Study On Condition Forecast And Diagnosis Of Reactor Based On Dissolved Gas-in-oil

Voltage reactor is one of the key equipment to EHV and UHV system, It is operation reliability directly impacts on the safe operation of power systems. Therefore, this paper chromatography using oil analysis DGA, the use of dissolved gas content of historical data predict the future of the reactor operation, and fault diagnosis.The research in this paper, the traditional grey forecasting model on the basis of the unbiased grey forecasting model introduced to the reactor's condition to forecast, through the example of unbiased grey forecasting model of validity. The use of this article based on extension theory, matter-element model of the neural network reactor fault diagnosis method. The method is based on the integration of complementary through, the extension theory of matter-element model and neural network theory combine to overcome a neural network to learn shelters, and other defects. Through comparing the extension neural network with the traditional three-ratio method diagnosis results, the accuracy of diagnostic methods and the effectiveness of multiple fault diagnosis was verified.