Analog circuit failure diagnosing method

Abstract

The invention discloses an analog circuit failure diagnosing method which comprises the following steps of: 1, acquiring data, i.e., acquiring test variables, acquiring voltage or current variables at two or more testable nodes of a circuit to be tested; 2, testing that a maximum information entropy of the variables approaches a characteristic direction vector; 3, searching a maximum non Gaussian property characteristic direction vector of the test variables; and 4, calculating projection values of the test variables at the maximum non guassian property characteristic direction vector. The invention has the advantages of simpleness in realization and low false rate, solves the problem of performance reduction or failure of a system designed by observation variables on the basis of Gaussian property assumption, and is especially suitable for non Gaussian property distribution conditions.

Description

technical field [0001] The invention relates to a fault diagnosis method for an analog circuit. Background technique [0002] The fault diagnosis of analog circuits is essentially equivalent to the problem of pattern recognition, and the key lies in the search for feature extraction techniques. Therefore, how to compress the original features of circuit states from high-dimensional feature space to low-dimensional feature space and extract effective fault features In order to improve the fault diagnosis rate, it has become an important research topic. [0003] When diagnosing analog circuit faults, for specific pattern recognition problems, if the collected raw data is directly sent to the classifier for classification, the calculation amount of the classification is often very large, the realization is difficult, and the classification effect is not good. , the false positive rate is high. Therefore, in the specific execution process, it is often necessary to first perfor...

AI Technical Summary

Problems solved by technology

Among them, the feature extraction method based on statistical theory plays an important role in the process of fault diagnosis of analog circuits. The commonly used methods include principal component analysis, factor analysis, etc. The distribution of the problem makes the calculation of the optimal transformation matrix difficult

When using wavelet transform to extract fault features of analog circuits, the key is to choose which type of wavelet base to choose. At present, there is no perfect theoretical guidance, and it is mostly determined based on experience or experiments. There are often errors and a high misjudgment rate.

Images