Combined Carbon Capture and Utilization With Peer-to-Peer Energy Trading for Multimicrogrids Using Multiagent Proximal Policy Optimization

Abstract

Microgrids integrated with distributed renewable energy are regarded as a crucial evolution toward economical and environmentally sustainable power systems. Carbon capture and utilization (CCU) technologies and peer-to-peer (P2P) energy trading schemes are two potential strategies for mitigating carbon emissions by capturing the emitted CO$_{2}$ or trading surplus renewable energy, respectively. Hence, a collaborative energy scheduling model that combines CCU with P2P energy trading is needed under the coupling of multiple energy domains, including electricity, CO$_{2}$, and natural gas. In this article, we investigate a novel multimicrogrid framework that jointly considers CCU and P2P trading, aimed at reducing costs and mitigating carbon emissions. Correspondingly, an energy-coupled decision-interdependent multimicrogrid energy scheduling problem is developed that involves the stochastic system states, such as intermittent renewable generation and unpredictable loads. We regard each microgrid as an agent and adopt a multiagent proximal policy optimization (MAPPO) algorithm for distributing the interdependent energy scheduling actions to each agent. This algorithm can cope with the high-dimensional continuous action space and find the energy coordination policy without requiring system future statistical information. In particular, we introduce the centralized training with decentralized execution (CTDE) mechanism, which alleviates the nonstationarity of the environment via centralized training and alleviates the curse of dimensionality via decentralized execution. Simulation results demonstrate that the proposed joint CCU-P2P energy coordination model and the CTDE-based MAPPO algorithm outperform other models in achieving economic and environmental benefits.