The following sections describe the experimental method adopted in this paper. The definitions of the parameters used throughout this paper are given in Notation.
The schematic diagram of the hollow cylinder torsional shear apparatus is represented schematically in Fig. 1. The apparatus was modified from the previous one (Toyota et al., 2001a) in order to reduce the stress and strain non-uniformity. The new size of the specimen is the inner diameter, di =50 mm, the outer diameter, do =80 mm and the height, h =160 mm. The changes in the inner and outer diameters during experiment are calculated through the volume changes, which are measured by differential pressure transducers (DPT) in the hollow cylinder and the inner cell, respectively (see Fig. 1). Water may begin to cavitate as the water pressure approaches zero absolute pressure. Therefore, the axis-translation technique (Bishop and Donald, 1961), which is used to measure pore water pressure less than zero absolute was adopted in this system using a ceramic disk. Load cells, which are used to measure axial load and torque, are installed inside the triaxial cell to remove the influence of cell pressure and friction of the rod. The measuring and automatic control system is the same as that shown in Toyota et al. (2001b).
The cohesive soil used is a lateritic soil, which is referred to as Yoneyama sandy silt. The average sand, silt and clay content are 19.4%, 57.0% and 23.6%, respectively. Plasticity index tests give a liquid limit of 52.4% and a plastic limit of 29.9%. The soil and de-aired water were mixed and stirred well to make slurry, the water content of which was about 80%. Pre-consolidation using this slurry was performed one-dimensionally in a mold under a vertical pressure of 45kPa. The water content of the pre-consolidated soil block is about 40%. The hollow cylinder specimen was made by trimming from this soil block. The drying portion of the moisture characteristic curve obtained from the pressure plate method is shown in Fig. 2. The experiment was carried out under pnet =100kPa using triaxial apparatus. This result suggests that the air entry value of this soil under pnet =100kPa is approximately 50kPa.
1. The specimen was saturated (B-value greater than 0.95) by the vacuum saturation procedure (Rad and Clough. 1984) and consolidated isotropically under a certain confining pressure, pc' . The consolidation time, which was decided by the 3t method, was 3 hours.
2. The specimen was dehydrated through the ceramic disk under a constant pnet by applying a certain ua from the top of specimen. The dehydration was continued until the drainage became less than 0.1cc/day. Fig. 3 shows the volumetric strain of the specimen and the drainage volume of water during the dehydration process. It is clear from Fig. 3 that the drainage was continuing after the change of the volumetric strain had almost stopped.
3. Torsional shear tests, under a constant pnet , α=45o and b =0.5 were carried out under constant suction conditions with strain control until the failure. The shear strain rate, εs , was 0.0012%/min, which was decided from preliminary tests conducted with different strain rates.