Advances in Transportation Geotechnics Ellis, Yu, McDowell, Dawson & Thom (eds) 2008 Taylor & Francis Group,

p, London, ISBN 978-0-415-47590-7

Large-scale direct shear testing of geocell reinforced soil

C.S. Wang & Z.X. Hou

ABSTRACT: The shear strength test results of geocell reinforced soil that are carried out with large-scale direct shear tests are reported in this paper. Three types of specimens which are unreinforced silty gravel soil, geocell reinforced silty gravel soil, and geocell reinforced cement stabilizing silty gravel soil were investigated. The comparison of large-scale shear test results and triaxial compression test results involving unreinforced silty gravel soil was conducted to evaluate the influence of testing methods. The test results showed that the unreinforced soil and geocell reinforced soil gave similar nonlinear features on the behavior of shear stress and displacement, while there was a quasi-elastic characteristic for the geocell reinforced cement stabilizing soil in the case of higher normal stress. With the reinforcement of geocell, the cohesive strength of silty gravel soil was found considerably increased, but the friction angle of the geocell reinforced soil did not change obviously.

INTRODUCTION

Geocell reinforced soil has gained a considerable popularity in highway subgrade engineering in the recent past. It has been found to be useful on reinforcement of embankment, steep slopes, retaining walls and abutment backfills because of improving the loadbearing capacity, increasing the strength and stiffness, reducing settlement, and saving cost and time in construction (Bathurst & Jarrett 1988, Cancelli et al. 1993, Sekine et al. 1994, Bathurst & Crowe 1994, Mhaiskar & Mandal 1996, Rajagopal et al. 1999, Latha et al. 2006, Mengelt & Edil 2006). Since the first use of cellular confinement systems to improve road bases over weak subgrades (Bathurst & Jarrett 1988), the stability of geocell reinforced soil has required considerations on the interface shear strength between soil and geocell. A test method for determining the interface shear capacity of geosynthetic reinforced soil was first introduced by ASTM D 5321-92 (1992), Standard Test Method for Determining the Coefficient of Soil and Geosynthetic or Geosynthetic and Geosynthetic Friction by the Direct Shear Method and currently is the revised edition ASTM D 5321-02 (2002). The methods of test introduced above are now used to provide the shear parameters of a geosynthetic against soil, or a geosynthetic against another geosynthetic, under a constant rate of deformation. It is applicable for all kinds of geosynthetics. However, limited investigations were reported in literature on

the shear behavior of geocell reinforced soil. (Bathurst & Karpurapu 1993, Rajagopal et al 1999, Cheng & Jonathan 1993, Yan, et al. 2006, Jing, et al. 2006). This paper reports the results from serials of largescale direct shear tests carried out on silty gravel soil confined with or without multiple geocells. The comparisons of large-scale shear test with triaxial compression test for the same type of soil are conducted to evaluate the influences of testing method on the shear strength indices as well. Meanwhile, the testing results of geocell reinforced cement stabilizing silty gravel soil are studied and analyzed as well.

EXPERIMENT METHOD

2.1 Test instrumentations The general test arrangement of large-scale direct shear test is shown in Figure 1. The test system was built in the structure engineering laboratory at South China University of China. The shear box includes two stacked parts. The inner size of each part has a dimension of 500 mm 500 mm 200 mm (length width height). Each part can be disassembled easily for installing and unloading soil specimen. During the directly shearing test, the bottom half was fixed on the ground, and the top half was driven by a 30-tonne capacity horizontal loading system with force transducers collecting data constantly by computer software.

759
2008 Taylor & Francis Group, London, UK

This soil can be classified as GM according to the Unified Soil Classification System. 2.2.2 Geocell The geocell reinforcement material used in the test was a commercially available industrial product manufactured by Netlon China Limited. It was polypropylene stripes welded together to give an open-cell construction that had a cell area of 625 cm2 and depth of 10 cm. The cellular materials came in panels expanded to cover an area of 4 m 12.5 m. The properties of geocell are summarized in Table 2.
Figure 1. General arrangement of large-scale direct shear test.

