Research on Machine Learning Based Fault Diagnosis Methods for Power Transformers

With the rapid development of science and technology, the power system has become an indispensable part of people's daily life. Transformers play an irreplaceable role in the power system, and their operation status directly affects the reliability of the power supply of the grid. In order to improve the accuracy of oil-immersed dissolved gas analysis to diagnose transformer faults, a transformer fault diagnosis method based on the hybrid grey wolf optimization algorithm (HGWO-SVM) is proposed from the perspective of improving the classification and recognition performance of the support vector machine (SVM) and mining the deep feature information embedded within the data of the dissolved gas in the oil. Elite backward learning, nonlinear control parameters and Levy flight strategy are adopted to improve the grey wolf optimization algorithm, and the penalty factors and kernel parameters of SVM are efficiently optimized to ensure the best classification performance of SVM. The results of constructing the HGWO-SVM fault diagnosis model show that the feature extraction of data can effectively improve the diagnosis accuracy and computational efficiency, which provides theoretical support for the subsequent research.