Machine learning-guided design and development of metallic structural materials

Yu, Jinxin; Xi, Shengkun; Pan, Shangke; Wang, Yongjie; Peng, Qinghua; Shi, Rongpei; Wang, Cuiping; Liu, Xingjun

doi:10.20517/jmi.2021.08

Cited by 12 publications

(8 citation statements)

References 47 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, it is very challenging to obtain an alloy with desired properties by a "trial and error" method due to the complexity of chemical compositions and phases. Compared with the experimental method, machine learning (ML) provides a new approach to accelerating the discovery of new materials by building the relationship between targeted properties and various materials descriptors [104][105][106][107][108] . Until now, many works have been conducted to predict the possible phases in HEAs using ML algorithms, including logistic regression, random forest, decision tree, K-nearest neighbor, support vector machine and artificial neural network (ANN) approaches [109][110][111] .…”

Section: Machine Learning For Alloy Design and Ammentioning

confidence: 99%

New trends in additive manufacturing of high-entropy alloys and alloy design by machine learning: from single-phase to multiphase systems

Zhou¹,

Zhang²,

Wang³

et al. 2022

J Mater Inf

View full text Add to dashboard Cite

Alloys with excellent properties are always in significant demand for meeting the severe conditions of industrial applications. However, the design strategies of traditional alloys based on a single principal element have reached their limits in terms of property optimization. The concept of high-entropy alloys (HEAs) provides a new design strategy based on multicomponent elements, which may overcome the bottleneck problems that exist in traditional alloys. To further maximize the capability of HEAs, a novel additive manufacturing (AM) technique has been utilized to produce HEA components with the desired structures and properties. This review considers a new trend in the AM of HEAs, i.e., from the AM of single-phase HEAs to multiphase HEAs. Although most as-printed single-phase HEAs show superior tensile properties to as-cast ones, their strength is still not satisfactory, especially at elevated temperatures. Thus, multiphase HEAs are developed by introducing hard second phases, such as L12, BCC, carbides, oxides, nitrides, and so on. These phases can be introduced to the matrix using in situ alloying during AM or the subsequent heat treatment. Dislocation strengthening is considered as the main reason for improving the tensile properties of as-printed single-phase HEAs. In contrast, multiple strengthening and toughening mechanisms occur in as-printed multiphase HEAs, which can synergistically enhance their mechanical properties. Furthermore, machine learning provides an effective method to design new alloys with the desired properties and predict the optimal AM parameters for the designed alloys without tedious experiments. The synergistic combination of machine learning and AM will significantly speed up scientific advances and promote industrial applications.

show abstract

Section: Machine Learning For Alloy Design and Ammentioning

confidence: 99%

New trends in additive manufacturing of high-entropy alloys and alloy design by machine learning: from single-phase to multiphase systems

Zhou¹,

Zhang²,

Wang³

et al. 2022

J Mater Inf

View full text Add to dashboard Cite

show abstract

“…Alternatively, machine learning can be used in alloy design, usually by using a large experimental input dataset. (Un-)desirable quantities can then be related to the composition of the alloy and a predictive tool relating composition to properties can be derived [5]. In [6], a large database of superalloys is used to predict the formation of phases based on physical quantities that are calculated for the given alloy composition (melting temperature, the atomic radius difference, the structural entropy).…”

Section: Introductionmentioning

confidence: 99%

Understanding element solution energies in nickelbase alloys using machine learning

Bäker

2023

Mater. Res. Express

View full text Add to dashboard Cite

The design of nickelbase superalloys requires to tune the content of diﬀerent phases with a composition of Ni3A, either to strengthen the alloy (γ′-phase, Ni3Al, γ′′-phase, Ni3Nb) or to inﬂuence its grain size and avoid embrittlement (δ-phase, Ni3Nb, η-phase, Ni3Ti). Here, we use a machine-learning-inspired approach to understand the inﬂuence of elemental properties on the energy of an alloying element in the respective phases. It is shown that the energy in γ′′, δ, and η can be ﬁtted well using the Bader charge and the volume of the element in a nickel matrix. In the case of γ′, the small lattice mismatch requires a ﬁt not involving the volume, but the bond order instead. We also show that the frequently used Md-parameter can be predicted from the properties of an element in a pure nickel matrix. Finally, the physical basis of the results is discussed in detail.

show abstract

“…Many nickel-based and other superalloy developments are guided by high-throughput experiments [17][18][19] . Recently, researchers have explored machine learning strategies, such as active learning strategies, to iteratively conduct experiments facilitating the exploration of the search space [20][21][22] . To maximize the full potential of AM, alloy development for AM processing requires additional attention so that resultant AM parts meet desired specifications over a broad processing landscape.…”

Section: Introductionmentioning

confidence: 99%

Linking processing parameters with melt pool properties of multiple nickel-based superalloys via high-dimensional Gaussian process regression

Menon¹,

Mondal²,

Basak³

2023

J Mater Inf

View full text Add to dashboard Cite

A physics-based model is used to predict the melt pool properties in the laser-directed energy deposition of several nickel-based superalloys for different process parameters. The input space is high-dimensional, consisting of a common 19-dimensional composition space for each alloy and the process parameters (laser power and scan velocity). Gaussian Process-based regression frameworks are developed by training surrogates on data generated by a validated analytical model. These surrogates are thereafter used to predict and define relationships between the composition, resultant thermophysical properties, process parameters, and the subsequent melt pool property. The probabilistic predictions are augmented by uncertainty quantification and sensitivity analysis to substantiate the findings further.

show abstract

Machine learning-guided design and development of metallic structural materials

Cited by 12 publications

References 47 publications

New trends in additive manufacturing of high-entropy alloys and alloy design by machine learning: from single-phase to multiphase systems

New trends in additive manufacturing of high-entropy alloys and alloy design by machine learning: from single-phase to multiphase systems

Understanding element solution energies in nickelbase alloys using machine learning

Linking processing parameters with melt pool properties of multiple nickel-based superalloys via high-dimensional Gaussian process regression

Contact Info

Product

Resources

About