Algorithm Development in Materials Science and Engineering: Models and Algorithms for Microscale
Sponsored by: TMS Materials Processing and Manufacturing Division, TMS: Integrated Computational Materials Engineering Committee, TMS: Phase Transformations Committee, TMS: Computational Materials Science and Engineering Committee
Program Organizers: Mohsen Asle Zaeem, Colorado School of Mines; Garritt Tucker, Colorado School of Mines; Charudatta Phatak, Argonne National Laboratory; Bryan Wong, University of California, Riverside; Mikhail Mendelev, NASA ARC; Bryce Meredig, Travertine Labs LLC; Ebrahim Asadi, University of Memphis; Francesca Tavazza, National Institute of Standards and Technology

Wednesday 2:00 PM
February 26, 2020
Room: 31C
Location: San Diego Convention Ctr

Session Chair: Mohsen Eshraghi, California State University, Los Angeles; Mohammadreza Yaghoobi, University Of Michigan


2:00 PM  
“Sintering” Models and Measurements: Data Assimilation for Microstructure Prediction of Nylon Component SLS Additive Manufacturing: William Rosenthal1; Francesca Grogan1; Yulan Li1; Erin Barker1; Josef Christ1; Timothy Pope1; Tamas Varga1; Chris Barrett1; Mathew Thomas1; Noah Oblath1; Kevin Fox1; Malachi Schram1; Marvin Warner1; Amra Peles1; 1Pacific Northwest National Laboratory
    Selective laser sintering (SLS) printers drive high-throughput polymer additive manufacturing. However, thermal, feedstock, and exposure variations can introduce significant microstructure variability in the same batch of components. Phase-field models have been developed to simulate material microstructure evolution and kinetics during synthesis. We develop sensitivity analyses and introduce an adaptive sampling Bayesian algorithm to estimate significant parameters and uncertainties in a 3D phase-field model for nylon-12 polymer synthesis, including system free energy, interfacial energy, and sintering kinetics. In a high-throughput DIRAC-automated computational design loop, we validate the model through comparison to high-resolution 3D CT images of components built with varying orientations throughout the build chamber, as well as to partial sintering artifacts identified by laser exposure metadata. We quantify uncertainties in phase-field initial and operating conditions by developing a stochastic feedstock model from laser diffractometry and 3D CT imaging, and by analyzing real-time infrared thermographic movies taken throughout the build process.

2:20 PM  
Multi-scale Modeling of Solidification Microstructure during Powder Bed Fusion: Ryan Lenart1; Antonio Magana1; Mohsen Eshraghi1; 1California State University, Los Angeles
    Controlling solidification microstructure in additively manufactured metallic components can play a game-changing role in the way engineering components are designed and fabricated. However, experimental analysis of the melt pool is difficult because the size of the melt pool is very small and the heat source is moving very fast. A multi-scale model was developed to simulate the solidification microstructure in the melt pool of a powder bed fusion additive manufacturing process. A meso-scale thermo-fluid model was employed to simulate the molten pool geometry and dynamics. Then, the microstructural evolution was simulated by importing temperature data from the meso-scale model to a micro-scale model. The phase field method was used to model the evolution of the solidification microstructure while the lattice Boltzmann method was used to model solute transport in the liquid. The effect of process parameters on the resulting melt pool microstructure was studied.

2:40 PM  
Large Scale 3D Phase-field Sintering Simulations: Robert Termuhlen1; Hui-Chia Yu1; 1Michigan State University
    Sintering is a physical-metallurgy process, during which powder particles coalesce to form a connected solid mass without the melting of the particles. The phase field model (PFM) is an ideal technique for simulating materials science problems involving moving boundaries and has been applied to study the sintering phenomena. However, in typical PFM sintering simulations, each powder particle requires one order parameter, which leads to high computer memories to store multiple order parameter fields. We developed a method to dramatically reduce the memory demands, in which particles not in direct contact are assigned to the same order parameter. Furthermore, an additional label is assigned to the region occupied by each individual particle. Thus, the motion of each particle can still be explicitly tracked. This innovative method is implemented to simulate 3D nickel particle sintering during metal brazing processes.

3:00 PM  
Phase Field Modeling of Microstructure Evolution During Selective Laser Sintering and Post Aging: Yulan Li1; Erin Barker1; William Rosenthal1; Francesca Grogan1; Amra Peles1; 1Pacific Northwest National Laboratory
    Phase field models have been developed to simulate the microstructure evolution during selective laser sintering (SLS) process and post aging, respectively. As one of the fastest growing additive manufacturing techniques, the SLS uses a laser as the power source to sinter powdered material into a solid structure. In this work, we aim to investigate the effects of (1) the feedstock powder particle size distribution and packing and (2) the SLS process parameters such as temperature and temperature gradient on microstructures of the resulting component. The microstructure evolution during post aging in terms of pore structure and porosity volume fraction is simulated as well. Polymer nylon–12 is taken as the model material. Surface diffusion and condensation through particle surface melting and rigid movement dominate the sintering process while bulk diffusion and interfacial energy minimization determine the pore structure during aging. Discussion on comparison between simulations and experimental measurements will be presented.

