Optimized hybrid renewable–hydrogen microgrid with machine learning forecasting and demand response for sustainable campus electrification

The increasing need for reliable, low-carbon electricity in institutional facilities demands advanced hybrid microgrid architectures that integrate long-term storage and intelligent operational strategies. This study proposes and optimizes a renewable–hydrogen hybrid microgrid for large-scale campus electrification by integrating solar photovoltaic (PV), wind turbines (WT), battery energy storage system (BESS), electrolyzer-based hydrogen production, hydrogen storage, fuel cells (FC), and grid interaction within a unified techno-economic and operational framework. The innovation of this work lies in the coordinated integration of hydrogen storage, machine-learning-based renewable forecasting, multi-scenario optimization, dynamic voltage–frequency stability assessment, and demand-response analysis to enhance both economic feasibility and operational resilience. The forecasting model is applied to predict total renewable generation with high accuracy (R 2 = 1.00), enabling improved energy scheduling, hydrogen dispatch planning, and reduced grid dependency under resource variability. The optimized configuration achieves a cost of energy of $0.0352/kWh, a net present cost of $966,319, an 84.4% renewable fraction, and hydrogen production of 16,453 kg/year at a competitive levelized cost of $4.00/kg. Compared to conventional grid-dependent configurations, the proposed system significantly reduces CO 2 emissions and improves system reliability while achieving annual energy savings of 33,850.20 kWh through demand-response strategies. The results demonstrate that integrating hydrogen storage with intelligent forecasting and demand-side management provides a scalable and economically progressive pathway for sustainable campus microgrids.