Hybrid-electric powertrains offer a solution to significantly reduce aircraft emissions in flight. This study presents a method for automatically generating hybrid-electric architectures and optimizing two different objective functions by evaluating the control parameters of each unique architecture using reinforcement learning. Two ATR 72-600 configurations serve as reference aircraft, and three technology levels are considered. When maximizing the ratio of effective radiative forcing to payload mass, the results indicate that the optimal design is sensitive to both aircraft configuration and technology level; however, architectures fully powered by hydrogen fuel cells are preferred when feasible. When maximizing payload and applying the Flightpath 2050 sustainability goals as constraints, the optimal architecture shifts to one in which conventional jet fuel and hydrogen are combusted in a gas turbine to power the primary propulsive line, while the majority of the power is delivered by the fuel cells to an auxiliary propulsive line. Compared with a conventional architecture, this design reduces CO2 and NOx emissions by up to 74% and 86%, respectively, while reducing payload mass by only 24%.

https://github.com/mvandongeren/Conceptual-Design-of-Green-Propulsive-Systems-Using-Reinforcement-Learning.git