Bearing Fault Diagnosis Based on Improved Hilbert-Huang Transform

For empirical mode decomposition (EMD) method had endpoint effect and modal aliasing problems, in this paper an improved Hilbert-Huang transform (HHT) based on combination of symmetrical-mirror continuation and correlation coefficient method was put forward. This method can eliminate the modal aliasing and false intrinsic mode function (IMF) problems effectively. Through correction parameters threshold denoising method and the method of selecting sensitive IMF, the Hilbert spectrum can be more accurate for fault diagnosis. The experiment research confirmed that this method can diagnose the bearing fault signal reasons effectively. A new and effective method was proved for bearing fault diagnosis.