The Intelligent Optimization Of Feature Extraction For Multiple Soft Fault Diagnosis Of Analog Circuit

Analog circuit fault intelligent diagnosis technology is an important researchcontent in the field of circuit testing. Pattern recognition is the core problem of theprocedure of fault diagnosis. The key to this problem is optimizing the process ofselecting and extracting fault feature of analog circuits. The initial parameters forpattern recognition process often contain a lot of redundant information. If all ofthese parameters are used for fault diagnosis, the calculating amount will be too bigto calculate, that will reduce the efficiency of failure diagnosis, accuracy, even mayaffect the diagnosis. Therefore, it is particularly important for the work of faultfeature selection and extraction. The paper uses Wiener kernel to describe nonlinearanalog circuit, extracting and selecting Wiener kernel of the analog circuit faultfeatures by optimized particle swarm ant colony algorithm, in order to achieve thecircuit diagnosis efficiently and accurately.The paper selects the Wiener orthogonal series to describe analog when circuitsmodeling. Discussing the discrete circuit Wiener core acquisition method andindirect method, then model the circuit by the low order Wiener kernel as well as theprinciple and method of Wiener core based intelligent fault diagnosis of analogcircuit.The paper discussed the optimized and intelligent select and extract method tothe feature of measured circuit. Research the ant colony algorithm and particleswarm algorithm and analysis their principle and mathematical model firstly. Thenimproved hybrid particle swarm ant colony algorithm is put forward by combing thetwo algorithms and experiment simulation the result. The experimental validatingthis method is an effective method comparing with the ant colony algorithm and thePSO algorithm, hybrid particle swarm has better global optimization ability, highsearch efficiency and fast convergence speed. Then study extracting feature sectionmethod by hybrid particle swarm combined with the ant colony algorithm, andextracting fault feature parameters of the Wiener kernel of the analog circuit. The experimental validating the ant colony algorithm is an effective and accurate methodof extracting fault feature parameters of the analog circuit Wiener kernel. Thismethod can used in getting the characteristic parameters of fault diagnosis of analogcircuits intelligent.The paper designed the circuit hardware and the PC software of the faultdiagnosis of the analog circuit in the basis of theory study and simulation verify, andthe whole fault diagnosis system is formed. Using this system to diagnosis the multisoft fault of the measured circuit can get the state of the fault components of themeasured circuit fast and accurately.