Unsaturated Shear Strength Behavior under Unconsolidated Undrained Tests

Unsaturated Shear Strength Behavior under Unconsolidated Undrained Tests Majid Sokhanvar 1, Ir. Dr. Azman Kassim 2 1: Master of Engineering (Civil- Geotechnics), Faculty of Civil Engineering, Universiti Teknologi Malaysia (UTM), e-mail: maj_sok@yahoo.com 2: Associate Professor, Department of Geotechnics & Transportation, Faculty of civil Engineering, Universiti Teknologi Malaysia (UTM), e-mail: azmankassim@utm.my ABSTRACT In this study, unsaturated shear strength behavior of a tropical residual soil under different stress levels is investigated by using uncomplicated testing procedure. Existing triaxial tests use translation technique for determining unsaturated shear strength parameters but ordinary Unconsolidated Undrained triaxial tests were carried out due to lacking of the advanced testing unit in the laboratory. The Unconsolidated Undrained tests were carried out under different cell pressures at different suctions values to obtain the undrained compressive strengths of the specimens. Preliminary results of the consolidated isotropic undrained tests, show effective cohesion and effective angle of friction i.e., saturated shear strength parameters were 9 and 23, respectively. In Unconsolidated Undrained tests, the values of apparent shear strength at high stress levels range from 66.1 72.6. At low stress levels, the range of apparent shear strength values was obtained in between 53.1-57.5 kpa. The value of friction angle for the highest suction pressure tested in this study (300 kpa) was determined to be 9.9. This study illustrated that there is nonlinear relationship between the apparent shear strength and suction. KEYWORDS: Apparent Shear Strength, Friction Angle, Apparent Cohesion, Soil Water Characteristic Curve, Matric INTRODUCTION According to Rahardjo et al. (1995), residual soils cover more than three-quarts of the land area of Peninsular Malaysia. Many steep slopes in these residual soils often have a deep ground water table. Above those ground water tables, the soils are in unsaturated conditions. In unsaturated condition, the pore-water pressure is negative relative to atmospheric conditions. This negative pore-water pressure is called matric suction. It is well established that the stability of a natural or a cut slope in residual soils depends on the shear strength which is affected by the matric suction. The in-situ matric suction and the shear strength of soils are in turn affected by the climatic conditions, particularly rainfall distributions. In vadose zone, the zone above groundwater table, matric suction has a strong influence on shear strength behaviour. This extra attractive force i.e., matric suction is producing extra shear strength, i.e. apparent shear strength ( ) and friction angle with respect to suction ( ). The parameters and are named unsaturated shear strength parameters. Shear strength parameters are the key input parameters in any soil - 601 -

Vol. 18 [2013], Bund. C 602 stability analysis. According to Md. Noor (2011), unsaturated shear strength parameters are not constant variables, but vary with depth and suction. Laboratory works, despite of imposing extra time consuming and relatively higher expenses, are evidently providing the most appropriate mean for measuring the unsaturated shear strength parameters than empirical models. Problem Statement Existing laboratory tests for determining unsaturated shear strength parameters are base on measuring pore-air and pore-water pressures.i.e. translation technique. Those procedures are difficult to conduct, complicated, costly and time consuming. This study has been proposed a simple, low cost, and quick way for predicting unsaturated soil shear strength parameters by using uncomplicated testing procedure. Objectives of the Study In this study, three objectives are outlined as follows: 1) To determine the apparent shear strength ( )from unconsolidated undrained test at different stress levels. 2) To determine the friction angle ( ) from unconsolidated undrained test at different stress levels. 3) To investigate the relationship between apparent shear strength and stress level of the unsaturated residual soil. Scope and Limitation of the Study The results of this study were restricted to the undisturbed soil samples collected from 0.5m to 1m depth of the slope at latitude (+1 33 32.03 ) and longitude (+103 38 38.04 ) in front of P16 at Faculty of Electrical Engineering Universiti Teknologi Malaysia. In this thesis, consolidated isotropic undrained (CIU) tests were conducted based on BS 1377: Part 8:1990, Clause 7. Unconsolidated undrained (UU) triaxial tests used in this study were carried out following BS 1377: Part 7:1990, Clause 8. The only difference was that the unsaturated soil specimens were tested in their initial water contents and suctions. Lack of the advanced testing unit was the limitation of this project. LITERATURE REVIEW Shear Strength The shear strength of soil mass can be defined as the internal resistance per unit area that soil mass can offer to resist failure and sliding along any plane inside it. Terzaghi (1936) has presented the shear strength of soil ( ) in terms of effective cohesion and effective normal stress at failure as follows: = + ( ) tan (1) where and are the effective shear strength parameters (also called the effective cohesion and the angle of shearing resistance respectively); ( ) is the effective normal stress. In unsaturated condition, the shear strength of soil is influenced by matric suction. By extending Terzaghi s equation (1936), Bishop (1959) introduced a shear strength equation for unsaturated soil by combining two independent stress state variables, namely the net normal stress, ( ) and the matric suction, ( ) as follows: = + ( ) tan + ( ) [( )(tan )] (2)

