Discovering Community Structures in Power System Networks using Voltage - Reactive Power Sensitivity

For most large-scale systems, centrality is simply not feasible because it requires expensive and sophisticated data acquisition and communication networks. Moreover, with the integration of large-scale distributed energy resources into the power systems, it is desired to divide a one large system into multiple small systems to make the analysis and control feasible and comparatively easier. In this paper, an optimal power system network partitioning algorithm was presented which detects community structures (the partition of a group of network vertices into densely connected subgroups) in power system networks and divides the system into communities while minimizing intercommunity V-Q sensitivities and reactive power imbalance (because this partitioning was mainly done for decentralized voltage control). The presented algorithm was tested on three test power systems: 9bus, 39 bus and 118 bus systems.