Study on Engine Fault Diagnosis Based on the Abnormal Sound Analysis in Transformation Domain

Engine fault diagnosis based on abnormal sound signal analysis is a non-contact fault diagnosis methods, It has advantages that judge fault fastly and accuratly, convenient operation, high degree of automation. Based on this background. This paper research on the transform domain feature extraction of the engine abnormal sound signal and recognition problems. Mainly using the homomorphism analysis method to extract cepstrum features of different sound channel signal, using linear spectrum pairs analysis method to establish full pole model to the different sound channel, using the template matching method to the simulation results, that shows in this paper the cepstrum and the characteristic parameter can effectively identify various abnormal sound signal.