Transformer Fault Diagnosis Based On Grey Wolf Algorithm Optimizing Support Vector Machine

Power transformer is the core equipment in the system,which determines whether the power system can run stably.Finding abnormal operation of transformer in time can avoid serious fault of transformer and is beneficial to stable operation of power system.The more accurate and reliable method of transformer fault diagnosis is dissolved gas analysis in oil method.According to the relationship between the volume fraction of gas in oil and the fault type,the transformer fault diagnosis can be carried out.Traditional transformer fault diagnosis and artificial intelligence transformer fault diagnosis are based on the analysis of dissolved gas in oil.With the development and maturity of artificial intelligence,more and more artificial intelligence methods are applied to transformer fault diagnosis,and some achievements have been made.Support vector machine can solve not only machine learning problems with small samples,but also local minima and structural selection problems in neural networks.Therefore,the volume of dissolved gas in the transformer oil is used as the Support vector machine input to diagnose the transformer fault.The performance of the Support vector machine classification is mainly determined by the kernel function of the Support vector machine and its parameters.Therefore,the research of transformer fault diagnosis using Support vector machine is focused on the selection of kernel function and the optimization of kernel function parameters.In this paper,a transformer fault diagnosis model based on the Grey Wolf Optimization Algorithm is proposed,which has few adjustment parameters and good ability of parameter optimization.In order to verify the superiority of the proposed model,it is compared with the Support vector machine model and the Support vector machine model.By comparing the same data with different methods,it is found that the proposed model not only improves the accuracy of transformer fault diagnosis,but also shortens the fault diagnosis time compared with the other two methods,it has better diagnostic effect.This study provides a more efficient and accurate method for transformer fault diagnosis,and provides a new solution for transformer operation state prediction and fault diagnosis in the future.