Menou, E., Rame, J., Desgranges, C., Ramstein, G. & Tancret, F. (2019) Computational design of a single crystal nickel-based superalloy with improved specific creep endurance at high temperature. Comput. Mater. Sci. 170 UNSP 109194.
Added by: Richard Baschera (2019-12-17 10:13:28) Last edited by: Richard Baschera (2019-12-17 10:14:21) |
Type de référence: Article DOI: 10.1016/j.commatsci.2019.109194 Numéro d'identification (ISBN etc.): 0927-0256 Clé BibTeX: Menou2019 Voir tous les détails bibliographiques |
Catégories: ID2M Mots-clés: Artificial intelligence, calphad, chemical-composition, Machine learning, neural-network, Optimisation, optimization, power-plant applications, stress, Thermo-Calc, uhs stainless-steels Créateurs: Desgranges, Menou, Rame, Ramstein, Tancret Collection: Comput. Mater. Sci. |
Consultations : 1/360
Indice de consultation : 5% Indice de popularité : 1.25% |
Résumé |
A Gaussian process regression model for estimating the creep rupture stress of single crystal nickel-based superalloys is constructed. It is built and validated on data disclosed in patents as well as scientific and technical reports. This model is coupled with computational thermodynamics for the prediction of microstructural features, and a model for the estimation of density. Using this combination, the key characteristics of a large number of potential alloys are systematically computed as a function of their composition. Materials specifications targeted towards applications as turbine blades, which none of the alloys from the creep database follow, narrow down the search from 300 000 000 to 180 000 alloys. A large number of these candidates are predicted as featuring greater specific properties than existing single crystal nickel-based superalloys. The selection criteria used to isolate alloys for experimental validation are discussed, and lead to several alloys displaying good agreement between their predicted microstructure and estimated properties.
|