A Decomposition System to Solve AC Optimal Power Flow Problem

摘要

We develop a decomposition system to solve AC optimal power flow (ACOPF) problem iteratively. The decomposition system comprises one nonconvex ACOPF module, four convex ACOPF modules (SOC- ACOPF), one network partition module and one parallel computation module. All convex ACOPF modules of the decomposition system are derived from the nonconvex ACOPF module (Module-Q). The underlying idea of the developed SOC-ACOPF modules is to drop simplifying assumptions in Module-Q by convex relaxation and estimation methods. The network partition module is used to partition large networks to multiple sub-networks. Our numerical results show that the proposed modules and decomposition system can solve the ACOPF problem efficiently. We also develop a parallel computation module to accelerate the implementation of the decomposition system. The numerical results of our proposed parallel computation module show fast convergence. The computational efficiency is also improved by increasing the network partition depth in our decomposition system.