A Priority Scheduling Strategy of a Microgrid Using a Deep Reinforcement Learning Method

The fluctuation of renewable energy leads to a severe power fluctuation on the tie-line between microgrid (MG) and the power grid. A novel scheduling strategy is proposed in this paper to reduce power fluctuation based on a data-driven approach. The scheduling strategy is divided into two steps based on the scheduling priority of gas turbines, battery storage systems, and power-to-hydrogen devices. The aim is to minimize the penalties caused by the tie-line power fluctuation and operation costs of MG. And the scheduling optimization models are expressed as a Markov decision process. Simulation results verify the effectiveness of the proposed strategy.