Structural reliability analysis of bogie frame based on R-Vine copula model

To address the challenges posed by multi-source and high-dimensional variable dependencies in the service process of bogie frames, a method based on the R-Vine Copula model is developed. A finite element model of the bogie frame is developed, and its static strength under exceptional conditions is evaluated according to the UIC615-4 standard. A parameterized model is constructed using APDL scripting, followed by a design of experiments and Sobol sensitivity analysis to identify design variables exhibiting strong interactions. An optimal R-vine Copula structure tree is established, and the associated copula parameters are selected through maximum likelihood. Optimal Vine Copula functions are selected and validated using the AIC criterion. Variable correlation data consistent with the optimal marginal distribution functions are generated via the Vine Copula sampling algorithm. The failure function of the bogie frame serves for reliability evaluation in combination with an improved RBF surrogate model, considering both independent and correlated variable assumptions. Comparison of the reliability results demonstrates the effectiveness of the proposed method, providing a more realistic assessment of the bogie frame’s structural reliability under actual service conditions.