Unsaturated soil–geotextile interface behavior

Geotextiles and Geomembranes 29 (2011) 17e28

Contents lists available at ScienceDirect

Geotextiles and Geomembranes journal homepage: www.elsevier.com/locate/geotexmem

Unsaturated soilegeotextile interface behavior Charbel N. Khoury*, Gerald A. Miller 1, Kianoosh Hatami 1 School of Civil Engineering and Environmental Science, University of Oklahoma, 202 W. Boyd, Room 334, Norman, OK 73019, USA

a r t i c l e i n f o

a b s t r a c t

Article history: Received 21 August 2009 Received in revised form 15 June 2010 Accepted 16 June 2010 Available online 8 August 2010

The behavior of mechanically stabilized earth (MSE) structures under seasonal climatic variations, i.e. wetting and drying, is not well understood. Stability and serviceability of MSE walls and embankments can significantly depend on the soil-reinforcement (e.g., geosynthetics) interface shearing behavior in unsaturated conditions. This is especially true for reinforced soil slopes and embankments that have significant fines contents. This paper presents results of a laboratory study on the mechanical behavior of unsaturated soil-geotextile interfaces using a specially modified direct shear apparatus. Several suctioncontrolled laboratory tests were conducted to investigate the effect of soil suction on the soil-geotextile interface. Results of the study indicate that the peak shear strength of the soil-geotextile interface increases nonlinearly with the soil suction. On the other hand, while inconclusive, the effect of suction on the post-peak shear strength of the interface was negligible in some cases. An elastoplastic constitutive model was used to simulate the laboratory results. This study demonstrates that the constitutive model is capable of capturing the mechanical behavior of the unsaturated soil-geotextile interface subjected to constant suction. Both shearing and volume change responses were reasonably simulated by the model. Ó 2010 Elsevier Ltd. All rights reserved.

Keywords: Geotextiles Unsaturated interface Unsaturated soil Shear strength Interface constitutive model

1. Introduction and background Construction of mechanically stabilized earth (MSE) walls and reinforced soil slopes (RSS) has increased significantly worldwide and specifically in the United States. For instance, on average more than 850 000 m2 of MSE and 190 000 m2 of RSS are constructed annually in the United States (Berg et al., 2009). Therefore, it is crucial to understand the behavior of these structures recognizing that their design is influenced by the shear strength of the interface between reinforcement layers and soil. Although coarse-grained soils are recommended as backfills in MSE walls in North America (Elias et al., 2001, AASHTO, 2002), some industry design guides (NCMA, 2002) allow for the use of up to 35% fine-grained soils, provided that a properly designed drainage system is present. The British Standard (BS8006, 1995) also allows cohesive-frictional soils (i.e., soils with greater than 15% passing 63 mm sieve) to be used for wall backfill materials. Backfills with up to 50% fine-grained soils (i.e., passing sieve #200) are allowed in some guidelines for the construction of reinforced embankments and slopes (Elias et al., 2001). In many projects (e.g., Powel et al., 1999; Musser and

* Corresponding author. Tel.: þ405 325 9244; fax: þ405 325 4217. E-mail addresses: [email protected] (C.N. Khoury), [email protected] (G.A. Miller), [email protected] (K. Hatami). 1 Tel.: þ405 325 5911; fax: þ405 325 4217. 0266-1144/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.geotexmem.2010.06.009

Denning, 2005) low quality backfill soils have been used in slopes and highways due to scarcity and high cost of good backfill soils in local areas. Since fine contents as low as 6e10% can significantly reduce the permeability of soils (BS8006, 1995, Elias et al., 2001; Koerner, 2005) and since these structures are built under unsaturated conditions, a main concern in their stability analysis and design is the reduction of the soil-reinforcement interface shear strength as a result of wetting. Factors such as seasonal precipitation and variation of the ground water table can significantly alter the soil moisture condition and suction, and thus the interface behavior. For example, some case studies of failure or large deformations of MSE walls have been reported (e.g., Mitchell and Zornberg, 1995; Christopher et al., 1998; Koerner, 2005; Sandri, 2005; Lawson, 2005; Stulgis, 2005) where backfill soils were compacted wet of optimum or where the structures under construction were subjected to heavy rainfalls resulting in increase of pore water pressure, decrease in matric suction, and thus reduction in shear strength and excessive deformations. Matric suction in the soil is defined as uaeuw, where ua and uw denote the pore air pressure and pore water pressure, respectively (e.g., Fredlund and Rahardjo, 1993; Lu and Likos, 2004). Current laboratory techniques to determine the soil-geosynthetic interface strength include interface shear tests (ASTM, 2009; D5321) and pullout tests (ASTM D6706) on soil-geosynthetic specimens. Soil specimens are generally compacted at optimum moisture content and 95% of maximum dry density (e.g. as

18

C.N. Khoury et al. / Geotextiles and Geomembranes 29 (2011) 17e28

per the standard Proctor test e AASHTO T-99) in order to simulate field conditions during construction. However, no assessments are currently made on the influence of drying and wetting on the soilgeosynthetic interface properties. This constitutes an unsaturated soil problem in which the influence of soil suction on the shearing response of the soil-geosynthetic interface needs to be investigated. Several studies can be found in the literature on direct shear testing of soil-geosynthetic interfaces (e.g., Frost and Han, 1999; Goodhue et al., 2001; Gourc et al., 2004; Koerner, 2005, and Sia and Dixon, 2007 among others). However, very limited studies have been conducted on unsaturated soil-geosynthetic interfaces. Goodhue et al. (2001) carried out direct shear tests on sand and foundry sand samples with geosynthetic (i.e., geogrid, geotextile and textured geomembrane) specimens and found that the drained interface friction angles were in close agreement with the ascompacted friction angles of the foundry sand samples except when a significant amount of bentonite was added to the sand. Fleming et al. (2006) used a modified direct shear device with a miniature pore pressure transducer (PPT) that measured changes in the pore water pressure at smooth geomembraneesoil interfaces. However, they could only measure relatively low suction values (less than 30 kPa) due to limitation of their PPT device. In addition, they were able to predict the soil-geomembrane interface shear strength values using unsaturated soil mechanics theory only at low normal stress values. At higher normal stress values, the interface behavior appeared to be governed only by the magnitude of total normal stress. They also reported that a near saturated condition at the soil-geomembrane interface resulted in a lower strength. Sharma et al. (2007) used the same device described by Fleming et al. (2006), to measure suction close to the interface of the soil-geomembrane during shearing. They concluded that it is important to evaluate the shearing behavior at the interface between geomembrane and unsaturated soil at low matric suction values. Their results suggested that soil suction contributes to shearing resistance at low normal stress values but not as much at higher normal stresses.

