A New Vibration Diagnosis Method Based on the Neural Network and Wavelet Analysis 2003-01-0363
A new vibration diagnosis method for diesel engine based on the Neural Network and Wavelet Analysis is proposed in this paper. It is an on-line prognostics and diagnosis technique by picking-up the frequency characters of the vibration signals on the cylinder head or body of diesel engine. The experiments were carried out on the 295 Diesel Engine and a large number of training and testing samples were acquired. The Wavelet Noise Reduction is used to reduce the disturbance of background noises and the Wavelet Decomposition is applied to obtain the relevant characteristic vectors. Then a Radial Basis Function neural network is built. Using the experiment data to train it. After that, the neural network can be used to classify the faults of diesel engine according to the measured vibration signals on the cylinder head. The simulation and experiment results demonstrate that this method can efficiently diagnose and classify the faults. At the same time, this fault diagnosis method can be applied for vibration analysis for other complex machineries.