Unsupervised machine learning in fractography: Evaluation and interpretation

Tsopanidis, Stylianos; Osovski, S.

doi:10.1016/j.matchar.2021.111551

Cited by 15 publications

(4 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…While there are also some compelling approaches to microstructure segmentation using unsupervised learning [58][59][60][61][62], supervised learning is used for the most part. This makes sense, because complex microstructures in particular require the algorithm to learn more complex concepts that can no longer be distinguished by means of unsupervised learning.…”

Section: Overview Of ML Applications In Microstructure Analysismentioning

confidence: 99%

Overview: Machine Learning for Segmentation and Classification of Complex Steel Microstructures

Müller,

Stiefel,

Bachmann

et al. 2024

Metals

View full text Add to dashboard Cite

The foundation of materials science and engineering is the establishment of process–microstructure–property links, which in turn form the basis for materials and process development and optimization. At the heart of this is the characterization and quantification of the material’s microstructure. To date, microstructure quantification has traditionally involved a human deciding what to measure and included labor-intensive manual evaluation. Recent advancements in artificial intelligence (AI) and machine learning (ML) offer exciting new approaches to microstructural quantification, especially classification and semantic segmentation. This promises many benefits, most notably objective, reproducible, and automated analysis, but also quantification of complex microstructures that has not been possible with prior approaches. This review provides an overview of ML applications for microstructure analysis, using complex steel microstructures as examples. Special emphasis is placed on the quantity, quality, and variance of training data, as well as where the ground truth needed for ML comes from, which is usually not sufficiently discussed in the literature. In this context, correlative microscopy plays a key role, as it enables a comprehensive and scale-bridging characterization of complex microstructures, which is necessary to provide an objective and well-founded ground truth and ultimately to implement ML-based approaches.

show abstract

Section: Overview Of ML Applications In Microstructure Analysismentioning

confidence: 99%

Overview: Machine Learning for Segmentation and Classification of Complex Steel Microstructures

Müller,

Stiefel,

Bachmann

et al. 2024

Metals

View full text Add to dashboard Cite

show abstract

“…They were able to achieve a total mean Intersection over Union (IoU) that was > 91% for two different materials. In [42], unsupervised learning [43] was employed on a clustering algorithm in order to classify SEM images based on tungsten composition achieving accuracy that was greater than 90%.…”

Section: Related Workmentioning

confidence: 99%

Transfer Learning Methods for Fractographic Detection of Fatigue Crack Initiation in Additive Manufacturing

Anidjar,

Lang,

Mega

2024

IEEE Access

View full text Add to dashboard Cite

In recent times, there has been a growing surge of interest in additive manufacturing (AM) due to its ability to bring about cost savings and weight reduction in fabricated components. Nevertheless, AM materials are susceptible to imperfections that can severely undermine their ability to withstand fatigue. Consequently, before incorporating these materials into vital structural components, it is imperative to possess a thorough comprehension of their fatigue properties. This necessitates the execution of fatigue tests and manual examination of fracture surfaces. Within this research, we put forward a novel approach involving the development of a machine-learning model tailored to identify the starting point of fatigue cracks within specimens of Titanium Ti-6Al-4V, which were produced using selective laser melting (SLM). Moreover, this model also quantifies the distance of these initiation sites from the material's surface. The proposed method encompasses the segmentation of image sections that are devoid of initiation sites, succeeded by the detection of these sites within the remaining segments. Subsequently, established computer vision techniques are harnessed to compute the distance from the surface. The outcomes of this study underscore a remarkable potential for automating the process of fractographic analysis through the integration of machine learning and computer vision models. These strides in technology hold the promise of streamlining and expediting the fractographic analysis procedure, ushering in a new era of efficiency.

show abstract

“…So far, many algorithms have been employed to analyze plenty of data in materials science. [67][68][69][70] Among these algorithms, the frequency distribution statistics and unsupervised machine learning have been proved to exhibit outstanding accuracy and efficiency in narrowing the design ranges of alloys. In this work, the performances of these two algorithms were compared by selecting the composition ranges with superior microstructural stability and creep resistance, respectively.…”

Section: Unsupervised Machine Learning-assisted Alloy Designmentioning

confidence: 99%

High‐Throughput Method–Accelerated Design of Ni‐Based Superalloys

Liu

Wang

Wang³

et al. 2022

Adv Funct Materials

View full text Add to dashboard Cite

Ever-increasing demands for superior alloys with improved hightemperature service properties require accurate design of their composition. However, conventional approaches to screen the properties of alloys such as creep resistance and microstructural stability cost a lot of time and resources. This work therefore proposes a novel high throughput-based design strategy for high-temperature alloys to accelerate their composition selections, by taking Ni-based superalloys as an example. A numerical inverse method is used to massively calculate the multielement diffusion coefficients based on an accurate atomic mobility database. These coefficients are subsequently employed to refine the physical models for tuning the creep rates and structural stability of alloys, followed by unsupervised machine learning to categorize their composition and determine the range of the composition with optimal performance. By using a strict screening criterion, two sets of composition with comprehensively optimal properties are selected, which is then validated by experiments. Compared with recent data-driven methods for materials design, this strategy exhibits high accuracy and efficiency attributed to the high-throughput multicomponent diffusion couples, self-developed atomic mobility database, and refined physical models. Since this strategy is independent of the alloy composition, it can efficiently accelerate the development of multicomponent high-performance alloys and tackle challenges in discovering novel materials.

show abstract

Unsupervised machine learning in fractography: Evaluation and interpretation

Cited by 15 publications

References 22 publications

Overview: Machine Learning for Segmentation and Classification of Complex Steel Microstructures

Overview: Machine Learning for Segmentation and Classification of Complex Steel Microstructures

Transfer Learning Methods for Fractographic Detection of Fatigue Crack Initiation in Additive Manufacturing

High‐Throughput Method–Accelerated Design of Ni‐Based Superalloys

Contact Info

Product

Resources

About