Identification of multibody vehicle models for crash analysis using an optimization methodology

Abstract

This work proposes an optimization methodology for the identification of realistic multibody vehicle models, based on the plastic hinge approach, for crash analysis. The identification of the design variables and the objective function and constraints are of extreme importance for the success of the optimization. The characteristics of the plastic hinges are used as design variables while the objective functions are formulated with measures of the difference between the dynamic response of the model and a reference response. The sequential application of genetic and gradient-based optimization methods is used to solve the optimization problem constituting a systematic approach to the automatic identification of vehicle multibody models. The methodology is demonstrated with the identification of the multibody model of a large family car for side and front crash. The vehicle model is developed in the MADYMO multibody code which is linked with the optimization algorithms implemented in the Matlab Optimization Toolbox.