Description du projet
Apprentissage automatique et nanoparticules pour des avions compétitifs
Une concurrence croissante et des réglementations environnementales strictes poussent l’industrie aérospatiale à explorer des méthodes permettant de réduire le poids, la consommation de carburant et l’impact des vols sur l’environnement. Les panneaux composites rigidifiés dans lesquels des nanoparticules sont introduites dans des régions critiques exposées aux dommages offrent beaucoup de potentiel par rapport à l’utilisation de fixations mécaniques. Cependant, il n’existe pas encore de panneaux rigidifiés certifiés contenant des nanoparticules. En outre, les mécanismes des dommages sont complexes et les méthodes de prévision de la propagation des dommages et d’estimation du cycle de vie ne sont pas encore totalement établies. Le projet MACADAMIA, financé par l’UE, envisage d’aborder le problème en combinant des nanoparticules et des modèles basés sur la physique pour détecter l’évolution de la résistance et des dommages des panneaux rigidifiés tout en utilisant l’apprentissage automatique pour perfectionner leur estimation du cycle de vie.
Objectif
MACADAMIA ambitiously seeks a seamless integration of machine learning concepts with physics-based models to optimise aerospace stiffened panels for damage tolerance. As an innovative strategy to delay damage, nanoparticles will be added in failure-prone hot-spots of composite stiffened panels to serve as damage arrest features. The efficacy of machine learning when used in conjunction with advanced computational methods for data classification and prediction will be smartly leveraged to classify and predict damage mechanisms in aircraft structures, the understanding of which is critical to their safe implementation.
In aircraft, stiffened composite panels are popular alternatives to structures with mechanical fasteners because they retain strength while reducing weight and part count; but cost and weight savings cannot be fully realized until stiffened panels are certified without fasteners in primary load-bearing structures. It is estimated that a one-pound weight reduction on each aircraft in a commercial fleet would result in fuel savings of 14000 gallons/year, which also mitigates the environmental impact of flight. To strengthen the competitiveness of European aerospace technologies in compliance with evolving environmental regulations, it is vital to work towards accelerated certification of fastener-free composite panels. Major challenges to this goal are: i) damage mechanisms in stiffened panels are complex and coupled, making the evaluation of strength and durability difficult; ii) predictive models for life-cycle estimation have large uncertainty. MACADAMIA envisions an approach with carefully designed experiments for nanoparticle inclusion along with physics-based models to investigate strength and damage evolution in stiffened panels, and machine learning to further optimise them for longer useful life. Multidisciplinary concepts of structural mechanics, computational physics, nanotechnology and machine learning will be used to accomplish research plan.
Champ scientifique
- engineering and technologymaterials engineeringcomposites
- engineering and technologymechanical engineeringvehicle engineeringaerospace engineeringaircraft
- engineering and technologynanotechnologynano-materials
- engineering and technologyenvironmental engineeringenergy and fuels
- natural sciencescomputer and information sciencesartificial intelligencemachine learning
Programme(s)
Régime de financement
MSCA-IF - Marie Skłodowska-Curie Individual Fellowships (IF)Coordinateur
2628 CN Delft
Pays-Bas