Probabilistic CPT-Based Liquefaction Hazard Assessment Using Monte Carlo Simulation

Abstract

The common liquefaction assessment used in Indonesian practice does not include uncertainty in soil and earthquake parameters. This study aims to quantify liquefaction hazard risk in Indonesia by integrating the Idriss-Boulanger method with Monte Carlo Simulation (MCS) to account for uncertainties. Two distinct case studies in Indonesia were analyzed: an evaluation of the 2009 Padang earthquake and a prediction for the Universitas Riau campus development. Methods involved modeling stochastic variables, including cone resistance, sleeve friction, groundwater level, total vertical soil stress, earthquake moment magnitude, and peak ground acceleration, to calculate the probability of liquefaction (PL) across various iterations. Findings reveal that at the evaluation site, 74% of tested depths showed PL exceeding 65%, indicating a significant probability. This finding aligns with field-observed sand boils and settlement. In contrast, the prediction site showed predominantly PL between 15% and 35%; thus, costly ground improvement is unnecessary. This research found that cone resistance had the most significant influence on PL. Utilizing six stochastic variables, the minimum required number of iterations is 600,000. This research improves upon the traditional deterministic method widely used in Indonesian practice by providing a more reliable, non-conservative quantitative framework to decide whether ground improvement, as liquefaction mitigation, is required.