Variational Autoencoder Based Anomaly Detection for AIOps of Power Grid Supercomputing Center

With the fast development of power system and big data, the power grid supercomputing center faces the problem that index anomaly detection is not timely and accurate, which affects safe and stable operation of the power system. To solve the above problems, we propose to use a small variational autoencoder network for index anomaly detection of power grid supercomputing center for the first time. First, the index data is preprocessed and normalized, and then the normalized data is fed into the proposed small variational autoencoder for training and testing. Finally, the best F1-Score of 0.9091 is obtained on the publicly available GAIA index anomaly detection dataset, which can timely and accurately detect whether indexes of the power grid supercomputing center are abnormal.