None of the studies described included tests under suctioncontrolled conditions. Researchers at the University of Oklahoma (e.g., Hamid, 2005; Miller and Hamid, 2007) developed a direct shear device for testing of soilestructure interfaces under suctioncontrolled conditions. A series of soilesteel interface (rough and smooth plates) shear tests was carried out at different moisture contents along a drying path and subjected to different matric suction values (20 kPa, 50 kPa and 100 kPa). Hamid and Miller (2009) found that matric suction clearly influenced the strength of the interface between the soil and a structural counterface. Hatami et al. (2008), in a related study, conducted a few suction controlled soil-geotextile interface shear tests at 25 and 50 kPa suctions. Their study showed that the extended Mohr-Coulomb theory could be used to describe the increase in the unsaturated soil-geotextile interface shear strength due to suction. Hamid and Miller (2008) developed a constitutive model capable of modeling the behavior of interfaces in unsaturated soils subjected to constant net normal stress and suction values. The current paper builds upon this earlier work. The above survey of literature clearly indicates that there is a need to investigate the soil-geotextile interface shearing behavior in unsaturated soil conditions. The objective of this study was to address this need by modifying the direct shear test device (Miller and Hamid, 2007) to test soil-geosynthetic interfaces under suction-controlled unsaturated conditions. This paper reports the results from these suction-controlled tests using a fine-grained soil and a commercially available geotextile product (a woven polypropylene geotextile). Results for both unsaturated soil direct shear tests and soil-geotextile interface tests are presented and discussed. In addition, the elastoplastic constitutive model developed by Hamid and Miller (2008) was used to model the unsaturated soil-geotextile interface test results obtained. Ongoing research has a longer-term objective of improving and incorporating the constitutive model into advanced computational methods to study the response of unsaturated reinforced soil structures.

Fig. 1. Schematic cross-section of the test chamber and shear box.

C.N. Khoury et al. / Geotextiles and Geomembranes 29 (2011) 17e28

2. Suction-controlled direct shear tests on unsaturated soil-geotextile interfaces

a

19

100 90

2.1. Testing apparatus

2.2. Test materials One major challenge in studying unsaturated soils is the extremely long testing time. Currently, this problem was circumvented to some extent by testing an artificial soil with a range of matric suction values comparable to that of a silty soil. However, the particle size gradation resulted in a hydraulic conductivity greater than that of most natural fine-grained soils with comparable suction ranges. This allowed the tests reported in this paper to be conducted more rapidly than comparable natural soils. The test soil is a mixture of two commercially available manufactured soils, Sil-Co-Sil 250 manufactured by U.S. Silica Company and Glass Beads, Size BT-9, manufactured by Zero Products. The soil mixture consists of 75% ground silica and 25% glass beads. The mixture is nonplastic and has a grain size distribution similar to that of fine sandy silt having about 48% fine sand (0.075e0.425 mm), 46% silt (0.002e0.075 mm), and 6% clay size material (0.002 mm), as shown in Fig. 2(a). The soil has a maximum dry density (gd) of 103.6 pcf (16.3 kN/m3) and optimum moisture content (OMC) of 16.5% based on standard compaction. Soil Water Characteristic Curves (SWCCs), depict the relationship between the soil matric suction and (gravimetric or volumetric) water content. The SWCC in this study subjected to net normal stress (snet ¼ sneua, where sn is the normal stress and ua is the pore air pressure) values of 0 kPa and 150 kPa are presented in Fig. 2(b). The SWCCs were obtained using the device described by Miller et al. (2008). As shown in Fig. 2(b), two of the SWCC tests were conducted immediately after compaction at initial degree of saturation So of nearly 67% without saturating the samples. Similarly, the soil specimens in the direct shear tests were not saturated prior to the application of constant suction. However, one SWCC test was carried out when the soil was initially in a saturated condition (So ¼ 100%). Also superimposed on Fig. 2(b) is the volumetric water content and corresponding matric suction for direct shear and interface shear specimens prior to shearing.

% Finer by Weight

70 60 50 40 30 20 10 0 1

0.1

0.01

0.001

Particle Diameter (mm)

b

σn-ua= 150 kPa σn-ua= 0 kPa σn-ua= 0 kPa, So = 100 % Direct Shear Test Data Degree of Saturation (S%) 0

10

20

30

40

50

60

70

80

90 100 110

1000

100

ua-uw (kPa)

In Fig. 1 a schematic diagram of the modified direct shear test (DST) device (e.g., Hamid, 2005; Miller and Hamid, 2007) that was used to test unsaturated soil and soil-interfaces under suctioncontrolled conditions is shown. Details of the test apparatus are described by Miller and Hamid (2007). Briefly, the modified DST apparatus consists of an air pressure chamber, a control system, shear boxes for testing unsaturated soil and interfaces (e.g., steel/ geosynthetics), a precision stepper-motor pump to control (and measure) the pore water volume and pressure, drainage lines, high air entry porous disc (HAEPD) and a diffused air volume indicator (DAVI). Suction is controlled via the air pressure inlet and the pump to control the water pressure in the specimen using the axis translation technique (e.g., Fredlund and Rahardjo, 1993). The automated water pump is capable of maintaining target pressure within 1 kPa and detecting volume changes as low as 1 mm3. Water is transmitted from the controller to the soil specimen through drainage lines. The drainage lines are 3-mm diameter, high-pressure polyvinylidene flouride tubes that are connected to the HAEPD with an air entry value of 3 bars (304 kPa). Air pressure in the chamber is controlled using a regulator and monitored using a pressure gage with a resolution of 0.7 kPa. The air can access the soil specimen through openings in the shear box including a 0.3 mm gap that is created between the two halves of the box during testing, and a small space (of about 3 mm) surrounding the top cap.

75 % SCS; 25 % Beads

80

10

1

0

0.00

0.10

0.20

0.30

0.40

Vol. Water Content (θ) Fig. 2. a) Grain size distribution of the soil used in this study, b) Soil water characteristic curves.

The geosynthetic material used in the interface tests was a woven polypropylene (PP) geotextile with material properties shown in Table 1. In each test, a fresh geotextile specimen was cut to a diameter of 100 mm and attached to a stainless steel cylindrical block (Fig. 3). Preliminary tests were carried out to determine an optimal method and the type and amount of adhesive to firmly attach the geotextile specimen to the steel block without introducing any wrinkles into the geotextile or leakage of the adhesive material into the soil-geotextile interface. 2.3. Sample preparation The soil was mixed to the desired moisture content (w ¼ 17.2%  1%) and compacted to the required dry density (unit weight gd ¼ 15.4 kN/m3) by moist tamping (i.e., volume-based compaction) directly inside the shear box. The box was assembled by fixing the upper and lower halves of the shear box together using two screws (Figs. 1 and 3). All soil specimens were prepared to achieve nominally the same initial conditions with respect to their unit weight and moisture content.

C.N. Khoury et al. / Geotextiles and Geomembranes 29 (2011) 17e28

Table 1 Properties of the woven geotextile used in the shear tests. Property test

Protocol/ Specification

Value/description

Polymer type Fabric Mass per unit area (g/m2) Percent open area (%) O95 (mm), Apparent opening size (U.S. Sieve) Permittivity (s1) Puncture resistance (kN) Trapezoidal tearing strength (kN) Grab tensile strength (kN) Elongation (%) Survivability class Applications Wide-width ultimate tensile strength (kN/m)

e e ASTM D5261 CWO-22125 ASTM D4751

Polypropylene Slit-film, Woven 1 0.425 (40)

ASTM D4491 ASTM D4833 ASTM D4533 ASTM D4632 ASTM D4632 AASHTO M288 AASHTO M288 ASTM D4595

0.05 0.4 0.33 (MD), 0.33 (XD) 0.9 (MD), 0.9 (XD) 15 (MD), 10 (XD) 3 Separation, Stabilization 17.6 (MD), 21.0 (XD)

Source: IFAI, 2007.

