Abstract
DC microgrids are increasingly deployed to integrate renewable energy sources and storage systems with high efficiency and flexibility. Advanced control strategies are vital to ensure stable operation and optimal energy management. This paper presents an improved methodology for enhancing the performance of Hybrid Energy Storage Systems (HESS) in DC microgrids using a Deadbeat Model Predictive Control (DBMPC) scheme. The proposed approach combines the fast dynamic response of Deadbeat Control with the optimal prediction capabilities of Model Predictive Control (MPC), achieving faster settling times and reduced computational burden. The methodology is validated through extensive simulations and hardware experiments, comparing the DBMPC with conventional PI and standard MPC controllers. The results demonstrate that the proposed method significantly reduces dynamic errors, shortens settling time, and ensures accurate reference tracking without steady-state error. Hardware validation is conducted using a DC microgrid prototype, confirming superior system stability, improved response time, and effective energy management under various operating conditions. Additionally, the implementation requires only two current sensors and one voltage sensor, significantly lowering hardware costs compared to conventional methods. Owing to its low computational complexity, the proposed DBMPC is well-suited for real-time control of HESS in islanded microgrids, offering a cost-effective and robust solution for advancing renewable energy integration.