Distributed Model Predictive Frequency Control in Inverter-Based Microgrids Based on ADMM with Virtual Subsystems

Proceedings of the IEEE Conference on Decision & Control, including the Symposium on Adaptive Processes, pp. 5221–5227

Abstract

This paper presents a distributed Model Predictive Controller (dMPC) for frequency regulation in inverter-based microgrids, leveraging the Alternating Direction Method of Multipliers (ADMM) for decentralized optimization. The proposed approach addresses the scalability and communication challenges of centralized MPC by decomposing the global optimization problem into localized subproblems. The algorithm is validated using a microgrid scenario derived from the CIGRE medium-voltage benchmark network, adapted for islanded operation. The network configuration emphasizes local subsystem interactions, reflecting realistic operating conditions while focusing on the optimizer’s performance. Results demonstrate that the dMPC achieves near-identical performance to its centralized counterpart when sufficient iterations ensure convergence. Under limited computational resources, the controller reliably tracks optimal operating points when constraints are inactive but exhibits violations when constraints become active. Strategies such as improved initialization schemes and dynamically adjusted optimization parameters show promise in mitigating these limitations. The findings underscore effectiveness and robustness of ADMM in decentralizing complex control problems while ensuring coordination across subsystems.

Authors 2

1. Alexander Och Aachen Institute for High Voltage Equipment and Grids

   RWTH Aachen University

   Affiliation as printed

   Institute for High Voltage Equipment and Grids, Digitalization and Energy Economics, RWTH,Aachen,Germany

2. Andreas Ulbig Aachen Institute for High Voltage Equipment and Grids

   RWTH Aachen University

   Affiliation as printed

   Institute for High Voltage Equipment and Grids, Digitalization and Energy Economics, RWTH,Aachen,Germany