Horizontal Shear Stress (τ)

20

Target Suction

D

Shearing

Initial Suction

B O

C A

Seating Load Target σ

2.4. Test procedure The soil-geotextile interface shear test procedure used in this study is similar to that described by Hatami et al. (2008). The soil samples were first compacted on top of the geotextile that was mounted on a steel block. After compaction, the shear box containing the test specimen was placed in the DST air pressure chamber. The drainage line from the pore-water pressure controller was connected to the top platen inlet port of an HAEPD placed on the top of the compacted soil (Fig. 1). The HAEPD was saturated with de-aired water prior to testing. A vertical seating load with a magnitude not greater than 35 kPa was applied on the specimen to produce lateral stresses needed to stabilize the position of the upper half of the shear box when it was raised to introduce the gap needed before the shearing process. Fig. 4 shows an illustration of the stress loading pattern of the test. Approximately 1 h after the specimen was subjected to compression under the initial seating load (path OA, Fig. 4), the two screws holding the halves of the shear box together were removed from the pressure chamber. Then, a gap was created between the halves of the shear box by turning the four raising screws, which were then reversed to eliminate contact between the screws and the box. After creating a gap of about 0.3 mm, the air chamber lid was sealed with bolts. The target matric suction (path AB, Fig. 4) was then applied to the specimen by increasing the air and water pressures simultaneously via the axis translation technique. The test specimen was allowed to equilibrate under the applied suction value. Equilibrium was assumed completed when the change in the water volume became negligible. The net normal stress (snet ¼ sneua) was then increased to the target value for each test in increments of 35 kPa by applying vertical loads (path BC, Fig. 4). After the specimen was consolidated under the target

Fig. 4. Illustration of the loading path for the suction-controlled direct shear tests.

vertical stress, it was subjected to drained shearing (path CD, Fig. 4) while both the suction and vertical net normal stress were kept at their set levels. The direction of shearing was perpendicular to the machine direction of the geotextile and applied at a displacement rate of 0.005 mm/min up to nearly 10 mm displacement. During shearing, the horizontal load and the horizontal and vertical displacements were measured with LVDTs and recorded. 3. Test results and discussion 3.1. Effect of soil suction and net normal stress on the shearing behavior of soil and soilegeotextile interfaces 3.1.1. Suction and net normal stress equalization phases The tests performed in this study on both materials (i.e., soil and soil-geotextile interfaces) were carried out at different vertical net normal stresses, ranging from 50 kPa to 300 kPa, and suction values ranging from 0 kPa to 100 kPa. Fig. 5 shows a comparison of the consolidation responses of the soil and soil-geotextile interface shear tests at suction and net normal stress values of 100 kPa. Fig. 5 (a) shows the variation of the specimen vertical deformation, normalized to the specimen height (v/H0), versus time. Fig. 5(b) shows a comparison of changes in water content (Dw%) during equalization periods in both soil and soil-geotextile specimens. Results shown in Fig. 5(a) indicate that the magnitude of vertical strain is practically the same in the soil and soil-geotextile interface tests for given suction and net normal stress values. Similarly, it is

Fig. 3. Photographs of the geotextile specimen and the test cell.

C.N. Khoury et al. / Geotextiles and Geomembranes 29 (2011) 17e28

Due to (ua-uw)

0.000

v/H0

0.002 0.004 Due to (σn-ua)

0.006 0.008 0.010

Δw %

b

0 -2 -4 -6 -8 -10 -12 -14 -16

Soil Soil-geotextile

0

1500 3000 4500 6000 Time (min)

Fig. 5. Comparison of consolidation results for soil and soil-geotextile specimens at suction and net normal stress values of 100 kPa.

observed in Fig. 5(b) that the change in water content due to drainage in both tests is essentially the same (Dw z 13.5%). Soil-geotextile interface tests were carried out at suction values of 0 kPa, 25 kPa, 50 kPa and 100 kPa. Fig. 6(a) shows a typical plot of v/H0 versus time during equalization phases for suction magnitudes of 25 kPa, 50 kPa and 100 kPa, and net normal stress of 50 kPa. Note that the change in net normal stress needed to achieve the target net normal stress of 50 kPa was lower for the 100 kPa suction test as compared to tests at lower suction values. This was due to an increase in the seating load from 14 to 34 kPa for the 100 kPa suction test (see the legend in Fig. 6). As a result, the soil specimen compressed less when subjected to the full 50 kPa net normal stress as shown in Fig. 6(a). For a given change in net normal stress, the

a

magnitude of compression in the specimen was found to be essentially independent of the soil suction for the range of suction values examined in this study. Fig. 6(b) shows the influence of soil suction on the change in water content (Dw%) in the soil-geotextile interface specimens during equalization periods for the three suction values of 25 kPa, 50 kPa and 100 kPa. During the application of suction, water continuously flowed out of each specimen (drying path) to reach the target suction value. As expected, it was observed that the amount of water drained from the specimen increased as suction increased (e.g., Dw z 4%, 9% and 13.5% for the suction values 25 kPa, 50 kPa and 100 kPa, respectively). As shown in Fig. 2 (b), the equilibrium water contents for a given suction agree well with the soil water characteristic curves obtained independently. The effect of the net normal stress on the change in moisture content was found to be small for all suction values examined (i.e., about 0.1%). 3.1.2. Shearing phase After the suction and net normal stress equalization phases, the specimens were sheared at a rate of 0.005 mm/min. This slow rate was selected to avoid changes in pore pressures during shearing as recommended by Miller and Hamid (2007). The average shear stress (s) on the horizontal failure plane in each specimen and corresponding horizontal and vertical displacements were recorded throughout the test at 1 min intervals. Fig. 7 shows a comparison of shear stress and volume change responses of soil specimens subjected to different net normal stress values at a suction of 100 kPa. Results shown in Fig. 7(a) indicate that the soil shear strength increased with net normal stress. During shearing, the soil compressed slightly at the beginning but started to dilate as its

a

300 250 τ (kPa)

a

150

50 0

b

0.000

0.004

v/H0

v/H0

200

100

0.002

0.006 0.008

ua - uw = 25 kPa Δ (σn - ua) = 36 kPa

0.010 0 -2 -4 -6 -8 -10 -12 -14

ua - uw = 50 kPa Δ (σn - ua) = 36 kPa ua - uw = 100 kPa

-0.015 -0.010 -0.005 0.000 0.005 0.010 0.015

c

Δ (σn - ua) = 16 kPa

σn - ua = 50 kPa σn - ua = 100 kPa σn - ua = 150 kPa σn - ua = 300 kPa

0.00 -0.05

Δw (%)

Δw %

b

21

-0.10 -0.15 -0.20 -0.25

0

1500

3000

4500

6000

Time (min) Fig. 6. Influence of suction on the consolidation response of soil-geotextile specimen at suction values of 25 kPa, 50 kPa and 100 kPa and at net normal stress of 50 kPa.

