Natural Hazards Research Summit: Surrogate Models of Highway Bridges for Regional-Scale Simulations of Transportation Networks

This poster outlines the process of developing a surrogate model for structure-specific, site-specific assessments of highway bridge columns under seismic loading. This work includes a comparison of three surrogate models, linear regression, gaussian process regression, and probabilistic learning on manifolds, as well as an example use case which compares the surrogate performance against high fidelity simulations and HAZUS predictions. This poster was presented at the NHERI Natural Hazards Research Summit 2024 as an example of the ways high-fidelity regional modeling enabled by surrogates can improve our understanding of seismic risk.