Edge Computing-Aided Framework of Fault Detection for Traction Control Systems in High-Speed Trains
Hongtian Chen, Bin Jiang, Wen Chen, Ziheng Li
Abstract
Long-time operation of traction systems under severe working environment easily induces various faults in high-speed trains, and thereby puts passengers in an unsafe position. Thanks to widely equipped sensors in high-speed trains and the technical development of advanced data analysis techniques, real-time fault detection (FD) using only online data is an inevitable trend to improve safety and reliability of high-speed trains. In this paper, an edge computing-aided FD framework for traction systems in high-speed trains is proposed; the superior advantages include: 1) its implementation can be easily carried out without system information such as accurate mathematical models of traction systems; 2) it has the well-deserved expansibility without being redesigned for the control structure of traction systems; 3) it nearly does not occupy resource of traction control units, and it does not overburden the information transmission systems. This highly intelligent FD framework using edge computing is firstly speculated in details, and its core theory is then applied to a dSPACE platform of high-speed trains for verification.