A Full State-Variable Direct Predictive Control for Islanded Microgrids with Parallel Converters

In this work, we propose a high-quality control solution for islanded microgrids with multi-parallel power converters; it uses a full state-variable direct model predictive control (FSV-DMPC) and has a simple structure. Unlike the conventional cascaded control loops, the proposed FSV-DMPC solution tracks the optimal reference generated by a robust droop loop using a unified cost function. This proposal enables the FSV-DMPC to be inserted into the entire control framework with plug-and-play capability; it is robust to parameter variations, while also guaranteeing dynamics and stability. We conduct a deep analysis of the proposed approach, taking into account both the characteristics of the solution and the bounded stability of the system. Through comprehensive comparative studies with a classical double-loop linear controller, we validate that our solution achieves superior output voltage regulation during the load transients in terms of voltage error and settling time. Meanwhile, similar steady-state performances are accomplished for both methods. Finally, we verify our approach experimentally in different scenarios through a lab-constructed microgrid test-bench. Experimental data confirm that the proposed approach achieves excellent steady-state and transient performances, and obtains accurate load sharing.