(30% seminar) Safe Multi-Agent Deep Reinforcement Learning based Energy Management for Multi-Microgrid Energy Communities

Multi-microgrid (MMG) systems enhance efficient integration of the renewables and emerging loads into moder sustainable communities. By facilitating energy sharing among MGs, MMG system can enhance reliability, improve efficiency, and promote economic operation. In this presentation, a projection based safe multi-agent deep reinforcement learning (DRL) algorithm is developed for the decentralized energy management of multi-microgrid energy communities. The safe projection layer integrated into DRL training effectively restricts the agent’s actions to the feasible space, thereby preventing any violations of physical constraints during system operation. At last, the proposed method is validated in modified IEEE 33 bus and 118 bus system to demonstrate the effectiveness of the proposed method.