-0.30

0

2 4 6 8 Horizontal Displ. u (mm)

10

Fig. 7. Direct shear test results on soil specimens at 100 kPa matric suction.

22

C.N. Khoury et al. / Geotextiles and Geomembranes 29 (2011) 17e28

shear strength was mobilized (Fig. 7(b)). The soil dilation is observed to be more significant when subjected to smaller net normal stress values (e.g., snet ¼ 50 kPa). For a net normal stress of 300 kPa, dilation was followed by compression at large horizontal displacements. This dilation behavior was also observed in a compacted natural unsaturated lean clay (CL) soil by Miller and Hamid (2007) and Hamid and Miller (2009). In each test during shearing, a small amount of water drained out of the test specimen in order to maintain the soil suction at the target value. Results shown in Fig. 7(c) indicate a relatively small (<0. 25%) decrease in the specimen moisture content during the shearing phase. The soil-geotextile interface shear test results for 100 kPa suction subjected to different net normal stress values are shown in Fig. 8. Results presented in Figs. 7 and 8 show that overall, the peak and post-peak shear strength values of both the unsaturated soil and unsaturated soil-geotextile interface increased with the net normal stress. Both soil specimens and soil-geotextile interfaces showed a relatively insignificant initial vertical compression before they exhibited dilation during shearing. Results shown in Fig. 8(b) indicate that the soil-geotextile specimen stopped dilating once the strain softening response was completed and the shear stress reached the interface post-peak shear strength value. This behavior was also observed by Miller and Hamid (2007) for the response of an unsaturated rough interface. The dilation response observed in both the soil and soil-geotextile interface shear tests is attributed to the rearrangement of soil grains and sliding of soil particles over each other and over the geotextile surface. Similar to the soil specimen tests, small decreases in water content (i.e. change in Dw%) were detected during shearing of the soil-geotextile interface, a behavior that was also observed by Hamid and Miller (2009) for direct shear test on unsaturated soils and steel interfaces. This

a

behavior contradicts the expected behavior of saturated soils or interfaces, since for unsaturated conditions, changes in pore water pressure are in large part the result of changes to the airewater interfaces between particles (or menisci); these changes can thus be independent of total volume changes. Therefore, it is postulated that during dilation in unsaturated conditions the menisci between soil particles is disrupted causing a tendency for increasing pore water pressure and since these are drained constant-suction tests, water drained out of samples during shear to maintain the same values of pore water pressure and suction. Fig. 9 shows a comparison of shearing responses obtained for the soil and soil-geotextile interfaces for the magnitudes of suction and net normal stress equal to 100 kPa and 50 kPa, respectively. Results shown in Fig. 9 indicate that the soil-geotextile interface showed a more significant strain-softening behavior than the soil specimen. In general, a slight strain hardening response was observed after strain softening of soil specimens. Results in Fig. 8 suggest that strain softening could be important when the magnitudes of soil-geotextile interface pullout or sliding displacement prior to failure is expected to be significant. However, verification of this conclusion requires further studies involving larger-scale specimens. Fig. 10(a) shows the variation of shear stress with horizontal displacement for the soil-geotextile interface for suction values of 25 kPa, 50 kPa, and 100 kPa and net normal stress value of 100 kPa. It is observed that increasing suction from 25 kPa to 50 kPa in the specimen resulted in a detectable increase in the interface peak shear strength. However, the interface peak shear strength was observed to decrease slightly at suction of 100 kPa which indicates

a

100

250

τ (kPa)

80

τ (kPa)

200

20

50

0

b

0

Δw (%)

v/H0

-0.015 -0.010 -0.005 0.000 0.005 0.010 0.015

c

40

100

σn - ua = 50 kPa σn - ua = 100 kPa

c

σn - ua = 300 kPa

-0.015 -0.010 -0.005 0.000 0.005 0.010 0.015

0.00

-0.05 -0.10

-0.10 -0.15

-0.15 -0.20

-0.20

-0.25

-0.25

-0.30

-0.30

0

2 4 6 8 Horizontal Displ. u (mm)

10

Fig. 8. Soil-geotextile interface test results at 100 kPa matric suction.

Soil-Geotextile Soil

0.00

-0.05

Δw (%)

v/H0

b

150

60

0

2

4

6

8

10

Horizontal Displ. u (mm) Fig. 9. Soil direct shear and soil-geotextile interface shear test results at 100 kPa suction and 100 kPa net normal stress.

C.N. Khoury et al. / Geotextiles and Geomembranes 29 (2011) 17e28

a

Two additional tests were conducted on the soil specimens to check for repeatability of the results. One test at 50 kPa suction and 50 kPa net normal stress and another at 50 kPa suction and 150 kPa net normal stress were repeated. As shown in Fig. 11, similar behavior was exhibited during equalization and shearing. While some slight difference is observed in post-peak response, shearing behavior is quite similar up to the point of yielding. To be more precise, the difference in the interpreted shear strength was within 3 kPa (i.e. 2.5%) for a given net normal stress and suction value. These duplicated suction-controlled direct shear tests on soil specimens in addition to the consistent patterns observed in the Mohr-Coulomb failure envelopes (presented in the coming section) for both soil and soil-geotextiles suggest that the data are valid and repeatable.

100 τ (kPa)

80 60 40 20

b

0 -0.009 -0.006

v/H0

-0.003 0.000 0.003

Δw (%)

c

23

0.006

ua -uw = 25 kPa

0.009

ua -uw = 50 kPa

3.2. Mohr-Coulomb failure envelopes for peak shear strength of unsaturated soil-geotextile interfaces

ua -uw = 100 kPa

0.00 -0.05 -0.10 -0.15 -0.20 -0.25 -0.30 -0.35 -0.40 0

2

4

6

8

3.2.1. Background There is a variety of opinions regarding the proper selection of stress state variables for formulating the shear strength of unsaturated soil (e.g., Fredlund et al., 1978; Wheeler and Sivakumar 1995, Gallipoli et al., 2003; Khalili et al., 2004; Lu and Likos, 2004; Zhang and Lytton, 2006). A common formulation is based on the concept of two stress state variables (Fredlund et al., 1978), net normal stress (the difference between the total stress and pore air pressure) and

10 σn - ua = 50 kPa; Repeat Test σn - ua = 50 kPa

Horizontal Displ. u (mm) Fig. 10. Influence of suction on the soil-geotextile interface shear response at 100 kPa net normal stress.

