EFFECT OF CAPILLARITY ON COEFFICIENT OF ELASTIC UNIFORM COMPRESSION OF SILTY SAND BEDS

Abstract

The most critical step in the analysis and design of foundations subjected to dynamic loads is the correct evaluation of stiffness parameters of the soil. The dynamic stiffness of the soil is often expressed in terms of coefficient of elastic uniform compression, Cn or the shear modulus, G. A review of literature reveals that these stiffness parameters do not have unique value for a soil. Alarge number of factors influence them. Capillarity in the soil is one such factor, which can affect the value of Cn or G considerably. However, it has not received due consideration so far. In India, the cyclic plate load test method is very widely used to estimate the value of Cq in situ. In certain areas, like the Indo-Gangetic plain of Northern India, the test often gets carried out on capillary affected soil bed. The soil in this region is silty sand/fine sand and the water table remains at shallow depths (about 2-4 mbelow ground level). The results of any test conducted above the position of water table in such cases are likely to be affected by the effect of capillarity. However, in the absence of adequate knowledge, no consideration to the effect of capillarity is currently being given in the interpretation of such test data. In view of this, a comprehensive programme of experimental study on the effect of capillarity on Cu was planned. The following broad objectives were set out for the investigation. 1. To examine the effect of capillarity on C, obtained from the results of cyclic plate load test conducted on silty sand type of soil and 2. To suggest a procedure of interpreting the results of cyclic plate load test conducted on a soil bed affected by capillarity. The experimental investigation consists of a number of cyclic plate load tests on prepared beds of silty sand. A review of literature reveals that the behaviour of a sand deposit is very sensitive to the loading history of the deposit. Therefore, tests were conducted on normally consolidated and overconsolidated (preloaded) sand beds. Further, the thickness of capillary zone below the test plate was a variable in the study. The normally consolidated sand beds were prepared in a large tank of size 1480 mm x 1480 mm x 1500 mm. The tank is provided with two side tanks (500 mm x 600 mm x 1500 mm) which facilitate raising, lowering or maintaining a desired water level in the test tank. The tank was filled with dry sand using rainfall method and it was then submerged by raising water level in the test tank to the top of the sand bed. The water level was then lowered and maintained at a desired elevation by manipulating the water level in the side tanks. Thus a capillary zone of desired thickness was created between the water level and the top of test bed. A plate of size 300 mm x 300 mm was placed at the centre of lest bed and the test was conducted using plate load test device. A load was first applied to the plate and the total settlement of the plate was noted. This load was then released and the elastic rebound of the plate was recorded. The load was increased in steps and the process of loading and unloading was repeated. The value of Co was determined from the slope of the curve between loading intensity and the elastic rebound. Thus, values of Cu for different thickness of capillary zone below the test plate were obtained. A few tests were also carried out on submerged and dry condition of test bed for the purpose of comparison of test results. The results of these tests are presented and discussed to bring out in (a) the effect of capillarity on C (b) the role of thickness of capillary zone on Cu . For the study on the effect of capillarity on C in overconsolidated sands, a circular test tank of diameter 590 mm and height equal to 890 mm having provisions for creating desired capillary conditions in the soil was used. The tank was first filled with dry sand using rainfall method. The effect of overconsolidation was achieved by preloading the test bed.For this purpose a, plate of diameter slightly smaller than that of the testing tank was placed at the top of the test bed which was used as a base plate to transfer the preload.Thus the bed was subjected to a desired preloading pressure (in the present investigation a constant preload intensity of 50 kN/m2 was used). The test bed was then subjected to a cycle of submergence and drainage by raising and lowering of water level in the test tank. The capillary zone of desired thickness in the test bed was created using a procedure similar to that adopted in normally consolidated sand bed case. The applied preload and the base plate were then removed from the top of the sand bed and cyclic plate load test was carried out using a test plate of diameter 120 mm. A number of tests were conducted varying the thickness of capillary zone below the test plate. A few tests were also performed in submerged and dry condition of the test bed. A comparison of the values of CH under capillary condition (C ) with that under submerged condition (C . ) was made to study the effect of capillarity on Ca in overconsolidated (preloaded) sands. A review of literature suggests that mean effective principal stress ao in the soil is the factor which affects the value of Cu the most. Capillarity in the soil increases the value of o"o , thereby changing (increasing) the value of C . Similarly, overconsolidation induces additional in situ stress thus affecting ct . In order to incorporate the additional stresses due to IV capillarity and/or overconsolidation in the interpretation of the test data, appropriate values of ao under different conditions of test beds (i.e. submerged/capillary -normally consolidated/ overconsolidated) have been established. Experimentally obtained values of Cu under capillary conditions of soil (Cuco values) were then reduced (extrapolated) to give a value of C corresponding to mean effective principal stress equal to that occuring under submerged condition of soil. The extrapolated values of Cu thus obtained for submerged condition of soil were termed as C .(analytical) and were compared with that obtained experimentally for submerged condition of soil i.e. Cmui (experimental). The comparison of CatuA (analytical) with the CmaA (experimental) made it possible to evolve a procedure of accounting the effect of capillrity on C . Based on the above investigations carried out the following significant conclusions have been drawn: 1. Capillary affected sand beds exhibit higher values of C in comparison to submerged condition of the test bed - in both the normally consolidated and overconsolidated (preloaded) sands. 2. The ratio C /C . is found to increase with the thickness of capillary ucap usub » •> bed. For a 900 mm thick capillary bed the ratio C /C . is about r J neap usub 3.5 in normally consolidated sands. 3. The increase in ratio C /C . with the thickness of capillary bed ucap usub r J (expressed in dimensionless form as Hc /B; Hc = thickness of capillary bed, B = width of the test plate) is found to be less in preloaded sands than that in normally consolidated sands. 4. The effect of capillarity in the soil may vanish due to submergence of the bed. Therefore, the value of Cu obtained from a cyclic plate load test carried out on a capillary bed (Cnco ) should be moderated to obtain the value of C for a possible submerged condition of the soil (CiuuA ) as the safe performance of the foundation under the submerged condition also needs to be ensured. A procedure of moderating the test value C to obtain C has been suggested accounting for the change in ucap usub mean effective principal stress due to the possible submergence. The procedure has been found to predict well the value of Cmul> when applied to the results of test carried out in the present investigation