A Fully Distributed Consensus ADMM Approach for Loss Minimization in AC- DC Radial Distribution Networks

Abstract

In the smart grid era, AC-DC distribution networks are evolving for integrating massive amounts of distributed energy resources (DERs), especially solar power generation units, into distribution networks in the most energy-efficient way. How-ever, the integration of DERs inherently enhances the number of controllable grid edge resources, and that creates computation challenges in front of conventional centralized optimization approaches. In this regard, this article proposes a fully distributed active power loss minimization approach for AC- DC distribution networks based on the consensus Alternating Direction Method of Multipliers (ADMM). The original centralized problem is decomposed into three different sub-problems corresponding to the AC network, DC network, and voltage source converter. By solving the sub-problems in parallel and sharing the global information among different agents, optimum conditions are obtained through consensus. The sub-problems are designed as second -order cone relaxed convex programming to avoid the chance of sticking at the local optimum point. Demonstrating the simulation cases on 33-bus and 132-bus AC-DC radial test systems, it is established that the proposed strategy provides optimal decision faster compared to the centralized method. © 2024 IEEE.