/m3 17.3OMC (%) 18.0Cohesion (kPa) at OMC-2%, OMC and OMCþ2% 42.6, 29.3, 20.4Friction angle (�) at OMC-2%, OMC and OMCþ2% 29.6, 27.3, 27.1

Mirafi HP370Ultimate tensile strength (kN/m) 40Tensile strength at 5% strain (kN/m) 20

Note: (1) Maximum dry unit weight and OMC were determined using ModifiedProctor compaction effort (ASTM D1557 (2012)); (2) Geotextile properties weredetermined as per ASTM D4595 (ASTM, 2009) test protocol.

0

5

10

15

20

25

30

100 1000 10000

GW

C (%

)

Soil suction (kPa)

Fig. 3. Soil water characteristic curve for Chickasha soil using WP4 potentiometer.Dashed lines show GWC at OMC-2% and OMCþ2%.


D4595 test protocol (ASTM, 2009) and was compared with themanufacturer’s data (Table 1).

3.2. Suction sensors

The soil suction was initially measured using several differentmethods and sensors including thermal conductivity sensors(Fredlund and Wong, 1989; Fredlund et al., 2000), filter paper tests(ASTM, 2010), PST-55 psychrometer sensors and the WP4 potenti-ometer. However, based on the range ofmeasured suction values andthe accuracy and reparability of the test results, it was concluded thatthe PST-55 psychrometer and theWP4 potentiometer were themostsuitable “in-situ” sensor (i.e. inside the pullout box) and “off-site”equipment to measure the soil suction, respectively. Hence, theseinstruments are described briefly in the following sections.

3.2.1. PST-55 psychrometersPST-55 is an in-situ psychrometer which can measure soil suc-

tion for values up to 5000 kPa. Under vapor equilibrium conditions,the water potential of the PST-55 porous cup is directly related tothe vapor pressure of the surrounding air. This means that the soilwater potential is determined by measuring the relative humidityof the chamber inside the porous cup (Campbell and Gardner,1971). PST-55 psychrometers are commonly used in geotechnicalresearch projects. These sensors can lose their factory calibrationover time. Therefore, in this study they were calibrated using a1000 mmol/kg NaCl solution before they were used in the pullouttests. An HR-33T data logger was used to read thewater potential ofthe NaCl solution samples, and an ice chest provided a controlledtemperature and moisture environment for the calibration of thesensors. The sensors were submerged in NaCl solutions and kept inthe ice chest for 2 h to reach equilibrium (Wescor Inc. 2001). Then,each sensor was connected to the data logger (one at a time) andthe voltage representing the water potential of the control NaClsolution was read in microvolts (mV).

3.2.2. WP4 potentiometersThe WP4 equipment consists of a sealed block chamber equip-

ped with a sample cup, a mirror, a dew point sensor, a temperaturesensor, an infrared thermometer and a fan. The soil sample is placedin the sample cup and brought to vapor equilibrium with the air inthe headspace of the sealed block chamber. At equilibrium, thewater potential of the air in the chamber is the same as the waterpotential or suction of the soil sample.

Seventeen (17) 40mm (diameter) by 6mm (height) disk-shapedWP4 samples of Chickasha soil were prepared at different GWCvalues at the same dry unit weight as that in the laboratory pullouttests. The WP4 samples were placed in sealed disposable cups.Before testing each soil sample using WP4, a salt solution of knownwater potential (i.e. 0.5 molal KCl in H2O) was used to calibrate theWP4 sensor. For each test, the sample was placed inside the WP4sample cup and was allowed to reach temperature equilibriumwith the equipment internal chamber. The magnitude of the soilsuction was recorded once the displayed reading stabilized at aconstant value. Fig. 3 shows the Soil-Water Characteristic Curve(SWCC) for the Chickasha soil as was obtained from the WP4 tests.Results shown in Fig. 3 indicate that the suction in Chickasha soilvaries between 300 kPa and 1200 kPa for the range of GWC valuesbetween OMC-2% and OMCþ2%. This range of soil suction isconsistent with the values reported in the literature for lean clay(e.g. Cardoso et al., 2007; Nam et al., 2009).

3.3. Large-scale pullout tests

A series of large-scale pullout tests were carried out on thewoven geotextile in Chickasha soil. The tests were carried out atthree different GWC values (Table 2). The differences in themagnitude of geotextile pullout resistance among these cases wereused to determine a moisture reduction factor (MRF), denoted bym(u) in Eq. (2), to account for the loss of reinforcement pulloutresistance due to an increase in the soil GWC value.

3.3.1. Test equipmentThe nominal dimensions of the large-scale pullout test box used

in this study (Fig. 4) are 1800 mm (L)� 900mm (W)� 750mm (H).The size of the box and its basic components, including metalsleeves at the front end exceed the minimum requirements of theASTM D6706 test standard (ASTM, 2013). The boundary effects inthe test were further minimized by lining the sidewalls of the testbox with plastic sheets to reduce sidewall friction, and placingStyrofoam panels in contact with the soil inside the box to provideadded compressibility at the front boundary. A surcharge assemblyincluding an airbag and reaction beams across the top of the soilsurface was used to apply overburden pressures up to 50 kPa on thesoil-reinforcement interface. The pullout load on the reinforcementspecimen was applied using a 90 kN high-precision servo-controlled hydraulic actuator. In the tests carried out in this study,only one half of the box length (i.e. 900 mm) was used.

Table 2Large-scale pullout test parameters.