σn - ua = 150 kPa; Repeat Test σn - ua = 150 kPa

a τ (kPa)

150 100 50

v/H0

b

c

0 -0.025 -0.020 -0.015 -0.010 -0.005 0.000 0.005 0.010 0.0 -0.2 Δw (%)

a non-linear behavior. Generally, for most net normal stresses, the peak shear strength increased nonlinearly with suction. In addition, some results (e.g., data for 25 and 100 kPa suction shown in Fig. 10) suggest that increasing suction may not significantly increase the post-peak strength of the soil-geotextile interface. This is consistent with previous observations by Hamid and Miller (2009), but current results are somewhat inconclusive. Results in Fig. 10(a) show that the magnitude of horizontal displacement corresponding to the interface peak shear strength decreased slightly with the increase in suction. In addition, specimens subjected to a greater suction showed a more pronounced strain softening behavior. Volume change results shown in Fig. 10 (b) indicate a slight amount of initial contraction followed by dilation in the specimen during shearing. The specimens exhibited negligible further dilation after softening behavior was completed. Results shown in Figs. 8 and 10 are, in general, consistent with the observations reported by Hamid and Miller (2009) on the shear response of a rough soilesteel interface in natural, low-plasticity clayey silt. While artificial soil may not represent fine-grained plastic soil, interface behavior has been examined for an unsaturated natural lean clay (CL) soil in other related studies as mentioned previously in this section (e.g. Hamid and Miller, 2009). In general, similar behavior was observed although it is believed that suction will have a greater influence on the peak shear strength of soil with higher clay contents. Similar trends of shear strength and volume behavior are also expected for geosynthetics with various degrees of surface roughness. The surface roughness of the geotextile used in this study seems to fall slightly below the rough steel interface, and in between the smooth and the rough steel interfaces tested in a related study (Hamid and Miller, 2009).

200

-0.4 -0.6 -0.8 -1.0 0

2

4

6

8

10

Horizontal Displ. u (mm) Fig. 11. Comparison of direct shear test results on soil specimens with repeated tests at suction of 50 kPa and net normal stresses of 50 and 150 kPa.

24

C.N. Khoury et al. / Geotextiles and Geomembranes 29 (2011) 17e28

matric suction (the difference between the pore air and pore water pressure). While older versions of this theory assume a planar failure envelope, more recent studies (e.g., Fredlund and Rahardjo, 1993; Oloo and Fredlund, 1996; Vanapalli et al., 1996, and Kayadelen et al., 2007), indicate significant nonlinearity in the shear strength versus matric suction failure envelope. Hamid and Miller (2009) proposed the following interface shear strength equation (Eq. (1)), based on the constitutive model for shear strength of unsaturated soils presented by Fredlund et al., (1978):

ss ¼ c’a þ ðsn  ua Þtand’ þ ðua  uw Þtandb

(1)

where: c0 a ¼ adhesion intercept, sn ¼ normal stress on the interface at the failure, ua ¼ pore air pressure at failure, d0 ¼ friction angle between soil and counterface with respect to (sneua), uw ¼ pore water pressure at failure, db ¼ friction angle between soil and counterface with respect to (uaeuw), and uaeuw ¼ matric suction at failure. The parameters ca0 , d0 and db are analogous to c0 , 40 and 4b defined by Fredlund et al. (1978) for unsaturated soil shear strength, where 4b and 40 are the soil internal friction angles associated with the matric suction and the net normal stress variables, respectively. As noted previously, the parameter 4b is typically non-linear with respect to suction and can be represented as a non-linear function of volumetric water content (q) corresponding to the soil water characteristic curve (SWCC). For example, Vanapalli et al. (1996) presented an equation that indirectly relates the solidwater interfacial area to the change in matric suction from the SWCC, and used this relationship in the expression for 4b, where tan4b ¼ (qeqr)/(qseqr) tan40 . In this expression, q ¼ current volumetric water content, qr ¼ residual volumetric water content from SWCC, and qs ¼ volumetric water content from SWCC at 100% saturation. In other words, the shear strength can be predicted when the saturated shear strength parameters (40 and c0 ) and the

SWCC are available. In a similar fashion, Hamid and Miller (2009) obtained satisfactory results using the following equation: b

tand ¼ ðqs  qr Þ=ðq  qr Þtand

0

(2)

3.2.2. Test results A 3D plot of the extended Mohr-Coulomb failure envelope (Eq. (1)) for the soil-geotextile interface is shown in Fig. 12. To better appreciate the separate effects of suction and net normal stress (snet) on the shear strength, results are also presented in 2D plots as shown in Figs. 13 and 14(a), respectively. Fig. 13 shows the variations of peak shear strength (smax) versus net normal stress (i.e., failure envelopes) for the soil-geotextile interface at four suction values 0 kPa (i.e., saturated interface), 25 kPa, 50 kPa, and 100 kPa. Results shown in this figure represent the frontal planes of the extended Mohr-Coulomb failure envelopes for the soil-geotextile interface. The slope and the intercept of the failure envelopes on these frontal planes are denoted as d0 , and ca, respectively. The best-fit linear failure envelopes for all suction values shown in Fig. 13 are essentially parallel to each other with the interface friction angle value (i.e., d0 ) equal to 32 . Therefore, the increase in suction in the soilgeotextile interface tests resulted in an increase in the interface adhesion (ca). This observation, together with the fact that interface friction angle value did not increase with suction (Fig. 13), is consistent with the concept of “apparent cohesion” discussed in several numerical simulation studies (e.g., Cazzuffi et al., 1993; Rowe and Skinner, 2001; Skinner and Rowe, 2003, 2005; Hatami and Bathurst, 2005, 2006) to account for the effect of moisture in the backfill of MSE walls. However, the rate of increase in interface adhesion is non-linear with suction as illustrated in Fig. 14(a). The line intercept and slope in Fig. 14(a) represent the effective adhesion corresponding to zero net normal stress (snet ¼ 0 kPa), and interface friction angle with respect to suction (db). Basically, db showed a non-linear behavior with increase in suction; db decreases after suction exceeded the Air Entry Value (AEV) as shown in Fig. 14(a).

Fig. 12. Extended Mohr-Coulomb failure envelope from suction-controlled tests on soil-geotextile interfaces.

C.N. Khoury et al. / Geotextiles and Geomembranes 29 (2011) 17e28

300

250

ua-uw = 0 kPa ua-uw = 25 kPa ua-uw = 50 kPa

250

200

ua-uw = 100 kPa

200

δ' = 32

150

ca3 = 34 kPa

100

ca2 = 30.69 kPa ca1 = 23.4 kPa

50 ca

0

50

The soil friction angle 4b also exhibited a non-linear behavior as shown in Fig. 14(b). This nonlinearity is similar to that reported for other unsaturated soils and interfaces (e.g., Gan et al., 1988; Escario and Juca, 1989; Fredlund and Rahardjo, 1993; Hamid and Miller, 2009). In these studies, the soil friction angle 4b at low matric suction was equal to 40 but decreased at higher matric suctions. Fig. 15 shows a comparison of the failure envelopes, associated with net normal stress, for the soil and soil-geotextile specimens at

Soil (ua-uw = 100 kPa)

150 100

Soil: φ ' = 36 ; c = 48 kPa Geo: δ' = 32 ; c = 34 kPa

0

100 150 200 250 300 σn-u (kPa) a

Fig. 13. Failure envelopes for the soil-geotextile interface at 0 kPa, 25 kPa, 50 kPa and 100 kPa suction values.

Soil_Geotextile (u a-u w = 100 kPa)

50

ca' = 14 kPa

ca'

0

τ max (kPa)

τ max (kPa)

300

25

0

50

100 150 200 250 300 σn-ua (kPa)

Fig. 15. Comparison of failure envelopes for soil and soil-geotextile interface at 100 kPa suction.

