322 / 2021-04-30 13:51:55

Research on Fault Diagnosis of Wheelset Polygon Based on EEMD and Improved FastIca Algorithm

Wheel Polygon; Fault Diagnosis; EEMD; Improved FastIca

Special Sessions > Vibration detection and fault diagnosis of rail transit rotating machinery

The polygon wear of wheels will affect the comfort of passengers when the train is running, and even endanger driving safety in severe cases. A method of wheel polygon fault detection based on the combination of EEMD and improved fast independent component analysis (FastIca) is proposed. This method improves the convergence speed of the FastIca algorithm and improves the quality of fault signal separation. the method first uses EEMD to decompose the signal into multi-order IMF components. Secondly, the corresponding decomposed characteristic signals are selected to form the observation sequence. The improved FastIca algorithm is used to separate the observation sequence to obtain the source signal. Finally, the frequency spectrum and envelope spectrum of polygonal multi-order spectral lines are analyzed. The results show that the frequency at which the spectral peak of the polygonal wheel is consistent with the rotation frequency of the wheel appears on the envelope spectrum, and the frequency of the wheel rotation frequency is obviously appeared on the first-order component spectrum of the polygonal wheel. Through the verification and analysis of the measured road data, the algorithm can detect the wheel polygon very well, which provides a theoretical basis for wheel repair.