2.2.3 Cement Portland cement has been used as stabilizer to increase the strength of coarse soil for a long time. In this study, the cement stabilized soil was used as backfill materials together with geocell reinforcement in the abutment. Table 3 shows properties of the cement used in the study. 2.3 Test procedures 2.3.1 Geocell reinforced soil The test method for geocell reinforced soil in this paper was large-scale direct shear test. The procedures of large-scale direct shear test were described as following: First of all, the large-scale direct shear device must be calibrated to measure the internal resistance to shear inherent to the device. Then, the geocell reinforced soil specimens were prepared. The geocell reinforced soil specimens were molded within the shear box which had a volume of 0.1 m3 (500 mm long 500 mm wide 400 mm high). Generally, three or four specimens were required in one direct shear test. After finishing the specimen molding, a confining stress was applied vertically to the specimen, and the upper box was pushed horizontally at a rate of 1 mm/min. until the sample failed, or got to a general strain of 5%.Three specimens were tested at varying confining stresses to determine the shear strength parameters including the soil cohesion (c) and the friction angle (). 2.3.2 Unreinforced soil Large-scale direct shear tests and triaxial compression tests were carried out to investigate the shear feature of unreinforced soil and the influence of different testing methods. The procedure of unreinforced soil large-scale direct shear test was similar to the geocell reinforced soil large-scale direct shear test except without paving the geocell material. The triaxial compression tests were carried on through a LoadTrac II triaxial compression test apparatus made by Geocomp Company in the U. S. The size of each specimen was 61.8 mm 132.7 mm

Figure 2. Particle size distribution of backfill soil.

The vertical loading system was driven by a 100-tonne capacity MTS hydraulic actuator. There were two steel plates placed on the top of the soil specimen to bear the vertical load. Two lines of steel linear bearings were seated between these two steel plates in order to reduce friction of plates caused by the horizontal movement of shear box. The normal load was measured by an electronic load transducer. Two 0.01 mm-sensitivity dial indicators were located on the left side and right side of top-half shear box to measure the horizontal displacement manually. 2.2 Test materials 2.2.1 Soil The soil used in the tests was well graded with a maximum particle size of 40 mm, coefficient of curvature of 2.98, coefficient of uniformity of 20.6 and fines content less than 4% by weight. The particle size distribution is shown in Figure 2. The maximum dry density of the soil is 2.049 g/cm3 . The optimum water content is 9.057%. Soil properties are summarized in Table 1.

760
2008 Taylor & Francis Group, London, UK

Table 1.

Soil properties. Optimum water content % 9.057 Maximum dry density g. cm3 2.049 Unconfined compressive strength MPa 0.487

Liquid limit 38.8

Plastic Limit 26.2

Plasticity index 12.6

CBR % 19.8

Table 2.

Geocell properties. Unit elongation % 7.6 Breaking elongation % 9.8 Stripe tensile strength kN.m1 311 Welding tensile strength kN.m1 130

Product type TGLG-PP-100-500

Length mm 250

Width mm 25

Height mm 10

Thickness mm 0.07

Table 3.

Properties of Portland cement. Coagulating time Compressive strength Rupture strength R3 R28 MPa MPa 5.6 8.3

Cement type

Fineness %

Specific gravity Initial set Final set Invariability R3 R28 g.cm3 min. min. (Boiling) MPa MPa 3.1 212 280 Eligibility 31.3 59.2

Portland cement 3.8

Large than 0.08 mm sieve

(diameter height). Consolidated Undrained test method was taken with the triaxial apparatus. Three specimens were tested for each type of soil with confining pressures of 50 kPa, 200 kPa, and 400 kPa respectively. 2.3.3 Geocell reinforced cement stabilizing soil The test method for geocell reinforced cement stabilizing soil in this paper was large-scale direct shear test. A dosage of 5% cement (by dry soil weight) was added into the well soaked soil and mixed fully with the soil before the soil mixture was molded to the specimen of large-scale direct shear test. Other procedures were the same as described in section of geocell reinforced soil. 2.3.4 Unreinforced cement stabilizing soil For the purpose of investigating the influence of geocell on cement stabilizing soil, several unreinforced cement stabilized soil triaxial compression tests were carried on. The test procedures are similar to above. A total of 15 tests which include 9 large-scale direct shear tests and 6 triaxial compression tests were carried out in this study. Considering silty gravel and cement stabilizing silty gravel which reinforced with geocell or

not. If the soil is stabilized with cement, the dosage of cement is 5%.

