Distribution Network Anomaly Detection Algorithm Based on VAE

Abstract

With the booming development of economy and technology, China's electric industry has gradually realized intelligent, has developed into a comprehensive network including computer network, power network, information network. The current power information monitoring technology is mainly aimed at the power generation, transmission and transformation stage, and the lack of effective power information detection means in the power distribution stage. Therefore, a distribution network anomaly detection algorithm based on variational auto-encoder is proposed to solve the problem of anomaly detection of distribution terminal data. The input time series power load data is compressed and reconstructed, and the anomaly degree of samples is detected by reconstruction error. Experimental results show that the proposed algorithm has high detection rate and accuracy, as well as high robustness.