Vol. 18 [2013], Bund. C 603 where, is a parameter dependent on the degree of saturation. The value of varies from 1 for fully saturated soil to 0 for totally dry condition. Due to some difficulties in determining the parameter, Fredlund et al. (1978) proposed a practical formulation for unsaturated soil shear strength. The shear strength contributions from net normal stress and matric suction are characterized by the effective angle of shearing resistance ( ) and the angle of frictional resistance due to the contribution of matric suction ( ) as follows: = + ( ) tan + ( ) tan (3) To retain the shear strength equation in its conventional form, the effective cohesion and the strength contribution due to matric suction can be represented by a term known as total cohesion ( ), since this shear strength vanishes when the soil is fully saturated; it is termed as apparent shear strength. = + ( ) tan (4) where, = ( ) tan = + ( ) tan (5) At failure condition, Equation 3 can be rewritten as follows: = +( ) tan + ( ) tan (6) where, ( ) and ( ) are the net normal stress and the matric suction at failure condition, respectively. By extending the classic Mohr-Coulomb envelope, the three dimensional representation of Equation 6 can be visualized in Figure 1. From the Figure 1, it is apparent that: = = cos = ( ) cos (7) ( ) = = = ( 2 + ) ( ) sin (8) 2 2 Figure 1: Extended Mohr-Coulomb failure envelopes for unsaturated soil According to Vanapalli and Fredlundd (1997), by equating the Equation 2.6 and 2.7 using Equation 2.8, the parameter can be estimated as follows:

Vol. 18 [2013], Bund. C 604 tan = (cos +sin tan ) ( ) + ( + ) tan ( ) (9) = + ( ) tan (10) where, is the undrained compressive strength obtained from the UU triaxial test defined as half of the maximum deviator stress ( ). Nonlinear Shear Strength Envelope The shear strength versus matric suction envelope has been found to be nonlinear by Gan et al. (1988). According to Donald (1956), the slope of the shear strength versus matric suction envelope can also be negative. The nonlinearity of the shear strength versus matric suction envelope in low suctions is related to desaturation and in high suctions is related to dilation (Figure 2.2). Where the suction is lower than AEV of the soil, the soil is at or near saturation condition and the ai-phase consists of a few occluded bubbles (Corey, 1957). The soil would be expected to behave as though it was saturated. In other word, the negative pore-water pressure acts throughout the predominantly water filled pores as in the saturated soil condition. Consequently an increase in matric suction produces the same increase in shear strength as does an increase in net normal stress. As a result, the same values are obtained for and. In matric suctions higher than the AEV, the soil starts to desaturate. The negative pore-water pressure does not act throughout the entire pores as in the saturated soil condition. Therefore, the contribution of matric suction towards the strength of the soil is less than the contribution of the net normal stress at the same stress level. In other words, the increase in shear strength with respect to matric suction is less than the increase with respect to net normal stress. As a result the value becomes less than at high matric suctions as observed in Figure 2. Figure 2: Relationship between soil water characteristic curve and shear strength versus matric suction envelope. According to Gan and Fredlund (1996), when dilation during shear occurs, the local density decreases. Dilation leads eventually to a disruption of the water phase. Disruption of the water phase reduces the effectiveness of matric suction in generating normal face between the particles of the soil hence, the shear resistance will decrease. On the other hand, dilation increases shear

