Power Transformer fault Diagnosis based on Hybrid Intelligent Algorithm

Abstract

The gas content in the oil is used as the fault input characteristic for the power transformer. Still, the accuracy of the diagnosis results is not ideal, and such a model is unstable. This research proposes a hybrid intelligent fault diagnosis method based on the improved grey wolf algorithm and an optimized probabilistic neural network. Firstly, a strategy of three nonlinear control factors is introduced to fit the grey wolves’ search process. The weighted distance was modified to update the position information of grey wolf elements to avoid the algorithm falling into the local optimum. Secondly, the performance of the improved grey wolf algorithm was tested through six commonly used functions. The results show that the improved grey wolf algorithm has high convergence accuracy and stability in both multimodal and unimodal functions. Finally, the improved grey wolf algorithm and the probabilistic neural network were combined to diagnose the oil-immersed power transformer through hybrid intelligent algorithms. As a result, the fault diagnosis model proved valid for transformer fault diagnosis.

Keywords

• transformer
• fault diagnosis
• control factor
• weighted distance
• grey wolf algorithm
• probabilistic neural network