Interval methods for lack-of-knowledge uncertainty in crash analysis

This paper deals with lack-of-knowledge uncertainty in complex non-linear simulations on a component level, i.e., crashbox during frontal impact of vehicle. Specifically, the focus lies using interval field techniques to model uncertain boundary conditions simulations. The considered this work is unknown mechanical response from adjacent structure. be epistemic, representing case where structure at time analysis performed. In practice, refers situation still under development, e.g., different department or even outsourced. addition, safety critical performance both, and overall should guaranteed wide range circumstances, which are typically encountered real-life situations. Typically, car manufacturers use multidisciplinary optimisation identify designs that perform best all requirements deterministic sense, while minimising weight. Unfortunately, results such schemes known converge an often non-robust optimum. As result, may sensitive small changes input parameters conditions. answer these challenges, proposes approach accounts for lack-of-knowledge, structures, early design stage. accomplished by introducing spatially varying compliance elements connect structures. These have valued stiffness, varied along following realisations field. bounds interval-valued quantities interest, mean force peak force, identified differential evolution algorithm. method demonstrated four studies full overlap crash rectangular box, represents generic within front demonstrate applicability potential proposed method. last it shown can assessed increasing uncertainty.