Low-carbon Energy Sharing for Multi-energy Microgrid using Cooperative Reinforcement Learning

It is a significant challenge to reconcile the competing interests of individual microgrid units when energy is shared in a multi-energy system. Moreover, the majority of these systems solely focus on power sharing, without adequately addressing the associated carbon emissions. This paper puts forward a sharing approach that makes use of multi-agent collaboration to optimise energy sharing in multi-energy systems and achieve low-carbon operation. Firstly, a system model is constructed to describe the interrelationship between each microgrid and the energy-sharing platform. Subsequently, a shared pricing mechanism will be devised to incentivise each microgrid to prioritise its participation in the local sharing market. The shared pricing mechanism is employed to construct the microgrid utility maximisation problem and transform it into a Markov game process. A multi-agent cooperative approach is put forth as a means of optimising the energy-sharing strategy. Ultimately, the simulation results demonstrate that this energy-sharing strategy can satisfy the operational constraints, maintain equilibrium between supply and demand, reduce energy costs, and promote the economic and low-carbon operation of the microgrid.