Two-stage Flow-based Partition Framework for Unbalanced Distribution Networks

Multiarea parallel computing has emerged as a powerful tool for improving operation and control efficiency in the context of expanding scale of new power systems. However, existing partitioning methods cannot handle the problem of three-phase imbalance; results are not stable enough and have not been verified in large-scale distribution networks. Driven by this motivation, this paper presents a two-stage algorithm that can be applied to large-scale distribution networks to produce non-overlapping and overlapping zoning results. The algorithm is split into two sections: sketchy partitioning and multi-region merging. A voltage/admittance matrix must first be used to determine electrical distance between nodes. The Louvain algorithm is used to obtain the result of coarse partition. The flow algorithm or hierarchical flow method is then employed, depending on limitations of the number of partitions, to obtain sub-regions with comparable scales. Finally, IEEE 123-node test feeder and IEEE 8500-node test feeder are used to validate the algorithm. Simulation results and comparisons show the proposed partition algorithm can produce more reliable and efficient partition results by contrast with existing methods, in case of unbalanced three-phase and large-scale distribution networks.