ANALOG CIRCUIT FAULT DIAGNOSIS METHOD BASED ON CROSS WAVELET FEATURES

摘要

An analog circuit fault diagnosis method based on cross wavelet features. The method comprises the following steps: inputting an excitation signal to an analog circuit under test, and collecting time domain response output signals to form an original data sample set; dividing the original data sample set into a training sample set and a test sample set; performing cross wavelet decomposition on both the training sample set and the test sample set to respectively acquire wavelet cross spectra of the training sample set and the test sample set; applying bidirectional two-dimensional linear discriminant analysis to process the wavelet cross spectra of the training sample set and the test sample set, and extracting fault feature vectors of the training sample set and the test sample set; submitting the fault feature vectors of the training sample set to a support vector machine for training an SVM classifier, and constructing a support vector machine fault diagnosis model; and inputting the fault feature vectors of the test sample set into the model to perform fault classification. By using the present invention, an analog circuit fault can be identified efficiently, and the precision of diagnosis of the analog circuit fault can be significantly improved.