Additive Manufacturing Modeling, Simulation, and Machine Learning: Microstructure, Mechanics, and Process: Microstructures & Defects I
Program Organizers: Jing Zhang, Indiana University – Purdue University Indianapolis; Brandon McWilliams, US Army Research Laboratory; Li Ma, Johns Hopkins University Applied Physics Laboratory; Yeon-Gil Jung, Korea Institute of Ceramic Engineering & Technology

Monday 2:00 PM
October 10, 2022
Room: 303
Location: David L. Lawrence Convention Center

Session Chair: Li Ma, Johns Hopkins University Applied Physics Laboratory; Jing Zhang, Indiana University - Purdue University Indianapolis; Brandon McWilliams, CCDC Army Research Laboratory; Yeon-Gil Jung, Changwon National University


2:00 PM  
Quantitative Analysis of Computed Tomography Characterization of Porosity in AM Ti64 Using Serial Sectioning Ground Truth: Bryce Jolley1; Michael Uchic1; Daniel Sparkman1; Christine Henry1; Michael Chapman2; Edwin Schwalbach1; 1Air Force Research Laboratory; 2UES, Inc.
    X-Ray Computed Tomography is a widely-used method for nondestructive characterization of internal porosity in complex-shaped parts produced by Additive Manufacturing. However, there are a host of factors that can significantly affect the accuracy of computer-based analysis of porosity from XCT data. This presentation will explore a quantitative assessment of XCT porosity-characterization workflows for a 10 mm diameter cylindrical Ti-64 sample. The same sample has been characterized using four different XCT systems to explore the sensitivity to the XCT experimental, reconstruction, and segmentation workflows, after which automated serial sectioning was employed to estimate the true internal porosity distribution within a 2 mm thick subregion of the cylinder with ~2 micrometer spatial resolution. Notably, titanium alloy ball bearings were adhered to the surface of the cylindrical sample to facilitate automated registration of the multi-modal datasets, thus enabling both global and local analysis of the changes in pore measurements.

2:20 PM  
Uncertainty Quantification in Process-Structure-Properties Simulations of Additive Manufactured Ti-6Al-4V: Joshua Pribe1; Brodan Richter2; Patrick Leser2; Saikumar Yeratapally1; George Weber2; Andrew Kitahara1; Edward Glaessgen2; 1National Institute of Aerospace; 2NASA Langley Research Center
    Metal additive manufacturing (AM) enables rapid fabrication of parts with complex geometries. However, AM parts typically have heterogeneous microstructures that depend on a range of processing parameters. The resulting mechanical property variations inhibit confidence in the structural performance of AM parts, particularly in fatigue where local material behavior is critical. This work leverages process-structure-properties simulations and uncertainty quantification to predict fatigue indicators in AM Ti-6Al-4V. Microstructures are generated using kinetic Monte Carlo simulations that include a calibrated analytical solution for the temperature field and predict crystallographic texture. Micromechanical stress and strain fields are predicted using an elasto-viscoplastic fast Fourier transform formulation. Uncertainty in the fatigue indicators is analyzed with respect to build parameter variations, material property uncertainty, and inherent randomness in microstructure and texture. The key outcome is more robust AM process-structure-properties linkages, supporting development of a computational materials-informed approach to qualification and certification of structural AM parts.

2:40 PM  
Analyzing Uncertainty in Modeled Additive Process-Microstructure-Property Relationships Using the ExaAM Framework: Matthew Rolchigo1; John Coleman1; Robert Carson2; Gerry Knapp1; Alex Plotkowski1; Scott Wells3; Samuel Reeve1; Lyle Levine4; Jim Belak2; John Turner1; 1Oak Ridge National Laboratory; 2Lawrence Livermore National Laboratory; 3Purdue University; 4National Institute of Standards and Technology
    Modeling of additive process-microstructure-property relationships necessitates understanding and quantifying the sources of uncertainty in each model, as well as the result of input parameter uncertainty propagation on model results. Using pre-Exascale hardware and codes developed and modified as part of the ExaAM project, we use Tasmanian sparse grids to generate a range of values for several uncertain input parameters for AdditiveFOAM modeling of heat transport and ExaCA modeling of as-solidified grain structure, generating a large ensemble consisting of hundreds of simulated microstructures. Microstructures will be compared to an additive benchmark dataset for Inconel 625, and sensitivity of several aspects of the grain structures will be quantified as functions of model and input parameter uncertainties. ExaConstit will then be used to model the variation in constitutive properties using representative regions of the resulting microstructures, further linking the sensitivity of properties to processing and microstructure. Supported by ECP (17-SC-20-SC).

3:00 PM  
Thermal Modeling of Laser Powder Bed Fusion Additive Manufacturing of Refractory Materials: Li Ma1; Gianna Valentino1; Mitra Taheri2; Morgana Trexler1; 1Johns Hopkins University Applied Physics Laboratory; 2Johns Hopkins University
    Refractory metals, e.g., tungsten, are an extraordinary class of materials, known for their exceptional properties including high melting temperatures, thermal conductivity, stiffness and strength. While these materials are desirable for high temperature and extreme environment applications, their limited room temperature ductility makes fabrication and implementation challenging. Additive manufacturing (AM) has many potential advantages, particularly in the processing of refractory metals. However, the process parameters needed to fabricate defect-free parts, as well as the effects of AM processing on material properties, are not fully known. In this work, we will computationally study the effect of laser processing parameters on the solidification behavior of tungsten alloys. Single track melt pools will be modeled using computational fluid dynamics software, FLOW3D, and validated with AM experiments across a range of processing parameters (e.g., laser power, velocity). The results from this work will enable rapid development of AM processing parameters for refractory metals and alloys.

