Reconstruction of Microstructural and Morphological Parameters for RVE Simulations with Machine Learning

Pütz, Felix; Henrich, Manuel; Roth, Andreas; Könemann, Markus; Münstermann, Sebastian

doi:10.1016/j.promfg.2020.04.193

Cited by 11 publications

(5 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In Figure 3, the aspect ratio of 1 corresponds to almost perfectly round shapes, whereas the higher aspect ratio corresponds to the adequately elongated grain or inclusion. Considering this criterion, the aspect ratio of ferrite grains in Figure 3a predominantly lies between 1-2, except for some grains with a high degree of elongation of >4, which is the normal ferrite microstructure reported in previous literature [42,43]. On the other hand, the inclusions in Figure 3b-d are largely round (aspect ratio = 1), with some particles slightly elongated.…”

Section: Statistical Analysis Of the Ebsd Datasupporting

confidence: 64%

Influence of Non-Metallic Inclusions on Local Deformation and Damage Behavior of Modified 16MnCrS5 Steel

et al. 2022

View full text Add to dashboard Cite

This work investigates a ferrite matrix with multiple non-metallic inclusions to evaluate their influence on the global and local deformation and damage behavior of modified 16MnCrS5 steel. For this purpose, appropriate specimens are prepared and polished. The EBSD technique is used to record local phase and orientation data, then analyze and identify the size and type of inclusions present in the material. The EBSD data are then used to run full phase crystal plasticity simulations using DAMASK-calibrated material model parameters. The qualitative and quantitative analysis of these full phase simulations provides a detailed insight into how the distribution of non-metallic inclusions within the ferrite matrix affects the local stress, strain, and damage behavior. In situ tensile tests are carried out on specially prepared miniature dog-bone-shaped notched specimens in ZEISS Gemini 450 scanning electron microscope with a Kammrath and Weiss tensile test stage. By adopting an optimized scheme, tensile tests are carried out, and local images around one large and several small MnS inclusions are taken at incremental strain values. These images are then processed using VEDDAC, a digital image correlation-based microstrain measurement tool. The damage initiation around several inclusions is recorded during the in situ tensile tests, and damage initiation, propagation, and strain localization are analyzed. The experimental results validate the simulation outcomes, providing deeper insight into the experimentally observed trends.

show abstract

Section: Statistical Analysis Of the Ebsd Datasupporting

confidence: 64%

Influence of Non-Metallic Inclusions on Local Deformation and Damage Behavior of Modified 16MnCrS5 Steel

et al. 2022

View full text Add to dashboard Cite

show abstract

“…Due to the poor resolution of tomographic images, 3D X-ray diffraction measurements were defined as ground truth during the data training. Pütz et al [ 32 ] investigated advanced high-strength steels by using representative volume elements with periodic material structures using virtual (artificial) microstructures. Since the input parameters are the most critical part of generating the RVEs, an ML algorithm was trained to reproduce input data parameters equivalent to the real microstructural and morphological parameters.…”

Section: Machine Learning Applied For Image Segmentationmentioning

confidence: 99%

Many-Scale Investigations of the Deformation Behavior of Polycrystalline Composites: I—Machine Learning Applied for Image Segmentation

Schneider

Prabhu

Hoess³

et al. 2022

Materials

View full text Add to dashboard Cite

Our work investigates the polycrystalline composite deformation behavior through multiscale simulations with experimental data at hand. Since deformation mechanisms on the micro-level link the ones on the macro-level and the nanoscale, it is preferable to perform micromechanical finite element simulations based on real microstructures. The image segmentation is a necessary step for the meshing. Our 2D EBSD images contain at least a few hundred grains. Machine learning (ML) was adopted to automatically identify subregions, i.e., individual grains, to improve local feature extraction efficiency and accuracy. Denoising in preprocessing and postprocessing before and after ML, respectively, is beneficial in high quality feature identification. The ML algorithms used were self-developed with the usage of inherent code packages (Python). The performances of the three supervised ML models—decision tree, random forest, and support vector machine—are compared herein; the latter two achieved accuracies of up to 99.8%. Calculations took about 0.5 h from the original input dataset (EBSD image) to the final output (segmented image) running on a personal computer (CPU: 3.6 GHz). For a realizable manual pixel sortation, the original image was firstly scaled from the initial resolution 10802 pixels down to 3002. After ML, some manual work was necessary due to the remaining noises to achieve the final image status ready for meshing. The ML process, including this manual work time, improved efficiency by a factor of about 24 compared to a purely manual process. Simultaneously, ML minimized the geometrical deviation between the identified and original features, since it used the original resolution. For serial work, the time efficiency would be enhanced multiplicatively.

show abstract

“… The electron backscatter diffraction (EBSD) of steel DP800 where the x-axis is the rolling direction (RD) and the y-axis is the sheet normal (SN) [ 20 ]. …”

Section: Figurementioning

confidence: 99%

Generating Input Data for Microstructure Modelling: A Deep Learning Approach Using Generative Adversarial Networks

et al. 2020

Self Cite

View full text Add to dashboard Cite

For the generation of representative volume elements a statistical description of the relevant parameters is necessary. These parameters usually describe the geometric structure of a single grain. Commonly, parameters like area, aspect ratio, and slope of the grain axis relative to the rolling direction are applied. However, usually simple distribution functions like log normal or gamma distribution are used. Yet, these do not take the interdependencies between the microstructural parameters into account. To fully describe any metallic microstructure though, these interdependencies between the singular parameters need to be accounted for. To accomplish this representation, a machine learning approach was applied in this study. By implementing a Wasserstein generative adversarial network, the distribution, as well as the interdependencies could accurately be described. A validation scheme was applied to verify the excellent match between microstructure input data and synthetically generated output data.

show abstract

Reconstruction of Microstructural and Morphological Parameters for RVE Simulations with Machine Learning

Cited by 11 publications

References 19 publications

Influence of Non-Metallic Inclusions on Local Deformation and Damage Behavior of Modified 16MnCrS5 Steel

Influence of Non-Metallic Inclusions on Local Deformation and Damage Behavior of Modified 16MnCrS5 Steel

Many-Scale Investigations of the Deformation Behavior of Polycrystalline Composites: I—Machine Learning Applied for Image Segmentation

Generating Input Data for Microstructure Modelling: A Deep Learning Approach Using Generative Adversarial Networks

Contact Info

Product

Resources

About