Probabilistic liquefaction-induced lateral spread hazard analysis incorporating the spatial variability of soil parameters

Abstract

Liquefaction is a phenomenon which tends to cause damage to buildings and structures founded on liquefiable soils. However, existing probabilistic liquefaction-induced lateral spread hazard analysis (PLLSHA) is based on empirical lateral spread (LS), in which the spatial variability of soil parameters is ignored in liquefaction hazard evaluations. This paper thus proposes a PLLSHA method which can consider the spatial variability of soil parameters. The response spectrum matching method was used to select 15 ground motions from NGA-West2 database under the earthquake scenario with magnitude of 7.0 and rupture distance of 10 km. Based on the Cholesky midpoint method, random fields for the relative densities of the liquefiable layer with different mean values, coefficients of variation, horizontal and vertical correlation lengths were generated and then embedded into the sloping grounds containing liquefiable layers. A set of random variables for relative densities of the sloping ground is also generated for comparative purpose. The finite element software OpenSees is used to establish the two-dimensional sloping ground model and conduct several dynamic response analyses based on the ground motions selected. The mean LS on surface ground are monitored and recorded as indicators of liquefaction consequence for analysis. An assumed site source with variable magnitudes and 10 km from sloping ground is used for conducting the PLLSHA. Result shows that the spatial variability of relative density of soils has a great influence on the liquefaction hazard results. Specifically, the liquefaction-induced LS hazard would be greater when the mean value of the relative densities or vertical correlation lengths are smaller, as well as when the coefficient of variation or horizontal correlation lengths are larger. When a random variable model is utilized, the liquefaction-induced LS hazard is significantly overestimated. This study would provide a theoretical reference for lateral spread hazard analysis of liquefiable soil layers exhibiting strong spatial variability.