Research On Transformer Fault Diagnosis Method

As the core equipment in our country’s power system,transformer working state is bound up with the stability of power grid.If the transformer happen abnormal or fault situation,it will have a negative impact on the power supply of the power grid,thus resulting heavy loss in the economy.Therefore,the study of efficacious transformer fault diagnosis ways is great important to the normal operation of the grid.Aiming at the problem of transformer fault diagnosis’ s low accuracy,a transformer fault diagnosis way which is improved shuffled frog leaping algorithm(ISFLA)to optimize the parameters of support vector machine(SVM)mode is proposed.And complete the construction of the fault diagnosis system.The main research contents are as follows:(1)Researches on the fault mechanism and key factors of the transformer.Through the study of transformer faults,the formation mechanism of dissolved gas in transformer oil is analyzed,the reason of dissolved gas generation is explained,and the corresponding relationship between the content and type of characteristic gas and transformer internal fault is mainly studied.(2)Improve and test the shuffled frog leaping algorithm.Aiming at the shortcomings of the shuffled frog leaping algorithm,proposes an an improved shuffled frog leaping algorithm.The algorithm uses a combination of uniform randomization and anti-approximate opposition to improve population initialization;The differential mutation operator is introduced to improve the local update strategy,the Levy mutation operator is used to instead of the random update operation,and the cosine adaptive skeleton differential evolution algorithm is embedded into the local search.Through simulation experiments,it is verified that the performance of ISFLA is improved greatly.(3)Research on transformer fault diagnosis method based on ISFLA-SVM.ISFLA is introduced into SVM to realize the optimization of SVM model parameters,and the fault diagnosis model is established.Through analyzing the method of selecting fault characteristics of transformer,select characteristic gas content and ratio as fault samples for experiment.The results show that the optimized SVM model can accurately diagnose transformer faults.(4)The fault diagnosis system of transformer is realized.With GUI and SQL server19 as development tools and database,the overall architecture design and human-computer interface development of transformer fault diagnosis system are completed.The experimental test results verify the feasibility of the system.