INVESTIGATION OF A VALIDATION APPROACH CONSIDERING UNCERTAINTIES FOR COMPONENTS IN AUTOMOTIVE CRASH SIMULATION MODELS

In the design of safe cars for crash, critical components should deform or fail in a certain manner. Uncertainties can influence the components, such they do not show the desired behavior. The design of such components can be improved via credible simulation models that are able to predict the effect of uncertainties. Our goal in this publication is to undertake the validation of crash simulation models in presence of uncertainties. Investigation of validation approaches for this application is challenging because crash experimental data are limited and inherently noisy. We use synthetic experimental results to address these challenges. We apply V&V 20 standard for the validation of a crash box model. Ultimately, we evaluate the validation results in terms of correctness and conservativeness; additionally, we examine some of the underlying assumptions in the validation approach.