GrainPaint: A multi-scale diffusion-based generative model for microstructure reconstruction of large-scale objects
Source abstract
Abstract
Simulation-based approaches to microstructure generation can suffer from a variety of limitations, such as high memory usage, long computational times, and difficulties in generating complex geometries. Generative machine learning models present a way around these issues, but they have previously been limited by the fixed size of their generation area. We present a new microstructure generation methodology leveraging advances in inpainting using denoising diffusion models to overcome this generation area limitation. We show that microstructures generated with the presented methodology are statistically similar to grain structures generated with a kinetic Monte Carlo simulator, SPPARKS.
Source: openaire · Quality: candidate unverified · Source record ↗
Research context
OpenAlex topics: Image Processing and 3D Reconstruction · 3D Shape Modeling and Analysis · Generative Adversarial Networks and Image Synthesis
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
References (83)
Show source-provided references
1. US NSTC · 2011
2. Materials genome initiative strategic plan (2014)
3. Materials genome initiative strategic plan (2021)
4. Horstemeyer · 2012
5. Allison · 2013
6. New frontiers for the materials genome initiative10.1038/s41524-019-0173-4 ↗
7. Computational microstructure characterization and reconstruction: Review of the state-of-the-art techniques10.1016/j.pmatsci.2018.01.005 ↗
8. Kingma · 2013
9. NVAE: A deep hierarchical variational autoencoder
10. Improved training of wasserstein gans
11. Mirza · 2014
12. A computational building block approach towards multiscale architected materials analysis and design with application to hierarchical metal metamaterials10.1088/1361-651X/accfb5 ↗
13. P. Fernandez Zelaia, S. Thapliyal, R. Kannan, P. Nandwana, Y. Yamamoto, V. Paquit, A. Nycz, M.M. Kirka, Denoising Diffusion Probabilistic Models for Generative Alloy Design. Available At SSRN 4698278.
14. Digital polycrystalline microstructure generation using diffusion probabilistic models10.1016/j.mtla.2023.101976 ↗
15. Conditional diffusion-based microstructure reconstruction10.1016/j.mtcomm.2023.105608 ↗
16. Denoising diffusion algorithm for inverse design of microstructures with fine-tuned nonlinear material properties10.1016/j.cma.2023.116126 ↗
17. A data-driven framework for designing microstructure of multifunctional composites with deep-learned diffusion-based generative models10.1016/j.engappai.2023.107590 ↗
18. Predictive microstructure image generation using denoising diffusion probabilistic models10.1016/j.actamat.2023.119406 ↗
19. Statistically conditioned polycrystal generation using denoising diffusion models10.1016/j.actamat.2024.119746 ↗
20. Designing bioinspired composite structures via genetic algorithm and conditional variational autoencoder10.3390/polym15020281 ↗
21. Variational autoencoder-based topological optimization of an anechoic coating: An efficient-and neural network-based design
22. Towards inverse microstructure-centered materials design using generative phase-field modeling and deep variational autoencoders10.1016/j.actamat.2023.119204 ↗
23. Exploration of optimal microstructure and mechanical properties in continuous microstructure space using a variational autoencoder10.1016/j.matdes.2021.109544 ↗
24. Machine learning generative models for automatic design of multi-material 3D printed composite solids10.1016/j.eml.2020.100992 ↗
25. Designing complex architectured materials with generative adversarial networks10.1126/sciadv.aaz4169 ↗
26. Design of architectured composite materials with an efficient, adaptive artificial neural network-based generative design method10.1016/j.actamat.2021.117548 ↗
27. Reconstruction of the meso-scale concrete model using a deep convolutional generative adversarial network (DCGAN)10.1016/j.conbuildmat.2023.130704 ↗
28. Generation of highly realistic microstructural images of alloys from limited data with a style-based generative adversarial network10.1038/s41598-023-27574-8 ↗
29. Synthesizing controlled microstructures of porous media using generative adversarial networks and reinforcement learning10.1038/s41598-022-12845-7 ↗
30. On the use of artificial neural networks in topology optimisation10.1007/s00158-022-03347-1 ↗
31. Arjovsky · 2017
32. Improved techniques for training gans
33. Diffusion models beat gans on image synthesis
34. Denoising diffusion probabilistic models
35. R. Rombach, A. Blattmann, D. Lorenz, P. Esser, B. Ommer, High-resolution image synthesis with latent diffusion models, in: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022, pp. 10684–10695.10.1109/CVPR52688.2022.01042 ↗
36. Ramesh · 2022 · 3
37. Denoising diffusion algorithm for inverse design of microstructures with fine-tuned nonlinear material properties10.1016/j.cma.2023.116126 ↗
38. Multi-scale topology optimization with neural network-assisted optimizer
39. IH-GAN: A conditional generative model for implicit surface-based inverse design of cellular structures10.1016/j.cma.2022.115060 ↗
40. Optimizing diffusion to diffuse optimal designs
41. Machine learning-based inverse design of auxetic metamaterial with zero Poisson’s ratio
42. A. Lugmayr, M. Danelljan, A. Romero, F. Yu, R. Timofte, L. Van Gool, RePaint: Inpainting using Denoising Diffusion Probabilistic Models, in: 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR, 2022, pp. 11451–11461.10.1109/CVPR52688.2022.01117 ↗
43. A novel method to design biomimetic, 3D printable stochastic scaffolds with controlled porosity for bone tissue engineering10.1016/j.matdes.2022.110857 ↗
44. A data-driven approach for process optimization of metallic additive manufacturing under uncertainty10.1115/1.4043798 ↗
45. Optimizing the microstructure and processing parameters for lithium-ion battery cathodes: A use case scenario with a digital manufacturing platform10.1002/ente.202201032 ↗
46. 3D Denoising Diffusion Probabilistic Models for 3D microstructure image generation of fuel cell electrodes10.1016/j.commatsci.2024.113596 ↗
47. Multi-plane denoising diffusion-based dimensionality expansion for 2D-to-3D reconstruction of microstructures with harmonized sampling10.1038/s41524-024-01280-z ↗
48. Generating 3D images of material microstructures from a single 2D image: a denoising diffusion approach10.1038/s41598-024-56910-9 ↗
49. S. Plimpton, C. Battaile, M. Chandross, L. Holm, A. Thompson, V. Tikare, G. Wagner, E. Webb, X. Zhou, C.G. Cardona, et al., Crossing the Mesoscale No-Man’s Land Via Parallel Kinetic Monte Carlo, Sandia Report SAND2009- 6226, 2009.
