Research On Transformer Fault Diagnosis Method Based On Improved Seagull Algorithm Optimized Probabilistic Neural Network

With the rapid development of social economy,there is an increasing demand for electrical energy in various fields.As a key equipment of the power system,it is vital to ensure the safe operation of power transformers.Therefore,it is significant to study the real-time monitoring and fault diagnosis method of transformer operation status.Dissolved Gas-in-oil Analysis(DGA)is a very efficient means of diagnosing transformer faults.Traditional transformer fault diagnosis methods,including the three-ratio method,the characteristic gas method and the David’s triangle method,are all based on DGA.With the maturity of artificial intelligence algorithms,the integration of artificial intelligence technology with DGA has significantly improved the accuracy of transformer fault diagnosis.To overcome the problems of fuzzy fault boundaries and low diagnostic accuracy of traditional transformer fault diagnosis methods,this paper establishes a transformer fault diagnosis model based on DGA with an improved seagull algorithm optimised probability neural network(PNN).The main research work of this paper is as follows:(1)The mechanisms of dissolved gas generation in transformer oil and the types of transformer faults are analysed in depth,and the correspondence between different fault types and characteristic gases is summarized.(2)To address the problem of redundancy in the original DGA data affecting the diagnostic accuracy.Kernel Principal Component Analysis(KPCA)was adopted to achieve non-linear dimensionality reduction by extracting principal components from the original data,the effectiveness of KPCA was verified by comparing the accuracy and computation speed of the fault diagnosis model with the original feature vector.(3)Based on the impact of the smoothing factor of PNN on the performance of the fault diagnosis model,the Improved Seagull Optimization Algorithm(ISOA)is established to optimize the transformer fault diagnosis model of PNN,and three strategies of Tent chaotic mapping,nonlinear inertia weight and Gaussian variation are introduced to improve the traditional seagull algorithm,several benchmark functions are selected for testing to verify the effectiveness of the improved strategies,in this paper,the KPCA-ISOA-PNN transformer fault diagnosis model is compared with other transformer fault diagnosis models and the simulation results show that the former has superior performance with an accuracy of 94.44%.(4)The software and hardware design of the online monitoring and fault diagnosis system for dissolved gas in transformer oil is studied to achieve real-time monitoring of dissolved gas content information in transformer oil and fault diagnosis of transformers,which provides a strong guarantee for safe operation of transformers.