Vol. 18 [2013], Bund. C 605 strength because of interlocking effects; on the other hand, dilation during shear tends to decrease the contribution of matric suction to shear strength because of a disruption of the water phase.irfan (1988) attributed the compressive behavior during shear above the critical normal stress to the destruction of the primary relict structures, the destruction of the secondary bonding, and the collapse of weak fabric elements. Therefore in high suctions, dilation below the critical stress level leads to decreasing the apparent shear strength. Increasing the apparent shear strength above the critical stress is due to compression behavior. METHODS Sampling Sampling is carried out in order to conducting laboratory tests. Samples obtained for testing should be representative of the ground from which they are taken. Since, undisturbed specimens provide better representation of field conditions than remolded specimens; in this study undisturbed soil samples were used. These samples were prepared from the residual soil from the slope in front of P16 at Faculty of Electrical Engineering of university Teknologi Malaysia, Johor. Undisturbed soil samples were taken based on BS5930: 1981. Firstly, standard metal tube (38mm diameter by 150 mm height) was pushed into the ground without rotation. Then, specimen slowly was taken outside to extracting and trimming. Samples were 38mm in diameter and 76 mm in height. All samples were taken from the same level of the ground in order to be similar as possible. Laboratory Tests From the sieve analysis, the soil was identified as fine grained. Furthermore, several laboratory tests were carried out to classify and to determine the index properties of the soil.based on the BSCS, the soil can be classified as MHS (i.e. sandy silt of high plasticity) with initial void ratio ( ) of 1.44. The initial water content ( ) was performed in accordance with BS 1377: Part 2: 1990, Clause 3. Its value was 43.9%.The soil water characteristic curve (SWCC) was obtained by pressure plate extractor test in accordance with the procedure in ASTM D6836-02 (ASTM, 2002) by Universiti Pertanian [Figure 4.1 (a)]. Data To create suction data for unsaturated soil specimens, soil water characteristic curve (SWCC) and moisture content concept were used. Since, SWCC was based on volumetric water content versus suction [Figure 4.1 (a)] and the equation between volumetric water content and moisture content is: = (11) 1+ where, = volumetric water content (%); = gravimetric moisture content (%), based on moisture content test in this study, initial moisture content ( ) is 43.9 %; = specific gravity of soil solids; in this study, =2.65 (based on laboratory tests conducted by Universiti Pertanian Malaysia); = void ratio, in this study initial void ratio( ) is 1.44 (based on laboratory tests conducted by Universiti Pertanian Malaysia). So, we can find target moisture content ( ) values related to target suctions using Equation 11 and SWCC. Since, weight of specimen is: = + (12) where, = weight of solid soils in the specimen; = weight of water in the specimen On the other hand;

