The large-scale integration of electric vehicle (EV) and exchange stations (EC) into distribution networks introduces strong spatiotemporal load fluctuations and charging capacity constraints, leading to frequent voltage violations and reduced control flexibility. Traditional centralized control approaches face critical limitations, including high communication latency and computational complexity. To address these challenges, this paper proposes a Hybrid Intelligence (HI)-driven framework for distribution networks, which explicitly considers EV/EC charging power limits, cluster-level resource balance, and voltage security constraints. By incorporating spatiotemporal characteristics with intelligent optimization techniques, a Variant Monte Carlo Sampling (VMCS) algorithm is developed to generate the initial node partitions. These partitions are further refined using a Capacity-Corrected K-means Extension combined with Simulated Annealing Optimization (CCE-SAO), resulting in an optimized cluster configuration. A two-layer control architecture, termed “spatiotemporal collaborative optimization—distributed iteration,” is established to effectively address the drawbacks of traditional static clustering methods in large-scale systems, such as vulnerability to local optima and limited adaptability. This enhances global optimization under complex operational scenarios. Simulation results on the IEEE 33-bus and IEEE 123-bus test systems show that the proposed HI-based method effectively improves voltage quality. In the IEEE 33-bus system, the voltage deviation at the most fluctuating node is reduced by 30% compared with conventional K-means clustering and by 40% compared with centralized control under peak load conditions, validating the effectiveness of the proposed framework for future distribution networks with large-scale EVEC integration.
Zhang et al. (Thu,) studied this question.