Abstract
Analyzing the uncertainty of electric vehicle (EV) charging behavior is of great significance for the safety, stability, and reliability of power distribution networks. Cluster analysis can effectively and conveniently analyze EV charging behavior. This paper proposes a hybrid birch-k-means algorithm to aggregate and analyze a large amount of EV charging data. Both clustering characteristics and behavioral features are considered in the analysis of charging behavior. Firstly, the birch-k-means algorithm is used to cluster nearly 10,000EV data points, resulting in four cluster groups. The clustering patterns are then analyzed and derived. The accuracy and practicality of the obtained clustering patterns are verified through an analysis of overall behavior. The four identified clusters offer a technical foundation for establishing scientific models and implementing optimized control strategies for electric vehicle charging infrastructure.