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 | Li, Bowen | - |
| dc.identifier.uri | https://theses.lib.polyu.edu.hk/handle/200/11521 | - |
| dc.language | English | en_US |
| dc.publisher | Hong Kong Polytechnic University | en_US |
| dc.rights | All rights reserved | en_US |
| dc.title | Machine learning approach for predicting the tensile strength of high entropy alloys | en_US |
| dcterms.abstract | Researchers showed great passion for high entropy alloys (HEAs) in the past decade for its broad compositional space, relatively simple phase structure and outstanding properties. Like most metal materials, finding a way to promote strength while maintaining satisfactory plasticity has attracted great concerns. Composition design induced multiple toughening mechanisms, combined with appropriate processing is promising. Meanwhile, an advanced data-driven machine learning (ML) assisted materials design strategy is developing and show great potentials. In this research, a ML framework was performed to predict the tensile strength (UTS) of HEAs with purpose of rapid search and lead of next-generation experimental alloys. The ML models were constructed based on datasets with 261 samples from HEAs experiments, and 57 related feature candidates were presented. A three-step feature selection were adopted after carefully considering the multiple validation methods and algorithms available. Redundant and irrelevant features were eliminated by CFS based RReliefF and MDI features importance ranking. A wrapper method finally selected the best feature subsets for datasets with processing attributes and as cast HEAs on two separate models, namely the Support Vector Machine with puk kernel and Random Forest Regression. Sets of better hyperparameters were given according to the grid search. Finally, the outperform model was applied to guide the composition design and heat treatment process of HEAs, demonstrated great predictive capacity. Also, challenges in building more robust and applicable model, and understanding the connection between the data, the model and the performance expressed were proposed. | en_US |
| dcterms.extent | xii, 121 pages : color illustrations | en_US |
| dcterms.isPartOf | PolyU Electronic Theses | en_US |
| dcterms.issued | 2021 | en_US |
| dcterms.educationalLevel | M.Sc. | en_US |
| dcterms.educationalLevel | All Master | en_US |
| dcterms.LCSH | Alloys -- Testing | en_US |
| dcterms.LCSH | Machine learning | en_US |
| dcterms.LCSH | Hong Kong Polytechnic University -- Dissertations | en_US |
| dcterms.accessRights | restricted access | en_US |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| 5933.pdf | For All Users (off-campus access for PolyU Staff & Students only) | 18.75 MB | Adobe PDF | View/Open |
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/200/11521