3 TEST RESULTS AND COMPARISONS 3.1 Large-scale direct shear tests

The shear stress-displacement behaviors observed from large-scale direct shear tests on different reinforcement soil are shown in Figures 3a, b,c. The results of unreinforced soil show that the shear stress increases with increasing shear displacement when the normal stress is 200 kPa. As increasing the normal stress to 400 kPa and above, the shear stress-displacement response is softening pattern which shows an increase in shear stress with increasing shear displacement in the beginning and a decrease in shear stress with increasing shear displacement in the end. In the case of geocell reinforced soil (Fig. 3b), all the curves of shear stress-displacement with different normal stresses appear in softening pattern. There is a characteristic that the shear stress increases to a peak with increasing the shear displacement from zero to a certain magnitude and thereafter, it decreases gradually

761
2008 Taylor & Francis Group, London, UK

Figure 4. Shear capacity envelops.

Figure 3. Shear stress- displacement behavior of different reinforcement soil.

to the residual shear stress with the displacement on the increase. The reason for this correspondence is that there is a considerable degree of interlocking on dense soil and there is an additional friction on the interface between soil and geocell reinforcement. Therefore, before shear failure can take place this interlocking

and additional friction must be overcome in addition to the frictional resistance at the points contact. As the interlocking and additional friction is progressively overcome, the shear stress necessary for additional deformation decreases. However, the result of geocell reinforced cement stabilizing soil shows different feature when the normal stress comes up to 1000 kPa (Fig. 3c). The shear failure takes place with a relatively high shear stress and low deformation. It is clear from the data in this figure that there is a quasielastic characteristic on the behavior of shear stress and displacement. This is mainly due to the 5% cement stabilizing soil possessing a considerable degree of rigidity. Each plot in Figure 3 shows that there is a transformation in shear stiffness. It increases with increasing the normal stress for all tests. The linear shear capacity envelops deduced from the peak or maximum recorded shear stress versus normal stress data for different reinforcement soils plot together as illustrated in Figure 4. It shows that there is a good correspondence for unreinforced soil and geocell reinforced soil. But in the case of geocell reinforced cement stabilizing soil, there is some scatter of the data points about the linear regression line. This is due to the unavoidable small variations in the setup of the shear surface between the two parts of shear boxes exactly at the interface between geocell reinforcement soil and soil. The correlation coefficient of the linear regression on geocell reinforced cement stabilizing soil is 0.91, and on unreinforced soil and geocell reinforced soil both are 0.99. Based on the linear regression lines presented in Fig. 6 the parameters of shear strength which include cohesion c and friction angle can be derived. The results show that there is a considerable increase on cohesion from unreinforced soil to geocell reinforced soil. Comparing the cohesion of unreinforced soil with that of geocell reinforced soil, cohesion of geocell reinforced soil increases 244%, and the cohesion of geocell

762
2008 Taylor & Francis Group, London, UK

Figure 6. Failure strength envelopes.

Figure 5. Triaxial behavior of principle stress difference versus vertical strain.

reinforced cement stabilizing soil is 11 times as big as that of unreinforced soil. However, the friction angle does not change dramatically. Based on the data presented on Figure 4 it can be argued that the geocell reinforced soil develops a large amount of cohesion on the interface shear strength but the friction angle. 3.2 Comparison of results with triaxial compression tests

Figure 5 shows the behavior of principle stress difference versus vertical strain of consolidated-undrained triaxial compression tests for unreinforced soil and unreinforced cement stabilizing soil. Comparing the results of unreinforced soil with that of unreinforced cement stabilizing soil, it seems that these two types of soil have different features of stress- strain. With the same vertical strain of 2%, the principal stress difference of unreinforced cement stabilizing soil reaches the peak point of the curve, but the principal stress difference of unreinforced soil increases

with increasing vertical strain and after that, the stiffness turns to less. Nevertheless, the linear shear failure strength envelopes deduced from the maximum recorded effective principal stress ratio was plotted as Figure 6. The p and q in this diagram are the mean effective normal stress and effective shear stress. It appears a very good correlation for both unreinforced soil and unreinforced cement stabilizing soil. Based on the regression linear lines and test data, the consolidated undrained shear strength parameters are interpreted and summarized on Table 4. It is observed that the effective cohesion of unreinforced cement stabilizing soil is nearly 6 times as big as that of unreinforced soil while the frictional angle increases 19.2%. Comparing the results of triaxial compressive shear test to that of large-scale direct shear test on unreinforced soil, it shows that there is a significant difference in the value of the cohesion. The triaxial compressive shear tests gave a cohesion of 22.2 kPa, and the large-scale direct shear tests exhibited a relatively higher result of 40 kPa. However, there is only a marginal decrease in the friction angle which is 34.7 and 36.6 respectively. This result can be attributed to the different specimen size that was used in the tests. With a dimension of 500 mm 500 mm 400 mm (length width height), the large-scale shear test specimen has much more area of soil confined by shear box. Therefore, the degree of interlocking in the soil can be developed to a higher lever which results in the increase of cohesion. Based on the test results in this study, it may be concluded that the cohesion of soil is mainly responsible for the increase of shear strength due to the influence of test methods. Comparing to the results of triaxial compression test, the large-scale direct shear test provides a higher cohesion and an approximate friction angle.

