WP1 focused on upgrading high-fidelity physical models and adjoint solvers to enable advanced multidisciplinary design optimization (MDO). Developments included improved modelling of wing–engine–exhaust interaction and ice microphysics for realistic contrail prediction, as well as multi-fidelity approaches for microchannel flows in optimized heat exchangers. Advanced adjoint methods addressed laminar–turbulent transition, aero-propulsive coupling, broadband noise, and unsteady simulations at realistic time scales using compression techniques. The Fourier Time-Spectral Method and its adjoint were developed for efficient nonlinear time-periodic simulations. The open-source MDO platform GEMSEO was enhanced with new functionalities, including uncertainty management and distributed workflow capabilities. WP2 assessed the applicability and limits of machine learning in aerospace design. Work included integrating ML-based turbulence closures into CFD solvers, replacing selected high-fidelity models with neural-network surrogates, and enabling geometry reconstruction from inspection data such as images and point clouds. Additional efforts improved manufacturing tolerances and accelerated digitalization workflows to support digital twin applications. The methods were validated on real aerospace components, including engine parts, with real-time prediction and AI-based computer vision tools. WP3 demonstrated aircraft-level multidisciplinary optimization on two high-aspect-ratio wing configurations defined by Airbus and Dassault Aviation: the DLR-F25 and a natural-laminar-flow business jet. Complete MDO chains were established and optimization studies confirmed the feasibility and benefits of these configurations. Transition models combined with adjoint-based gradients enabled extended laminar flow on the business jet wing. For the DLR-F25, aero-structural optimizations using multiple MDO formulations captured trade-offs between aerodynamic drag and structural mass, highlighting the decision-support value of MDO. WP4 investigated advanced propulsion concepts targeting significant fuel-burn reduction, namely an Ultra-High-Bypass-Ratio turbofan and an Unducted Single Fan configuration. Multidisciplinary optimizations were carried out on the UHBR fan and high-pressure turbine, combining aerodynamic, structural, and thermal analyses to demonstrate the applicability of advanced MDO frameworks to next-generation engine architectures. WP5 addressed uncertainty quantification and robust optimization to ensure reliable entry into service of new aircraft and engine components. Robust design studies covered a high-aspect-ratio wing under twist and camber uncertainties, a laminar business jet wing accounting for surface roughness variability, and engine components including an Unducted Single Fan blade, an Ultra-High-Bypass-Ratio fan, and a high-pressure turbine blade. With around ten uncertain parameters per case, the resulting designs showed reduced sensitivity and improved operational stability. WP6 developed digital twin methodologies for future engine components by combining advanced modelling and machine-learning approaches. Manufacturing variations of fan blades were analysed using industrial data, and an inverse mapping procedure morphed CAD geometries to match high-precision measurements, enabling large-scale CFD simulations and comparison with engine test data. In parallel, topology optimization was applied to design an innovative heat exchanger with improved heat transfer, reduced pressure losses, and lower mass. Optimized designs were manufactured via metallic additive manufacturing and experimentally tested, showing good agreement between simulations and measurements and confirming the reliability of the digital twin framework.
