A practical approach to train-bridge system performance evaluation with consideration of random uncertainty and weighted evaluation indexes

Abstract

In this paper, a practical methodology for evaluating the dynamic performance of train-bridge systems is presented, which incorporates both random uncertainty and the weighting of evaluation indexes. Firstly, a train-bridge stochastic model is established by combining an existing train-bridge deterministic model with the Generalized F-discrepancy (GF-discrepancy) minimized point selection strategy and the probability density evolution method (PDEM). The model enables the accurate and efficient quantification of the variability of any dynamic response quantity of interest in the train-bridge system. Building on this foundation, an assessment model based on the multi-layer fuzzy comprehensive evaluation method is developed to determine the possibility of the train-bridge stochastic system in different performance states with consideration of the weighted evaluation indexes. In the assessment model, the main criteria of the train-bridge system are collected, classified, and further structured a logical hierarchical framework, and the weightings of the criteria are solved by analytic hierarchy process method and entropy weighting method. As demonstrated by a numerical example, the proposed method can properly consider system uncertainties, appraisal hierarchy, and the weightings of evaluation indexes, which helps identify critical evaluation indexes and improve evaluation reliability. The proposed approach can provide a useful tool for the performance evaluation of train-bridge systems.