Partial Discharge Localization Method of Dry-type Air-core Reactor Based on Irradiance Analysis
The localization of partial discharges in the dry-type air-core reactor is of vital significance for ensuring power quality and saving maintenance cost.Aiming at the problem of low accuracy of the existing partial discharge localization methods in the practical application of reactors,this paper proposes a partial discharge localization method for the dry-type air-core reactor based on irradiance analysis.Firstly,it establishes a three-dimensional optical simulation model according to the typical structure of dry-type air-core reactor and analyzes the irradiance distribution characteristics of different positions of the discharge light source at the bottom of the encapsulation.Meanwhile,the features of irradiance distribution matrix are extracted based on the maximum variance theory to construct an optical simulation original set.Additionally,the paper uses a cross validation method to optimize the parameters of the random forest(RF)model and the back propagation(BP)neural network,and the test set to evaluate the localization effect of the model.Finally,the proposes localization method is verified empirically by using the dry-type air-core reactor partial discharge experimental platform.The results show that the localization accuracy of both models is above 90%.Compared with the BP neural network,the localization accuracy of RF classifier is higher and the predict time is shorter,which verifies the effectiveness of discharge localization method based on irradiance analysis.
万方数据
维普
