Research On Key Technology Fault Diagnosis Module Base On Embed

Analog-circuit fault diagnosis is a significant field which has already formed a set of theory and methodology in recent years. However, as its multiplicity and complexity, it needs further development and improvement.This paper presents intelligent fault diagnosis using neural network and genetic algorithm optimize neural network, especially focus on Backpropagation neural networks(BPNN). It is also to improve and realize the classical techniques, by building up practical electric circuit and diagnosing artificial faults with the BP network. The testing shows that it can classify the type of faults accurately for pattern recognition by selecting the correct BP Network Topology and setting the appropriate parameters, when acquiring the main signal by testing precisely at the same time. Referring to other intelligent diagnosis methods, this investigation combines the neural network and genetic algorithm into GA-BP network, taking advantage of global detecting of genetic algorithm, to enhance the ability of global detecting. Finally simulating in Pspice software, this approach indeed develops the performance of network recognition.According to the methodology of fault diagnosis design, board S3C210 is selected for buil-in system based on ARM920T kernel of Sumsung Company. In this board, a driver is programmed to display results on LCD screen for fault diagnosis with intelligent algorithm. Meantime, in order to develop a better operation interface for fautl diagnosis, Linux operate system and MINIGUI are transplated into the board to improve its utility.