Cellular automata simulations of texture evolution in flat-top laser additively manufactured Hastelloy-X

Fabien Briffod; Phuangphaga Daram; Tomonori Kitashima; Makoto Watanabe

doi:10.1016/j.jmrt.2026.06.251

Journal article Cellular automata simulations of texture evolution in flat-top laser additively manufactured Hastelloy-X

; ; ;

Download file (8.96 MB)
Download as zip (7.76 MB)

Collection

Citation

Fabien Briffod, Phuangphaga Daram, Tomonori Kitashima, Makoto Watanabe. Cellular automata simulations of texture evolution in flat-top laser additively manufactured Hastelloy-X. Journal of Materials Research and Technology. 2026, 43 (), 3228-3244. https://doi.org/10.1016/j.jmrt.2026.06.251

(BibTeX)

Description:

(abstract)

The present study investigates the influence of process parameters, nucleation density, and stochastic effects on the microstructure evolution of Hastelloy X samples fabricated by selective laser melting on near single crystal Ni substrates using a flat top laser. Experiments and cellular automata simulations were performed to analyze crystallographic texture and M-index evolution along the build height as function of laser power, scan speed and heterogeneous nucleation density. The results show that higher linear energy densities promote epitaxial growth, stabilizing the texture, while lower energy densities lead to finer grains and increased high-angle grain boundaries due to enhanced nucleation. A nucleation density of 50$\times$10$^{12}$ m$^{-3}$ provided the most reasonable agreement in terms of texture strength evolution with most experimental observations while higher nucleation densities lead to a weakening of the texture combined with a faster transition from the initial $\{001\}_{z}\langle 110 \rangle_{x}$ texture of the substrate to a $\{001\}_{z}\langle 100 \rangle_{x}$ texture. Additionally, microstructure variability due to stochastic nucleation was found to be process-dependent, with some conditions exhibiting stable, repeatable texture evolution, while other conditions showed high sensitivity to random seed selection. These findings highlight the need for probabilistic modeling approaches to improve the predictive accuracy of AM microstructure simulations.

Rights: