Fault Prediction Model of High-Power Switching Device in Urban Railway Traction Converter with Bi-Directional Fatigue Data and Weighted LSM

• Main Authors: Lei Wang, Shenyi Liu, Ruichang Qiu, Chunmei Xu
• Format: Article
• Language: English
• Published: MDPI AG 2019-01-01
• Series: Applied Sciences
• Subjects: fault prediction; fatigue model; accelerated fatigue test; high-power switching device
• Online Access: https://www.mdpi.com/2076-3417/9/3/444

The switching device is relatively weakest in the traction converter, and this paper aims at the fault prediction of it. Firstly, the mathematical distribution is analyzed based on the results that were obtained in electro thermal simulation and a single-directional accelerated fatigue test. Then, the accelerated fatigue test with bi-directional fatigue current is proposed, the data from which reflects the accelerating effect from FWD on the device aging process. The analytical model of fatigue process is fitted with the data that were obtained in the test. In order to shorten the test time consumption, we propose a weighted least squares method (LSM) to fit the failure data. Finally, the prediction model is presented with the consideration of fatigue signature and Arrhenius temperature factor.