Quasi-steady state: a real behaviour?: Reply

Abstract

Discussions 189 The Authors would like to thank the Discusser for his interest in the paper. The Discusser disagrees with the main conclusion in the paper that the QSS behaviour or the postPT increase in shear stress in undrained triaxial test on loose saturated sand may be a test-induced behaviour. It is useful to reiterate some of the arguments for this con clusion, which are as follows: (1) In undrained triaxial tests on loose Unimin sand, constant pore pressure, constant shear stress, constant effective normal stress, constant sample vol ume and constant strain rate are reached over a strain range between 9‐15%; thereafter, the shear resistance continuously increases with further straining (e.g., Fig. 4). In other words, the requirements of the definition of the steady state are fully satisfied with in this strain range; however, the post-PT increase in shear resistance conflicts with the concepts of the critical void ratio and the steady state. (2) The drained be haviour of loose sands corresponding to the undrained QSS behaviour is that after peak, the shear resistance remains constant, while the sample volume remains constant only temporarily and then continuously increases (e.g., Fig. 7). (3) The post-PT increase in shear resistance in undrained tests or the continuous increase in sample volume in drained tests on loose sands usually commences at a strain of about 15% (this study; Ishihara 1993). This strongly suggests that the increase is likely an error caused by the triaxial test system. (4) For a given consolidation void ratio, the higher the friction force between end platens and a sample, the larger is the shear resistance of the sample (e.g., Figs. 9 and 13). (5) The radial dilation of a specimen with lubricated end platens is much larger than that without lubrication (e.g., Fig. 12). The results of the drained tests indicate, as stated in the paper, that the deviator stress and the volumetric strain may reflect responses within different zones of a loose sand specimen at large strain. The measured deviator stress reflects the shear resistance controlled locally by the failure zone, and remains constant at large strain because the effective confining stress maintains constant during drained shear, and the failure zone has reached an ultimate state. The measured volumetric strain, however, reflects the global volume change, and continuously changes at large strain since zones other than the failure zone have not reached an ultimate state after peak. In the case of undrained triaxial tests, the volume change is restrained during shear and the dilation tendency of global volume at large strain results in continuous decrease in pore pressure, and then causes the post-PT increase in deviator stress, even though the failure zone may have reached an ultimate state as in the drained triaxial tests. The above results also indicate that the end restraint due to the fiction between end platens and a specimen can significantly affect the dilation at its ends. Therefore, it is reasonable to question whether the post-PT increase in deviator stress during undrained shear is a real behaviour of loose sands. The Discusser argues that the results of the hollow cylinder torsional tests shown in Fig. 13 disproves the assumption that the post-PT increase in deviator stress is mainly caused by the end restraint. The reasons for his comment are that when b =