100 kPa suction. Results in Fig. 15 show that the soil internal friction angle, 40 , is slightly greater than the soil-geotextile interface friction angle, d0 . However, at zero net normal stress and suctions less than 25 kPa, the adhesion intercept (ca) of the interface is greater than the cohesion intercept (c) for soil as shown in Fig. 14(b). The soil-geotextile adhesion intercepts (ca) were found to be smaller than the soil cohesion intercepts (c) for suction values greater than 25 kPa. The difference between (ca) and (c) may be related to the amount of soil entrenchment in the rough surface of the geotextile that occurs under various levels of suction. 4. Constitutive modeling and predictions of unsaturated interface test results 4.1. Constitutive model for unsaturated interfaces Since unsaturated soil-interface behavior resembles the behavior of unsaturated soil, Hamid and Miller (2008) exploited their similarities to develop a constitutive model capable of modeling the behavior of interfaces in unsaturated soils under constant net normal stress and suction. Hamid and Miller’s model extended the Navayogarajah et al. (1992) model to unsaturated interfaces. The modified model is based on the disturbed state concept, incorporating two independent stress variables, net normal stress and matric suction. Using these stress state variables, Hamid and Miller (2008) proposed yield, potential, and hardening functions for unsaturated soil-steel interfaces as follows:

F ¼ s2 þ aðsÞ½snet þ RðsÞn gðsÞ½snet þ RðsÞ2

(3)

where, F ¼ yield function, s ¼ shear strength, snet ¼ sneua ¼ net normal stress, R(s) ¼ bonding stress which is the increase in the strength of the unsaturated interface with the increase in suction defined as RðsÞ ¼ lðsÞðua  uw Þ þ l1 Rn þ l2 ; the plot of R(s) versus (uaeuw) gives the slope l(s) and intercept l*. For each interface, l* is plotted versus Rn which provides parameters l1 (slope) and l2 (intercept). The parameter Rn is the roughness ratio defined as Rn ¼ Rmax/D50, Rmax is the maximum peak-to-trough dimension on the interface surface and D50 is the grain size diameter corresponding to fifty percent finer. g(s) ¼ material parameter that defines the limiting state of stress as follows:

Fig. 14. Failure envelopes associated with suction in the zero net normal stress plane for: a) the soil-geotextile interface [to determine effective adhesion (c0 a) and friction angle (db)], and b) both the soil and the soil-geotextile interface.

gðsÞ1=2 ¼

sp ¼ mp1 þ mp2 Rn snet þ RðsÞ

where, sp ¼ peak shear strength.

(4)

26

C.N. Khoury et al. / Geotextiles and Geomembranes 29 (2011) 17e28

The intercept and slope of g(s)1/2 versus Rn yield the material constants mp1 and mp2, respectively. Parameter a(s) is a hardening parameter that defines the evolution of the yield surface during deformation (Eq. (5)), (s) indicates dependence of the parameters on matric suction, and n is a phase change parameter related to a state of stress at which the material passes through a state of zero volume change.

b
* aðsÞ ¼ gðsÞexpðaxv Þ xD * xD ; xD

aðsÞ ¼ 0;

xD < x*D xD 

x*D

9 = ;

Table 2 Model parameters for the unsaturated soil-geotextile interface in this study. Parameters

Value

x*D ðmmÞ

x*D1 ðmmÞ x*D2 ðmmÞ mp1 mp2

g(S) n

k

(5)

k1 k2 m01 m02

m0 R(s), (kPa)

l (s)

l1 l2

*

(6)

where, aQ(s) was defined as




aQ ðsÞ ¼ aðsÞ þ aph ðsÞ 1 

aðsÞ ai


 D 1k 1 Du

(7)

where, Q ¼ potential function, k ¼ material parameter (non-associative parameter) and is related to the normalized roughness, net normal stress, and suction. D ¼ damage function, aph is the value of a(s) at the phase change point and ai the value of a(s) at the initiation of the non-associativeness, defined as ai ¼ gðsÞ ðsnet þ RðsÞÞ2n , and Du ¼ sp  sr =sp ; sp and sr are the peak and residual shear stresses, respectively. The constitutive model expressed by Eqs. (3)e(7) was applied to the results of the unsaturated soil-geotextile interface tests carried out in this study. 4.2. Application of the constitutive model to the unsaturated soilegeotextile interface test results The model parameters for the soil-geotextile interface tests in this study were obtained using the same procedure as described by Hamid and Miller (2008). The normalized surface roughness (Rn) value for the soil-geotextile interface was estimated to be 4.2 in this study. For the soil used, D50 ¼ 0.071 mm. The roughness of the geotextile surface is somewhere in between rough and smooth surfaces as defined by Hamid and Miller (2008). For this study, Rmax was found to be approximately equal to 0.3 mm by crudely measuring the geotextile peak-valley distance. The constitutive model parameters determined from the experimental data for sneua of 50, 100 and 150 kPa and uaeuw of 25, 50 and 100 kPa are summarized in Table 2. Fig. 16(a) and 17(a) compare typical predicted and measured results of shear stress (s) versus horizontal shear displacement (u) for the soil-geotextile interface examined in this study subjected to different suction and net normal stress values. The comparison between the experimental and predicted results reveals that the model is capable of predicting the behavior of the unsaturated soilgeotextile interface with reasonable accuracy. For instance, the predicted peak shear strengths and corresponding displacements were within 5% and 12% of experimental results, respectively. The following observations with respect to the model predictions can be made:

Notes: For the preliminary modeling, R(s) was assumed to be a linear function of suction. All parameters are dimensionless unless indicated otherwise.

1. The peak shear strength of soil-geotextile interface increased with suction. 2. The post-peak shear strength did not increase notably with suction. 3. The peak and post-peak shear strength increased with the net normal stress. 4. Strain softening was more pronounced at higher suction and higher net normal stress values. Fig. 16(b) and 17(b) compare the predicted and measured results of vertical strain (v/H0) versus horizontal shear displacement (u) for different suction and net normal stresses. Important volume change behavior of unsaturated soil-geotextile interfaces was captured using this model: 1. Unsaturated soil-geotextile interface predictions generally followed the overall trend (mostly dilation) of volume change behavior. However, in most cases experimental results showed a slight amount of compression preceding dilation.

a

120 100 τ (kPa)

Q ¼ s2 þ aQ ðsÞ½snet þ RðsÞn gðsÞ½snet þ RðsÞ2

a b Kn (kPa) Ks (kPa)

b

80 60 40

Experimental ua -uw = 25 kPa

20

Model Predictions ua -uw = 25 kPa Experimental ua -uw = 50 kPa

0 -0.015

Model Predictions ua -uw = 50 kPa

-0.010 -0.005 v/H0

Parameters a, b, and xD are functions of R(s) and roughness ratio R R Rn. xv ¼ jdvp j, xD ¼ jdup j, dvp and dup are the plastic displacements normal and tangential to the shearing surface, respectively, * and xD is the value of xD when shear stress reaches its peak value. A non-associative flow rule was adopted in the model to correlate the volume change behavior and loading. By modifying the growth function, a(s) in the yield surface, a potential function (Q) was proposed as follows:

0.1714 0.0715 0.3995 0.0639 2.3 0.0962 0.0323 0.4682 0.0755 0.1454 13.268 77 34 4.1 1050 282

0.000 0.005 0.010 0.015

0

2

4

6

8

10

Horizontal Displ., u (mm) Fig. 16. A comparison of experimental shear response results with model predictions for soil-geotextile interface at a net normal stress of 100 kPa.

