Computational Thermodynamics and Kinetics: Solidification/Additive Manufacturing
Sponsored by: TMS Functional Materials Division, TMS Materials Processing and Manufacturing Division, TMS Structural Materials Division, TMS: Chemistry and Physics of Materials Committee
Program Organizers: Vahid Attari, Texas A&M University; Sara Kadkhodaei, University Of Illinois Chicago; Eva Zarkadoula, Oak Ridge National Laboratory; Damien Tourret, IMDEA Materials Institute; James Morris, Ames Laboratory

Wednesday 8:30 AM
March 2, 2022
Room: 255C
Location: Anaheim Convention Center

Session Chair: Kubra Karayagiz, Worcester Polytechnic Institute; Damien Tourret, IMDEA Materials Institute


8:30 AM  
Multiscale Dendritic Needle Network Study of the Effect of Buoyant Liquid Flow on Dendritic Growth Kinetics: Thomas Isensee1; Damien Tourret2; 1IMDEA Materials Institute & Universidad Politécnica de Madrid; 2IMDEA Materials
    The growth of metals and alloys from the liquid is heavily disturbed by gravity-induced buoyant melt flow induced by the rejection of heat and/or solute during solidification. Depending on the direction of growth and that of gravity, hydrodynamically stable or unstable regimes can be established, which lead to steady or oscillatory growth dynamics, and also affect length scale (e.g. dendritic spacings) selection. We use the multiscale Dendritic Needle Network model to investigate dendrite growth dynamics under various gravity conditions, including a hydrodynamically unstable regime resulting in oscillatory dendritic growth velocities. This oscillatory growth behavior was recently observed experimentally during the directional solidification of nickel-based single crystal superalloys [G. Reinhart, et al., Acta Materialia 194 (2020) 68-79]. Our DNN simulations shall unveil the fundamental mechanisms behind these oscillations, pinpoint their specific conditions of occurrence, and therefore help the design of single-crystal either promoting or avoiding such unstable regimes.

8:50 AM  
Machine Learning-assisted High-throughput Exploration of Interface Energy Space in Multi-phase-field Model with CALPHAD Potential: Vahid Attari1; Raymundo Arroyave1; 1Texas A&M University
    Computational methods are increasingly being incorporated into the exploitation of microstructure–property relationships for microstructure-sensitive materials design. We propose non-intrusive materials informatics methods for high-throughput exploration and analysis of a synthetic microstructure space using a machine learning-reinforced multi-phase-field modeling scheme in the framework of ICME. We study the interface energy space as one of the most uncertain inputs in phase-field modeling and its impact on the shape of a growing secondary phase between solid and liquid phases. We use variational autoencoder, a deep generative neural network method, and label spreading, a semi-supervised machine learning method for establishing correlations between inputs and outputs of the model. A structure map in the interface energy space is developed that shows $\sigma_{SI}$ and $\sigma_{SL}$ alter the shape of the intermetallic synchronously where an increase in the latter and decrease in the former changes the shape from dewetting structures to wetting.

9:10 AM  
Capturing the Undercooling for Solidification of Inoculated FCC Metals: Mark O'Masta1; Eric Clough1; John Martin1; 1HRL Laboratories LLC
     Inoculation of metals and alloys to promote heterogeneous nucleation for controlled grain structure and size is widely used from casting to additive manufacturing. This talk presents a computation study investigating the thermodynamic factors controlling incipient solidification on lattice mismatched substrates. Molecular dynamic (MD) simulations analyze the nucleation of FCC metals, and strain relief mechanisms that can substantially reduce the required undercooling for heterogeneous nucleation [1]. Energy based calculations are used to analyze prominent relief mechanisms, and their implications on nucleant potency and selection. The talk will be motivated by recent results from a combinatorial study, using selective laser melting (SLM), of aluminum functionalized with foreign particles [2]. [1] O’Masta, M.R., et al. Comput. Mater. Sci. 192 (2021): 110317. https://doi.org/10.1016/j.commatsci.2021.110317.[2] Martin, J.H., et al. Acta Mat. 200 (2020): 1022–37. https://doi.org/10.1016/j.actamat.2020.09.043.

9:30 AM  Cancelled
Simulation of Evolving Microstructure in Additive Manufacturing of Al-Si Alloys Using Phase Field Modeling: Abdur Al Azad1; Philip Cardiff1; David Browne1; 1University College Dublin
    A phase field model of alloy solidification, incorporating thermal effects due to the release of latent heat at the growing solid-liquid interface, has been developed. The model was initially used to study the growth from a spherical solid nucleus inside a domain of undercooled melt of Al-Si binary alloy, using different intensities of computational perturbation and crystal surface energy anisotropy. With the increase of perturbation intensity, dendritic interface morphology emerges and growth becomes more developed, with additional secondary and tertiary arms appearing. Furthermore, the anisotropy has a significant effect on the kinetics of dendrite growth – tips accelerate with increasing anisotropy. For applications in laser-melted additive manufacturing, the effects of typical processing parameters such as temperature gradient, scan speed and cooling rate, on simulated solidification morphology are reported.

9:50 AM Break

10:10 AM  
Modelling Informed Strategy for the Additive Manufacturing of High-strength Al-alloys: Giuseppe Del Guercio1; William Reynolds1; Adam Clare1; Marco Simonelli1; 1University of Nottingham
    Material properties of components manufactured by laser powder bed fusion are heavily influenced by the thermal gradients and solidification rates experienced during processing. These factors cause high strength aluminium alloys to suffer from detrimental cracking phenomena. Controlling the thermal gradients and solidification rates are key in eliminating cracking and increasing their mechanical performance. Here we show how the calculation of processing parameters deliberately conceived to lower thermal gradients, eliminates cracking from a popular high strength aluminium alloy, AA2024. This work shows how the computation from an inverse solution method can lead directly to parameters that are able to achieve crack free 99.6% dense AA2024. In addition, we show how cracking can be predicted in high strength aluminium alloys and the mechanical performance improvements for crack-free AA2024. These results show how the pairing of computational modelling and materials science can result in the crack-free AA2024 components without chemistry modification.

10:30 AM  
Simulation of Molten Pool Dynamics during Metallic Additive Manufacturing: Lu Wang; Wentao Yan1; 1National Univeristy of Singapore
    The molten pool flow, particularly the keyhole effect, plays a critical role in defect formation in additive manufacturing and welding processes. In this study, we derive an evaporation model for metal alloys considering the gas flow structure and material composition and implement it in a multi-physics thermal-fluid flow model, which utilizes the volume of fluid (VoF) in the finite volume method (FVM) to capture free surfaces and the ray-tracing method to track multi-reflections of the laser within the keyhole. Furthermore, a Thermoelectric Magnetohydrodynamic (TEMDH) model is developed by incorporating the electrodynamic model with the Seebeck effect into the multi-physics thermal-fluid flow model to manipulate the molten pool flow and grain growth. Both models are validated against the experiment results.