Additive Manufacturing of Metals: Applications of Solidification Fundamentals: Micro-scale Modeling
Sponsored by: TMS Materials Processing and Manufacturing Division, TMS: Additive Manufacturing Committee, TMS: Solidification Committee
Program Organizers: Alex Plotkowski, Oak Ridge National Laboratory; Lang Yuan, University of South Carolina; Kevin Chaput, Northrop Grumman; Mohsen Asle Zaeem, Colorado School of Mines; Wenda Tan, The University of Michigan; Lianyi Chen, University of Wisconsin-Madison
Monday 2:00 PM
March 15, 2021
Room: RM 4
Location: TMS2021 Virtual
Session Chair: Lianyi Chen, University of Wisconsin; Wenda Tan, University of Utah
2:00 PM
3D Analysis of Grain Morphologies and Solidification Texture in AM 316L: David Rowenhorst1; 1Naval Research Laboratory
The morphologies and shapes of grains in additively manufactured material are dominated by the local growth conditions and the competition between neighboring grains. Most actual builds further complicate this by altering the spatial arrangement of the melt-pools from layer-to-layer, meaning that grains that had a favorable orientation on one layer are poorly aligned for the next. Using serial-sectioning, we will show how the 3D grain morphologies are dominated by localized growth dynamics throughout a powder-bed fusion 316L build. By analyzing over thirty-thousand grains and their local crystallographic 3D orientation we show that even within this changing competitive growth environment, <001> dominate, but lead to a large disparity in grain size and morphologies.
2:20 PM
A Multi-scale Modeling Approach to Microstructure Prediction for Powder Bed Fusion Additive Manufacturing Processes Through Phase Field and Cellular Automata Methods: Daniel Dreelan1; Abdur Rahman Al Azad1; Alojz Ivankovic1; Philip Cardiff1; David Browne1; 1University College Dublin
A phase field (PF) model is developed to simulate the growth and morphology of a representative number of solidifying grains, predicting microsegregation and local impingement dynamics. The effects of various processing-related parameters, including local temperature gradient and cooling rate, on cell/dendrite morphology and growth velocity are predicted. At the scale of the printed components, fast and efficient envelope-type cellular automata (CA) techniques are developed to simulate the nucleation and growth of all grains in 3D. This global CA model predicts the number and morphology of grains, the location of their boundaries, and crystallographic texture. The dendrite kinetics controlling the growth of the grains are imported into the CA model from the micro-scale PF predictions. The variation of the predicted microstructure with process parameters such as the heat source intensity and speed are presented, and the relevance of the simulations to defect formation is outlined.
2:40 PM
CA Model Sensitivity to Material Parameters, Nucleation, and Thermal Conditions Across AM Process Space: Matthew Rolchigo1; Alex Plotkowski2; John Coleman2; Jim Belak1; 1Lawrence Livermore National Laboratory; 2Oak Ridge National Laboratory
The balance between epitaxial and nucleated grain growth during additive processing of alloys is a function of many processing and material parameters, and uncertainty in parameters governing this balance contribute to uncertainty in grain structure prediction. For representative thermal conditions, CA model predictions of the columnar-to-equiaxed transition are validated with predictions from alloy solidification theory. Variability in CA-predicted microstructures for these representative conditions is shown to be non-uniformly distributed as functions of thermal gradient, interfacial response, and nucleation parameters. By coupling the CA model to OpenFOAM predictions of a specific Additive process, microstructure prediction uncertainty associated with the thermal model is examined and compared to uncertainty regarding heterogeneous nucleation event density and potency. This understanding of the inherent uncertainty in CA predictions of grain structure due to uncertain inputs and the model itself will be important for validating the model with experimental results and interpreting the validity of model predictions.
3:00 PM
Controlling Additive Manufacturing Processes with Magnetic Fields: Andrew Kao1; Teddy Gan1; Xianqiang Fan2; Catherine Tonry1; Ivars Krastins3; Peter Lee2; Koulis Pericleous1; 1University of Greenwich; 2UCL; 3University of Latvia
Strong thermal gradients in additive manufacturing (AM) form thermoelectric currents in the melt pool. Applying an external magnetic field generates a Lorentz force driving flow, a phenomenon known as Thermoelectric Magnetohydrodynamics (TEMHD). Theoretical studies by the authors have shown that TEMHD can significantly alter the melt pool dynamics and consequently, solute redistribution and microstructure evolution. In this work the effect of the magnetic field is investigated combined with scanning strategies, such as hatching. The results show that TEMHD has the potential to alleviate defects in AM for example disrupting epitaxial growth. The study uses a coupled meso-micro approach. On the meso-scale an enthalpy method resolves the electromagnetic, thermal and hydrodynamic problem. The CA method is used to predict the microstructure.The change in melt pool shape was experimentally validated using correlative imaging, including fast infra-red and optical imaging. Metallography was performed to quantify the impact on grain size and microstructure.
