A Multidisciplinary Design Framework for Hydrogen Fuel Cell-Powered Aircraft

Hydrogen-powered aircraft represent a promising pathway toward net-zero aviation by eliminating carbon emissions at the source, but their integration introduces fundamental design challenges. Low-TRL enabling technologies such as cryogenic storage, fuel distribution, and thermal management require high-fidelity physics to ensure feasibility, while strong couplings between aerodynamic, propulsion, and thermodynamic disciplines make performance and safety dynamic mission-dependent properties. In addition, the wide architectural freedom enabled by hydrogen propulsion creates a fragmented and largely infeasible design space that conventional Multidisciplinary Design Optimization (MDO) approaches cannot explore efficiently. To address these challenges, this work proposes a unified approach that combines an aircraft design framework for the optimization of conventional and non-conventional power architectures with a Non-Myopic Multipoint Multistart Multifidelity Bayesian Optimization (NM4-BO) algorithm. The framework supports multifidelity modeling to efficiently employ high-fidelity analyses to characterize low-maturity technologies and captures complex, time-dependent interdisciplinary couplings through a dynamic mission analysis. NM4-BO allocates computational effort adaptively through non-myopic learning, multifidelity sampling, and trust-region feasibility control. Together, these elements maintain dynamic multidisciplinary coupling without prohibitive computational cost, capture sharp feasibility boundaries arising from nonlinear subsystem interactions, and enable efficient navigation of sparse high-dimensional design spaces. Application to a fuel cell–powered regional aircraft demonstrates the framework’s ability to identify feasible, energy-efficient architectures while maintaining full physical consistency and computational scalability.