Peak envelope spectrum Fourier decomposition method and its application in fault diagnosis of rolling bearings.

• A method called PESFDM is proposed to improve the performance of FDM. • An improved "locmaxmin" spectrum segmentation method is used in PESFDM. • The segmentation boundaries of PESFDM are more effective. • Compared with other methods, PESFDM gets a better diagnostic effect. The Fourier decomposition method (FDM) is a useful method for signal decomposition. However, FDM has the problems of over decomposition and time-consuming. Therefore, a peak envelope spectrum Fourier decomposition method (PESFDM) is proposed in this paper. Firstly, the upper peak envelope is used to process the signal spectrum and obtain the peak envelope spectrum. Secondly, the segmentation boundaries are adaptively determined on the peak envelope spectrum using a modified "Locmaxmin" spectrum segmentation method. The bands between the segmentation boundaries are then reconstructed into several single-component signals. Finally, the proposed method is used to decompose the simulated and actual bearing fault signals. The results are compared against the other four signal processing methods: empirical mode decomposition, empirical wavelet transform, FDM, and adaptive power spectrum Fourier decomposition method. The results show that the proposed method has a more effective spectrum segmentation and a better effect of fault diagnosis. [ABSTRACT FROM AUTHOR]