Distributed Cluster Control Optimizes Microgrid Power Allocation with Communication Weights

A new paper on arXiv (arXiv:2607.15322) introduces a distributed cluster economic dispatch (ED) scheme for cross-regional microgrids. The work, authored by Yalin Zhang, Zhongxin Liu, Yulin Chen, Donglian Qi, and Zengqiang Chen, addresses a key limitation of current consensus-based ED methods: they cannot guarantee differential demand between subgrids within a large-scale microgrid aggregated by a virtual power plant (VPP).

The researchers propose a method based on eigenvector centrality to design communication weight matrices for directed and connected graphs. This allows each cluster of agents (representing subgrids or distributed generators) to have the same eigenvector centrality value, enabling cluster-specific consensus. A leader-follower cluster consensus controller then drives the marginal cost (MC) of each generator to achieve multiconsensus, effectively allocating power across distributed generators (DGs) according to economic efficiency. Additionally, the VPP collects power deficits from each subgrid and distributes them to utility grids based on predetermined ratios, maintaining supply-demand balance.

Simulations confirm the scheme's effectiveness. The approach is significant because it preserves the autonomy of subgrids while achieving global economic optimization. By using a directed communication network and requiring leader information for only a few clusters, the scheme is practical for real-world microgrid deployments where communication is asymmetric and not all nodes have equal access. This work bridges graph theory and power systems, offering a scalable solution for cross-regional energy management.