A Study of Intelligent Analysis of Abnormal Power Consumption Behavior Based on Daily Load Curve

Abnormal power consumption not only causes great safety risks to the power system, but also hinders the intelligent development of power enterprises. This study proposed a multi-feature fusion algorithm from the perspective of unsupervised learning, constructed a feature model of electricity consumption patterns, and verified the reliability of the algorithm through experiments. The research results show that the accuracy rate of the detection result of the multi-feature fusion algorithm plus power consumption mode feature model is as high as 95%, and the AUC value of its ROC curve is the largest. When the threshold is between 0.2 and 0.3, the algorithm works best. At the same time, the cluster analysis of users' electricity consumption patterns shows a good detection effect of abnormal electricity consumption behavior. This study has reference significance for the intelligent management of power systems and provides technical ideas for the safe development of power grids.