C.N. Khoury et al. / Geotextiles and Geomembranes 29 (2011) 17e28

100 90 80 70 60 50 40 30 20 10 0 -0.015

3)

τ (kPa)

a

b

Experimental σn - ua = 100 kPa Model Predictions σn - ua = 100 kPa Experimental σn - ua = 50 kPa Model Predictions σn - ua = 50 kPa

-0.010 v/H0

4)

5)

-0.005 6)

0.000 0.005

0

2

4

6

8

10

Horizontal Displ., u (mm) Fig. 17. A comparison of experimental shear response results with model predictions for soil-geotextile interface at suction of 100 kPa.

7) 2. Specimens did not show any dilation after the softening behavior was completed. 3. The magnitude of dilation decreased under greater net normal stress values. While negligible, dilation magnitude was generally consistent with experimental results. 5. Conclusions This study was carried out to investigate the shear response and preliminary constitutive modeling of unsaturated soil-geotextile interfaces. To this end, a series of suction-controlled direct shear tests were first carried out on an unsaturated soil (fine grained artificial soil). Interface shear tests were then carried out on unsaturated soil-woven geotextile interfaces. Both types of tests were carried out at different suction (0 kPa, 25 kPa, 50 kPa and 100 kPa) and net normal stress (50 kPa, 100 kPa, 150 kPa and 300 kPa) values. A constitutive model was used to simulate the mechanical behavior observed in the experimental results. While the model is still under development, the results presented in this paper have shown that the model is overall capable of capturing the salient response features of unsaturated soil-geotextile interfaces. Some of the findings and conclusions of this study are summarized below: 1) The modified direct shear test device for unsaturated soilgeotextile interfaces performed well and proved to be a suitable device for such testing. The net normal stress and suction were applied at set target values without difficulty. This is verified by the fact that the amounts of water drained from comparable soil and soil-geotextile specimens (i.e., change in moisture content), and the magnitude of specimen compression subjected to the same suction and net normal stress values were the same during the equalization phases. 2) In general, the soil-geotextile interface exhibited slightly higher amounts of strain softening during post-peak shearing as compared to soil specimens. Both types of tests showed a slight initial compression before dilation. Once the strain softening was completed, unsaturated soil-geotextile interfaces showed no further dilation. While small, the amount of dilation in soil specimens seemed slightly higher than that of soil-geotextile specimens. This may be attributed to the

8)

27

compressibility of the geotextile or possible entrenchment of soil particles in the geotextile openings. Small decreases in water content in both soils and soil-geotextile specimens were detected during the shearing process. A behavior which can be described by disruption of the menisci between soil particles causing a tendency for increasing pore water pressure (i.e. decrease in suction), thus resulting in water draining out of the sample. Increasing net normal stress and suction in the soil-geotextile tests resulted in an increase in the interface peak shear strength. However, the rate of increase was non-linear with suction. Increase in suction resulted in a reduction in the magnitude of horizontal displacement at peak shear strength, and a more pronounced strain softening behavior. Increase in suction of the soil-geotextile interface tests resulted in an increase in the interface adhesion. The interface friction angle with respect to net normal stress (d0 ) remained essentially constant at greater suction values. However, the interface friction angle with respect to suction (db) increased nonlinearly with suction. This finding is consistent with the observations on unsaturated soils and interfaces in the previous studies cited in this paper. The interface friction angle (d0 ¼ 32 ) was smaller than the internal friction angle (40 ¼ 36 ) of comparable soil. However, this was not true for the effective adhesion (ca0 ) and cohesion (c0 ), at zero suction. The adhesion intercepts (ca) were higher than cohesion intercepts (c) for suctions less than 25 kPa. However, the opposite was observed as suction exceeded 25 kPa. The constitutive model was capable of capturing both the peak and the post-peak shear strength responses of soil-geotextile interfaces to the increase in the applied suction and net normal stress. In addition, specimen volume change behavior in the interface shear tests was predicted with reasonable accuracy using this model.

The current findings of the study and its future developments will have important applications in the design of reinforced soil structures constructed with marginal soils and subjected to variations in their soil moisture content (or suction). More significant dependence of the soil-geotextile interface response on suction is expected for clayey soils with higher plasticity than the type of soil investigated in this study. In arid or semi-arid areas, it is highly unlikely that reinforced earth slopes and embankments with significant fines content become saturated. For these types of structures, unsaturated soil mechanics theory could be used to quantify the increase or decrease in the soil shear strength due to changes in matric suction. Acknowledgements Equipment used in the research was developed with support of funding from the National Science Foundation (NSF) under Grant No. 0079785. The authors are grateful to the NSF for the support. Special thanks to Michael F. Schmitz of the University of Oklahoma for his exceptional fabrication skills, dedication, and good ideas. The first two authors would also like to thank the Oklahoma Aeronautic Commission (OAC) for financial support. References American Association of State Highway and Transportation Officials (AASHTO), 2002. Standard Specifications for Highway Bridges, seventeenth ed.. American Association of State Highway and Transportation Officials, Washington, DC, USA.

