Fault Diagnosis Of Analogcircuit Based On BP Neural Network

With the development of modern industry, the traditional fault diagnosismethod reflects the limitations. And the development of artificial intelligencetechnology is for fault diagnosis of analog circuits opened a new train ofthought and method, especially the development of artificial neural networkresearch in recent years to a great role for the development of analog circuitfault diagnosis of the.At present in analog circuit fault diagnosis using BP neural network, BPneural network is used to have the following advantages: the utility model hasthe advantages of analog circuit fault diagnosis is measured after diagnosisspeed, strong real-time performance, the reason is this method makes use ofthe information processing ability of neural network is highly parallel.In this paper, the theoretical basis of BP algorithm to analyze theoperation mechanism of BP algorithm, the study of traditional BP algorithmto solve the problem of fault diagnosis of analog circuit performance, animproved method and analog circuit fault diagnosis problem in Matlabruntime environment.This article will be "analog circuit fault diagnosis--BP algorithm-optimized fault feature extraction–Mat lab--BP algorithm," the main line.The full text is divided into the following sections: Introduction⑴BPalgorithms and basic principles; research methods⑵fault feature extraction;⑶analog circuit based on BP algorithm to solve the fault diagnosis problem;⑷BP algorithm by comparing the experimental data and the optimized BPalgorithm performance.This paper used LED dimming lamp circuit as an example, studies thecomposition of diagnosis system based on BP network diagnosis process andputs forward the corresponding.At the same time, the data preprocessingmethod commonly used is research, through2groups of experimental design are verified the traditional BP network and performance comparison of BPnetwork improved the additional momentum method based on the feasibilitystudy of the BP network, and improved based on PCA for the experimentalcircuit fault diagnosis.