Fault Diagnosis of Power Transformer Based on Improved Neural Network
Hailong Ma, Huaning Song, Chengjv Meng, Renli Wang · 2024
To detect the latent fault of a power transformer before it is too late and improve the accuracy of fault diagnosis, this paper proposes a fault diagnosis method for power transformers based on a BP neural network with improved particle swarm optimization (IPSO). First of all, to solve the problem that particle swarm optimization (PSO) is prone to premature and oscillatory phenomena, an improved particle swarm optimization is proposed by randomly varying the inertia weight of PSO and nonlinearly changing the learning factor of PSO so that it can jump out of the local optimum. Then, IPSO is used to globally optimize the initial weights and thresholds of the BP neural network, and the optimized BP neural network is applied to power transformer fault diagnosis. Finally, by training and testing the transformer fault sample data of substations and power plants and comparing it with BPNN and PSO-BPNN models, further verification is carried out by using the data of six groups of power grid companies that have clearly defined the cause of transformer failure. The results show that the IPSO-BPNN model performs better than BPNN and PSO-BPNN in power transformer fault diagnosis. The method presented in this paper has important reference significance for intelligent fault diagnosis of power transformers and other high-voltage electrical equipment.