| The operation state of transformer is closely related to the safety of power grid.In order to ensure the healthy operation of power grid,real-time monitoring and fault diagnosis of transformers are needed.Therefore,it is important to study the fault diagnosis methods of transformers to ensure the reliability of power supply.In order to improve the accuracy of transformer fault diagnosis,this paper presents a diagnosis method based on Improved Particle Swarm Optimization(IPSO)to optimize RBF neural network.Firstly,a transformer fault diagnosis model based on RBF network is established,which utilizes the good pattern recognition ability and strong self-learning ability of RBF network,and can accurately express the mapping relationship between dissolved gases in oil and faults.Secondly,in order to improve the difficulty of determining the connection weight and width vectors of RBF network and improve the accuracy of RBF diagnosis model,this paper optimizes the weight and width of RBF network by using the characteristics of global optimization of particle swarm optimization.In order to enhance the optimization ability of PSO,an improved strategy for inertia weights is proposed,and Rastrigrin function is used to test the feasibility of the improved algorithm.Finally,the IPSO-RBF transformer fault diagnosis model is established,and the fault data of State Grid Corporation are used for simulation and case analysis.The results show that the improved particle swarm optimization is more powerful than the standard particle swarm optimization,and the premature convergence of the standard particle swarm optimization is successfully avoided.The diagnostic accuracyof RBF diagnostic model is 81.1%,while that of IPSO-RBF diagnostic model is 96.7%,and the results are consistent with the actual faults in the case verification.Thus,the diagnostic accuracy of IPSO-RBF diagnostic model has been significantly improved and can be used in engineering practice. |
PDF Full Text Request