763
2008 Taylor & Francis Group, London, UK

Table 4.

Shear strength parameters with different test methods. large-scale direct shear test triaxial compression test c kPa 22.2 130.36 34.7 42.7

Type of reinforcement soil Unreinforced soil Geocell reinforced soil Geocell reinforced cement stabilizing soil Unreinforced cement stabilizing soil

c kPa 40 137.6 433.9

36.6 37.6 43.6

CONCLUSIONS

This paper reports the shear strength test results of geocell reinforced soil carried out with large-scale direct shear tests. Unreinforced silty gravel soil, geocell reinforced silty gravel soil, and geocell reinforced cement stabilizing silty gravel soil were investigated. The comparisons of large-scale shear tests with triaxial compression tests for the same type of soil are conducted to evaluate the influences of testing methods on the shear strength indices as well. The following conclusions can be drawn from the results reported here: (1) The use of large-scale direct shear test device provides a valuable method to gain the shear strength parameters for geocell reinforced soil. (2) The behaviors of stress-strain on both unreinforced soil and reinforced soil with large-scale direct shear tests are nonlinear. When the normal stress comes up to 1000 kPa, there is a quasi- elastic characteristic on the behavior of shear stress and displacement for the geocell reinforced cement stabilizing soil. The shear stiffness increases with increasing the normal stress for all tests. (3) With the reinforcement of geocell, the cohesive strength of silty gravel soil considerably increases and it increases much more highly for the cement stabilizing soil. (4) The friction angle of the geocell reinforced soil does not change obviously. The failure envelops of both the unreinforced soil and the reinforced soil are nearly parallel to each other. REFERENCES
Bathurst, R.J. & Jarrett, P. 1989. Large-scale model tests of geocomposite mattresses over peat subgrades. ASTM

Transportation Research Record 1188, Transportation Research Board, Washington, D. C.: 2836 Bathurst, R.J. & Kappurapu, R. 1993. Large-scale triaxial compression testing of geocell-reinforced granular soils. Geotechnical Testing Journal, 16: 296303 Bathurst, R.J. & Crowe, R.E. 1994. Recent case histories of flexible geocell retaining walls in North America. The International Symposium on Recent Case Histories of Permanent Geosynthetic-reinforced Soil Retaining Walls, Tokyo, Japan Cancelli, A. et al. 1993. Index and performance tests for geocells in different applications. ASTM Special Technical Publication, 1190: 6475 Cheng, S.C. & Jonathan. 1993. Symposium on geosynthetic soil reinforcement testing procedures. ASTM SpecialTechnical Publication, 1190: 243 Jing, H. & Yu, M. 2006. Application of geocell to strengthen expressway subgrade in desert. Journal of Changan University (Natural Science Edition), 26: 1518 Latha, G. et al. 2006. Experimental and theoretical investigations on geocell-supported embankments. International Journal of Geomechanics, 6: 3035 Mengelt, M. et al. 2006. Resilient modulus and plastic deformation of soil confined in a geocel. Geosynthetics International, 13: 195205 Mhaiskar, S.Y. & Mandal, J.N. 1996. Investigations on soft clay subgrade strengthening using geocells. Construction and Building Materials, 10: 281286 Rajagopal, K. et al. 1999. Behavior of sand confined with single and multiple geocells. Geotextiles and Geomembranes. 17: 171184 Sekine, E. et al. 1994. Study on properties of road bed reinforced with geocell. Quarterly Report of RTRI (Railway Technical Research Institute) (Japan), 35: 2331 Yan C. et al. 2006. Testing and application of geocell in loess slopes of highway. Chinese Journal of Rock Mechanics and Engineering, 25, n SUPPL. 1

764
2008 Taylor & Francis Group, London, UK