3:20 PM Break

3:40 PM  
Parent Grain Reconstruction Using Orientation Imaging Microscopy and Deep Learning: Patxi Fernandez-Zelaia1; Andres Marquez Rossy1; Quinn Campbell1; Andrzej Nycz1; Chris Ledford1; Michael Kirka1; 1Oak Ridge National Laboratory
    We present a deep convolutional neural network model which estimates parent grain austenitic structure from observed child phase martensite electron backscatter diffraction micrographs. The model was trained using only four micrographs by using a novel data augmentation strategy that exploits properties of the orientation representation. Despite the model being trained on a traditionally fabricated martensitic alloy, when tested on additively manufactured material, which exhibits vastly different austenitic grain structure, the model performs exceptionally well. The model efficacy depends on the microstructure length scale and receptive field of the vision model. This work demonstrates that modern machine learning vision models are well suited for analyzing the complex spatial-orientation structure found within orientation imaging micrographs. The model evaluates much faster than traditional approaches which may enable rapid characterization of additively manufactured materials.

4:00 PM  
CFD Simulations of Spatter Removal in a Laser Powder Bed Fusion Machine: Nicholas Obrien1; Syed Zia Uddin1; Satbir Singh1; Jonathon Malen1; Jack Beuth1; 1Carnegie Mellon University
    In Laser Powder-Bed Fusion (LPBF) the melt pool can eject spatter on the scale of 20 µm to upwards of 100 µm, causing defects in the final part when it lands back in the build area. To prevent this, argon is blown parallel to the build plate, entraining spatter particles as they appear. In this work, computational fluid dynamics (CFD) simulations are employed to investigate the argon flow and its interaction with spatter ejected during the LPBF process. Spatter particles with various diameters and velocities are released at multiple locations in the machine. It is found that gravity and drag forces can cause spatter particles to either land on the powder bed or be carried from the machine. Simulation results are compared to spatter particle distributions visualized via infrared imaging. Design changes to the machine are explored to improve spatter removal.

4:20 PM  
A Parametric Molecular Dynamics Study of Additive Nanomanufacturing: Effects of Size, Misorientation, and Temperature on Sintering Characteristics : Dourna Jamshideasli1; Shuai Shao1; Masoud Mahjouri-Samani1; Nima Shamsaei1; 1Auburn University
    Optimizing properties of additively manufactured "biodegradable papertronics" dictates the fine balance of laser settings to maximize the sintering of nanoparticles and minimize damage to substrates. In this study, molecular dynamics is utilized to understand the sintering of silver and copper double-nanoparticles at different temperatures. The effects of nanoparticle size, nanoparticles’ size ratios, misorientation angle (including both twist and tilt), sintering temperature, and material type on the double-nanoparticles’ sintering characteristics such as neck size and center-to-center distances are quantified. Moreover, the evolution of particle shape and defect contents during the sintering process are captured. The changes in neck size in different case studies are compared to and improved upon, existing mathematical models.

4:40 PM  
Additive Manufacturing Beyond the Gaussian Beam: Insights from Mesoscale Modeling Studies: Daniel Moore1; Theron Rodgers2; Heather Murdoch3; Fadi Abdeljawad1; 1Clemson University; 2Sandia National Laboratories; 3Army Research Laboratory
    In laser-powder-bed-fusion additive manufacturing (LPBF AM), the microstructures that form are sensitive to the AM process parameters. Recent studies have explored laser beam types beyond the commonly used Gaussian beam in LPBF. However, the impact of beam shape on the evolution of temperature fields and its spatial gradients and the net effect on microstructure development remains poorly understood. Based on a recently developed mesoscale model that accounts for thermal transport and microstructure formation during AM, we explore the evolution of temperature fields and grain microstructures during LPBF using ring and Bessel beams. Sensitivity analysis of the parameters controlling the beam type will be presented. It is shown that changing the laser beam shape results in large variations in the spatial distribution of temperature fields and resulting microstructures. Broadly speaking, our work will expand the parameter space for microstructural design during AM to include laser beam profiles.

5:00 PM  
Increasing the Service Life of the Trimming Punch Using Nimonic Cutting Edge: Miroslav Urbanek1; 1COMTES FHT a.s.
    Repairing and depositing of functional edges and surfaces of forming tools can be efficiently performed by additive production using Directed Energy Deposition (DED). DED can be used to deposit directly on the tool and then mill it to the required quality. This paper focuses on tools for trimming membranes and flashes after closed die forging. Nickel alloys such as Nimonic 80A are suitable for these applications, but their price is higher than steel, so it is advisable to apply this material in a small volume to the cutting edge. The research focused on the characterization of the powder, tuning of depositing parameters, and measurement of mechanical properties using miniaturized test specimens. Furthermore, metallographic analysis and hardness measurements were performed. The trimming process was simulated using FE analysis in order to optimize the size of the trimming edge. At the same time, life tests were performed in an industrial application.