The decomposition and computation method for distributed optimal power flow based on message passing interface (MPI)

This paper investigates the decomposition and computation method for the distributed optimal power flow (DOPF) based on message passing interface (MPI) framework in large-scale interconnected power grids. Firstly, a DC-DOPF model and an AC-DOPF model are introduced respectively. Next, a new equivalent decomposition model to be used to solve DC-DOPF and AC-DOPF is proposed. It decomposes the OPF computation of large power grid into the sub-problems of interconnected multiple regions. Then, two different decomposition methods, i.e., partial duality and auxiliary problem principle (APP), are used to solve interconnected DC-DOPF and AC-DOPF, respectively. DC-DOPF and AC-DOPF are modeled as multiple base-cases OPF to get the runtime status in real time. Finally, several experiments are implemented based on multiple interconnected IEEE RTS-96 regions and IEEE 118 test regions. The computational results illustrate that the proposed decomposition and computation methods for DOPF based on MPI are applicable and effective, and it can be as a useful computation method for interconnected power system.

Research highlights
► Equivalent decomposition model to be used to solve DC-DOPF and AC-DOPF.
► Computation method for the distributed optimal power flow based on MPI.
► Auxiliary problem principle (APP) used to solve interconnected DOPF.