Abstract: A new way is offered for gear fault diagnosis. A new approach, designated as Hilbert-Huang transform is introduced into gear fault diagnosis. Then the concept of local Hilbert energy spectrum is proposed and two fault diagnosis methods named frequency family separation based on EMD and Hilbert energy spectrum are established according to the characteristic of gear fault vibration signal. The frequency families are separated effectively by applying empirical mode decomposition to the gear fault vibration signals. The fault information of the gear vibration signals can be extracted from the local Hilbert spectrum from which the distribution of the energy of the gear fault vibration signal with the change of time and frequency. These two methods have been applied to gear fault diagnosis. The results of the application to gear fault diagnosis show that frequency family separation based on EMD and Hilbert energy spectrum can extract gear fault information effectively.

Key words: Frequency family separation, EMD, Gear fault diagnosis, Hilbert-Huang transform, Local Hilbert energy spectrum