Seismic fragility analysis of geosynthetic-encased stone column composite foundation

The geosynthetic-encased stone column (GESC) is a cost-effective and efficient soil treatment method widely adopted in seismic regions. Its pivotal role in fortifying foundations enhances the safety of the supported structures. This study focuses on the fragility analysis of GESC composite foundations under seismic loads. The application of Logistic regression, a machine learning technique, derives parameterized fragility functions. These parameterized fragility functions elucidate the effects of GESC diameter, encasement tensile strength, column friction angle and soil variability on the GESC composite foundation's failure probability. The results demonstrated the efficiency of Logistic regression in predicting GESC foundation failure probability, boasting an impressive determination coefficient of up to 99.8 when compared with the Monte Carlo Simulations (MCS). Enhancing the tensile strength of the encasement and the column's friction angle reduces the failure probability in GESC foundations. Also, increasing the replacement ratio reduces the failure probability of the GESC foundation. However, for a constant replacement ratio and the same geosynthetic encasement, the seismic bearing capacity of a composite foundation reinforced by GESCs with small diameter is superior to that reinforced by GESCs with large diameter. Notably, inherent soil variability escalates the failure probability of GESC foundations, emphasizing the importance of selecting high-performance GESCs, especially when treating soil with strength that is characterized with a higher coefficient of variation (COV).