Test information

Soil Chickasha soilGeosynthetic reinforcement Mirafi HP370, woven PPOverburden pressure, kPa 10, 20, 50Gravitational water content (GWC) OMC-2%, OMC, OMCþ2%


3.3.2. InstrumentationDifferent instruments were used in the pullout tests to measure

the movement of geotextile reinforcement and the matric suctionin the soil. Deformation of the geotextile reinforcement wasmeasured using four (4) wire-line extensometers attached to

Vertical columns to limit horizontal deformation of test box

Geotextile reinforcement attached to actuator

Styrofoam panels to fill space above soil specimen

Air tube to fill up air bag

Fig. 4. One of the pullout test boxes at

Pullout Direction

Fig. 5. (a) Wire-line extensometers attached to the geotextile reinforcement. L is distance fpullout test box.

different locations along its length (Fig. 5a). A Geokon Earth Pres-sure Cell (EPC) was used to verify the magnitude of the overburdenpressure on the soil-geotextile interface from the airbag that wasplaced on the top of the soil (Fig. 5b).

PST-55 psychrometers were placed 25 mm above and below thesoil-geotextile interface to measure and monitor the soil suctionnear the soil-reinforcement interface. The locations of WP4 sam-ples and those of the in-situ PST-55 psychrometers are shown inFig. 6 using white and black circles, respectively.

3.3.3. Test procedureThe soil was air dried and its larger clumps were broken into

smaller pieces. It was then grinded into smaller pieces and passed

Plastic sheets to minimize sidewall friction

900 mm

1800 mm750 mm

the OU Geosynthetics Laboratory.

(a)

(b)

rom the front end of the geotextile; (b) Earth pressure cell on the top of the soil in the


through a #4 sieve using a soil processor. Afterward, the soil wasmixed with water to reach the desired GWC value for each test. Thewet soil was stored in sealed buckets for at least 24 h to reachmoisture equilibrium. The soil GWC value in each bucket wasmeasured by testing one soil sample using the oven drying method.The above procedure was repeated for every pullout test.

After the soil was processed and was ready to be placed in thetest box, the pullout box was lined with plastic sheets to preservethe soil water content and tominimize the friction between the soiland the sidewalls during each test. Next, the soil was placed andcompacted in the test box in nine 50 mm lifts. The thickness of thesoil layers below and above the geotextile reinforcement was230mmwhich exceeded the minimum depth of 150 mm accordingto the ASTM D6706 test protocol (ASTM, 2010a,b). The soil wascompacted to 95% of its maximum dry unit weight (i.e.gd ¼ 16.4 kN/m3). The instrumented geotextile was placed at themid-height of the box. The pullout box containing compacted soil atits target GWC value was sealed with plastic sheets on the top. Thesoil was left for at least 24 h until the psychrometer sensors reachedequilibrium with their surrounding soil and for three to fouradditional days until the soil GWC inside the pullout test box sta-bilized. In all pullout tests, rectangular Styrofoam blocks with di-mensions 900 mm (L), 228 mm (H) and 140 mm (T) were placed infront of the soil specimen above and below the 200mm-widemetalsleeves, which helped further minimize the influence of the frontboundary condition on the soil-geotextile interface. The pulloutforce was applied on the geotextile reinforcement at a targetdisplacement rate of 1 mm/min according to the ASTM D6706 testprotocol.

3.4. Results of large-scale tests and discussion

3.4.1. Water content and suctionFigs. 7 and 8 show distributions of the soil GWC and suction in

each layer for the large-scale pullout tests carried out at differentGWC and subjected to 50 kPa overburden pressure. The mean andCoefficient of Variation (COV) values for these parameters werecalculated for the fifth layer (lift) in the pullout box (i.e. for the soillayer in contact with the geotextile reinforcement) to examine theproximity of their as-placed and target values. Table 3 shows themean and COV values for the GWC and suction in the fifth layer inlarge-scale pullout tests. The accuracy of the soil suction valuesfrom the PST-55 psychrometers was also examined by comparingthem with the readings from the WP4 potentiometer as shown in

Fig. 6. Schematic diagram of the large-scale pullout test box setup (not to scale).Notes: (1) Black and white circles represent the locations of PST-55 sensors and soilsamples for the WP4 sensor, respectively; (2) The distance between the sensors and theinterface is 25 mm; (3) The sleeves above and below the geotextile layer are 200 mmwide.

Table 4. The GWC and suction COV values for all test cases as givenin Figs. 7 and 8 and Table 3 are overall reasonable and indicate thatthe soil moisture condition was fairly uniform and consistentthroughout the large-scale test models.

3.4.2. Reinforcement strain and interface strengthFig. 9 shows the strain distributions over the length of geotextile

reinforcement at maximum pullout force based on the extensom-eter results for the test case under 50 kPa overburden pressure. Thestrain near the front end of the geotextile reinforcement wascalculated by subtracting the displacements measured at thelocation of Extensometer 1 from those measured at the front end ofthe geotextile exiting the soil. The displacements at the front endwere determined by subtracting the calculated elongation of the in-air portion of the geotextile specimen from the actuator displace-ment. Results in Fig. 9 indicate that strains in the geotextile rein-forcement are greater at higher overburden pressures and lowerGWC values (i.e. higher soil suction). Fig. 10 shows the pullout testdata and results of interface shear strength (s) for the Chickasha soilfor different magnitudes of GWC and overburden pressure. The svalues were calculated by dividing the pullout force for each case bythe in-soil area of the geotextile specimen (i.e. two times the geo-textile area). The target GWC values in the pullout tests includeOMC-2% (16%), OMC (18%) and OMCþ2% (20%) (see Table 1). InFig. 8a-c, the measured pullout force is plotted as a function of theactuator displacement. Results shown in Fig. 10 quantify the in-crease in the reinforcement pullout resistance in Chickasha soilwith overburden pressure for a given GWC value. It should be notedthat for the test at OMC-2% subjected to 50 kPa overburden pres-sure (Fig. 10a) it was found that the geotextile had been rupturedbefore pullout. Therefore, the pullout force at failure was estimatedusing the trends in the corresponding test data at OMC andOMCþ2%. As expected, increasing suction led to a higher maximumreinforcement pullout resistance in otherwise identical test speci-mens (Fig. 10d and e). Results shown in Fig. 10d indicate thatapparent adhesion increases at lower GWC due to higher suction.This observation is consistent with those reported by Khoury et al.(2011) from suction-controlled interface testing of fine-grained soilspecimens.

