Abstract:

In order to improve the fault diagnosis accuracy of analog circuit, the authors proposed an analog circuit fault diagnosis model based on wavelet packet analysis and relevance vector machine. Firstly, different fault output signals of analog circuit were collected and decomposed by wavelet packet to extract normalized energy features of signal, and then the feature vectors were input to relevance vector machine to train and establish analog circuit fault diagnosis model to realize the classification and identification, and finally the simulation example was used to test the performance. The results show that compared with other analog circuit fault diagnosis models, the proposed model not only improves the fault diagnosis accuracy rate but also increase the fault diagnosis speed of analog circuit.

Key words: analog circuit fault, wavelet packet analysis, relevance vector machine, classification identification