Research on Automatic Identification of Bad Data in Power System Based on Entropy Method and LightGBM

If there is bad data in the data, it may cause dispatchers to make wrong decisions, increase the risk of equipment failure, and threaten the safe and stable operation of the power system. Therefore, this paper proposes a research on automatic identification of bad data in the power system based on entropy method and LightGBM. Based on sliding window for power system data field filtering processing, and based on stacked autoencoder deep neural network for power data feature extraction; Calculate the entropy value of key nonlinear features of power system bad data based on entropy method, and implement automatic identification of bad data in power system based on LightGBM. The experimental results show that this method has a high effective screening ratio for bad data in the power system, and a high matching correlation value for identifying bad data types. Under different proportions of bad data, the proposed method has a recognition time of less than 50ms, high recognition efficiency, and better application effect in identifying bad data in the power system.