Results shown in Fig. 10a-d represent the frontal planes ofextended Mohr-Coulomb failure envelopes for the soil-geotextileinterface at different GWC and suction values. These failure enve-lopes can be considered to be practically linear for all GWC casesexamined. The interface strength results, i.e. the values for the slope(tan d0) and the intercept (ca) of the failure envelopes on thesefrontal planes are summarized in Table 5. Abu-Farsakh et al. (2007)studied the effect of the GWC on the interaction between threecohesive soils and a woven geotextile reinforcement material. Theyfound that an increase in the molding GWC of the soil from 24% to33% caused 43% reduction in the interface shear resistance. Thedata summarized in Table 5 are overall consistent with Abu-Farsakhet al.’s observations. For instance, the pullout resistance (Pr) atOMCþ2% is between 17% and 35% lower than the correspondingvalue at OMC-2% depending on the overburden pressure. A smallerconfining pressure resulted in a greater reduction in pullout resis-tance for a given increase in the soil GWC value.

Results in Fig. 10e show the failure envelopes of the three-dimensional extended Mohr-Coulomb failure surface on thelateral plane for the soil-geotextile interface as a function of the soilsuction. The line intercept and slope represent the effective adhe-sion at zero overburden pressure (sn ¼ 0 kPa) and interface frictionangle with respect to suction (db), respectively. The data shown inFig. 10e indicate that the interface friction angle with respect tosuction for the Chickasha soil-geotextile tested is negligible (it isless than 1�; note the significantly different scales of the horizontal

(a)

4

8

12

16

20

24

1 2 3 4 5 6 7 8 9 10

GW

C (%

)

Soil lift number in pullout box

GWC for each bucket

Mean GWC per layer

(b)

4

8

12

16

20

24

1 2 3 4 5 6 7 8 9 10

GW

C (%

)


GWC for each bucket

Mean GWC per layer

18%

v: 50 kPa

COV: 1.5%

(c)

4

8

12

16

20

24

1 2 3 4 5 6 7 8 9 10

GW

C (%

)


GWC for each bucket

Mean GWC per layer

20.1%

v: 50 kPa

COV: 1.7%

Fig. 7. Distributions of the soil GWC with depth in the pullout box for different pullouttest cases. Notes: (1) One soil sample was taken from each bucket to test its GWC value;(2) The number of soil samples from each soil lift in the pullout box is given in Table 5(caption); (3) The horizontal line indicates the target GWC for each test case; (4) Thevertical dashed line shows the location of the soil-geotextile interfaces; (5) The meanand COV values reported in the legends are calculated for the fifth layer (i.e. soil-geotextile interface) data only.


and vertical axes in the figure). These results indicate that as theoverburden pressure increases, interface adhesion and conse-quently, the interface shear strength increases. The extendedMohr-Coulomb envelope in Fig. 11 shows the variation of the interfaceshear strength (s) with the values of soil suction and overburden

(a)

0

400

800

1200

1600

1 2 3 4 5 6 7 8 9 10

Tota

l suc

tion

(kP

a)Soil lift number in pullout box

Soil suction distrinution per layer

Mean soil suction per layer

OMC-2%

v : 50 kPa

COV: 3.4 125 kPa

(b)

0

400

800

1200

1600

1 2 3 4 5 6 7 8 9 10

Tota

l suc

tion

(kP

a)


Soil suction distribution per layer


OMC

σv : 50 kPa

COV: 8.2%, Ψ = 590 kPa

0

400

800

1200

1600

1 2 3 4 5 6 7 8 9 10

Tota

l suc

tion

(kP

a)


Soil suction distribution per layer


OMC+2%

v : 50 kPa

COV: 11 298 kPa

(c)

Fig. 8. Distributions of the soil suction with depth in the pullout box from WP4 atdifferent GWC. The number of soil samples from each soil lift in the pullout box isreported in Table 5.

Table 3Mean and COV values for the GWC and suction in the fifth layer (in contact withgeotextile) in large-scale pullout tests.

Target u (%) sn (kPa) Mean J (kPa) COV(J) (%) Mean u (%) COV(u) (%)

16 (OMC-2%) 10 1236 7.3 15.7 1.520 1196 5.9 15.8 1.150 1125 3.4 16.0 0.7

18 (OMC) 10 513 5.7 18.5 0.920 570 6.7 18.1 1.150 590 8.2 18.0 1.5

20 (OMCþ2%) 10 304 11.2 20.1 1.820 352 9.9 19.6 1.650 298 11.0 20.1 1.7

0

5

10

15

20

25

30

0 100 200 300 400 500 600S

train

(%)

Distance on the geotextile from front end of the soil (mm)

OMC-2%, 10 kPa

OMC, 10 kPa

OMC+2%, 10 kPa

Locations where potentiometers were attached to the geotextile

Geotextile reinforcement

0

5

10

15

20

25

30

0 100 200 300 400 500 600

Stra

in (%

)


OMC-2%, 20 kPa

OMC, 20 kPa

OMC+2%, 20 kPa

30OMC-2%, 50 kPa

(a)

(b)


pressure at the soil-reinforcement interface. Taken together, theresults based on the description of soil shear strength using twostress state variables (i.e. soil suction and net normal stress) aspresented in Figs. 10 and 11 and Table 5 are in good agreement withthose reported by Hatami et al. (2010a) and Khoury et al. (2011) onother marginal soils.

