Identification of crashworthiness indicators of column energy absorbers with triggers in the form of cylindrical embossing on the lateral edges using artificial neural networks

Abstract

The paper presents the possibility of neural network application in order to identify the most advantageous design variants of column energy absorbers in terms of the achieved energy absorption indicators. Design variants of the column energy absorber made of standard thin-walled square aluminium profile with triggers in the form of four identical cylindrical embossments on the lateral edges were considered. These variants differ in the diameter of the trigger, its depth and position. The geometrical parameters of the trigger are crucial for the energy absorption performance of the energy absorber. The following indicators are studied: PCF (Peak Crushing Force), MCF (Mean Crushing Force), CLE (Crash Load Efficiency), SE (Stroke Efficiency) and TE (Total Efficiency). On the basis of numerical studies validated by experimentation, a neural network has been created with the aim of predicting the above-mentioned indices with an acceptable error for an energy absorber with the trigger of specified geometrical parameters and position. The paper demonstrates that the use of an effective multilayer perceptron can successfully speed up the design process, saving time on multivariate time-consuming analyses.