A unified deep-forecasting and IGDT framework for robust operation of Multi-Energy Microgrids with hydrogen systems and electric vehicle integration

This paper proposes a forecasting-driven uncertainty management framework for a hydrogen-enabled Multi-Energy Microgrid (MEMG) integrating a 14-bus IEEE power network, a District Heating Network (DHN), and a Gas Distribution System (GDS). The framework incorporates renewable energy resources, Combined Heat and Power (CHP) unit, Power-to-Hydrogen-and-Heat (P2HH) technology, and Plug-in Electric Vehicles (PEVs) to simultaneously satisfy electricity, heat, and gas demands while enhancing operational flexibility. Unlike conventional IGDT-based energy management approaches that rely on predefined nominal demand profiles, the proposed framework integrates CNN–LSTM demand forecasting with IGDT-based optimization by using the forecasted demand as the nominal operating profile for uncertainty-aware scheduling. A two-stage methodology is developed. First, a hybrid CNN–LSTM model is employed to forecast electricity, heat, and gas demands. Subsequently, Information Gap Decision Theory (IGDT) is used to manage forecasting-driven uncertainty through risk-averse and risk-seeking decision-making strategies. The forecasting model is implemented in Python, while the operational optimization problem is formulated as a Mixed-Integer Nonlinear Programming (MINLP) model and solved in GAMS. Simulation results demonstrate satisfactory forecasting performance, with sMAPE values below 15% for all energy carriers. The results further show that, compared with the risk-neutral strategy, the risk-averse strategy increases total cost, electricity transaction cost, and emission cost by 28%, 51%, and 46%, respectively, in exchange for enhanced robustness against uncertainty. Conversely, the risk-seeking strategy reduces the total cost and emission cost by 40% and 85%, respectively, while the electricity transaction balance shifts from a net cost to a net revenue due to increased electricity exports under favorable operating conditions. These findings demonstrate the effectiveness of the proposed forecasting-driven IGDT framework in balancing economic performance and operational robustness in hydrogen-enabled multi-energy microgrids under uncertainty.