Spectral clustering-based partitioning of volt/VAR control areas in bulk power systems

This study proposes a novel method to partition volt/VAR control areas in bulk power systems using spectral clustering. By revealing the internal connectivity structure of the network considering the power-flow-based volt/VAR model, an undirected edge-weighed graph is formed according to the defined electrical distance. On the basis of the graph theory, a novel approach is proposed through appropriate volt/VAR control area number selection, quality factor design, subspace projection, and spectral clustering. The approach is to partition the system into multiple volt/VAR control areas that have strong internal connections but weak external connections. Meanwhile, the proposed approach maintains the reactive power balances in each individual area and the minimum reactive power flow between those areas to facilitate area voltage regulations. Simulation results based on the New England 39 and IEEE-118 test systems demonstrate the effectiveness of the proposed method. In addition, the Union for the Coordination of the Transmission of Electricity system is further employed to evaluate the performance in practical bulk systems. These results show that the proposed approach can partition volt/VAR control areas appropriately in the power system.