Research on gas sensor fault diagnosis based on wavelet analysis and VLBP neural network

For four types of common faults of gas sensor, namely offset, impact, drift and periodic types, on the basis of wavelet analysis and VLBP neural network, a method of the gas sensor fault diagnosis is proposed based on the pattern matching and classification of characteristic energy spectrum extracted by wavelet packet decomposition and VLBP neural network. Characteristic energy spectrum of each singularity can be obtained through the wavelet packet decomposition of output signal of gas sensor. After processing, it, as the characteristic vector for training VLBP neural network, adopts the momentum VLBP algorithm for updating weight values at each sample point; and, through learning a variety of characteristics data of fault model, it is used to determine the fault type of sensor. The experiment result shows that, this method can be effectively applied to the fault diagnosis of gas sensor.