Fault Diagnosis of Track Circuit Based on KELM Optimized by SSA Algorithm
Yuefan Zhang, Yunshui Zheng · 2022
Aiming at the diversity of ZPW2000A jointless track circuit faults, an intelligent diagnosis method based on sparrow search algorithm(SSA) optimizing kernel extreme learning machine(KFLM) is proposed. Firstly, according to the basic structure of ZPW2000A jointless track circuit, parameters collected by the track circuit centralized signal monitoring(CSM) system are chosen as the fault data sample set. Secondly, SSA algorithm is used to optimize the kernel function parameter and the penalty coefficient of the KELM to obtain best fault classification accuracy. Finally, SSA-KELM model with optimal parameters is used to identify specific fault types. The Matlab simulation results show that the fault diagnosis accuracy of SSAKELM model reached 98.13%, which is superior to KELM model and PSO-KELM model. SSA-KELM model can effectively and accurately identify the fault types of track circuit, which provides a new solution for the development of intelligent railway fault diagnosis.