Hydrogen is regarded as one of the most promising energy in the future, and the intra-day scheduling of HFSs has received more and more attention. However, the uncertainty of the hydrogen loads keeps increasing with the development of hydrogen-based vehicles. To deal with the uncertainty, a Proximal-Policy-Optimization-based (PPO-based) intra-day scheduling method of hydrogen fueling station (HFS) is presented. The PPO-based method can well manage the continues operation of the HFS. Moreover, the actions are reduced to make full use of the computing resources and the advantages of both reinforcement learning and optimization model can be combined. The results show that the PPO-based method can well manage the energy in the random environment and has the potential to face the strong uncertainty.

energy storage; reinforecement learning; intra-day scheduling; hydrogen fueling station; hydrogen