Probabilistic Liquefaction Assessment Based on an In-situ State Parameter Considering Soil Spatial Variability and Various Uncertainties

Abstract

The state parameter (ψ) is based on a framework of critical state soil mechanics, which reflects the influences of soil compactness and stress level and has significant advantages for liquefaction analysis. In engineering practice, a simplified procedure based on the cone penetration test (CPT) to assess the liquefaction potential in the field involves model and parameter uncertainties. Furthermore, the use of limited CPT measurements involves project-specific test uncertainties, and the spatial variability of soil properties has a remarkable effect on soil liquefaction. To tackle these challenges, this study develops a ψ-based liquefaction probability framework to integrate the prior information of project-specific CPT, and uses limited tests to characterize the 2D spatial distribution of liquefaction potential with proper consideration of various uncertainties and soil spatial variability. The framework is developed based on simulation of Gaussian stationary random fields and Markov Chain Monte Carlo simulation. The proposed method is demonstrated using real CPT data from the 2011 Tohoku-Oki earthquake in Japan. The results indicate that the ψ-based evaluation performs well and can reasonably assess the liquefaction phenomenon of heterogeneous soils. This study incorporating the spatial variability of ψ into the liquefaction probability framework provides valuable information for geotechnical engineering design.