Physics-based and data-driven modeling for stability evaluation of buried structures in natural clays

This study presents a hybrid framework to predict stability solutions of buried structures under active trapdoor conditions in natural clays with anisotropy and heterogeneity by combining physics-based data-driven modeling. Finite-element limit analysis (FELA) newly developed anisotropic undrained shear (AUS) failure criterion is used identify the underlying mechanisms as well develop numerical (physics-based) database numbers for both planar circular trapdoors. Practical considerations are given three linearly increasing strengths compression, extension, direct simple AUS material model. The obtained compared validated published literature. A multivariate adaptive regression splines (MARS) algorithm further utilized learn act fast FELA surrogates evaluation. current MARS-based modeling provides relative importance index accurate design equations that can be confidence practitioners.