Table 6 shows aF* (Eq. (1)) values calculated from all large-scalepullout tests in Chickasha soil. Example calculations for a accordingto the FHWA guidelines (Berg et al., 2009) are shown in Fig. 12.Based on the intersection of the horizontal asymptote with the y-axis, the design value for a from pullout tests in Chickasha soil wasfound to be 0.5 (Fig. 12e). This value of a indicates a fairly extensiblegeotextile material and a linear strain distribution along its length.It is also comparable to the value reported by Hatami et al. (2010a)for the same geotextile material tested in Minco silt (i.e. a ¼ 0.59)and the value a ¼ 0.6 recommended by FHWA for geotextiles (Berget al., 2009). F* values were calculated using Eq. (1).

3.5. Small-scale tests

In addition to large-scale pullout tests, a series of small-scalepullout and interface shear tests were performed on the samesoil that was used in the large-scale pullout tests. In addition, thesetests were carried out at the same soil GWC, unit weight (i.e. 95% ofmaximum dry unit weight from modified Proctor tests) and over-burden pressure magnitudes as those in the large-scale pullouttests (Table 7). One main advantage of small-scale pullout tests isthat once they are calibrated against the large-scale tests, they canbe carried out at significantly larger numbers to test a large varietyof soil types, GWC values and overburden pressures. This will helpto develop a better understanding of the influence of the soil GWCand matric suction on marginal soil-geotextile interfaces using a

Table 4Comparison of suction values in Chickasha soil as measured using psychrometers(in-situ) and WP4 (offsite equipment).

Target u (%) sn (kPa) Mean J (kPa) WP4a Mean J (kPa) PST-55b

16 (OM-2%) 10 1243 92120 1200 90650 1135 910

18 (OMC) 10 520 48820 570 52050 580 543

20 (OMCþ2%) 10 311 30320 360 33550 282 310

a Mean values were calculated using four undisturbed samples from the fifthlayer (in contact with geotextile) for each pullout test as shown in Fig. 6.

b Mean values were determined using three PST-55 psychrometers placed in thefifth soil layer (Fig. 6).

multi-scale laboratory testing approach. It is worth noting that boththe soil particle size and the asperities of the geotextile reinforce-ment are orders of magnitude smaller than the dimensions of thesmall-scale text box. Therefore, the test results are not believed tobe negatively impacted by scale effects. However, proper boundary

0

5

10

15

20

25

0 100 200 300 400 500 600

Stra

in (%

)


OMC, 50 kPa

OMC+2%, 50 kPa

(c)

Fig. 9. Axial strain distributions in geotextile reinforcement subjected to pullout loadfrom large-scale pullout tests on Chickasha soil at different GWC values.

(c)(a)

(b)

(e)

(d)

l

llll

l l

Fig. 10. Pullout test data and interface strength results from large-scale pullout tests for Chickasha soil at different GWC values: (a)-(c) Load-displacement data; (d) Failure en-velopes for the soil-geotextile interface on frontal plane; (e) Failure envelopes for soil-geotextile interface on lateral plane. Note: in (a), dashed line indicates the estimated pulloutfailure.


conditions are required to help calibrate these test results againstthose from large-scale pullout tests.

The small-scale pullout and interface shear tests were carriedout using a direct shear testing (DST) machine shown in Fig. 13. Thesoil in both tests was placed in a 60 mm� 60 mm square test cell of

Table 5Interface strength properties from large-scale pullout tests in Chickasha soil.

Target u (%) sn (kPa) Mean u (%)a Mean J (kPa)b

16 (OMC-2%) 10 16.0 115320 16.0 115150 16.0 1135

18 (OMC) 10 18.3 55020 18.2 56650 18.1 576

20 (OMCþ2%) 10 20.3 28620 20.0 31250 20.2 290

a Mean values were calculated using 45 GWC samples for each pullout test (5 sampleb Mean values were determined from SWCC for Chickasha soil (Fig. 3) based on GWC

the DST machine. Two rectangular blocks of Styrofoam with di-mensions 60 mm (L), 12 mm (H) and 9 mm (T) were used in thesmall-scale pullout tests in front of the soil specimen to provide acompressible boundary condition similar to that in the large-scalepullout box. The Styrofoam blocks were placed in the upper and

Pr (kN/m) smax (kPa) d0(�) Ca (kPa)

29.6 24.3 17.3 21.634.8 28.545.2 37.1

24.8 20.3 15.4 18.129.7 24.438.7 31.7

19.1 15.7 14.7 15.328.8 23.633.8 27.7

s from each of the nine 2-inch soil lifts).values determined for each test (i.e. 45 data points).

Fig. 11. Extended Mohr-Coulomb envelope from large-scale pullout tests.


lower halves of the test cell in front of the soil. A 20 mm(W) � 40 mm (L) geotextile specimenwas used in each pullout test.The linear scale factor between the small-scale and large-scalepullout tests was 1:15. In the small-scale pullout tests, the geo-textile specimen was pulled out of a fixed test cell filled withChickasha soil at a speed of 0.06 mm/min (i.e. 1/15 of 1 mm/minnominal rate at large scale). In the interface shear tests, the lowerbox of the DST machine was pushed horizontally at a speed of0.06 mm/min to apply a shear load on the soil-geotextile interface.These two tests are described in more detail in the followingsections.

