Document Type : Research Articles
Authors
1 Department of Electrical Engineering, Faculty of Engineering and Technology, University of Mazandaran, Babolsar, Iran
2 Babolsar, IRAN
3 AAU Energy, Aalborg University, DK-9220 Aalborg, Denmark
Abstract
Effective energy management in renewable-based microgrids demands advanced optimization and forecasting strategies capable of addressing the intrinsic uncertainties of solar generation and load demand. This paper proposes a comprehensive day-ahead scheduling framework based on Model Predictive Control (MPC) to enhance the operational efficiency of a grid-connected microgrid comprising photovoltaic units, microturbines, fuel cells, and battery energy storage. The MPC structure incorporates 24-hour-ahead forecasts of load and solar irradiance generated through multiple deep learning architectures, enabling dynamic adaptation to rapidly changing environmental and consumption conditions. To solve the underlying optimization problem, a novel Quadratic Interpolation Optimization (QIO) algorithm is employed, offering improved convergence behavior and robustness in comparison with conventional metaheuristic methods. The integrated forecasting–optimization methodology ensures optimal dispatch of distributed resources, effective utilization of storage systems, and economically efficient power exchange with the utility grid. Extensive simulations validate the superiority of the proposed MPC-QIO framework in reducing operational costs, enhancing renewable energy penetration, and maintaining system reliability even under forecasting uncertainties. The results confirm that the synergy between deep learning-based prediction and advanced optimization significantly strengthens the controllability and resilience of modern microgrids, highlighting the method’s strong potential for real-world deployment.
Keywords
- Energy Management
- Energy Storage Systems
- Microgrid
- Model Predictive Control
- Quadratic Interpolation Optimization
Main Subjects
)