Online Noninvasive Technique for Condition Monitoring of Cooling System in Urban Rail Vehicles

Abstract

The open self-ventilated cooling system is an important component to dissipate heat in urban rail vehicles. While after long term operation, the cooling system will deteriorate, which ultimately affects the reliable operation of traction converter. In this paper, we propose a method to monitor the health status of cooling system in urban rail vehicles using the temperature response curve of negative temperature coefficient (NTC) thermistor built in IGBT module. Back Propagation Neural Network (BPNN) is employed to predict the reference curve of healthy state under different working conditions and Discrete Fréchet Distance (DFD) algorithm is applied to achieve the difference comparison. The proposed approach considers the dynamic response process and is suitable for online monitoring. Advantages of the proposed approach include no need to install additional measurement circuit and to heat the module to thermal equilibrium. Moreover, this method enables the concurrent monitoring of multiple components in the cooling system. Simulation analysis and experimental tests are performed to verify the accuracy and effectiveness.