The 7th International Congress on 3D Materials Science (3DMS 2025): Methods of Materials Simulation and Modelling in 3D & 4D I
Program Organizers: Henry Proudhon, Mines Paris Centre Des Materiaux; Can Yildirim, European Synchrotron Radiation Facility
Monday 9:00 AM
June 16, 2025
Room: Platinum Ballroom 2
Location: Anaheim Marriott
Session Chair: Aly Badran, GE Aerospace Research
9:00 AM Invited
The Limits of Predictability in 4D Microstructural Simulations: Elizabeth Holm1; Meizhong Lyu1; 1University of Michigan
A longstanding goal of microstructural simulation is to reproduce experimentally observed outcomes in 4D (space and time). Disagreement between experiment and simulation has generally been attributed to shortcomings in the computational instantiation; however, even after several decades of improving the physical bases of computational models, a perfect match has not been achieved. In this study, we examine the sources of uncertainty in simulation and experiment using polycrystalline grain growth as an exemplar. In nominally identical molecular dynamics simulations, growth trajectories of individual grains can vary significantly due to discrete events that have cascading effects on the microstructural ensemble. Thus, microstructural events are surprisingly sensitive to atomic scale processes, which imparts an innate uncertainty in 4D simulations. While this prevents prediction of an individual grain trajectory, a range of possible trajectories can be determined for each grain; this offers a more robust, statistically-based way to compare simulation and experiment.
9:30 AM
A Polycrystal Generator for Large-Scale Atomistic Calculations: Younggak Shin1; Vichhika Moul2; Keonwook Kang1; Byeongchan Lee2; 1Yonsei University; 2Kyung Hee University
Metals typically have a polycrystalline structure in which grain boundaries play a pivotal role in mechanical properties such as the yield strength, i.e. the Hall-Petch and the inverse Hall-Petch relationships. Atomistic calculations have been extensively used to understand these relationships, but now facing an unforeseen challenge: present atom-generation codes are serial codes, where the on-board memory limits the atomistic system size. In addition, reading and writing a file containing billions of atoms takes prohivitively long time in a serial I/O.Here we present a polycrystal generator called PolyPal, which is a fully parallized code with virtually perfect scalability. The software has been tested up to 10 billion atoms, which takes only minutes to generate all the positions, and write them to files. We also highlight the key features in generating different textures and alignments of microsctures to mimic the experimental specimen fabracated in various processes.
9:50 AM
Dislocation Contrast Simulations in Dark-Field X-Ray Microscopy to Identify Individual Dislocations: Sina Borgi1; Grethe Winther1; Henning Friis Poulsen1; 1Technical University of Denmark
Exploring dislocation identification in bulk deformed crystals using dark-field X-ray microscopy (DFXM) and forward modeling. A covariance matrix of weak beam images shows how variations in dislocation properties (line direction, Burgers vector, and slip plane) affect diffraction contrast, offering insights into DFXM’s sensitivity in distinguishing between different dislocations. In experimental and simulated weak beam images, dislocation distortion fields match closely, confirming DFXM's capability to identify slip planes and Burgers vectors. This is further illustrated by a dislocation undergoing double cross-slip, where the experimental and simulated Burgers vector align, reinforcing DFXM’s accuracy. These advances in simulation and experimental validation establish DFXM as a powerful tool for exploring 3D dislocation dynamics. The method has potential for integration with discrete dislocation dynamics to follow dislocation networks under in situ strain, and using forward modeled images serving as a foundation for supervised machine learning in the DFXM data analysis pipeline.
10:10 AM
Graph Neural Networks for Generalizable Machine Learning of 3D Microstructure--Property Relationships in Polycrystals: Guangyu Hu1; Gyu-Jang Sim2; Myoung-Gyu Lee2; Marat Latypov1; 1University of Arizona; 2Seoul National University
3D characterization provides unique datasets for understanding process--structure--property relationships in alloys. Graphs offer a reduced-order representation of the 3D polycrystalline microstructure that enables efficient machine learning of these relationships with graph neural networks (GNNs). In this contribution, we will present our recent advances in data-efficient training of GNNs that can generalize predictions well beyond their training sets. Specifically, we will present GNNs that predict (i) anisotropic mechanical properties of textured polycrystals in new loading directions; (ii) grain-level fatigue indicator parameters in microstructure volume elements order of magnitude larger than those used for training; and (iii) properties of polycrystals for metals and alloys not included in the training set. This contribution intends to spark 3DMS community's interest in the GNN computational platform that could significantly benefit from experimental datasets for model training and fine-tuning.
10:30 AM
Simulation of Plastic Strain Localisation in Polycrystals by 3D Discrete Dislocation Dynamics: Baptiste Joste1; Riccardo Gatti2; Henry Proudhon1; Benoit Devincre2; 1Centre Des Matériaux; 2CNRS
Understanding deformation leading to polycrystalline material failure is a major challeng in materials science. The origin of plastic strain localization and deformation propagation mechanisms remain unclear. To tackle those issues, we propose a multi-scale simulation approach using 3D Discrete Dislocation Dynamics (DDD) and the Finite Element Method (FEM) to model mesoscopic dislocation behavior. First, our results emphasize the impact of policrystalline microstructure on plastic deformation localization, revealing intragranular stress concentrations due to elastic incompatibilities. During plastic deformation, we obsereved that specific mechanisms such as cross-slip or collinear annihilation facilitate the initiation of plastic deformation within these stress concentration regions a. Furthermore, ,we implemented a local rule to model dislocation transmission, to decrease strain hardening and promotes slip localization. Our findings highlights that plastic deformation transmission is crucial to observe plastic strain localization in polycrystals.