Study On Fault Feature Selection And Extraction Of Nonlinear Analog Circuit

Circuit is widely existed in life, industrial, medical, aerospace, military andother fields. Nowadays it has become a significant element in the development ofscience and technology. With the improvement of the technology, the volume ofcircuit becomes smaller. The reduction of tested nodes which are used for the circuitincreases the difficulty of diagnosis. Therefore, it becomes a hot research topic.Currently, the methods of intelligent diagnosis have become the main researchorientation. For the feature extraction difficulties of the nonlinear analog circuit, thispaper will use Volterra series to analyze the nonlinear analog circuit and extract thefeature parameters of the circuit. This paper uses the feature selection and featureextraction method which is based on the annealing ant colony algorithm.This paper studies the theory of Volterra functional series, uses the Volterrakernel to extract and analyze the feature of the measured circuit, and gets two kindsof acquisition method of the Volterra kernel;For the Volterra kernel intelligent diagnosis method of the circuit, the featureextraction is the most significant. The current methods have a certain effect, but it isstill not ideal. Therefore, this paper studies the principle of the analog annealingalgorithm and the ant colony algorithm. According to their merits and drawbacks, weintroduce the annealing ant colony mixed optimization algorithm, and validate that itcan improve the optimization efficiency through the simulation and comparisonexperiments. On the basis of the above research contents, this paper presents thefeature selection and feature extraction method based on the annealing ant colonymixed algorithm, and validates the feasibility of this method.The theory study needs to be tested by practice. This paper designs theintelligent fault diagnosis system which is used for experiment, and completes theproduction of hardware and software. We embedded the feature selection and featureextraction optimization method of the Volterra kernel based on the annealing ant colony algorithm. The experiment results which are diagnosed through the classicalcircuit validate the correctness and feasibility of this paper’s method.