Elastomeric bearing performance prediction and damage effect evaluation using machine learning

Abstract Elastomeric bearings are vital components in bridge engineering, ensuring structural integrity by accommodating movements due to loads, temperature variations, and seismic activities. This research integrates Finite Element Analysis (FEA) with machine learning techniques, to enhance the predictive accuracy of elastomeric bearing performance under varying conditions. The study focuses on modeling elastomeric bearings with and without steel shim offsets and analyzing their deformation, stress, and strain responses to a compressive load. Furthermore, the research introduces damage to the bearings to simulate real-world conditions, followed by machine learning-based predictions. The results indicate that larger bearings exhibit less deformation and stress, suggesting improved load distribution and structural stability. The findings contribute to optimizing the design and durability of bridge bearings by providing a comprehensive methodology that combines traditional FEA with advanced data-driven approaches.