Model predictive control for microgrid applications

Abstract

Nowadays, renewable energy and plug-in energy technology are developed at a rapid speed. Microgrid (MG) is a small power system that can employ renewable energy generation and distributed generation for supplying power to multiple loads. However, MG has to focus on managing the uncertainty due to renewable energy and multiple loads. This dissertation reviews the research and development history of multi-microgrids model predictive control, introduces the mathematical model of multi-microgrids and the concept of distributed energy network trading model, studies and modifies the multi-microgrids coordinated energy dispatch strategy with distributed model predictive control (MPC). A two-layer control system is employed in this dissertation where the upper layer achieves electricity exchange between the distributed network control centre and microgrids, and the lower layer guarantees the electricity power flow balance economically through the local controllers. Moreover, we modify and upgrade the conditional probability distribution model to calculate better control performance and apply the demand-side response technique in the existing literature. This improved strategy provides better performance in terms of reliability, stability, and economy of multi-microgrids power systems. Finally, this distributed MPC strategy is established, and the simulation is carried out for verification. Detailed analysis and comparison of the dispatch flexibility, the operating mode of the energy storage system, the time domain analysis of the energy exchange dispatching, the operating costs and the load shifting level of two different dispatch strategies. The case study shows that the improved two-layer distributed MPC strategy in this dissertation can decrease the load shifting level with similar operating costs compared with the original strategy. It can also provide a reliable and stable dispatching performance of the system.