50. Parallel simulation via SPPARKS of on-lattice kinetic and Metropolis Monte Carlo models for materials processing10.1088/1361-651X/accc4b ↗
51. Simulation of metal additive manufacturing microstructures using kinetic Monte Carlo10.1016/j.commatsci.2017.03.053 ↗
52. A Monte Carlo model for 3D grain evolution during welding10.1088/1361-651X/aa7f20 ↗
53. Predicting mesoscale microstructural evolution in electron beam welding10.1007/s11837-016-1863-8 ↗
54. Fast three-dimensional rules-based simulation of thermal-sprayed microstructures10.1016/j.commatsci.2021.110437 ↗
55. Three-dimensional simulation of grain growth in a thermal gradient with non-uniform grain boundary mobility10.1016/j.scriptamat.2008.05.039 ↗
56. Computer simulation of normal grain growth in three dimensions10.1080/13642818908220181 ↗
57. A constant-time kinetic Monte Carlo algorithm for simulation of large biochemical reaction networks10.1063/1.2919546 ↗
58. Luo · 2022
59. Towards coherent image inpainting using denoising diffusion implicit models
60. Ronneberger · 2015
61. Parallel simulation via SPPARKS of on-lattice kinetic and Metropolis Monte Carlo models for materials processing10.1088/1361-651X/accc4b ↗
62. A density-based algorithm for discovering clusters in large spatial databases with noise
63. Effects of lattice anisotropy and temperature on domain growth in the two-dimensional potts model10.1103/PhysRevA.43.2662 ↗
64. The computer simulation of microstructural evolution10.1007/s11837-001-0063-2 ↗
65. A Bézier curve fit to melt pool geometry for modeling additive manufacturing microstructures10.1016/j.cma.2023.116208 ↗
66. Microstructure-based modeling of laser beam shaping during additive manufacturing10.1007/s11837-023-06363-8 ↗
67. Solidification and crystallographic texture modeling of laser powder bed fusion Ti-6Al-4V using finite difference-Monte Carlo method10.1016/j.mtla.2024.102279 ↗
68. Simulation of powder bed metal additive manufacturing microstructures with coupled finite difference-Monte Carlo method
69. Adams · 2015
70. Soylemez · 2018
71. Deep learning10.1038/nature14539 ↗
72. Scientific machine learning through physics–informed neural networks: Where we are and what’s next10.1007/s10915-022-01939-z ↗
73. Kokkos: Enabling performance portability across manycore architectures
74. Kokkos: Enabling manycore performance portability through polymorphic memory access patterns10.1016/j.jpdc.2014.07.003 ↗
75. Kokkos 3: Programming model extensions for the exascale era10.1109/TPDS.2021.3097283 ↗
76. The kokkos ecosystem: Comprehensive performance portability for high performance computing10.1109/MCSE.2021.3098509 ↗
77. Iyer · 2019
78. Part-scale microstructure prediction for laser powder bed fusion Ti-6Al-4V using a hybrid mechanistic and machine learning model
79. Direct numerical simulations in solid mechanics for understanding the macroscale effects of microscale material variability10.1016/j.cma.2015.01.017 ↗
80. Direct numerical simulations in solid mechanics for quantifying the macroscale effects of microstructure and material model-form error10.1007/s11837-016-1857-6 ↗
81. Direct numerical simulation of mechanical response in synthetic additively manufactured microstructures10.1088/1361-651X/aac616 ↗
82. Modeling mechanical behavior of an additively manufactured metal structure with local texture variations: A study on model form error10.1088/1361-651X/aaf107 ↗
83. Automatic finite element meshing of planar Voronoi tessellations10.1016/S0013-7944(01)00124-2 ↗