HYBRID ANALYSIS OF LIQUEFACTION POTENTIAL AND SITE ATTENUATION EFFECTS USING EMPIRICAL AND NONLINEAR NUMERICAL APPROACHES: A CASE STUDY OF RSU ANUTAPURA DURING THE 2018 PALU EARTHQUAKE

Abstract

The Palu earthquake of September 28, 2018 (M_w 7.5) triggered massive liquefaction that had a destructive impact on critical infrastructure, including the Anutapura General Hospital (RSU) Building. Conventional liquefaction potential evaluations generally rely on empirical procedures based on the Standard Penetration Test (SPT), which are effective in detecting early vulnerability (triggering). However, this method has a fundamental limitation because it is unable to generate dynamic surface acceleration time-histories for structural design. This research proposes an in-depth case study using a hybrid approach that combines empirical methods (Simplified Procedure) with 1D Non-Linear Site Response Analysis (NL-SRA). Computational modeling is developed in the OpenSees framework using the PressureDependMultiYield02 (PDMY02) constitutive model to simulate cyclic mobility and effective stress dissipation due to pore water pressure generation under undrained conditions. The analysis results on the soil profile of Anutapura General Hospital confirmed the occurrence of extreme liquefaction in shallow to medium sand layers (depth 5–31 m), which is characterized by a drastic decrease in the empirical Safety Factor (FS) to 0.49. Numerically, the NL-SRA model validates that the loss of shear stiffness in the liquefied layer triggers an attenuation (de-amplification) phenomenon of peak seismic acceleration at the surface, which is accompanied by an elongation of the dominant wave period. This study concludes that the integration of empirical methods and advanced non-linear constitutive models is absolutely necessary to capture the wave modification effects due to soil liquefaction, while producing realistic seismic demand as a precise basis for Soil-Structure Interaction (SSI) analysis.