Beam Element Model Optimization Applying Artificial Neural Networks on BIW Concept Design

摘要

Vehicle body-in-white crash models are important for crashworthiness analysis. Conventional finite element methods usually deal with a large-sized computational model and thus hinder efficient design evaluation. The proposed beam element method, with a significant reduction of model size and computation time, is capable of extracting essential safety dynamic characteristics. An artificial neural network is employed and the recurrent back-propagation learning rule trains the network to obtain optimized beam element features. Our analysis shows that the optimized beam element model can accurately capture the frontal crash characteristics of the impacting structures, and predict the vehicle body-in-white crash performance in conceptual design stage.