Fault Diagnosis of Engine Based on Wavelet Packet and RBF Neural Network

To solve the problem of fault diagnosis for engine,due to the complexity of the equipments and the particularity of the operating environments, generally speaking, there is no one-to-one correspondence between the characteristic parameters and status, so, the methods of diagnosis are very complicated. Because the vibration signals of the engine are usually nonlinear and non-stationary signals, traditional signal processing methods can not get perfect results, to overcome the deficiency of existing methods, In this paper, a new approach for fault diagnosis of engine based on wavelet packet and RBF neural network is proposed, using wavelet packet analysis as the pre-processing means of RBF neural network. First of all,take a 3-layer wavelet packet transformation for the fault signals and then provide the fault eigenvectors for the RBF network. Finally, use the RBF neural network to construct then on-linear mapping between fault types and fault eigenvectors,and then use the well trained network diagnosis the fault for engine.