Predictor-Corrector Accelerated ADMM Method for the SOCP-Relaxed AC-DC Optimal Power Flow

Abstract

AC-DC distribution networks are becoming pertinent to allow large penetration of distributed energy resources (DERs) with lesser power conversion loss. The conventional centralized optimal power flow (OPF) strategies pose scalability and computation difficulties for AC-DC networks with multiple controllable assets. In this regard, distributed methods especially the Alternating Direction Method of Multipliers (ADMM) algorithm have proved themselves superior by decomposing the centralized problem into multiple sub-problems and solving them in parallel by enabling neighborhood coordination. However, vanilla ADMM often requires a substantial number of macro-iterations (˜102) for convergence. To mitigate this, this article utilizes the concept of predictor-corrector acceleration in the vanilla ADMM framework for solving the second-order cone programming (SOCP) relaxed AC-DC OPF problem in a lesser number of iterations and the solution time. Further improvement in the solution time is obtained by merging the concept of the adaptive penalty with the accelerated ADMM framework. Simulation results on 33 and 132-bus AC-DC radial test systems demonstrate the proposed strategy's effectiveness in providing faster optimal decisions compared to the existing studies. © 2024 IEEE.