Full metadata record
| DC Field | Value | Language |
|---|---|---|
| dc.contributor | Department of Mechanical Engineering | en_US |
| dc.contributor.advisor | Shi, S. Q. (ME) | en_US |
| dc.creator | Xiong, Jie | - |
| dc.identifier.uri | https://theses.lib.polyu.edu.hk/handle/200/11128 | - |
| dc.language | English | en_US |
| dc.publisher | Hong Kong Polytechnic University | en_US |
| dc.rights | All rights reserved | en_US |
| dc.title | Materials informatics and its application to metallic materials design and discovery | en_US |
| dcterms.abstract | There is a pressing need to shorten the development period for new materials. However, the lack of predictive tools for uncovering the relationships between compositions and desired properties limits the further application of these materials. In this study, machine learning (ML) approaches were developed and applied to predict and understand the properties of steels, metallic glasses, and complex concentrated alloys. Three-hundred and sixty data samples on four mechanical properties of steels—fatigue strength, tensile strength, fracture strength, and hardness—were selected from the Japan National Institute of Material Science (NIMS) database, comprising data on carbon steels and low-alloy steels. Five ML algorithms were utilized to predict the mechanical properties of the three-hundred and sixty samples, and random forest regression (RFR) showed the best predictive performance. Feature selection conducted by RFR and symbolic regressions (SR) revealed the four most important features that most influence the mechanical properties of steels: the tempering temperature of steel, and the alloying elements of carbon, chromium and molybdenum. Mathematical expressions were generated via SR, and the expressions explicitly predicted how each of the four mechanical properties varied quantitatively with the four most important features. | en_US |
| dcterms.abstract | A dataset consisting of 6,470 unique alloy compositions and 5,933 critical casting diameters of alloys were collected from publicly available literature. The random forest (RF) algorithm and a three-step feature selection method were employed to predict and describe the glass-forming ability (GFA) and critical casting diameters of alloys, the predictions of unseen data were found to be in good agreement with most experimental values. In addition, 669 data on critical transformation temperatures were collected, the RFR models performed much better than conventional GFA criteria. A further 278 data samples on elastic properties of bulk metallic glasses (BMGs) were collected as well, and the built RFR models performed well with correlation coefficients greater than 0.98. Using symbolic regressions, we determined that an alloy with a large critical casting diameter would likely have a high mixing entropy, a high thermal conductivity, and a mixing enthalpy of approximately -28 kJ/mol, and that a BMG with a small average atomic volume would likely have a high elastic modulus. A total of 557 samples of various chemical compositions, comprising 61 amorphous, 167 single-phase crystalline, and 329 multi-phases crystalline complex concentrated alloys (CCAs) were collected. Three classification models are developed with high accuracies to categorize and understand the formed phases of CCAs. Two regression models are also constructed to predict the hardness and ultimate tensile strength of CCAs, and the correlation coefficient of the RFR model is greater than 0.9 for both of two targeted properties. Furthermore, the Shapley additive explanation (SHAP) values are calculated, and accordingly the four most important features are identified. A significant finding from the SHAP values is that there exists a critical value in each of the top four features, which provides an easy and fast assessment in the design of improved mechanical properties of CCAs. This research displays the efficacy of ML in furnishing a mechanistic understanding and enabling the prediction of properties of metallic materials. The present work demonstrates the great potential of ML algorithms such as RF, artificial neural network (ANN), and SR in the design of advanced metallic materials. | en_US |
| dcterms.extent | xxvi, 196 pages : color illustrations | en_US |
| dcterms.isPartOf | PolyU Electronic Theses | en_US |
| dcterms.issued | 2021 | en_US |
| dcterms.educationalLevel | Ph.D. | en_US |
| dcterms.educationalLevel | All Doctorate | en_US |
| dcterms.LCSH | Materials -- Mathematical models | en_US |
| dcterms.LCSH | Materials -- Data processing | en_US |
| dcterms.LCSH | Machine learning | en_US |
| dcterms.LCSH | Hong Kong Polytechnic University -- Dissertations | en_US |
| dcterms.accessRights | open access | en_US |
Copyright Undertaking
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item:
https://theses.lib.polyu.edu.hk/handle/200/11128