Direct Denoising of Fault Signal for Train Bogie Bearing Under Speed Change Condition

Due to wheel-rail coupling and other factors, the fault feature of train bogie bearings can be easily masked by noise. The superposition of speed changes under fault shock further causes more difficulty for the direct denoising of signals. Therefore, a denoising framework that can directly process the original vibration signal under speed change has been proposed in this study. The framework includes a variable peak search approach and an adaptive least square (ALS) algorithm. The established variable peak search can capture the transient shock at variable speed. Besides, the proposed ALS algorithm can adaptively search the optimal parameters for the best denoising results. Simulations and actual equipment tests have demonstrated that the proposed framework can effectively reduce noise which has a satisfactory result under speed change conditions. Meanwhile, the performance is superior to other approaches showing its potential for applications on rail transit trains. Especially our study may apply to other typological vehicles under speed change conditions.