In this section
@inproceedings{Schaefer:2026:10.2514/6.2026-1714,
author = {Schaefer, JW and Di, Fiore F and Wu, B and Mainini, L},
doi = {10.2514/6.2026-1714},
title = {Physics-Aware Multi-Source Modelling for the Design Optimisation of Fuel Cells for Sustainable Aviation},
url = {http://dx.doi.org/10.2514/6.2026-1714},
year = {2026}
}
TY - CPAPER
AB - Low temperature proton exchange membrane fuel cells represent a promising pathway toward zero emission aviation, yet their design at the scales and power densities required for flight remains constrained by limited experimental data, limited available numerical models, high computational cost of numerical models, and numerical fragility of detailed multi-physics models. In this work, a physics-aware multi source surrogate modelling framework is proposed to support the design optimisation of aviation relevant fuel cell systems under severe data scarcity and variable model reliability. The approach combines low-fidelity and high-fidelity numerical models within an autoregressive multi-fidelity formulation in which the relative influence of each information source is evaluated and adjusted dynamically across the design space. The method is applied to the optimisation of a low-temperature proton exchange membrane fuel cell with parallel flow bipolar plates, in which a one dimensional through plane low-fidelity model and a three dimensional multiphase high-fidelity model are combined to target maximum specific power through joint optimisation of operating conditions and channel length. Results demonstrate that the proposed modelling approach enables stable and physically meaningful design exploration in regimes where classical single fidelity and fixed hierarchy multi-fidelity approaches are misled by unreliable data. When embedded within a global optimisation framework, our physics-aware multi-source model identifies feasible high performance designs validated by high-fidelity simulation, highlighting its potential as a robust tool for early stage design of emerging aerospace energy systems.
AU - Schaefer,JW
AU - Di,Fiore F
AU - Wu,B
AU - Mainini,L
DO - 10.2514/6.2026-1714
PY - 2026///
TI - Physics-Aware Multi-Source Modelling for the Design Optimisation of Fuel Cells for Sustainable Aviation
UR - http://dx.doi.org/10.2514/6.2026-1714
ER -