3.5.1. Small-scale pullout testsThe soil in the small-scale pullout tests was prepared using the

same process as was followed for the large-scale tests: The clay wasfirst processed, then passed through a #4 sieve (i.e. 4.75 mmaperture size) and mixed with water to reach the target GWC valuewhich was measured using the oven drying method preceding andfollowing each test. The soil was placed in the bottom half of thetest cell in four lifts at the target GWC value and each lift wascompacted to the final thickness of 3 mm. The geotextile specimenwas attached to a custom-made clampmounted on the test box andwas embedded 40 mm inside the test cell. A U-shaped metal spacer(open on the front side) was used to maintain a gap within thepullout slot to prevent any frictional contacts within the test cellframe during the pullout process. The top half of the box was filledwith 4 more layers of the soil, each compacted to 3 mm thickness.

Table 6Calculated values of aF* from large-scale pullout tests in Chickasha soil.

Target u (%) sn (kPa) Pr (kN/m) smax (kPa) aF*

16 (OMC-2%) 10 29.6 24.3 2.4320 34.8 28.5 1.4350 45.2 37.1 0.74

18 (OMC) 10 24.8 20.3 2.0420 29.7 24.4 1.2250 38.7 31.7 0.64

20 (OMCþ2%) 10 19.1 15.7 1.5720 28.8 23.6 1.1850 33.8 27.7 0.56

3.5.2. ResultsFig. 14 shows the plots of pullout force versus actuator

displacement for the small-scale pullout tests in Chickasha soil.Soil-geotextile interface strength properties obtained from thesmall-scale tests are summarized in Table 8.

The pullout test results given in Table 8 and Fig. 14 show a clearinfluence of the soil overburden pressure and GWC on the soil-geotextile interface strength and pullout resistance. It is observedthat the pullout force increases with overburden pressure. Theinterface adhesion contributing to the geotextile pullout resistancedecreases by 31% as the soil GWC increases from OMC-2% toOMCþ2%. The interface friction angle also decreases by 40% fromOMC-2% to OMCþ2%. These results are consistent with those ob-tained by the authors in a previous study on Minco silt (Hatamiet al., 2010a,b). Results in Figs. 14d and 8d indicate that the inter-face adhesion from both small-scale and large-scale pullout testsdepends on the soil GWC and it is consistently larger for greater soilsuction values (i.e. lower GWC). These results also indicate that themagnitudes of soil-geotextile interface adhesion from small-scalepullout tests are greater than those from the corresponding large-scale tests. This could be attributed to the smaller size andgreater boundary effects in the small-scale tests. Consequently, acalibration (or scale) factor needs to be determined and applied tothe small-scale test results before they can be used for practicalapplications.

The data in Fig. 14e indicate that the interface shear strengthincreases with overburden pressure as a result of increase ininterface adhesion ðC0

aÞ. The results shown in Fig. 14e indicate thatthe interface friction angle with respect to suction (db) for theChickasha soil-geotextile tested is less than 2�.

3.5.3. Interface shear testsA series of interface shear tests was carried out on the woven

geotextile with Chickasha soil at different GWC values to determinethe strength properties of the Chickasha soil-geotextile interface. Ageotextile specimen was attached to a 60 mm � 60 mm aluminumpanel and was placed on the top of a stack of aluminum panels inthe bottom half of the test cell in the DST machine. The top half ofthe test cell was filled with four 3 mm-thick compacted layers ofChickasha soil, similar to the small-scale pullout tests.

3.5.4. ResultsFig. 15 shows theMohr-Coulomb envelopes from interface shear

tests at different GWC values. The results show that the soil-geo-textile interface strength increases consistently with the over-burden pressure and with the soil matric suction. According toFig. 15a, the interface friction angle was found to decrease by 15%from OMC-2% to OMCþ2%. Results in Fig. 15b indicate that theinterface friction angle with respect to matric suction on the lateralplane is less than 1�, which is consistent with the data from large-scale pullout tests (Fig. 10e). Fig. 16 shows the extended Mohr-Coulomb envelopes from small-scale pullout and interface sheartests. The plots of Mohr-Coulomb envelopes in Fig. 16 show that theadhesion values calculated from pullout tests are greater whichcould be attributed to this fact that, in pullout tests, as opposed tothe interface shear tests, the geotextile is stretched during the test.This could result in the enlargement of the geotextile openingswhich, in turn, could allow the fine-grained soil to penetrate intothe plane of the geotextile. Similar to geogrids but at a smaller scale,the soil within the openings of the geotextile subjected to over-burden pressure could exhibit some passive resistance against thepullout force which could be responsible for the larger adhesionintercept that is observed for the pullout test results as compared tothe interface shear data. Table 9 and Fig. 17 summarize the datafrom all laboratory tests carried out in this study.

(a)

(b)

(c)

(d)

(e)

llll

l

l

l l

ll

l

l

l

Fig. 12. Calculation of pullout parameters for Mirafi HP370 geotextile reinforcement from large-scale pullout test data in Chickasha soil at OMCþ2% subjected to 50 kPa overburdenpressure: (a) Pullout force versus actuator displacement; (b) Pullout force versus extensometer displacement; (c)-(e) Procedure to determine F* and a using large-scale pullout tests;Note: in (b), EX1 and EX4 are near the front end and the tail end, respectively. In (c), solid and dashed lines indicate actual and interpolated data, respectively.