Vol. 18 [2013], Bund. C 606 = =43.9% (13) Substitution of Equation 13 in Equation 12 resulted in: = (14) 1 + 0.439 In addition, = (15) Hence, target weight related to target moisture content is as follows: = + = + = 1 + 0.439 (1 + ) (16) Unsaturated soil specimens were put inside the oven for decreasing their weights to reach to their target weights. This work maybe needs to weigh the specimens in several times to obtain their target weights. The specimens in their target weights would have their target suctions. Triaxial Tests Triaxial tests in this study consist of consolidated isotropic undrained tests and unconsolidated undrained tests. For doing these triaxial tests, Digital 5: 25-3518 machine has been used. For these triaxial tests, soil specimens with 38mm in diameter and 76mm in height were used. Consolidated Isotropic Undrained Tests In this study, three unconsolidated isotropic undrained (CIU) tests using normal triaxial testing apparatus based on BS 1377: Part 8:1990, Clause7 were carried out to determine saturated shear strength parameters of the soil (, ). The CIU tests were carried out in three stages consist of saturation, consolidation and shearing stage. Unconsolidated Undrained Tests In this thesis, twenty unconsolidated undrained (UU) tests using normal triaxial testing apparatus, under different cell pressures (20, 50, 100, 200 kpa) and different suctions (20, 50, 100, 200, 300 kpa) were carried out to obtain the undrained compressive strengths( ) of unsaturated soil samples. Creating suction date for unsaturated soil specimens has been explained in part 3.3. The UU test procedure outlined for saturated soils (BS 1377: part 7:1990; clause8) was applied for the UU tests on unsaturated soils in this project. The only difference was that the unsaturated soil specimen was tested in its initial water content and suction i.e. the initial matric suction was not changed prior to commencing the test. Despite of the fact that the suction may decrease due to the increase of pore-water pressure during shearing stage of the soil sample, it is assumed that the variation of suction is negligible. Unsaturated Shear Strength Parameters According to Vanapalli and Fredlund (1997), the angle of frictional resistance due to the contribution of matric suction ( ) and strength due to the matric suction, apparent shear strength ( ), for undrained conditions were estimated from these formulas: tan = (cos +sin tan ) + ( + ) tan ( ) ( ) = + ( ) tan

Vol. 18 [2013], Bund. C 607 where, ( ) is the undrained compressive strength that was obtained from UU test; & are the effective shear strength and effective friction angle respectively, that were obtained from CIU tests; ( ) is the matric suction of soil sample (Values of suctions in this study are 20, 50, 100, 200, 300 kpa); ( ) is the matric suction at the failure condition[( ) = ( )]. RESULTS Summary of the soil properties extracted from SWCC (Figure 3) is tabulated in Table 3. (a) (b) Figure 3: Soil Warter Characteristic Curve (SWCC): (a) based on volumetric water content versuss suction; (b) based on gravimetric moisture content versus suction Table 1: Soil properties extracted from SWCC Soil properties Volumetric water Condition conten (%) ( ) Saturation Air-entry value Residual condition 0.44 0.405 0.2777 0.1 9 50 Based on the conducted CIU tests on saturates specimens, the effective cohesion ( ) is 9kPa and the effective internal friction angle ( ) is 23. Table 2 shows results for UU tests.

Vol. 18 [2013], Bund. C 608 Sample Table 2: Undrained compressive strengths for UU tests for several ranges of suctions and cell pressures Cell pressure ( ) ( ) Undrained compressiv e strength ( ) Sample Cell pressure ( ) ( ) Undrained compressiv e strength ( ) 1A 20 20 39.5 1C 100 20 92 2A 20 50 60 2C 100 50 124 3A 20 100 93 3C 100 100 158 4A 20 200 95 4C 100 200 163 5A 20 300 91 5C 100 300 164 1B 50 20 59.5 1D 200 20 157 2B 50 50 86 2D 200 50 197 3B 50 100 119 3D 200 100 232 4B 50 200 121 4D 200 200 235 5B 50 300 118 5D 200 300 238 Results of unsaturated shear strengths have been brought in Table 3. Table 3: Unsaturated shear strength parameters (, ): (a) Specimens under cell pressure 20 kpa; (b) Specimens under cell pressure 50 kpa; (c) Specimens under cell pressure 100 kpa; (d) Specimens under cell pressure 200 kpa Sample ( ) ( ) ( ) 1A 20 23.40 17.66 2A 50 23.96 31.22 3A 100 23.78 53.07 4A 200-8.81 54.39 5A 300 8.11 51.74 (a) Sample ( ) ( ) ( ) 1B 20 24.61 18.16 2B 50 28.10 35.70 3B 100 25.89 57.54 4B 200 14.00 58.87 5B 300 9.07 56.88 (b) Sample ( ) ( ) ( ) 1C 20 25.28 18.45 2C 50 31.49 39.63 3C 100 6.10 62.13 4C 200 15.76 65.44 5C 300 10.78 66.10 (c) Sample ( ) ( ) ( ) 1D 20 26.62 19.02 2D 50 36.13 45.50 3D 100 30.82 68.66 4D 200 17.13 70.65 5D 300 11.98 72.64 (d)

