Natural Hazards Research Summit 2024: Neural Network Applications in Bridge Seismic Resilience Evaluation

Assessing seismic resilience in critical infrastructures like highway bridges is crucial for emergency planning and maintaining network integrity post-hazard events. While machine learning techniques show promise in earthquake engineering, their application to bridge resilience remains under-explored. This study presents the feasibility of applying various neural network architectures to predict the resilience index for a cohort of concrete box-girder bridges with different structural characteristics and exposed to different levels of seismic peak ground acceleration events. The investigation underscores the effectiveness of neural network algorithms in modeling complex relationships between structural characteristics and seismic events, offering valuable contributions to resilience estimation and rapid decision-making processes.