Modelling of martensite decomposition in Ti6Al4V alloy during heat treatments after additive manufacturing

A. BOCCARDO; J. SEGURADO; Z. ZOU; M. SIMONELLI; M. TONG; S. LEEN; D. TOURRET

Congreso; 11th European Solid Mechanics Conference (ESMC 2022); 2022

University of Ireland - Galway

The additive manufacturing technique allows producing Ti6Al4V components of complex topology with the advantage of reducing/eliminating the number of part ensembles. The main application fields of such components are the biomedical, aeronautical, and aerospace, in which the mechanical performance as yield strength, elongation to failure, and fatigue life, take a critical role. The mechanical response is closely related to the resulting microstructure features as grain morphology, size, and crystal orientations, all of them determined during the printing processing. The selective laser melting technique, characterized by high cooling rates, allows for manufacturing high dimensional precision parts but a brittle behavior is obtained due to the presence of α’ martensite. In order to tailor the mechanical response, the martensite is decomposed into a more stable α+β structure by means of annealing heat treatments, after the printing processing, below/above the β transus temperature depending on the required mechanical performance. Parameters such as annealing cycle numbers, annealing temperature, and dwelling time allow modifying the α and β volume fractions and α grain morphology and size (ultra-fine, fine, and coarse grains). Numerical simulations allow accurately predicting the influence of processing parameters on the microstructure features and, by further coupling with mechanical models, then predicting the resulting mechanical response. Phase-field modeling, based on strong physical phenomena, predicts the evolution of alloying element concentration and phase morphology by means of Cahn-Hilliard and Allen-Cahn differential equations, respectively. This work presents a new phase-field model to predict the martensite decomposition and it is applied to a single annealing post-heat treatment. The coupled system of partial differential equations is numerically solved by means of the spectral Fourier method by assuming a periodic boundary condition and initial condition from experimental EBSD maps. The proposed semi-implicit algorithm is computationally implemented by means of Python and graphical processing units through PyCUDA and Scikit-CUDA packages. The α and β equilibrium volume fractions at different annealing temperatures and α grain shape and morphology at different dwelling times are predicted and compared with experimental results.