Integration of a variety of signal characteristics of analog circuit fault diagnosis
|School||Huazhong University of Science and Technology|
|Course||Communication and Information System|
|Keywords||Analog Circuit Fault Diagnosis Waveform classification Multidimensional vector fuzzy sets Transient fault dictionary|
The rapid development of electronic technology, the circuit scale and integration has been an unprecedented development, the analog circuitry and analog-digital hybrid circuit for a long time is still irreplaceable part, so analog circuit fault diagnosis study has important theoretical significance and application value. There is no effective applied to the analog circuit fault diagnosis set of methods appears. This paper describes the analog circuit fault diagnosis purpose and significance of the study, and the analog circuit fault diagnosis history and current status are described. Studied the signal waveform feature extraction and classification method, in a detailed analysis of the signal waveform classification in various fields on the basis of research and application, focusing on the waveform template matching and autocorrelation method of analysis. Platform for the current shortcomings of the DC fault dictionary, in order to increase the signal discrimination, and taking into account the actual circuit in the data acquisition automated test, we proposed a fusion of various signal characteristics simulated failure diagnosis method of signal characteristics include extraction scheme, the waveform classification and recognition programs, multi-dimensional vectors into fuzzy sets, transient fault dictionary generation and optimization, as well as the organizational structure of the standard interface file. Finally, in the \Using autocorrelation analysis and curve fitting with the zero crossing, histogram statistics, discrete cosine transform method of extracting signal features, and feature extraction based on the results, using the created waveform templates, similarity is calculated, a combination of statistical analysis of the waveform classification. Based on the signal characteristics and the waveform feature feature vector components and complete division of fuzzy sets multidimensional vector and transient faults dictionary generation and optimization. After introduction of the ATE test framework generates an analog circuit fault diagnosis standard interface file, the file will be applied to such a standard interface for automated testing and the actual circuit diagnosis. This paper first presents the characteristics of the signal and generates a feature vector composed of multidimensional fuzzy set method, and the concept of transient fault dictionary, which combines the DC fault dictionary is easy to implement and applications, while the AC fault dictionary can be more effectively diagnose some sensitive components advantages. Through the DC and transient fault dictionary comparison, transient fault dictionary significantly better than DC fault dictionary.