4. Moisture reduction factor, m(u)

Fig. 18 shows the variations of m(u) for the Chickasha soil-woven geotextile interface as a function of the soil GWC at

Table 7Small-scale pullout and interface test parameters.

Test information Chickasha soil

Type of small-scale test Pullout, Interface shearGeosynthetic reinforcement Mirafi HP370, woven PPSoil specimen dimensions 60 mm (W) � 60 mm (L) � 24 mm (H),

& 60 mm (W) � 60 mm (L) � 12 mm (H)Geosynthetic reinforcement

dimensions20.3 mm (W) � 40.6 mm (L)& 60 mm (W) � 60 mm (L)

Overburden pressure, kPa 10, 20, 50Gravitational water

content (GWC)OMC-2%, OMC, OMCþ2%

different overburden pressures from all three categories of testscarried out in this study. In the calculation of m(u), the interfacestrength at u ¼ OMC-2% is taken as the reference value (Hatamiet al., 2010a,b).

Results shown in Fig. 18 indicate that construction of reinforcedsoil slopes and embankments on the wet side of OMC or wetting ofthe soil-geotextile interface during construction or service life ofthe reinforced soil structure (as compared to e.g., the case of OMC-2%) could result in considerably lower pullout resistance of thegeotextile reinforcement. The calculated amounts of reduction ininterface strength from OMC-2% to OMCþ2% as obtained from thelarge-scale and small-scale test data are between 17% and 42%depending on the test cases and overburden pressures applied onthe specimens. Results shown in Fig. 18 indicate that the variationof m(u) with the soil GWC could be approximated as linear forpractical purposes for the range of GWC values examined in thisstudy.

Fig. 13. Small-scale pullout tests in Chickasha soil using a DST machine.

(a) (c)

(b) (d)

(e)

llll

ll

l

l

l

Fig. 14. Pullout test data and interface strength results from small-scale tests for Chickasha soil and comparison of failure envelopes for soil-geotextile interface at different GWCvalues: (a)-(c) Load-displacement data; (d) Failure envelopes for soil-geotextile interface on frontal plane; (e) Failure envelopes for soil-geotextile interface on lateral plane. Note:Suction values were calculated from the SWCC.


Table 8Interface strength properties from small-scale pullout tests.

Target u (%) sn ( kPa) u (%) J (kPa) smax ( kPa) d0 (�) Ca (kPa)

16 (OMC-2%) 10 16.3 1036 56.5 30.2 56.720 15.9 1164 76.350 16.0 1131 83.8

18 (OMC) 10 17.8 633 46.8 23.8 43.520 17.9 613 53.950 18.3 538 65.2

20 (OMCþ2%) 10 20.0 310 40.6 18.1 38.920 19.9 320 47.550 19.8 331 54.7

Note: Suction values were calculated from the SWCC.

Fig. 16. Extended Mohr-Coulomb envelope from small-scale pullout and interfaceshear tests.


5. Conclusions

The primary objective of this study was to develop a moisturereduction factor (MRF), denoted as m(u), for the pullout resistanceof geotextile reinforcement for the design of reinforced soilstructures with marginal soils. Based on the results of this study,the current FHWA design equation for pullout capacity of

(a)

0

20

40

60

80

100

0 20 40 60

She

ar s

treng

th (k

Pa)

Normal stress (kPa)

OMC-2%

OMC

OMC+2%δ' = 21ºca = 34.8 kPa

δ' = 20ºca = 27 kPa

δ' = 18ºca = 21.2 kPa

(b)

0

20

40

60

80

100

0 500 1000 1500

She

ar s

treng

th (k

Pa)

Suction (kPa)

10 kPa20 kPa50 kPa

δb = 1ºca

' = 32.3 kPa

δb = 0.9ºca

' = 29.0 kPa

δb = 0.8ºca

'= 18.6 kPa

Fig. 15. Mohr-Coulomb envelopes for Chickasha soil-geotextile interface from interfaceshear tests: (a) Envelopes on the frontal plane; (b) Envelopes on the lateral plane. Note:Suction values were calculated from the SWCC.

Table 9Summary of the results from all laboratory tests performed in this study.

Type of test Large-scalepullout

Small-scalepullout

Interface shear

Target u (%) sn (kPa) d0 (�) Ca (kPa) d0 (�) Ca (kPa) d0 (�) Ca (kPa)

16 (OMC-2%) 10 17.3 21.6 30.2 56.7 21.0 34.82050

18 (OMC) 10 15.4 18.1 23.8 43.5 20.0 27.02050

20 (OMCþ2%) 10 14.7 15.3 18.1 38.9 18.0 21.22050

geotextile reinforcement wasmodified to explicitly account for theinfluence of the marginal soil gravitational water content (GWC)on the soil-reinforcement interface strength. The m(u) values weredetermined through a series of multi-scale pullout and interfacetests in a silty clay in the laboratory.

The test results indicated that the change in the soil suction as aresult of variation in its GWC value could have a significant

0

10

20

30

40

0

20

40

60

12 16 20 24

Inte

rface

fric

tion

angl

e (°

)

Inte

rface

adh

esio

n(k

Pa)

Normal stress (kPa)

Interface adhesion

Interface friction angle

Large-scale pullout tests

Small-scale interface shear tests

Small-scale pullout tests

Fig. 17. Comparison of large-scale and small-scale pullout and interface test data.

