Wavelet analysis is a newly developed theory, which overcomes the disadvantages of traditional Fourier analysis. It has superior localized features in both the times and the frequencies domain and can be applied to signal processing, image processing and speech analysisThe paper gives a description of the basic theory of wavelet analysis, makes a research on the wavelet packet's applications to fault diagnosis for analog circuit and proposes the energy fault diagnosis algorithm based on the wavelet packet transform as extraction optimal features. It overcomes the complexity of the fault model in the analog system, which makes it difficult to quantify simply and eliminates the defects of components tolerances and nonlinear effects. Thus it proposes the BP neural network fault diagnosis algorithm based on the incompletion wavelet packet transform and energy normalization as preprocessors. After preprocessed, the sample signal will be sent to BP neural network to be trained which effectively reduces the numbers of the inputs and the hidden layer nodes, thereby reducing the size of the neural network, degrading its complexity and minimizing its training time. It identifies fault location with accuracy. Finally the paper studies the method for the fuzzy neural networks fault diagnosis combined by the neural networks and fuzzy rules. Employing the wavelet packet transform, a fuzzy neural networks fault diagnosis algorithm based on wavelet packet transform is presented. Simulation results show that this algorithm structure is simple and easy for auto test and improve the accurate rate of fault diagnosis for analog circuit.
|