Work

A reinforcement learning–based two-level low-carbon economic dispatch of integrated energy systems with liquid CO2 energy storage

Year: 2026
Type: article
Abstract: Integrated energy systems (IES) face increasing challenges in achieving low-carbon and economically efficient operation because of the strong inter-temporal coupling of long-duration energy storage and the limited coordination of carbon and oxygen utilization. However, the coordinated integration of long-duration liquid CO₂ energy storage (LCES) scheduling with carbon–oxygen recycling has received limited attention. To address this gap, this study develops a bi-level low-carbon economic dispatch framework. First, a coupled LCES–OCC–power-to-gas (P2G)–methanol synthesis system is established, in which the oxygen generated by P2G is supplied to oxygen-enriched combustion capture (OCC) and captured CO₂ is coordinated between LCES and methanol synthesis, forming a closed carbon–oxygen utilization loop. Second, a PPO-based bi-level scheduling architecture is developed for long-horizon LCES operation, with proximal policy optimization (PPO) determining the charge/discharge strategy and a lower-level mixed-integer linear programming (MILP) model performing constraint-based IES dispatch. Case studies show that the fully integrated configuration reduces total operating cost by 22.47% and actual carbon emissions by 15.94% compared with a battery-only configuration, while eliminating wind and solar curtailment. Compared with the configuration without LCES, the total operating cost is further reduced by 4.12%. Under the same integrated system, PPO reduces operating cost by 7.7% relative to the no-RL MILP strategy and by 1.29% relative to deep deterministic policy gradient (DDPG), while reducing actual carbon emissions by 4.32% compared with DDPG. Monte Carlo analysis achieves 97.7% feasibility with a 3.3% coefficient of variation in actual carbon emissions. These results demonstrate the potential of coordinated LCES and carbon–oxygen utilization for simultaneously improving the economic, renewable-accommodation, and low-carbon performance of IES.
Language: English
Open Access status: gold
Estimated APC paid: $2,690
OpenAlex ID: W7225295311
FWCI:
Cites: 37
Cited by:
Related to: 10