Comparison of Pattern Recognition Techniques of Loss of Excitation and Power Swing Using Machine Learning

Recently, as it is time to replace hydraulic power equipment, modernization projects are being promoted. In addition, industrial competitiveness is being secured through localization of core technologies for main and auxiliary devices in the field of hydropower and pumped storage. In order to pioneer new overseas market, it is necessary to proactively apply AI techniques to improve the next-generation hydropower generator system. In this paper, to attempt an intelligent IED, we propose a comparison of pattern recognition techniques using machine learning for loss of excitation and power swing, and verify the performance by comparing the pattern recognition results of two AI techniques, such as SVM and LSTM.