Application of wavelet packets and GA-BP algorithm in fault diagnosis for diesel valve gap abnormal fault

Analysing vibration signal is an effective important method for diesel engine fault diagnosis, and its key techniques are feature extraction and pattern recognition. In this paper, wavelet packet decomposition algorithm as an effective method for fault feature extraction is used to decompose the vibration signals, and its percentage of energy band wavelet packet and wavelet packet energy spectrum entropy are regarded as diagnostic feature vectors. At the same time, in the process of pattern recognition, a mixed neural network training algorithm-GA-BP algorithm was used to recognize the fault pattern in fault diagnosis of valve gap abnormal fault. This method can effectively and reliably be used in the fault diagnosis of valve gap abnormal fault by comparing the two algorithms and analyzing the results of real examples. This method can also effectively be used in other fields.