Energy-Power Distributionally Robust Optimization for Multi-energy Microgrid Considering Long Short-term Hybrid Energy Storage

For the autonomous operation demand of multi-energy microgrid, a rolling optimization method of energy-power matching for pre-week/day-ahead/real-time multi-energy microgrid is proposed based on long short-term hybrid energy storage (hydrogen storage, electric storage and thermal storage). On one hand, the rolling optimization of the long-term (week-ahead) energy balance model and the short-term (day-ahead and real-time) power balance model addresses the problem of large errors in the long-time-scale prediction of renewable energy and load. On the other hand, a data-driven two-stage distributionally robust optimization model is used to portray the intra-day source-load bilateral uncertainty. The 1-norm and ∞-norm are comprehensively used to constrain the uncertainty probability distribution confidence sets for ensuring the robustness of the multi-energy microgrid operation while avoiding the conservatism of the operation scheme. At the same time, the second-stage problem of the day-ahead model can be decomposed into multiple small-scale subproblems and can be processed in parallel by the column-and-constraint generation algorithm without complex pairwise computation. Finally, the arithmetic case analysis results verify the effectiveness of the proposed model and algorithm. © 2023 Automation of Electric Power Systems Press. All rights reserved.