An effective approach for reliability-based robust design optimization of uncertain powertrain mounting systems involving imprecise information

Abstract

The uncertain parameters of automotive powertrain mounting systems (PMSs) may involve imprecise information (e.g., incomplete, different and conflicting information) in engineering practice. An effective approach is proposed for the reliability-based robust design optimization (RBRDO) of uncertain PMSs involving imprecise information. In the proposed approach, the imprecise information of uncertain parameters is firstly addressed and combined based on evidence theory, and the uncertain parameters are treated as evidence variables. Then, an uncertainty analysis method named evidence perturbation-central difference method (EPCDM) is derived to fast estimate the mean intervals, standard deviation intervals, and the belief and plausibility measures related to system inherent characteristics. A reference method named evidence-Monte Carlo method (EMCM) is developed to verify the effectiveness of EPCDM. Next, to conduct robustness design, the weighted sum of the lower bounds of means and the upper bounds of standard deviations of system inherent characteristics are taken to construct optimization objective; while to perform reliability design, the belief measures related to system inherent characteristics are used to create reliability constraints. Afterwards, a nested RBRDO model is established to explore the optimum design of the PMS, which considers both reliability and robustness simultaneously. The nested PBRDO can be effectively simplified based on EPCDM. The effectiveness of the proposed approach is finally demonstrated by the application example.