New Microgrid Structure, Based on Z-Source Converter Series Coupling and Managed by a Markov Decision Process

Renewable energies are environmentally friendly and, when integrated in DC microgrids, they can solve the energy problems of rural areas. As their main problems are: the randomness they present, the coupling between the different sources and between them and the storage systems, as well as the management of the whole set. Therefore, a new structure has been proposed, based on the coupling of sources and storage systems by means of impedance source (Z-source) DC converters connected in series and managed by an intelligent Markov decision-making process based on reinforcement learning. In order to validate the system, the results of a 5000 W laboratory prototype have been provided, consisting of the coupling in a branch of three different renewable sources, two storage systems and a controller, all managed by an intelligent system based on reinforcement learning. A management efficiency of between 88% and 97% was achieved. The proposed structure allows the coupling and optimal management of renewable sources so that they work in isolation and, if necessary, interconnected with other branches of a DC microgrid. Its main application is in remote and residential areas.