Probabilistic Assessment of Spatial Distribution of Soil Liquefaction Potential Using Cone Penetration Test

In engineering practice, soil liquefaction potential is usually evaluated following the cone penetration test (CPT)-based simplified liquefaction triggering procedure. It is widely accepted that CPT-based liquefaction triggering correlation developed from the limited case histories database contains significant model uncertainty. In addition, characterization of spatial distribution of potentially liquefied soils is critical for liquefaction risk assessment and mitigation. However, soil liquefaction potential at untested locations interpolated from limited CPT measurements may involve considerable interpolation uncertainty. On the other hand, soil spatial variability can greatly affect the consequences of liquefaction. Neglecting these uncertainties and soil spatial variability may lead to unreliable or even biased liquefaction assessment results. To tackle this challenge, this paper presents a CPT-based probabilistic liquefaction assessment method in a vertical cross-section, which is able to simultaneously consider above-mentioned uncertainties and soil spatial variability. The presented method is demonstrated and validated using real CPT data from Wildlife Liquefaction Array (WLA), USA. The illustration example indicates that the presented method can properly characterize the spatially distributed soil liquefaction potential and estimate the probability of liquefaction at each point within a cross-section.

Keywords: Liquefaction potential, Uncertainty, Spatial variability, Compressive sampling