Optimization Method With Prediction-Based Maintenance Strategy for Traction Power Supply Equipment Based on Risk Quantification

Abstract

A prediction-based maintenance (PBM) strategy can save resources while simultaneously guaranteeing the traction power supply system (TPSS) operates in a stable and reliable condition. Based on risk quantification, this paper presents a method to assess maintenance needs and optimize the PBM strategy for traction power supply equipment (TPSE). This method utilizes historical maintenance information and the fault record of the TPSE as inputs to a fault prediction model based on a Bayesian classifier. By comparing the prediction results with the actual equipment operation state, statistical parameters are obtained to calculate the reliability and economic indices for the maintenance strategies. The comprehensive maintenance risk is then quantified by integrating the dynamic inspection and fault risk costs for the TPSE. The minimum comprehensive risk is selected as the optimization objective, and the criterion for fault prediction spurs from the optimization results are used to form the PBM strategy. Practical maintenance information and fault record data of 27.5-kV vacuum circuit breakers for a TPSS are used to verify the proposed method. The results show that, by fully utilizing the current data, this method can predict equipment faults by considering multiple factors, and it can quantify the risk for a particular PBM strategy, which is also applicable to similar large-scale traction power supply facilities. To improve the PBM strategy, the approach optimizes fault prediction criteria to achieve a minimum comprehensive maintenance risk. The proposed method can provide effective data and confidence for decision makers in implementing PBM for TPSS.