Distributionally robust decarbonizing scheduling considering data-driven ambiguity sets for multi-temporal multi-energy microgrid operation

As concerns about environmental sustainability continue to grow, the demand for effective low-carbon energy management becomes increasingly pressing. This study presents a novel framework for multi-temporal multi-energy microgrids (MMGs), integrating advanced low-carbon technologies to meet this imperative. The framework ensures flexible operations to navigate uncertainties stemming from renewable energy sources (RES) and fluctuating energy demand. Facilitating multi-energy transactions, encompassing gas and power exchanges in both markets, the model accommodates uncertainties from RES and demand fluctuations. Objectives include reducing carbon emissions and improving economic efficiency. To address uncertainties in the MMG system, a data-driven distributionally robust optimization (DRO) method is employed. Day-ahead scheduling utilizes a two-stage three-level approach, deploying the column-and-constraints generation (C&CG) algorithm, showcasing the efficiency of DRO in minimizing energy waste and carbon emissions while remaining cost-effective. Practicality is demonstrated through real-time intra-day scheduling using the model predictive control (MPC) algorithm, building upon hourly day-ahead results. The effectiveness of both strategies is evaluated using empirical data from an MMG based on the IEEE 33-bus test system. This cost-saving framework not only achieves a significant carbon reduction of 10.6 % but also provides reliable and adaptable solutions, effectively addressing real-world variations in renewable energy and mitigating potential risks.