A dynamic partitioning method for microgrid clusters considering load balancing

With the advancement of the "dual carbon" strategic goals and the green transition of the energy structure, it is necessary to study dynamic partitioning methods for microgrid clusters in order to address the increasing complexity of microgrids and optimize the control architecture. However, the current dynamic partitioning metrics for microgrid clusters have not yet been clarified. This paper thus proposes a dynamic partitioning method for microgrid clusters based on the enhanced elite retention genetic algorithm with consideration of load balancing. Firstly, the regulation characteristics of the internal resources within the microgrid cluster are described. Then, a comprehensive index system for dynamic partitioning is established for microgrid clusters, which includes three metrics for microgrid cluster partitioning: modularity, voltage regulation capability, and regional load balancing. An enhanced elite retention genetic algorithm is used to solve the fitness of the comprehensive index. Finally, a simulation is performed on an improved IEEE-123 node multi-microgrid interconnection scenario model. The results show that the solution time of the enhanced elite retention genetic algorithm is reduced by 27.7% compared to the traditional genetic algorithm, and the performance of the dynamic partitioning method with weight updates is improved by 6.4% compared to the fixed partitioning method.