Community Detection in Energy Networks based on Energy Self-Sufficiency and Dynamic Flexibility Activation

Abstract: The global energy transition towards distributed, smaller-scale resources, such as decentralized generation and flexible assets like storage and shiftable loads, demands novel control structures aligned with the emerging network architectures. These architectures consist of interconnected, self-contained clusters, commonly called microgrids or energy communities. These clusters aim to optimize collective self-sufficiency by prioritizing local energy use or operating independently during wide-area blackouts. This study addresses the challenge of defining optimal clusters, framed as a community detection problem. A novel metric, termed energy modularity, is proposed to evaluate community partitions by quantifying energy self-sufficiency within clusters while incorporating the influence of flexible resources. Furthermore, a highly scalable community detection algorithm to maximize energy modularity based on the Louvain method is presented. Therefore, energy modularity is calculated using linear programming or a more efficient simulation-based approach. The algorithm is validated on an exemplary benchmark grid, demonstrating its effectiveness in identifying optimal energy clusters for modern decentralized energy systems.