Research On Fault Diagnosis Of Power Transformer Based On Infrared Image Data Fusion

With the development of the power industry,the demand for power in various fields is constantly increasing.The stable and safe operation of the power grid can ensure the rapid development of other industries.The transformer is very important in the transmission of electric energy,and its operation is closely related to the grid system.How to quickly and accurately determine the fault of the transformer is a problem that needs to be solved urgently.Due to the intricate causes of transformer faults,the use of a single diagnosis method is prone to problems such as misdiagnosis and low diagnosis accuracy,which affect the diagnosis results.In order to accurately and comprehensively diagnose the fault status of transformer equipment,this paper has carried out the research of transformer fault diagnosis method based on infrared image data fusion.The main contents of the paper are as follows:1)Preprocessing and feature extraction of transformer infrared images.By analyzing the noise of the transformer infrared image,the denoising effects of three different methods are compared.The K-means clustering segmentation algorithm is used to divide the denoised transformer infrared image into different levels,which lays a foundation for extracting different temperature characteristics.Correlate the gray value of the infrared image of the transformer with the temperature value,directly convert the gray value of the infrared image into a temperature value,and construct a temperature characteristic quantity that can express the fault state of the transformer.2)Research on transformer fault diagnosis method.Aiming at the problem of varying degrees of influence of different parts of the transformer on the transformer,the temperature characteristics of different parts are used for weighted fusion processing,as the input for diagnosing the state of the transformer,and the level of the fault state of the transformer equipment is used as the output,and the transformer is designed The fault diagnosis model proposes to apply BP neural network and support vector machine to transformer fault diagnosis.PSO algorithm and genetic algorithm are used to optimize BP neural network and support vector machine to improve the accuracy of diagnosis.3)A fusion diagnosis method for transformer fault decision-making based on D-S evidence theory is proposed.The diagnosis output results of the PSO optimized BP neural network and the genetic algorithm optimized support vector machine are used as the input of the decision-making fusion diagnosis method.According to the fusion rules of D-S evidence theory,the fusion results are obtained and verified by experiments.Through the training and testing of the PSO optimized BP neural network diagnosis model and the genetic algorithm optimized support vector machine diagnosis model,and the use of DS evidence theory fusion rules for decision-making fusion,the results show that the influence of different parts of the transformer on it is solved The problems of varying degrees improve the accuracy of fault diagnosis,avoid the shortcomings of misdiagnosis by a single diagnosis method,and illustrate the feasibility of this method in transformer fault diagnosis.