Clustering-based Decentralized Optimization Approaches for DC Optimal Power Flow

This paper studies two decentralized scheme to solve DC optimal power flow (DC-OPF). The first scheme considers the decomposition of DC-OPF based upon augmented Lagrangian relaxation and uses alternating direction method of multipliers (ADMM) algorithm to solve the consensus optimization problem. An adaptive penalty method is proposed for the ADMM algorithm to improve the convergence performance. The second scheme utilizes Karush-Kuhn-Tucker (KKT) conditions and solves the coupled linear equations of DC-OPF directly. We show the impact of different cluster formations on both schemes. Both schemes are evaluated in terms of flexibility, robustness and iteration time using the IEEE 14-bus test system.