Reliability Analysis of Stone Columns Improved Ground for Mitigation of Liquefaction

Abstract

Stone columns are widely used to improve ground performance and mitigate liquefaction in weak or loose soils. However, conventional deterministic design methods fail to account for the uncertainties in soil and column properties, which can significantly affect performance under seismic loading. To address this gap, this study presents a reliability-based framework for assessing the effectiveness of stone columns in reducing liquefaction potential. The analysis considers key random variables – seismic acceleration (a_max), saturated unit weight (γ_sat), stone friction angle (φ^′_c ), shear wave velocity (V_s), and column diameter (D_c) – under normal and lognormal distributions. Probabilistic methods, including First Order Second Moment (FOSM), Point Estimation Method (PEM), and Monte Carlo Simulation (MCS), were applied to evaluate the reliability index (β). Results show that increasing parameter uncertainty leads to a significant reduction in β, even when the deterministic safety factor exceeds unity (≈1.50). Sensitivity analysis reveals that V_s and γ_sat have the greatest influence on reliability, while D_c and a_max exhibit minimal impact. These findings underscore the necessity of incorporating probabilistic approaches in design and provide insights for simplifying reliability assessment by reducing the number of random variables without compromising accuracy.