Research on distributed photovoltaic cluster partition and dynamic adjustment strategy based on AGA

Researching cluster partitioning and adjustment methods is essential for effectively implementing cluster control strategies and ensuring the safe operation of power grids amid challenges like reverse power flow and voltage violations resulting from large-scale distributed photovoltaic grid integration. The paper comprehensively evaluates factors including electrical distance, supply-demand balance, and power transmission. It introduces a method for partitioning and dynamically adjusting distributed photovoltaic clusters, based on an improved adaptive genetic algorithm (AGA). Firstly, a comprehensive performance index is introduced based on modularity, incorporating intra-cluster power supply rate and inter-cluster power transmission. Next, an unweighted adjacency matrix is employed to encode chromosomes, and an adaptive optimization parameter adjustment strategy is introduced to enhance the AGA, ensuring both direct node connectivity within clusters and improved global solution performance during optimization. By integrating these comprehensive indicators with AGA, a cluster partitioning optimization model is formulated based on the comprehensive indicator system. The IEEE 33-node system is utilized for cluster partitioning simulation analysis with this model. The outcomes demonstrate that, in contrast to the single modularity index-based cluster partitioning approach, the proposed comprehensive indicator method enhances the intra-cluster active and reactive power supply rates by approximately 20.34% and 28.61%, respectively, while decreasing inter-cluster power transmission by roughly 24.98%. Furthermore, in scenarios of insufficient photovoltaic power output, the cluster scheme, after dynamic strategy adjustment, notably mitigates system voltage fluctuations and network losses.