ML-EMPOWERED DAMAGE-STATE-CLASSIFICATION-DRIVEN SEISMIC FRAGILITY ANALYSIS METHOD FOR CORRODED BRIDGES

This dataset and trained artificial neural network (ANN) model support research on the seismic fragility assessment of corroded bridge structures. The uploaded files include: Bridge Damage State Dataset.xlsx: A spreadsheet file containing input features, i.e., parameters of corroded bridges and seismic intensity measures (e.g., PGA, PGV, Sa), and output target, i.e., corresponding damage state derived from nonlinear time history analyses (1=None, 2=Slight, 3=Moderate, 4=Extensive, 5=Complete). ANN_fragility_model.mat: A MATLAB .mat file containing the trained ANN model designed to classify damage states based on corroded bridge parameters and seismic intensity measures. The model enables efficient prediction of seismic damage states for aging bridge structures without the need for computationally expensive physics-based simulations. These resources can be reused or extended for fragility curve generation, model benchmarking, or large-scale seismic risk assessments.