Benders decomposition algorithm for reference network

Abstract

A reference network is topologically identical to the existing network in which the generators and loads are not changed while the transmission line capacity will be optimized. By applying a reference network, the optimal transmission capacities of each line could be identified, as it takes annual operational and transmission costs into consideration. A reference network provides a standard to evaluate the real network objectively. The congestion costs can be quantified and the optimal investment proposal can also be obtained based on load forecasting. Furthermore, the weakness of the system can be detected by comparing those in the reference and the actual network. However, for a large-scale power system, it is extremely complicated to build the reference network, as the optimal capacities for each line must be determined considering various contingency events. In this paper a simplified Benders decomposition algorithm is presented and applied to solve this model. Simulation results show that this technique can reduce computation time and improve convergence performance.