Machine-learning assisted laser powder bed fusion process optimization for AlSi10Mg: New microstructure description indices and fracture mechanisms

Liu, Qian; Wu, Hongkun; Paul, Moses J.; He, Peidong; Peng, Zhongxiao; Gludovatz, Bernd; Kruzic, Jamie J.; Wang, Chun H.; Li, Xiao Peng

doi:10.1016/j.actamat.2020.10.010

Cited by 189 publications

(34 citation statements)

References 75 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It is worth noting that machine learning-based process optimization approaches have also been reported in recent years [60][61][62]. However, since different MPEAs have distinct physical and chemical characteristics and, thus, different responses to laser and electron beam, carefully designed experiments, physics-based computational modeling, and data-driving machine learning approaches should be combined and wisely integrated to obtain the optimized PBF process window for specific MPEAs [63].…”

Section: Powder Bed Fusion (Pbf)mentioning

confidence: 99%

“…Firstly, it can enable the exploration of new MPEAs and the characterization of their solid solution structure and the evaluation of their AM processability by analyzing their physical and chemical properties [178][179][180]. Secondly, large optimal processing windows for AM of MPEAs can be established by machine learning methods which will allow the tailoring of their microstructures and properties, which has been recently demonstrated for conventional alloys [62,63]. Thirdly, exploring the relationships between the alloy composition, processing, microstructure, and properties for AM MPEAs would generally require a huge number of experiments, which poses a significant challenge in terms of needed time and resources.…”

Section: Machine Learning For Am Of Mpeasmentioning

confidence: 99%

See 1 more Smart Citation

Review: Multi-principal element alloys by additive manufacturing

Ferry

Kruzic

et al. 2022

J Mater Sci

Self Cite

View full text Add to dashboard Cite

Multi-principal element alloys (MPEAs) have attracted rapidly growing attention from both research institutions and industry due to their unique microstructures and outstanding physical and chemical properties. However, the fabrication of MPEAs with desired microstructures and properties using conventional manufacturing techniques (e.g., casting) is still challenging. With the recent emergence of additive manufacturing (AM) techniques, the fabrication of MPEAs with locally tailorable microstructures and excellent mechanical properties has become possible. Therefore, it is of paramount importance to understand the key aspects of the AM processes that influence the microstructural features of AM fabricated MPEAs including porosity, anisotropy, and heterogeneity, as well as the corresponding impact on the properties. As such, this review will first present the state-of-the-art in existing AM techniques to process MPEAs. This is followed by a discussion of the microstructural features, mechanisms of microstructural evolution, and the mechanical properties of the AM fabricated MPEAs. Finally, the current challenges and future research directions are summarized with the aim to promote the further development and implementation of AM for processing MPEAs for future industrial applications.

show abstract

Section: Powder Bed Fusion (Pbf)mentioning

confidence: 99%

Section: Machine Learning For Am Of Mpeasmentioning

confidence: 99%

Review: Multi-principal element alloys by additive manufacturing

Ferry

Kruzic

et al. 2022

J Mater Sci

Self Cite

View full text Add to dashboard Cite

show abstract

“…Hence, we briefly summarize related work on machine learning methods (e.g., neural networks, Support Vector Machines, k-means clustering, random forests, generative networks, and more) applied to several diverse challenges in molecular and materials science fields. In particular, active research areas for ML in materials science include (but are not limited to): accelerated materials design and property prediction [40][41][42][43][44][45][46], process optimization [47,48], discovery of structure-property relationships [49,50], construction of potential energy surfaces for molecular dynamics simulations [51][52][53], prediction of atomic scale properties [54], text mining for knowledge extraction [55], microstructure and materials characterization [10,11,25,30,[56][57][58], and generation of synthetic microstructure images [31]. Such applications span multiple length scales and a variety of material systems (metals and alloys, oxides, polymers) [59].…”

Section: Related Workmentioning

confidence: 99%

The Adoption of Image-Driven Machine Learning for Microstructure Characterization and Materials Design: A Perspective

Baskaran¹,

Kautz²,

Chowdhary³

et al. 2021

Preprint

View full text Add to dashboard Cite

The recent surge in the adoption of machine learning techniques for materials design, discovery, and characterization has resulted in an increased interest and application of Image Driven Machine Learning (IDML) approaches. In this work, we review the application of IDML to the field of materials characterization. A hierarchy of six action steps is defined which compartmentalizes a problem statement into well-defined modules. The studies reviewed in this work are analyzed through the decisions adopted by them at each of these steps. Such a review permits a granular assessment of the field, for example the impact of IDML on materials characterization at the nanoscale, the number of images in a typical dataset required to train a semantic segmentation model on electron microscopy images, the prevalence of transfer learning in the domain, etc. Finally, we discuss the importance of interpretability and explainability, and provide an overview of two emerging techniques in the field: semantic segmentation and generative adversarial networks.

show abstract

“…Hence, AM technology has been expected as an ideal method to fabricate parts with sophisticated shape [6][7][8][9]. With the rapid development of AM technology, more and more methods have sprung up, such as Selective Laser Sintering (SLS), Selective Laser Melting (SLM), Three-dimensional Printing (3DP), Direct Ink Writing (DIW), Digital Light Processing (DLP) and so on [10][11][12]. SLM and SLS are always used to produce metal.…”

Section: Introductionmentioning

confidence: 99%

“…SLM and SLS are always used to produce metal. Owing to the high melting temperature of ceramic, and the inherent high thermal gradient and residual thermal stress of the laser-based AM technologies, SLM and SLS are unsuitable for ceramic [11,13]. 3DP and DIW are the most widely applicable AM technologies, and it can be used in all material system, including metal, ceramic and polymer.…”

Section: Introductionmentioning

confidence: 99%

Digital Light Processing 3D Printing of Surface-oxidized Si3N4 Coated by Silane Coupling Agent

Yang

Sun

Huang

et al. 2021

Preprint

View full text Add to dashboard Cite

Due to high light absorption and high refractive index of silicon nitride (Si3N4) ceramic, it is difficult to prepare Si3N4 slurry with favourable curing ability, high solid loading and low viscosity at the same time, and thus the high quality Si3N4 parts are hard to fabricate via Digital Light Processing (DLP). In this paper, oxidation process was used to enhance the curing behavior of Si3N4 slurry, and then silane coupling agent (KH560) was used to improve the rheological properties of the oxidized Si3N4 slurry. The effect of Si3N4 slurry with oxidation and modification on its rheological behavior, light absorption, curing ability and stability have been systematically investigated. A Si3N4 slurry with abundant photocuring ability, high solid loading, low viscosity and reliable stability was fabricated by the oxidized (1h) and KH560 (1wt.%) modified Si3N4 powders, and the dense Si3N4 ceramic parts were produced by this slurry via DLP 3D printing. Subsequently, the influence of oxidation and modification on microstructure, mechanical properties and thermal conductivity have been investigated. Finally, the performances of the DLP 3D printed samples are close to the isostatic cool pressing (ICP) fabricated samples. Consequently, DLP has well potential to fabricate high-performance Si3N4 ceramics.

show abstract

Machine-learning assisted laser powder bed fusion process optimization for AlSi10Mg: New microstructure description indices and fracture mechanisms

Cited by 189 publications

References 75 publications

Review: Multi-principal element alloys by additive manufacturing

Review: Multi-principal element alloys by additive manufacturing

The Adoption of Image-Driven Machine Learning for Microstructure Characterization and Materials Design: A Perspective

Digital Light Processing 3D Printing of Surface-oxidized Si3N4 Coated by Silane Coupling Agent

Contact Info

Product

Resources

About