(a)

μ10 (ω) = -0.0885ω + 2.42μ20 (ω) = -0.0430ω + 1.67μ20 (ω) = -0.0633ω + 2.01

Large-scale pullout tests, 10 kPa



(b)

μ10 (ω) = -0.0704ω + 2.12μ20 (ω) = -0.0944ω + 2.47μ50 (ω) = -0.0868ω + 2.37

0.0

0.3

0.6

0.9

1.2

14 16 18 20 22 24

μ (ω

)

0.0

0.3

0.6

0.9

1.2

μ (ω

)

ωcompaction (%)

14 16 18 20 22 24ωcompaction (%)

Small-scale pullout tests, 10 kPa



(c)

μ10 (ω) = -0.105ω + 2.68μ20 (ω) = -0.0758ω + 2.20μ50 (ω) = -0.0818ω + 2.31

0.0

0.3

0.6

0.9

1.2

14 16 18 20 22 24

μ (ω

)

ωcompaction (%)

Small-scale interface tests, 10 kPa



Fig. 18. Moisture reduction factor for the woven geotextile in Chickasha soil: (a) Large-scale pullout tests; (b) Small-scale pullout tests; (c) Small-scale interface shear tests.


influence on the soil-geotextile interface strength. The shearstrength of the soil-geotextile interface at OMCþ2% was between17% and 42% lower than that of an otherwise identical interface atOMC-2%. The results reported in this paper were obtained usingone combination of marginal soil and woven geotextile only.Similar tests and analyses are underway to develop MRF values for

other combinations of marginal soil and geotextile reinforcementfor field applications.

It should be noted that the behavior of a marginal soil which isinitially placed and compacted at OMC-2% (with a flocculatedstructure) and is wetted to OMCþ2% is different from the samemarginal soil placed and compacted at OMCþ2% (with a dispersedstructure; e.g. Lambe, 1958; Fredlund and Rahardo, 1993). Conse-quently, the values of m(u) for a soil initially compacted at a lowerGWC and then wetted to a larger target GWC value are expected tobe somewhat different from those given in Fig.18. Nevertheless, thebroader conclusions of the study and their implications to thedesign of SRW walls with marginal soils are believed to remainvalid.

Acknowledgments

The authors would like to acknowledge the funding and supportfrom the Oklahoma Department of Transportation (ODOT), theOklahomaTransportation Center (OkTC) and TenCate Geosyntheticsfor the study reported in this paper. Contributions of Mr. MichaelSchmitz at the Fears Structural Laboratory and undergraduate stu-dents: Brandi Dittrich, John Tucker, Kyle Olson, Carlos Chang, ThaiDinh and Jesse Berdis in this project are also acknowledged.

References

Abu-Farsakh, M., Farrag, K., Almohd, I., Mohiuddin, A., 2005. Bearing and frictionalcontributions to the pullout capacity of geogrid reinforcements in cohesiveBackfill. In: Proceeding of Geo-frontiers. ASCE, Austin, TX, USA. In: http://ascelibrary.org/proceedings/resource/2/ascecp/166/40787/22_1?isAuthorized¼no.

Abu-Farsakh, M., Coronel, J., Tao, M., 2007. Effect of soil moisture content and dryDensity on cohesive Soil-Geosynthetic interactions using large direct sheartests. J. Mater. Civ. Eng. ASCE 19 (7), 540e549.

Anubhav, P.K., 2010. Modeling of soil-woven geotextile interface behavior fromdirect shear tests results. Geotext. Geomembr. 28, 403e408.

ASTM International, 2006. Standard Test Methods for Amount of Material in SoilsFiner than No. 200 (75-mm) Sieve. American Society for Testing and Materials,West Conshohochen, PA, USA.

ASTM International, 2007. Standard Test Method for Particle Size Analysis of Soils.American Society for Testing and Materials, West Conshohochen, PA, USA.

ASTM International, 2009. Standard Test Method for Tensile Properties of Geo-textiles by the Wide-Width Strip Method. American Society for Testing andMaterials, West Conshohochen, PA, USA.

ASTM International, 2010. Standard Test Method for Measurement of Soil Potential(Suction) Using Filter Paper. American Society for Testing and Materials, WestConshohochen, PA, USA.

ASTM International, 2011. Standard Test Method for Direct Shear Test of Soils underConsolidated Drained Conditions. American Society for Testing and Materials,West Conshohochen, PA, USA.

ASTM International, 2012. Standard Test Methods for Laboratory CompactionCharacteristics of Soil Using Modified Effort. American Society for Testing andMaterials, West Conshohochen, PA, USA.

ASTM International, 2013. Standard Test Method for Measuring GeosyntheticPullout Resistance in Soil. American Society for Testing and Materials, WestConshohochen, PA, USA.

Berg, R.B., Christopher, B.R., Naresh, C.S., 2009. Design and Construction of Me-chanically Stabilized Earth Walls and Reinforced Soil Slopes. Federal HighwayAdministration, Washington, DC, USA. FHWA-NHI-10-024.

Campbell, G.S., Gardner, W.H., 1971. Psychometric measurement of soil water po-tential: temperature and bulk density effects. Proc. Soil. Sci. Soc. Am. 35, 8e11.

Cardoso, R., Romero, E., Lima, A., Ferrari, A., 2007. A comparative study of soilsuction measurement using two different high-range psychrometers. Exp.Unsatur. Soil. Mech. 112, 79e93.

Christopher, B.R., Zornberg, J.G., Mitchell, J.K., 1998. Design guidance for reinforcedsoil structures with marginal soil backfills. In: Proceedings of the 6th Interna-tional Conference on Geosynthetics, Atlanta, GA, USA, 1, pp. 797e804.

