An efficient algorithm for generating diverse microstructure sets and delineating properties closures

Johnson, Oliver K.; Kurniawan, Christian

doi:10.1016/j.actamat.2018.01.004

Cited by 12 publications

(3 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Regarding homogenized microstructural features, in Iraki et al (2021), Latin hypercube design is used to generate a data set of textures for cold rolled steel sheets. Even special sampling heuristics have been developed for generating sets of microstructure features, like in Johnson and Kurniawan, 2018. Also, adaptive sampling techniques are used in materials design, however, in the sense of an optimization aiming to identify microstructures with targeted properties. In Liu et al (2015a) and Paul et al (2019), specific machine learning-based optimization approaches are presented that efficiently guide sampling to regions in the space of microstructures, where microstructures with desired properties are expected to be located.…”

Section: Related Workmentioning

confidence: 99%

Efficient Exploration of Microstructure-Property Spaces via Active Learning

et al. 2022

View full text Add to dashboard Cite

In materials design, supervised learning plays an important role for optimization and inverse modeling of microstructure-property relations. To successfully apply supervised learning models, it is essential to train them on suitable data. Here, suitable means that the data covers the microstructure and property space sufficiently and, especially for optimization and inverse modeling, that the property space is explored broadly. For virtual materials design, typically data is generated by numerical simulations, which implies that data pairs can be sampled on demand at arbitrary locations in microstructure space. However, exploring the space of properties remains challenging. To tackle this problem, interactive learning techniques known as active learning can be applied. The present work is the first that investigates the applicability of the active learning strategy query-by-committee for an efficient property space exploration. Furthermore, an extension to active learning strategies is described, which prevents from exploring regions with properties out of scope (i.e., properties that are physically not meaningful or not reachable by manufacturing processes).

show abstract

Section: Related Workmentioning

confidence: 99%

Efficient Exploration of Microstructure-Property Spaces via Active Learning

et al. 2022

View full text Add to dashboard Cite

show abstract

“…Generation methods and application of digital microstructure in metals are reviewed in 2 . Synthetic microstructure was introduced and successfully implemented in several studies [3][4][5][6][7][8][9] . In our study we present an extension of the existing methodology by constructing a representative volume element (RVE) that encompasses pertinent features such as texture information, grain morphology, and an explicit passive layer.…”

Section: Introductionmentioning

confidence: 99%

Multiparadigm modeling of texture effects on pitting corrosion in ferritic stainless steel

Jamebozorgi

Rasim

Schröder

2023

Preprint

View full text Add to dashboard Cite

Texture has long been recognized as a critical factor influencing the physical processes and properties of condensed matter. In this study, we implemented a multiparadigm approach and introduce a robust methodology to investigate the effects of texture in corrosion, which overcomes the common limitations of quantum-based computational approaches in terms of time and system size. Our approach provides the same level of accuracy as atomistic calculations but with significantly less computational cost. The methodology, based on the finite element method (FEM), employs 3D digital representations of polycrystalline microstructures. As a proof of concept, we apply our approach to the case of pitting corrosion in ferritic stainless steel. As shown in the literature irregular pit growth patterns through pitting corrosion are primarily caused by texture. Our study reveals that texture has a significant impact on the pitting corrosion rate leading to a wide range of irregular pit growth patterns in polycrystals. Our findings are supported by atomistic calculations and experimental literature, demonstrating the validity of our approach.

show abstract

“…We have recently devised an alternate method called hierarchical simplex sampling (HSS) [38] that leverages the geometric structure of microstructure hulls in the Dirac basis to efficiently generate samples in both the interior and covering the surface of the microstructure hull in such a way that the resulting points span the properties closure. Figure 2 illustrates the improved performance of HSS for a two-dimensional properties closure consisting of the texture sensitive properties for the present design problem.…”

mentioning

confidence: 99%