Intelligent diagnosis method for plant machinery using wavelet transform, genetic programming and possibility theory

This paper proposes an intelligent diagnosis method for plant machinery using wavelet transform (WT) genetic programming (GP) and possibility theory. The WT is used to extract feature spectra of each machine state from measured vibration signal for distinguishing faults. Excellent symptom parameters (SP) for detecting fault states are automatically generated by GP. The membership functions of symptom parameters are established using possibility theory for resolving the ambiguous diagnosis problems. The methods proposed in this paper are verified by applying them to the fault diagnosis of gear equipment.