Reliability Research Of Power Transformer Based On Grey Theory

The transformer is the key equipment in the power grid, with the advance of networkintelligence, the important position of transformer is more and more prominent in the grid. Thesecurity and stability of the power system are directly affected by operation reliability oftransformers, thus researching the reliability of the power transformer has important practicalsignificance.Firstly, the significance and necessity of studying the reliability of power transformers areexpounded in this paper. The researching status of transformer at home and abroad is expoundedand then the main content of the research is given.Secondly, the gray system theory knowledge is introduced in this paper. Modeling processof gray relational analysis method and gray forecasting model of GM (1,1) are expounded. Theproblem of discrimination coefficient in Grey relational analysis is discussed, the applicableconditions of gray GM (1,1) prediction model and related parameters are analyzed. For thecharacteristics of the modeling of gray relational analysis and gray GM (1,1) prediction model,the strategy conception that optimization algorithm will be applied to gray system theory is clear.The reliability of transformer insulation fault is researched in this paper. Transformerinsulation fault tree is finished by FTA method, and the probability of failure for the events offault tree is calculated. Reliability of transformer insulation fault is analyzed by gray correlationmethod, for the problem of discrimination coefficient in Grey relational analysis method, thecoefficient is optimized by particle swarm optimization and this method is applied to thereliability analysis of transformer insulation fault, which got the better the results and the firstfour factors causing the transformer insulation fault are clear.Finally, the transformer reliability prediction is researched in this paper. For the problem ofconstruction of background value in traditional gray forecasting model, the model parameters areoptimized by fish swarm algorithm method. The failure gas content is predicted throughoptimized model parameters and the transformer state in the next period of time is predicted bymerging three-ratio method. Instance results show that the feasibility and effectiveness of thefish swarm optimization grey prediction model are verified.