查看文章

Railway bearing is one of the important parts which constructs boogie structure in a passenger train. This part should be monitored from bearing failure phenomena at any time for passenger safety during traveling. This study presents an effective denoising noisy signal for bearing condition monitoring. A noisy signal was created first, then separated by Empirical Mode Decomposition (EMD) to be intrinsic mode decompositions (IMFs). From IMFs, the noise can be detected, and then it was removed. Following, IMFs which contain no noise was then reconstructed to be a new signal. The Hilbert-Huang spectrum (HHT) spectrum of reconstruction signal was generated by applying Hilbert transform. HHT of the reconstruction signal was then compared to the HHT baseline spectrum and HHT contained noise. The result showed that the proposed technique works well for analyzing signals. Without reconstruction technique, the railway bearing condition was difficult to be revealed by the HHT spectrum.