3:20 PM
Optimizing and Validating the Cellular Automata Finite Element Model for Additive Manufacturing: Kirubel Teferra1; David Rowenhorst1; 1United States Naval Research Laboratory
The present work details an implementation of the Cellular Automata Finite Element (CAFE) model that introduces key steps to improve its computational efficiency in order to simulate the additive manufacturing (AM) solidification process for large polycrystalline domain sizes. The CAFE model, which representing growing dendrites by their envelope geometry, is suitable for AM, where rapid solidification leads to a cellular and nearly planar solidification front. However, the highly localized and translating melt pool renders the CAFE model highly computationally inefficient, since only a small part of the simulation domain is active for any given time step. Key algorithmic improvements are made, and its parallel scalability is demonstrated. This enables simulating large 3D domains such that pole figure plots achieve convergence. Validation studies performed show an excellent comparison of grain morphology and texture to experimental characterization data for two different laser scan patterns for laser powder bed fusion 316L stainless steel.
3:40 PM
Prediction of Columnar-to-equiaxed Transition in Single Tracks during Laser Powder Bed Fusion Additive Manufacturing: Lang Yuan1; Adrian Sabau2; David StJohn3; Arvind Prasad3; Peter Lee4; 1University of South Carolina; 2Oak Ridge National Laboratory; 3The University of Queensland; 4University College London
Highly customized microstructures of the printed alloy formed during laser powder bed fusion additive manufacturing can be produced by tailoring the transition from columnar to equiaxed grains (CET). In this study, classic analytical solutions for both thermal diffusion and solidification are applied to solidifying melt pools to evaluate the formation of CETs for a wide range of combinations of laser power, speed and substrate temperature. This high-throughput tool effectively scans process parameters and provides potential indicators for target microstructures. Process parameters that the analytical solutions suggest may result in the desired columnar or fully equiaxed microstructure are examined in detail using a massively parallelized microstructural solidification model to reveal both nucleation and grain growth during multiple-layer solidification. The Interdependence theory and observations from controlled LPBFAM experiments are discussed to interpret the numerical predictions and provide insights for controlling CET during the LPBFAM processing.
4:00 PM
Effect of Kinetic Anisotropy on Microstructure Development during Simulated Powder Bed Fusion of 316L Stainless Steel: Alexander Chadwick1; Peter Voorhees1; 1Northwestern University
During additive manufacturing, typical melt pools may have hundreds of grains in contact with the solid-liquid interface. Each grain can have orientation-dependent solidification kinetics, which eventually leads to competitive grain growth and selection of preferred growth directions. Using the phase-field method and a Rosenthal solution, we have developed a physics-based (i.e., non-stochastic) model for the evolution of thousands of grains in three-dimensions for solidification in the absolute stability regime. We perform large-scale three-dimensional simulations of AM for a 316L surrogate to understand the complex interplay between the kinetic anisotropy, grain size, and laser power and speed on the resulting microstructure. From these simulations, we find that anisotropy has a strong effect only near the rear of the melt pool, but it directly impacts the time scale over which a grain solidifies and then becomes inactive. We also examine the resulting three-dimensional morphologies of individual grains after multiple laser passes.
4:20 PM
Microstructure Prediction Framework for Additively Manufactured Metals: Andrew Polonsky1; Narendran Raghavan2; McLean Echlin3; Michael Kirka2; Ryan Dehoff2; Jonathan Madison1; Tresa Pollock3; 1Sandia National Laboratories; 2Oak Ridge National Laboratory; 3University of California, Santa Barbara
The interplay between processing conditions and performance of additively manufactured (AM) parts remains an area of intense study. The wide scope of processing parameters is integral to the versatility of AM techniques, but can also create highly variable material properties. Here we present a framework for the accurate prediction of AM microstructures without extensive parametric studies. Existing models designed to predict the columnar-to-equiaxed transition (CET) are calibrated via targeted experiments using TriBeam tomography for three-dimensional characterization and thermal modelling using TRUCHAS on a sample of Inconel 718. The calibrated model is then validated against additional experimental microstructures of the same alloy. Implications of the predictive framework as well as its applicability on novel alloy formulations will also be dis