Multiple occurrences of liquefaction, termed as “reliquefaction,” have been observed during the recent earthquakes in Japan (2011) and New Zealand (2010-2011). These field observations indicated that once the soil liquefies, the reliquefaction resistance may become weaker despite an increased density due to the postliquefaction reconsolidation. This reduction in reliquefaction resistance is linked to the fabric anisotropy developed during liquefaction or large preshearing, but quantifying its effect still remains a challenge and is not yet fully understood. In this thesis, a comprehensive and fundamental experimental program was designed to investigate the role of fabric anisotropy on the reliquefaction resistance of Toyoura sand using a hollow cylinder torsional shear apparatus. The re...[ Read more ]

Multiple occurrences of liquefaction, termed as “reliquefaction,” have been observed during the recent earthquakes in Japan (2011) and New Zealand (2010-2011). These field observations indicated that once the soil liquefies, the reliquefaction resistance may become weaker despite an increased density due to the postliquefaction reconsolidation. This reduction in reliquefaction resistance is linked to the fabric anisotropy developed during liquefaction or large preshearing, but quantifying its effect still remains a challenge and is not yet fully understood. In this thesis, a comprehensive and fundamental experimental program was designed to investigate the role of fabric anisotropy on the reliquefaction resistance of Toyoura sand using a hollow cylinder torsional shear apparatus. The results were analyzed from various standpoints, and the major findings are summarized as follows:

In order to explore the role of induced fabric anisotropy, reliquefaction tests were conducted on Toyoura sand of different densities (i.e., D_r=45% and 70%), which experienced different residual shear strains (γ_res= 0.4%-5.0%)and reconsolidated at different states of A (i.e., zero shear stress state, after stress reversal) and B (i.e., zero shear strain state). The results showed that a medium preshearing (i.e., γ_res= 0.4%) increases reliquefaction resistance significantly, while the reliquefaction resistance decreases when the Toyoura sand experiences a large preshearing (i.e., γ_res ≥ 2.0%). For all tests conducted in this study, γ_res of 1.0%~2.0% can be considered as a threshold above which the reliquefaction resistance of sand tends to decrease. It is also worth noting that specimens reconsolidated at state A have less reliquefaction resistance at the same residual strain levels compared to the specimens reconsolidated at B, which can be explained by a higher degree of induced anisotropy in samples reconsolidated at A. However, in loose Toyoura sands, there was no significant difference between specimens reconsolidated at A and B once they were largely presheared. All in all, the effect of reconsolidation state in loose sands is not as significant as that in dense ones.

Moreover, to highlight the role of inherent anisotropy, reliquefaction tests were conducted on loose Toyoura sand prepared with different methods of dry deposition (DD) and moist tamping (MT). Medium preshaken MT Toyoura sands showed much higher reliquefaction resistance than those of DD ones. Regardless of the reconstitution method, medium preshaken sands reconsolidated at state A showed a lower reliquefaction resistance than those reconsolidated at state B. However, the effects of the initial fabric and reconsolidation state are lost once the loose Toyoura sand is largely persheared. Therefore, reliquefaction resistance is almost similar for loose Toyoura sands, which presheared largely irrespective of reconstitution method and reconsolidation state.

Based on experimental data, an energy-based method was developed to quantify the liquefaction and reliquefaction resistance of Toyoura sand. It was found that the dissipated energy (i.e., capacity energy) to reliquefaction was mostly greater than those of liquefaction tests regardless of relative density, inherent/induced anisotropies, and cyclic stress ratio (CSR). It was also found that most of the sands reconsolidated at point B showed consistently higher capacity energy than those reconsolidated at point A. This indicates a systematic difference in the soil fabric developed at points A and B from an energy-based standpoint. Furthermore, capacity energy to liquefaction or reliquefaction can be uniquely determined irrespective of cycle number, N, and CSR, except for a few cases of loose virgin and medium preshaken MT Toyoura sand. Furthermore, a new energy-based model with one calibration parameter was developed for predicting excess pore water pressure (EPWP) buildup during liquefaction and reliquefaction for the first time. Finally, correlations between the dissipated energy and cyclic resistance ratio (CRR₂₀) were established based on different preshearing histories. The results provided a linkage between the energy-based method and the conventional stress-based method for evaluating the liquefaction and reliquefaction resistance of sands.