Research on Error Evaluation of Capacitive Voltage Transformer Based on EMD-LSTM

Abstract

Capacitive voltage transformer (CVT) is a kind of electrical isolation and metering equipment. At present, the most widely used type of gateway voltage transformer in high-voltage and ultra-high voltage power grids is capacitive voltage transformer. The evaluation of its error determines the fairness and quality of the electric energy market. The safety of the power system. In this paper, by analyzing the error characteristics of the gateway voltage transformer, the error formation mechanism is analyzed and determined, and the deep learning algorithm is used to evaluate the error. Before the evaluation, the empirical mode decomposition method is used to decompose the characteristics of the monitoring data of the voltage transformer, and the components containing different time scale information are extracted. Each component is predicted using the long-short-term memory (LSTM) network, and finally the prediction result is obtained. At the same time, to avoid sudden changes caused by fluctuations and noise, multi-step averaging is used to reduce the possibility of misjudgment. The final false alarm rate is reduced to 0.1790%, improving the performance of the prediction.