3:20 PM  
A New Phase-field Model with Anisotropic Interface Width for the Highly Anisotropic Growth of Ice Dendrites: Gilles Demange1; Renaud Patte1; Helena Zapolsky1; 1University of Rouen
    The growth of ice crystals is a famous example of highly anisotropic dendrite growth: ice dendrites display at the same time a faceted aspect, and an extreme aspect ratio associated with a flat morphology. This flat shape is very hard to reproduce with usual models for solidification. To address this challenging issue, we present in this work an innovative procedure for slowly growing interfaces, which equips our new phase-field model (PFM) for the highly anisotropic growth of ice crystals [1]. This procedure relies on the fine tuning of the interface width in the PFM, which in turn accounts for the strongly reduced particle diffusion and attachment at the crystal surface in the slow growth direction. We show that this approach is capable of reproducing some fine features of ice crystal growth, such as terrace growth.

3:40 PM Break

4:00 PM  
Direct Consideration of Vacancies in CALPHAD Modelling of Zirconium Carbide: Theresa Davey1; Ying Chen1; 1Tohoku University
     Zirconium carbide is of interest in nuclear and aerospace industries due to its extremely high melting point. Its properties are strongly affected by significant structural vacancies. Conventional CALPHAD-type phase diagram models do not directly consider such defects, and the widely-used C-Zr phase diagram [1] has been shown to be intrinsically incompatible with our physical understanding of structural point defects [2]. This work uses state-of-the-art first-principles calculations of defect-related properties [3,4] to inform development of specific Gibbs energy models for cases where many structural point defects are present. Incorporating such information directly into the thermodynamic database produces a more physically consistent description and may allow further predictive ability. [1] A Fernández Guillermet. Journal of Alloys and Compounds, 217:69–89, 1995. [2] T Davey. PhD thesis (Imperial College London), 2017. [3] AI Duff, et al.; Physical Review B, 91(21):214311, 2015. [4] TA Mellan, et al., Physical Review B, 98(17):174116, 2018.

4:20 PM  
Multi-scale Modelling of Coarsening Process in the Ag-Cu Alloy: Bence Gajdics1; Helena Zapolsky2; Zoltán Erdélyi1; Gilles Demange2; János Tomán1; 1University of Debrecen; 2University of Rouen
    We report a new multi-scale procedure based on the recently developed Stochastic Kinetic Mean Field (SKMF) approach, combined with the Phase Field model (PFM) and CALPHAD database, to study the nucleation-growth-coarsening process in the Ag-Cu alloy. The SKMF approach reproduces the nucleation and early growth of copper precipitates in the silver matrix, and the PFM then simulates the coarsening of the microstructure. The length and time scales of SKMF and PFM are connected by matching the interface profiles and growth velocity of an isolated growing precipitate as simulated by both methods. Moreover, the free energy used in the PFM is derived from the interaction energies used in SKMF. Two implementations are proposed. First, the post-nucleation microstructure as provided by SKMF is used as the initial condition for subsequent PFM simulations. Second, only the particle size distribution and particle density are transferred to PFM, thereby giving access to bigger systems.

4:40 PM  
PRISMS-Plasticity: An Open-source Crystal Plasticity Finite Element Software: Mohammadreza Yaghoobi1; Sriram Ganesan2; Srihari Sundar2; Aaditya Lakshmanan2; Aeriel Murphy-Leonard1; Shiva Rudraraju3; John Allison1; Veera Sundararaghavan2; 1Materials Science and Engineering, University of Michigan, Ann Arbor; 2Aerospace Engineering, University of Michigan, Ann Arbor; 3Mechanical Engineering, University of Michigan, Ann Arbor
    An open-source parallel 3-D crystal plasticity finite element (CPFE) software package PRISMS-Plasticity is presented here as a part of PRISMS integrated framework. A highly efficient rate-independent crystal plasticity algorithm is implemented along with developing its algorithmic tangent modulus. A new twinning-detwinning mechanism is incorporated into the framework based on an integration point sensitive scheme. The integration of the PRISMS-Plasticity software with experimental characterization techniques using available open source software packages of DREAM.3D and Neper is elaborated. The integration of the PRISMS-Plasticity software with the information repository of Materials Commons is also presented. The parallel performance of the software is characterized which demonstrates that it scales well for large problems running on hundreds of processors. Various examples of polycrystalline metals with face-centered cubic (FCC), body-centered cubic (BCC), and hexagonal close-packed (HCP) crystals structures are presented to show the capability of the software to efficiently solve crystal plasticity boundary value problems.

5:00 PM  
Robust and Accurate Self-consistent Homogenization of Elasto-viscoplastic Polycrystals: Miroslav Zecevic1; Ricardo Lebensohn1; 1Los Alamos National Laboratory
    Robust and accurate homogenization schemes for elastic and rigid-viscoplastic polycrystals have been developed over the past several decades. However, homogenization in the elasto-viscoplastic (EVP) regime is a harder problem due to the interplay between elastic and plastic deformation, and has received considerably less attention. We have developed and compared three new self-consistent (SC) schemes for EVP polycrystals, based on two different solutions to Eshelby’s EVP inhomogeneity problem, which is at the basis of SC schemes. In addition, we have also investigated the feasibility of calculating intragranular stress fluctuations in the context of the new EVPSC formulations. Predictions were compared with each other, with experiments, and with full-field predictions. The elastic and viscoplastic limits were respected, and acceptable agreement was in general observed in the most challenging elastoplastic transition. Benefits and drawbacks of the proposed models are discussed, and main difficulties involved in the SC homogenization of EVP polycrystals are outlined.