Elias, V., Christopher, B.R., Berg, R.R., 2001. Mechanically Stabilized Earth Walls andReinforced Soil Slopes-design and Construction Guidelines. Federal HighwayAdministration, Washington, DC, USA. FHWA-NHI-00-043.

Fredlund, D.G., Morgenstern, N.R., Widger, R.A., 1978. The shear strength of un-saturated soils. Can. Geotech. J. 15 (3), 313e321.

Fredlund, D.G., Rahardo, H., 1993. Soil Mechanics for Unsaturated Soils, first ed. JohnWiley & Sons, Inc., Hoboken, NJ.

Fredlund, D.G., Shuai, F., Freng, M., 2000. Use of new thermal conductivity sensorfor laboratory suction. In: Rahardjo, H., Toll, D.G., Leong, E.C. (Eds.), Proceedings

http://ascelibrary.org/proceedings/resource/2/ascecp/166/40787/22_1?isAuthorized=no




http://refhub.elsevier.com/S0266-1144(14)00012-0/sref2





















http://refhub.elsevier.com/S0266-1144(14)00012-0/sref39a































of the Asian Conference on Unsaturated Soils, Singapore, ISBN 90-5809-139-2,pp. 275e280.

Fredlund, D.G., Wong, D.H., 1989. Calibration of thermal conductivity sensors formeasuring soil suction. Geotechnical Test. J. 12 (3), 188e194.

Hamid, T.B., Miller, G.A., 2009. Shear strength of unsaturated soil interfaces. Can.Geotech. J. 46 (5), 595e606.

Hatami, K., Miller, G.A., Garcia, L.M., 2010a. Use of MSE Technology to StabilizeHighway Embankments and Slopes in Oklahoma. Oklahoma Department ofTransportation, Tulsa, OK, USA. Final Report OTC REOS7-1-19.

Hatami, K., Garcia, L.M., Miller, G.A., 2010b. Influence of moisture content on thepullout capacity of geotextile reinforcement in marginal soils. In: 61st HighwayGeology Symposium, Oklahoma City, OK, USA.

Hatami, K., Garcia, L.M., Miller, G.A., 2011. A moisture reduction factor for pulloutresistance of geotextile reinforcement in marginal soils. In: Geo-Frontiers,Dallas, TX, ASCE Special Publication No. 211.

Hatami, K., Granados, J.E., Esmaili, D., Miller, G.A., 2013. Influence of gravimetricwater content on geotextile reinforcement pullout resistance in MSE wallswith marginal soils. Transp. Res. Rec. (2363), 66e74. http://dx.doi.org/10.3141/2363-08.

Keller, G.R., 1995. Experiences with Mechanically stabilized structures and nativesoil backfill. Transp. Res. Rec. (1474), 30e38.

Khoury, C.N., Miller, G.A., Hatami, K., 2011. Unsaturated soil-geotextile interfacebehavior. Geotext. Geomembr. 29, 17e28.

Koerner, R.M., 2005. Designing with Geosynthetics, fifth ed. Pearson, Prentice-Hall,New Jersey, USA.

Koerner, R.M., Soong, T.-Y., Koerner, G.R., 2005. Back drainage design and geo-composite drainage materials. In: NAGS 2005/GRI-19 Conference, December2005, Las Vegas, Nevada, USA.

Lambe, T.W., 1958. The engineering behavior of compacted clay. ASCE J. Soil. Mech.Div. 84 (1655), 1e35. SM2.

Liu, C.N., Ho, Y.H., Huang, j.W., 2009. Large scale direct shear tests of soil/PET-yarngeogrid interfaces. Geotext. Geomembr. 27, 19e30.

Miller, G.A., Hamid, T.B., 2005. Direct shear testing of interfaces in unsaturated soil.In: Proceedings of EXPERUS, International Symposium on Advances in Experi-mental Unsaturated Soil Mechanics, Trento, Italy. Taylor and Francis Group, plc,London, UK, pp. 111e116.

Mitchell, J.K., Zornberg, J.G., 1995. Reinforced soil structures with poorly drainingbackfills. Part II: case histories and applications. Geosynth. Int. 2 (1), 265e307.

Nam, S., Gutierrez, M., Diplas, P., Petrie, J., Wayllace, A., Lu, N., Munoz, J.J., 2009.Comparison of testing techniques and models for establishing the SWCC ofriverbank soils. Eng. Geol. 110, 1e10.

Ou, F.L., Cox, W., Collett, L., 1982. Rock aggregate management planning for energyconservation: optimization methodology. Transp. Res. Rec. (872), 63e69.

Palmeira, E.M., 2004. Bearing force mobilization in pullout tests in geogrids, geo-textiles and geomembranes. Geotext. Geomembr. 22, 481e509.

Sandri, D., 2005. Drainage recommendations for MSE walls constructed withmarginal fills. In: NAGS 2005/GRI-19 Conference, December 2005, Las Vegas,Nevada, USA.

Wescor Inc., 2001. HR-33T Dew Point Microvoltmeter Instruction/Service Manual.Wescor Inc., Logan, UT, USA.

Zheng, Y., Hatami, K., Miller, G.A., March 2013. Numerical modeling of the wetting-induced settlement of embankments. In: 2013 Geo-Congress, Oakland, CA.Paper #8.

Zornberg, J.G., Mitchell, J.K., 1994. Reinforced soil structures with poorly drainingbackfills. Part I: reinforcement interactions and functions. Geosynth. Int. 1 (2),103e148.



















http://dx.doi.org/10.3141/2363-08

http://dx.doi.org/10.3141/2363-08
















































influence of matric suction on geotextile reinforcement-marginal soil interface strength

Documents