Incipient fault characteristic extraction for gears by using MVMD and IDTW

The fault characteristics of incipient fault gears are relatively weak and easily buried by background noise. Moreover, the characteristics of fault gears exhibit similarities to those of normal gears, which may easily lead to misdiagnosis or missed diagnosis. To address these problems, an improved dynamic time warping (IDTW) method is proposed, and a methodology that combines IDTW and multivariate variational mode decomposition (MVMD) is also developed to extract the fault characteristics of incipient fault gears. In the developed methodology, MVMD is utilized to decompose the vibration signals of incipient fault and normal gears simultaneously to ensure the consistency of component frequency. Then, IDTW is employed to align the signal components that exhibit gear meshing frequency in normal and fault statuses to obtain the residual signal. Simulation and experimental results show that the developed method can effectively extract the incipient fault features of gears.