Stochastic analysis of a crash box under impact loading by an adaptive POD-PCE model

Structural and Multidisciplinary Optimization(2022)

引用 2|浏览16
Propagating uncertainty through a crash problem is very difficult due to non-linear and non-smooth behavior. The required number of model evaluations is often high, and therefore the computational cost is prohibitive. To deal with such problems, an adaptive meta-model is developed using a polynomial chaos expansion (PCE) and a proper orthogonal decomposition (POD). The adaptive meta-model is used for uncertainty quantification and for global sensitivity analysis of a crash box under impact loading. The time-dependent uncertain response quantities are expressed with the reduced POD modes. The predicted stochastic contact force and impactor velocity by the adaptive meta-model are quite close to the actual simulations. The time-dependent mean and standard deviation for all responses are predicted quite well with low number of model evaluations. Furthermore, it is found that the material property and the crash box thickness are the most influential parameters for the contact force, and the impactor mass is the most influential parameter for the total dissipated energy.
Adaptive meta-model,Uncertainty quantification,Global sensitivity analysis,Crash box,Polynomial chaos expansion,Proper orthogonal decomposition
AI 理解论文