Efficient energy generation and consumption is a key factor to achieve ambitious goals related to air pollution and climate change. Modern electricity networks can include different kind of sources, such as renewable energy sources (RES). Then, hybrid systems are obtained by combining several sources and storage types in the new concept called microgrid (MG). In order to draw the best performance from these hybrid systems, a proper design and operation is essential. The purpose of this paper is to present a detailed report to properly undertake the building and management of a hydrogen MG in a simple and reliable way to continue struggling for more comprehension on the MG operation modes and prevent the reported failures in the literature. The experimental platform developed will provide the valuable knowledge and solid guidelines for future test centers and demonstration plants. The MG, located in Seville, Spain, incorporates an electrolyzer, metal hydride storage, fuel cell, and a battery bank as main components. The developed MG laboratory has been successfully tested. The results indicate reliable operation incorporating the hydrogen and batteries as energy storage.

In this paper, a Model Predictive Control (MPC) approach is proposed to manage a grid-tied hydrogen microgrid $(\mu\mbox{G})$ . The $\mu\mbox{G}$ testbed is equipped with a 1-kW polymer electrolyte membrane (PEM) electrolyzer and a 1.5-kW PEM fuel cell as main equipment. In particular, we present a formulation that includes the cost of the electricity exported/imported, the aging of the components, and the operational constraints. The control objective is to satisfy user demand, as well as extend the lifespan of expensive equipment, as is the case of the fuel cell or the electrolyzer. $\mu\mbox{G}$ performance is investigated under realistic scenarios in three experiments. The experimental results illustrate how the proposed control system is able to manage the fuel cell and the electrolyzer through smooth power references, as well as to satisfy the power demanded. Finally, benchmarking is carried out between hysteresis band (HB) control and the proposed MPC in regard to efficiency and cost of the operation. The results obtained show that the MPC approach is more effective than HB for this type of $\mu\mbox{G}$ , with a reduction in operation cost of up to 30%.