Steady state behaviour of sands and limitations of the triaxial test.

Abstract

Steady state strength is an important parameter in evaluating the stability of sand deposits against flow sliding. Most loose sands exhibit the so called quasi-steady state behaviour in undrained triaxial tests, in which no unique ultimate shear resistance can be reached. This study focused on why the quasi-steady state behaviour occurs and how to determine the steady state strengths of sand samples with the quasi-steady state behaviour in triaxial tests. The main contributions in this study are as follows: (1) The quasi-steady state behaviour may not be an inherent behaviour of sands, but is primarily caused by end restraint which results from the friction between end platens and sample. The main effect of end restraint in triaxial tests on sands is decreasing pore pressure under undrained loading or increasing sample volume under drained loading. The QSS behaviour may also result from volume change during undrained loading. (2) There are four deformation stages in an undrained triaxial test on a loose sand sample: initial stage, collapse stage, critical stress stage and post failure stage. The strain rate in the test remains constant in a deformation stage, but is different in various deformation stages. The steady state strength occurs in the collapse stage. (3) A modified definition of the steady state of deformation is suggested as follows: The steady state of deformation for any mass of particles is that state in which the mass is continuously deforming at constant volume, constant normal effective stress, constant shear stress, and constant velocity. The constant shear stress is the minimum strength of the mass, which is dependent on the local void ratio within shear zone. (4) The steady state strength should be determined, based on the modified definition of the steady state, by considering all test information including shear stress, pore pressure, strain rate, and effective stress path under undrained loading condition or volume change under drained loading condition. (5) The error in steady state strengths determined in triaxial tests is affected by the coefficient of uniformity of a sand. The error for a sand with a 2.5 coefficient of uniformity can be 5 times that for a sand with a coefficient of uniformity lower than 1.8. (6) The volume changes of a moist tamped sand sample during saturation can be over 0.025 in terms of void ratio for the Unimin sand. (7) The volume change of a sand sample during undrained loading results from the compression of pore fluid and the variation of membrane penetration. The volume change due to the variation of membrane penetration is about 4 times that due to pore fluid compression. (8) A new method of area correction in triaxial tests was developed. The average diameter over the middle portion of a sample is dependent on axial strain, volumetric strain, end radial strain, and the length of the middle portion. A middle third average diameter is suggested to be used in the calculation of deviator stress in triaxial tests. The error in deviator stress due to using the conventional area correction can be over 10% at large axial strain under fixed end condition.