Towards a Centrality Control Center: Innovations in Complex Power Network Operation and Control

The operational efficiency and reliability of power systems have become increasingly critical with the growing integration of renewable energy sources and microgrids. In this context, the analysis of power systems as complex networks has gained significant attention. Building upon prior research, this paper explores operational centrality metrics derived from both AC and DC power flow models, comparing their efficacy in capturing network dynamics. While AC power flow provides a detailed and accurate representation of power system behavior, its computational complexity can be a limitation for large-scale systems. Conversely, DC power flow simplifies computations, allowing for faster analysis but with reduced accuracy in active power flow predictions. This study proposes two key applications for DC centralities: first, by replacing AC centralities with DC-based metrics, we show that the results remain consistent, validating DC centralities as an alternative for operational assessments. Second, we introduce a novel approach using DC operational betweenness centralities to predict total active power loss, and by correlating these with AC-based results, a deep neural network is trained to forecast losses with reasonable accuracy. The findings suggest that DC centralities can effectively capture overall trends in power loss despite inherent inaccuracies, offering a practical tool for optimizing power system operations.