Closure to 'Fault Rupture Propagation through Sand: Finite-Element Analysis and Validation through Centrifuge Experiments' by I. Anastasopoulos, G. Gazetas, M. F.Bransby, M. C. R. Davies, and A. El Nahas

摘要

In conclusion, although a simplified model without strain softening may in some favorable cases capture some aspects of the problem (such as the location of fault outcropping), it cannot achieve the observed localization of deformation within a realis-rntically narrow zone. Note that the comparisons presented herein refer to medium loose sand, in which the softening behavior is not highly pronounced. In dense sand, where the softening behavior is prevailing (see Fig. 5 in the original paper), the discrepancies can become even more dramatic. Note that this inadequacy had been demonstrated since 1974 by Scott and Schoustra (1974): using the FE method and an elastic-perfectly plastic constitutive soil model with Mohr-Coulomb failure criterion, they produced results contradicting reality and experiments. Most importantly, such simplified models cannot even qualitatively capture the stressing of a foundation mat due to faulting, not only leading to gross underestimation of bending moments, but also often failing to predict the mode of deformation: hogging versus sagging (see Fig. 8). As it has also been demonstrated by several other researchers (Bray 1990; Bray et al. 1994a,b; Roth et al. 1982; Loukidis 1999; Erick-son et al. 2001; Papadimitriou et al. 2007), for a realistic simulation of fault rupture propagation through soil, strain softening is absolutely mandatory.