Vol. 18 [2013], Bund. C 609 For convenience in survey, the results of Table 3 have been depicted in schematic in Figures 4 (a) and (b). (a) (b) Figure 4: Apparent shear strength ( ) versus suction ( ) : (a) in low cell pressures (20, 50 kpa); (b) in high cell pressures (100, 200 kpa) DISCUSSION Figures 4 (a) and (b) illustrate the nonlinear relationship between the apparent shear strength and suction. This nonlinear behavior is in agreement with the results of Gan et al. (1988). According to these Figures, in a similar suction with increasing the cell pressure the apparent shear strength increases. Comparison of the values of apparent shear strength of Figure 4(a) with Figure 4(b), shows that apparent shear strengths in high stress levels (envelope 3, 4) are more than their corresponding values in low stress levels (envelope 1, 2). The value of maximum apparent shear strength in high stress levels is 66.1 72.6. For low stress levels, value of maximum apparent shear strength is 53.1 57.5. In high suction pressures (200-300 kpa) in envelope 1, and 2 apparent shear strength tends to decrease. Increment of this decreasing in envelope 1 is more than envelope 2. In high suctions apparent shear strength in envelope 3 tends to be constant but in envelope 4 it tends to increase. These differences in soil behavior are related to dilation effect. According to Gan and Fredlund (1996), the local density decreases, when dilation during shear occurs. Finally, dilation leads to a disruption of the water phase. Disruption of the water phase reduces the effectiveness of matric suction in generating normal face between the particles of the soil. Hence, the shear resistance will decrease. Figure 5 illustrates dilation direction of the results of this study. According to Irfan (1988), compressive behavior during shear above the critical normal stress level (i.e., envelope 3) is due to the destruction of the primary relict structures, the destruction of the secondary bonding, and the collapse of weak fabric elements. Therefore in high suctions, dilation below the envelope 3 leads to decreasing the apparent shear strength. Increasing the apparent shear strength in the envelope 4 is due to compression behavior.

Vol. 18 [2013], Bund. C 610 Figure 5: Apparent shear strength envelopes with direction of dilation increasing According to Corey (1957), in low matric suctions, where the suction is lower than AEV, the soil is at or near saturation condition and the air-phase consists of a few occluded bubbles. The soil would be expected to behavee as though it was saturated. Consequently an increase in matric suction produces the same increase in shear strength as does an increase in net normal stress. As a result, the same values are obtained for φ and φ. According to Table 4.3, with interpolation for the suction =9 kpa in AEV, angle of frictional resistance due to the contribution of matric suction (φ ) is 23.16 which approximately equals to the value of effective friction angle (φ = 23 ). The soil startss to desaturate in matric suctions higher than the AEV. The negative pore-water pressure does not act throughout the entire pores as in the saturated soil condition. Hence, the contribution of matric suction towards the strength of the soil is less than the contributionn of the net normal stress at the same stress level. In other words, the increase in shear strength with respect to matric suction is less than the increase with respect to net normal stress. So, φ value becomes less than φ at high matric suctions as observed in Figures 4 (a) and (b). The value of φ for the highest suction pressure tested in this study (300 kpa) is equal φ =9.9. The slight inconsistency associated with sample preparation can be the main reason contributed in some increasing of φ values during increasing of suction pressures. CONCLUSION The following conclusionss can be drawn from this study: 1) The values of apparent shear strength at high stress levels (100 200 ) range from 66.1 72.6. At low stress levels (20 50 ), the range of apparent shear strength values was obtained in between 53.1 57.5. 2) Angle of frictional resistance due to the contribution of matric suction ( ) in air-entry value (AEV) is 23.16 which approximately equals to the value of effective friction angle obtained from consolidated isotopic undrainedd test ( = 23 ). At highh suction pressures,

