Bounds-based structure reliability analysis of bogie frame under variable load cases

Abstract

Estimating failure probabilities of variable load cases during the design process remains a challenging task in structural reliability analysis of bogie frame. To address this issue, a bound approach by structural reliability analysis is investigated based on modified surrogate model-based polynomial function in this paper. The initial polynomial response surface (PRS) surrogate model of bogie frame under variable load cases is first developed with the results of finite element analysis (FEA) and experimental of design (DoE), and the fitting accuracy is checked based on the analysis of variance (ANOVA) method as well. A modified PRS surrogate model is then proposed to characterize the functional relationship between design parameters and responses. The performance function of bogie frame under variable load cases is next established based on the stress-intensity interference theory. And the subset simulation (SS) method is employed to calculate its failure probability. Finally, the unimodal and bimodal bounds is introduced into the structure reliability analysis of bogie frame to solve the reliability under variable load cases. In this study, the proposed approach is further applied to a practical engineering example to demonstrate its efficiency and accuracy. Simultaneously, the importance of considering the uncertainties of the surrogate model is also shown and discussed by comparing with the Monte Carlo simulation (MCS) results in this example.