The Optimization And Realization Of Feature Extraction For Fault Diagnosis Of Analog Circuit Based On Wiener Kernel

Intelligent fault diagnosis of analog circuit occupies a test field of importantstatus based on accurate, efficient and intelligent character. Pattern recognition isthe core of intelligent diagnosis, feature extraction is the key of patternrecognition, can efficiently get measured circuit fault intelligent diagnosis systemof information directly determines the possibility of efficient fault diagnosis ofcircuit under test. Wiener series presented in this thesis was used to describenonlinear circuits, will feature extraction as an optimization problem, by ParticleSwarm algorithms to find optimal solutions of a colony, to achieve more efficientanalog circuit fault feature access.First introduced the theoretical basis of Wiener series and Wiener seriesdescribing nonlinear analog circuits, Wiener and describes discrete circuitmethod for obtaining nuclear, as well as the intelligent diagnostic system ofanalog circuits based on Wiener kernel.Analysis of Particle Swarm Optimization algorithm and Ant Colonyalgorithm, for the advantages and disadvantages of the two algorithms make AntColony hybrid Particle Swarm Optimization algorithm, improve the efficiency ofoptimization and MATLAB simulation of the rationality of the algorithm.Research based on Wiener kernel of principle and method of featureselection and extraction, on this basis, made based on Particle SwarmOptimization of Ant Colony Optimization hybrid algorithm for feature extractionmethod of Wiener, by Particle Swarm Optimization of Ant Colony Optimizationhybrid algorithm, effectively improve the efficiency of extraction and a casestudy on the methods.Design hardware and software of Intelligent fault diagnosis system, and PC design, realization of Ant Colony algorithm based on Particle SwarmOptimization of feature selection and extraction of Wiener kernel and instanceafter diagnostic contrast circuit to verify the feasibility of this method.Experiments show that presented particle swarm of Ant Colony algorithmbased on Wiener kernel feature extraction in fault diagnosis of analog circuitsoptimize, and adopted the typical circuit Diagnostics validated the feasibility ofthis method.