Abstract: Aiming at the problem that the probability distribution functions of some parameters in frontal crash contain uncertain interval variables, a hybrid uncertain reliability-based optimization model for vehicle frontal crash is constructed, and the optimal polynomial model based on error ratio selection technique is introduced to vehicle crash analysis. Due to the existence of interval parameters, the probabilistic constraints in inner layer are established by limiting the lower bound of reliability interval, thus ensuring the safety of body structure. An efficient shifting vector-based decoupling algorithm is adopted, which converts the nested optimization problem into deterministic optimization and sequential iterative process of hybrid reliability analysis, avoiding nested optimization between inner and outer layers, achieving efficient reliability optimization for vehicle frontal crash. The results indicate that after optimization, the total mass of bumper beam, energy-absorbing boxes and front longitudinal beams reduces by 2.4% and all the constraints of reliability indicators have been met, hence realizing the reliability optimization for the safety of vehicle body and occupants while ensuring the lightweighting of vehicle body.

Key words: reliability-based optimization of frontal crash, probability-interval mixed uncertainty, optimal polynomial model, shifting vector