Vol. 18 [2013], Bund. C 611 air enters the soil (i.e., soil is de-saturated) causing the decreases to a constant value lower than. The value of for the highest suction pressure tested in this study (300 kpa) is equal =9.9. 3) There is a nonlinear relationship between the apparent shear strength and suction in different cell pressures. The shear strength increases as suction and cell pressure increases. In high suctions (i.e., 200 to 300 kpa), increasing the dilation below the critical stress level (i.e., 100 kpa cell pressure in this study) leads to decreasing the apparent shear strength. Above the critical stress level in high suctions, apparent shear strength increases due to the compression behavior of unsaturated soil. REFERENCES 1. ASTM (2002). Standard Test Methods for Determination of the Soil Water Characteristic Curve for Desorption Using Pressure Extractor, D6836-02. West Conshohocken, PA, ASTM. 2. Bishop, A. W. (1959). The Principal of Effective Stress. Technische Ukebland.106 (39), 859-863. 3. British standards Institution (1981). Soil Sampling, Site Investigation. BS5930. London: British Standards Institution. 4. British standards Institution (1990). Consolidated-Undrained Triaxial Compression Test with Measurement of Pore Pressure. BS1377-8. London: British Standards Institution. 5. British standards Institution (1990). Determination of the undrained Shear Strength in Triaxial Compression without Measurement of Pore Pressure. BS1377-7. London: British Standards Institution. 6. British standards Institution (1990). Determination of Moisture Content. BS1377-2. London: British Standards Institution. 7. Corey, A. T. (1957). Measurement of Water and Air Permeability in Unsaturated Soils. Proc. of the Soil Science Society of America, 21(1),7-10. 8. Donald, J. B. (1956). Shear Strength Measurements in Unsaturated Non-Cohesive Soils with Negative Pore Pressures. Proceedings of the 2 Australia and New Zealand Conference on Soil Mechanics and Foundation Engineering, Christchurch, New Zealand, 200-205. 9. Fredlund, D. G., Morgenstern, N. R., and Widger, R. A. (1978). The Shear Strength of Unsaturated Soils. Canadian Geotechnical Journal. 15, 313-321. 10. Gan, J. K. M., and Fredlund, D. G. (1996). Shear Strength Characteristics of Two Saprolitic soils. Canadian Geotechnical Journal. 33(4), 595-609. 11. Gan, J. K. M., Fredlund, D. G., and Rahardjo (1988). Determination of the Shear Strength Parameters of an unsaturated Soil using the Direct Shear Test. Canadian Geotechnical Journal. 25, 500-510. 12. Irfan, T. Y. (1988). Fabric Variability and Index Testing of a Granitic Saprolite. Proceeding of 2 International Conference on Geomechanics in Tropical Soils, Dec. 12-14, Singapore, 1, 25-35. 13. Md.Noor, M. J. (2011). Understanding Rainfall-Induced Landslide. Universiti Teknologi Mara, Shah Alam: UiTM Press.

Vol. 18 [2013], Bund. C 612 14. Rahardjo, H., Lim, T. T., Chang, M. F., and Fredlund, D. G. (1995). Shear Strength Characteristic of a Residual Soil. Canadian Geotechnical Journal. (32), 60-77. 15. Terzaghi, K. (1936). The Shear Resistance of Saturated Soils. Proceedins of 1 st International Conference of Soil Mech.Found. Eng. Cambridge.1, 54-56. 16. Vanapalli, S. K., and Fredlund, D. G. (1997). Interpretation of Unsaturated Shear strength of Unsaturated Soils in terms of Stress State Variables. Proceedings of the 3 Brazilian Symposium on Unsaturated Soils, Tacio de Campos, Vargas, 35-45. 2013 ejge