A Simplified Method for Seismic Fragility Analysis of Geogrid-Encased Stone Column Composite Foundations

Geogrid-encased stone columns (GESCs) are increasingly utilized in earthquake-prone areas, yet a practical framework for quantifying their seismic fragility remains underdeveloped. This study develops a machine learning-based probabilistic assessment framework for the seismic fragility analysis of GESC composite foundations. The framework employs a validated analytical model for estimating the ultimate seismic bearing capacity, which is then integrated with the Monte Carlo simulation to account for the inherent uncertainties. A comprehensive parametric study is conducted to systematically evaluate the influence of seismic coefficients ( k h and k v ), column diameter ( D c ), tensile strength ( T ), area replacement ratio ( m ), shear strength ( φ c , c c , φ s , c s ), vertical load demand ( P v ), footing width ( B ), overburden pressure ( q 1 ), and shear strength utilization ratio ( n ). Results demonstrate that the parameters φ s and k v / k h are the most influential on the failure probability ( P f ), whereas D c and c c have a negligible effect. P f shows a positive correlation with T , m , φ c , c c , φ s , c s , B , and q 1 , and a negative correlation with k v / k h , D c , and P v . Since the seismic reliability is more critically influenced by native soil conditions than by column enhancements, the design shall focus on key parameters: high-strength geogrids, adequate foundation width, a high are replacement ratio, and a dense arrangement of small-diameter columns.