28

C.N. Khoury et al. / Geotextiles and Geomembranes 29 (2011) 17e28

American Society for Testing and Materials (ASTM), 2009. Geosynthetics, vol. 04.13. ASTM International, W. Conshohocken, PA, USA. Berg, R.R., Christopher, B.R., Samtani, N.C., 2009. NHI Courses No. 132042 and 132043. Design and Construction of Mechanically Stabilized Earth Walls and Reinforced Soil Slopes, Vols. 1&2. U.S. Department of Transportation Federal Highway Administration. Publication No. FHWA-NHI-10-024, November 2009. BS8006, 1995. British Standard, Code of Practice for Strengthened/reinforced Soils and Other Fills. BSI. Cazzuffi, D., Picarelli, L., Ricciuti, A., Rimoldi, P., 1993. Laboratory investigations on the shear strength of geogrid reinforced soils. In: Cheng, S.C.J. (Ed.), ASTM Special Technical Publication 1190: Geosynthetic Soil Reinforcement Testing Procedures, pp. 119e137. Christopher, B.R., Zornberg, J.G., Mitchell, J.K., 1998. Design guidance for reinforced soil structures with marginal soil backfills. In: Sixth International conference on Geosynthetics, Atlanta, GA, USA, 1, 797e804. Elias, V., Christopher, B.R., Berg, R.R., 2001. Mechanically Stabilized Earth Walls and Reinforced Soil Slopes-design and Construction Guidelines. FHWA-NHI-00-043, Federal Highway Administration, Washington, DC, USA. Escario, V., Juca, J.F.T., 1989. Strength and Deformation of Partly Saturated Soils. In: Proceedings of the 12th ICSMFE, 1, 43e46. Fleming, I.R., Sharma, J.S., Jogi, M.B., 2006. Shear strength of geomembrane-soil interface under unsaturated conditions. Geotextiles and Geomembranes 24, 274e284. Fredlund, D.G., Morgenstern, N.R., Widger, R.A., 1978. The shear strength of unsaturated soils. Canadian Geotechnical Journal 15 (3), 313e321. Fredlund, D.G., Rahardjo, H., 1993. Soil Mechanics for Unsaturated Soils. John Wiley, USA. Frost, J.D., Han, J., 1999. Behavior of interfaces between fiber-reinforced polymers and sands. ASCE Journal of Geotechnical and Geoenvironmental Engineering 125 (8), 633e640. Gallipoli, D., Gens, A., Sharma, R., Vaunat, J., 2003. An elasto-plastic model for unsaturated soil incorporating the effects of suction and degree of saturation on mechanical behavior. Géotechnique 53 (1), 123e135. Gan, J.K.M., Fredlund, D.G., Rahardjo, H., 1988. Determination of the shear strength parameters of an unsaturated soil using the direct shear test. Canadian Geotechnical Journal 25 (8), 500e510. Goodhue, M.J., Edil, T.B., Benson, C.H., 2001. Interaction of foundry sands with geosynthetics. Journal of Geotechnical and Geoenvironmental Engineering 127 (4), 353e362. Gourc, J.P., Reyes-Ramirez, R., Villard, P., 2004. Assessment of geosynthetics interface friction for slope barriers of landfill. In: Shim, J.B., Yoo, C., Jeon, H.-Y. (Eds.), GeoAsia 2004, Now and Future of Geosynthetics in Civil Engineering, Seoul, Korea, pp. 116e149. Hamid, T.B., 2005. Testing and modeling of unsaturated interfaces, Ph.D. Dissertation, School of Civil Engineering and Environmental Science, University of Oklahoma, Norman, Oklahoma. Hamid, T.B., Miller, G.A., 2008. A constitutive model for unsaturated interfaces. International Journal for Numerical and Analytical Methods in Geomechanics 32 (13), 1693e1714. Hamid, T.B., Miller, G.A., 2009. Shear strength of unsaturated soil interfaces. Canadian Geotechnical Journal 46 (5), 595e606. Hatami, K., Bathurst, R.J., 2005. Development and verification of a numerical model for the analysis of geosynthetic reinforced-soil segmental walls under working stress conditions. Canadian Geotechnical Journal 42 (4), 1066e1085. Hatami, K., Bathurst, R.J., 2006. A numerical model for reinforced soil segmental walls under surcharge loading. ASCE Journal of Geotechnical and Geoenvironmental Engineering 132 (6), 673e684. Hatami, K., Khoury, C.N., Miller, G.A., 2008. Suction-Controlled Testing of SoilGeotextile Interfaces. In: GeoAmericas2008. The First Pan American Geosynthetics Conference & Exhibition, Cancun, Mexico, 1236, 262e271. IFAI, 2007. Geosynthetics (Formerly GFR). Industrial Fabrics Association International, Oseville, MN, USA (24)6.

Kayadelen, C., Tekinsoy, M.A., Taskiran, T., 2007. Influence of matric suction on shear strength behavior of a residual clayey soil. Environmental Geology 53 (4), 891e901. Khalili, N., Geiser, F., Blight, G.E., 2004. Effective stress in unsaturated soils, a Review with New Evidence. International Journal of Geomechanics 4 (2), 115e126. Koerner, R.M., 2005. Designing with Geosynthetics, fifth ed.. Pearson Prentice Hall, Upper Saddle River, NJ, USA. Lawson, C., 2005. Geosynthetic Reinforced MSE Walls and Slopes with Fine Grained Fills: International Perspectives. NAGS 2005/GRI-19 Conference, Las Vegas, Nevada, USA. Lu, N., Likos, W.J., 2004. Unsaturated Soil Mechanics. Wiley, New Jersey, USA. Miller, G.A., Khoury, C.N., Muraleetharan, K.K., Liu, C., Kibbey, T.C.G., 2008. Effects of solid deformations on Hysteretic soil water characteristic curves: experiments and simulations. Water Resources Research Journal 44, W00C06. doi:10.1029/ 2007WR006492. Miller, G.A., Hamid, T.B., 2007. Interface direct shear testing of unsaturated soil. Geotechnical Testing Journal, ASTM 30 (3), 182e191. Mitchell, J.K., Zornberg, J.G., 1995. Reinforced soil structures with poorly draining backfills. Part II: case histories and applications. Geosynthetics International 2 (1), 265e307. Musser, S.W., Denning, C., 2005. Deep Patch Road Embankment Repair Application Guide. United States Department of Agriculture Forest Service. Technology & Development Program 7700. Transportation Management, October 2005, 0577 1204-SDTDC. Navayogarajah, N., Desai, C.S., Kiousis, P.D., 1992. Hierarchical Single surface model for Static and Cyclic behaviour of interfaces. Journal of Engineering Mechanics, ASCE 118 (5), 990e1011. NCMA, 2002. Design Manual for Segmental Retaining Walls, second ed. National Concrete Masonry Association, Herndon, VA, USA. Oloo, S.Y., Fredlund, D.G., 1996. A method for determination of 4b for statically compacted soils. Canadian Geotechnical Journal 33, 272e280. Powel, W., Keller, G.R., Brunette, B., 1999. Applications for geosynthetics on forest service low-volume roads. Transportation Research Record 1652, 113e120. Rowe, R.K., Skinner, G.D., 2001. Numerical analysis of geosynthetic reinforced retaining wall constructed on a layered soil foundation. Geotextiles and Geomembranes 19 (7), 387e412. Sandri, D., 2005. Drainage Recommendations for MSE Walls Constructed with Marginal Fills. NAGS 2005/GRI-19 Conference, Las Vegas, Nevada, USA. Sharma, J.S., Fleming, I.R., Jogi, M.B., 2007. Measurement of unsaturated soil-geomembrane interface shear-strength parameters. Canadian Geotechnical Journal 44, 78e88. Sia, A.H.I., Dixon, N., 2007. Distribution and variability of interface shear strength and derived parameters. Geotextiles and Geomembranes 25 (3), 139e154. Skinner, G.D., Rowe, R.K., 2003. Design and behaviour of geosynthetic reinforced soil walls constructed on yielding foundations. Geosynthetics International 10 (6), 200e214. Skinner, G.D., Rowe, R.K., 2005. A novel approach to the bearing capacity stability of geosynthetic reinforced retaining walls constructed on yielding foundations. Canadian Geotechnical Journal 42 (3), 763e779. Stulgis, R.P., 2005. Full-scale MSE Test Walls. NAGS 2005/GRI-19 Conference, Las Vegas, Nevada, USA. Vanapalli, S.K., Fredlund, D.G., Pufahl, D.E., Clifton, A.W., 1996. Model for the prediction of shear strength with respect to soil suction. Canadian Geotechnical Journal 33, 379e392. Wheeler, S.J., Sivakumar, V., 1995. An elasto-plastic Critical State Framework for unsaturated soil. Géotechnique 45 (1), 35e53. Zhang, X., Lytton, R., 2006. Stress State Variables for Saturated and Unsaturated Soils. In: Proc.. of UNSAT 2006, 2380e2391.