Performance-based probabilistic assessment of liquefaction-induced building settlements

Abstract

The current state of practice in the estimation of liquefaction-induced building settlements (LIBS) relies on pseudoprobabilistic approaches in which the estimation of ground motion intensity measures (IMs) is separated from the estimation of LIBS. In contrast, in a performance-based probabilistic approach, the estimation of the IM hazard is coupled with the estimation of the LIBS hazard. Thus, engineers can obtain LIBS estimates that are directly related to a selected design hazard level (or return period), which is more consistent with performance-based design. In this study, we present new developments for the performance-based probabilistic evaluation of LIBS, including 1) the performance-based assessment of LIBS considering the hazard from a single IM in the context of scalar probabilistic seismic hazard assessment (PSHA), 2) the performance-based assessment of LIBS considering the hazard from multiple IMs in the context of vector PSHA, 3) deaggregation of earthquake scenarios from LIBS hazard curves, 4) estimation of LIBS hazard curves in areas where earthquakes from multiple tectonic settings can occur (e.g., shallow crustal, subduction), and 5) treatment of uncertainties (i.e., aleatory and epistemic). The developments are implemented in a computational platform named "LIBS", which is fully coupled with PSHA assessments and facilitates the straightforward performance-based estimation of LIBS in engineering practice. Finally, we perform comparisons of performance-based and pseudoprobabilistic-based estimates of LIBS and share insights from the comparisons.