|Title:||Machine learning approach for predicting the elastic properties and glass forming ability of metallic glasses|
|Advisors:||Shi, S. Q. (ME)|
|Subject:||Hong Kong Polytechnic University -- Dissertations|
|Department:||Faculty of Engineering|
|Pages:||xv, 96 pages : color illustrations|
|Abstract:||In recent years, bulk metallic glasses (BMGs) have attracted more and more attentions because of their unique structural features, outstanding mechanical properties, and potential applications in various fields. However, it may take a long time to discover and optimize a potential glassy material via the traditional material design paradigms, due to the lack of underlying physics. In the past five years, the paradigm of "materials informatics" which is based on statistical algorithms and artificial intelligence has become popular in material design. The methodology of materials informatics is introduced in this research with the hope that the machine learning approach can accelerate the design and discovery of BMGs with desired mechanical properties and good glass forming ability (GFA). In this research, a data set of 219 BMGs with measured elastic properties (bulk modulus (K) and shear modulus (G)) was collected and trained through the machine learning (ML) approach for predicting the properties of BMGs. The selection of features was made with linear least square models. Among these features, four types of atomic radius were examined by the support vector machine regression (SVR) algorithms with radial basis function. Then, the best type of atomic radius was chosen by leave-one-out cross-validation. Moreover, the SVR algorithms with Pearson VII universal kernels were employed to optimize the ML models with the best atomic radius together with selected features. Based on the final ML models, predictions on K and G values were made, which showed a good agreement with most of experimental data. The ML models have also identified some errors reported in the literature. And the final ML models were also used to design BMGs with good GFA. This work indicates the great potential of ML in advanced materials design with target properties.|
|Rights:||All rights reserved|
Files in This Item:
|991022159053603411.pdf||For All Users (off-campus access for PolyU Staff & Students only)||2.93 MB||Adobe PDF||View/Open|
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: