The advent of renewable energy technologies and power electronics in low and medium voltage networks has led to the evolution of the current electrical system, with the development of smart distribution grids. These changes are still ongoing and will lead to further developments in the coming years, such as the creation of direct current (DC) distribution networks that can connect to the alternating current (AC) network at multiple nodes, strengthening the distribution grid.

Nevertheless, a stronger network infrastructure alone is not sufficient to support the imminent decentralization of the electrical system, and new coordination and control algorithms must be developed. In fact, the intermittent and non-deterministic nature of renewable sources and some new electrical loads increases network power variability, causing continuous power imbalances between generation and demand, which can lead to significant frequency deviations.

In this context, dividing the distribution network into cells represents a potential solution to reduce grid disturbances, minimize contingencies, and promote the integration of renewable generation within the distribution network. A cell is a portion of the distribution network that has one or more connection points with the main grid and other cells and is equipped with its own controller capable of locally compensating for active power variations compared to expected values, while also maintaining a reserve that can be used to provide additional services to the network or adjacent cells.

The work carried out focused on the design and validation through numerical simulations of a control structure for the coordination and management of cells into which a hybrid AC/DC distribution network is divided. The designed control structure is based on a supervised Model Predictive Control (MPC) that efficiently and quickly compensates for power imbalances caused by loads and renewable sources. This control is applied to each cell, balancing local power variability by utilizing flexibility reserves, represented by local microgrids. Each cell’s MPC also communicates with a central supervisor, which coordinates power flows in the direct current network in case of power shortages in the cells. The aim is to support the cells requiring power by directing it from those with availability, acting on the controllable interfaces of the direct current network.

The numerical results demonstrate the effectiveness of the approach, ensuring rapid action and excellent scalability properties. The designed control structure is highly flexible, allowing it to be integrated with various hybrid AC/DC network configurations. Moreover, the described control structure is perfectly adaptable to different types of generators and batteries and can be easily adjusted to various scenarios and model assumptions.