Distributed multi-layer control of hybrid AC/DC grids: Design and experimental validation

Abstract

As the electrical system becomes more and more decentralized, new control algorithms are necessary to manage the intermittent and non-deterministic production of non-programmable renewable sources, as well as the consumption of new loads like electric vehicles and heat pumps. Traditionally, electrical networks are controlled centrally, which provides full controllability of the system but introduces issues on scalability and complexity. This paper proposes a distributed multilayer control scheme based on model predictive control (MPC) applied to different portions of an electrical grid, optimizing power exchanges for balancing services. The first layer comprises local decentralized MPC controllers managing their areas, while the high-level distributed supervisor layer coordinates the exchange of flexibility between the network areas by acting on AC/DC converters. The overall distributed control architecture is applied and experimentally validated through the distributed energy resources test facility of RSE, showing enhanced performances in terms of prompt control action and compensation of the power disturbances.