Research On Fault Analysis And Diagnosis Of Regional Power Grid Equipment Based On Data Mining Technology

With the rapid development of domestic social economy,the continuous expansion of the scale of power grid system and the continuous enhancement of regional interconnection capacity,it not only improves the power supply capacity,but also increases the risk of safe operation of power grid.As the basis of power grid,the health status of power grid equipment is directly related to the safety of the whole power grid.Under the background of increasing power data,the original way of relying on manual information processing and fault analysis has become increasingly unable to meet the processing requirements of massive power grid information.It is urgent to study a method that can scientifically and quickly judge the equipment status based on the power grid equipment status information.This thesis will use the theory of data mining to establish the fault diagnosis model of power grid equipment,and apply it to practical work to find and eliminate the hidden dangers of equipment in advance.This thesis expounds the relationship between internal faults of power transformer and dissolved gas in oil,and analyzes the advantages and disadvantages of the original diagnosis methods.Under this theoretical system,a fault diagnosis model based on random forest algorithm is proposed,which takes the 15 pairs of non coding ratio of dissolved gas in oil as the model input and C4.5 algorithm decision tree to build a weak learner,and constantly adjust the number of decision trees and the maximum depth of decision trees to optimize the performance of the model.Based on the dissolved gas fault samples in transformer oil collected by Wuhai power supply company and relevant literature,the performance and feature selection effectiveness of different classification models and the proposed diagnosis model are compared and analyzed.The analysis results show that selecting the non coding ratio as the characteristic parameter can mine more transformer fault information,and the fault diagnosis model based on random forest algorithm has excellent performance in the evaluation indexes such as accuracy.Based on the above diagnosis model,the power grid equipment manager can judge the fault type inside the transformer by inputting the concentration parameters of each gas component in the test report,which improves the convenience of diagnosis.Through the application of typical transformer fault analysis scenarios in Wuhai power supply company in recent three years,it is proved that the model can accurately judge the fault and make up for the limitations of traditional diagnosis methods.In view of the technical and management problems found in the process of fault treatment,measures and suggestions are given,which provides a solution to improve the level of power grid management.