A software system for analog-circuit fault diagnostic based on wavelet analysis is presented in this paper. This system involves following three sections: the back-propagation neural networks, wavelet decomposition and some data preprocessor. In this paper, the technique of the wavelet transform as a preprocessor is introduced, and the neural network used for fault diagnosis is discussed. Base on windows, the proposed system has the capacity for circuit simulating, fault analysis and fault diagnosis through preprocessing the impulse response of the circuit. The software system is built up with an open architecture and high extensibility.
展开▼
