Simulation and control strategy of hydrogen refueling system for hydrogen powered unmanned aerial vehicles based on digital twin and intelligent optimization

Hydrogen powered unmanned aerial vehicles (UAVs), as one of the most promising models of new energy powered UAVs, are closely related to the concept of green aviation and low altitude economic scenarios. Fuel cells (FC), as an efficient and pollution-free new energy source, have attracted widespread attention from experts and scholars around the world. They are highly regarded for their cleanliness, low noise, long operating cycle, and high energy conversion efficiency. This article designs a simulation and control strategy for a hydrogen refueling system for a hydrogen powered UAV based on digital twin (DT) and deep learning (DL). Firstly, research the construction of a simulation model for the hydrogen production and energy replenishment system. At the same time, a dual delay depth deterministic policy gradient (TD3-PER) energy management strategy based on priority experience sampling was studied and designed. The strategy adopts the Double Delay Depth Deterministic Policy Gradient (TD3) algorithm to achieve more accurate continuous control. By combining the Priority Experience Sampling (PER) algorithm, the training of the strategy is accelerated while achieving better optimization performance. Simulation experiments show that this strategy can effectively improve the flexibility of scheduling hydrogen refueling systems for hydrogen UAVs.