Research on Intelligent Identification Method for Voltage Abnormal Data Based on Deep Learning

In the context of the new power system, this paper addresses the issue of low efficiency in identifying abnormal data within the massive load data of Shanghai distribution network. An intelligent identification method for voltage abnormal data based on deep learning is proposed. Firstly, the sequence characteristics of different electricity consumption data are extracted through a multi-level DBSCAN clustering analysis algorithm, and the data is classified into four types. On this basis, an intelligent anomaly detection model is constructed by combining the Bi-LSTM algorithm with the attention mechanism, and the model results are evaluated. An online system is designed and developed to visually display the identification results. Practical application results demonstrate that this method can quickly and accurately identify abnormal data in the massive voltage data of Shanghai distribution network, effectively improving the work efficiency of power grid companies and reducing the risk of grid failures.