Strategy State Maintenance And Fault Diagnosis Of Power Transformer

Power transformer is one of the most important equipments in the power system, and is also the link that has a high failure rate. Its running state is the basis for the safe operation of power grid. Therefore, detecting the transformer condition at any time, timely diagnosing and eliminating the trouble hidden faults in time, planning the reasonable maintenance strategy, have already become a significant theoretical and practical value subject in the power system.In view of the importance of the transformer, a study contents is mainly conducted in this paper as follow:In this paper, firstly, the power transformer fault type and the common methods of fault detection techniques have been mainly researched for further, the performance of transformer oil and the mechanism of gas production have been analysed, moreover, through a comprehensive analysis between them, a fault and the corresponding gas production characteristics has been educed.Secondly, the condition assessment about the parts of transformer body, bushing and core have been made separately to obtain a comprehensive state of the transformer, then judging the comprehensive state of transformers finally from integrating history, present, future state by Bayesian network. The health status of the transformers are divided into five states according to the various types of preventive test parameters, the various types of status information of the transformers are predicted, so that the relevant parameter data in the future time is educed; Then each of history, present, future experimental data is separately graded according to the hierarchical assessment model for getting the history, current and future health status of transformers. Judging the comprehensive state of transformers from integrating Bayesian network history, present, future state. Thus according to the state to make a maintenance strategy based on an comprehensive state of transformers. Finally,during the course of the transformer condition assessment, there may be found some hidden transformer faults, so that the decision tree ID3 and Bayesian network are integrated for comprehensive fault diagnosis of transformer for further in this paper, to determine the fault causes and possible fault positions. The synthesized diagnosis method of transformer faults based on the decision tree and Bayesian network, by the result of practical fault diagnosis examples, can make the classification accuracy of transformer faults higher and the classification more robust.