Computer-Aided Design Of Hydrogen-Resistant Superalloys – CADOHRS
Hydrogen embrittlement of nickel alloys can have severe consequences, e.g. in the nuclear or in the oil & gas sectors. The applicative aim of the project is to propose, through a computational materials design approach, new nickel alloys and superalloys with an improved resistance to hydrogen embrittlement (RHE), good mechanical performance and corrosion resistance, a low cost…
Alloys with unsurpassed combinations of –depending on cases– mechanical properties, oxidation resistance, processability and cost, have already been designed by associating data mining tools (machine learning), computational thermodynamics (CALculation of PHAse Diagrams – CALPHAD method), physical models and genetic algorithm multi-objective optimisation. However, no comprehensive criterion exists in the scientific literature to link the RHE to alloy composition through structure and/or microstructure. Nevertheless, literature suggests that MC carbides and g’ with a high g/g’ lattice misfit may trap hydrogen and increase the RHE, although there is no clear evidence. Besides, the electronic configuration, e.g. through the density of states at Fermi level, may also influence the RHE, although the existing model is limited in terms of accuracy and composition range applicability.
The aims of the project are therefore: (i) to investigate the above-mentioned potential routes to increase the RHE of nickel alloys, and (ii) to exploit the gained knowledge as new models or criteria for the conception of high-performance alloys. Our strategy will first involve the design, by computational thermodynamics, data mining and evolutionary algorithms, of model alloys displaying specific microstructures, i.e. containing either only MC carbides or g’ with targeted g/g’ lattice misfits. Other model alloys will also be designed on the basis of ab initio simulation using first principles, so as to obtain desired electronic configurations and structural interactions with hydrogen. Model alloys will then be fabricated and hot worked at laboratory scale, their microstructure will subsequently be characterised down to the nanometer scale (precipitates) and their behaviour in the presence of hydrogen will be investigated (permeation, thermal desorption, trapping, embrittlement…). This will bring new knowledge on the relations between structure, microstructure and the RHE of nickel alloys, which will be expressed as new criteria to enhance materials properties. These new criteria will be added to those already existing to predict microstructure (computational thermodynamics), mechanical properties (data mining), corrosion resistance and cost, to propose new alloys with improved performance (including a high RHE) by multi-objective optimisation. With this expected increase in the number of objectives to deal with simultaneously, we will also need to adapt or to change our optimisation algorithms. The resulting computationally designed alloys will finally be produced and wrought at laboratory scale and characterised in terms of microstructure, mechanical properties and resistance to hydrogen embrittlement.
To reach our goals, the consortium associates (1) a materials science laboratory with experts in ab initio simulation of cohesion and embrittlement in nickel alloys, in microstructural characterisation of metals as well as in computational alloy design using thermodynamics, physics, data mining and optimisation algorithms, (2) a computer science laboratory expert in the development of the latter methods, and (3) a laboratory specialised in the hydrogen embrittlement of alloys also possessing facilities for the fabrication of wrought metallic materials with a controlled composition.
Monsieur Franck TANCRET (INSTITUT DES MATERIAUX JEAN ROUXEL)
The author of this summary is the project coordinator, who is responsible for the content of this summary. The ANR declines any responsibility as for its contents.
ARMINES (SMS) ARMINES
LS2N Laboratoire des Sciences du Numérique de Nantes
IMN INSTITUT DES MATERIAUX JEAN ROUXEL
Help of the ANR 422,203 euros
Beginning and duration of the scientific project: November 2019 - 42 Months