Power Transformers Smart Configuration And Fault Prediction

Smart transformer is the electrical equipment of smart grid and intelligent substation,smart transformer level of smart related to the economy of the operating reliability ofintelligent substation and substation investment. Transformer fault prediction can findlatent faults as well as the development trend of the notice of faults of transformer.Research transformer fault prediction the state of the system of safe operation andtransformer maintenance is important. In this thesis, smart theory method for powertransformer and transformer fault prediction and fault diagnosis research and discussion,the main work is as follows:A. For the smart configuration of power transformers, we used two methods tostudy and explore from different angles. The first method is to first create a transformerMarkov model discussed based on long-term run-time of transformer to transformerreliability, the probability of each fault in the steady state, found that the largest possible tomonitor the occurrence of various fault line monitoring sensor, according to sensor theability of each failure on the importance of each sensor to sort in order to discuss the ismart configuration of the transformer. The second method is based on Life Cycle Cost ofthe transformer, based on the combination of the analytic hierarchy process and rough settheory to study the smart configuration of the transformer, the transformer reliability andeconomic indicators as the criteria, the establishment of the transformer intelligent theanalytic hierarchy model, the establishment of criteria and sub-criteria expertdecision-making table, using the weighted average of attribute importance to measure theimportance of each criterion in the hierarchy, as the criteria and sub criteria for the upperweights to get the optimal alternative configuration program.B. For transformer fault prediction, improved multivariable gray model GM (1, m),the development trend of a number of characteristics of gas at the same time to predict,considering the various factors change from a systems perspective, while the transformeroil characteristics of gases dissolved in the gray relational analysis to determine the degreeof close contact between them. Transform the original data sequence at the same time tooptimize the model background value of the model simulation accuracy and predictionaccuracy was significantly improved.C. For the characteristic quantities of gas oil fuzzy diagnostic model to study theprinciple of fuzzy comprehensive evaluation matrix initial value of the fuzzy relationshipbetween the ratio of coding and fault strike, the use of GM (1, m) model of gray predictionThe development trend of the prediction of transformer oil dissolved gas, and then as aninput to the output of the prediction model using fuzzy fault diagnosis of transformer failure, a transformer fault prediction and fault diagnosis.