Engineer data augmentation practices for machine learning in composites
Composites are vital in various sectors, providing materials with properties that can be adjusted according to individual requirements. However, composites are increasingly more advanced and require special steps to manufacture. Due to their various uses, especially in specialized industries such as...
Saved in:
Main Author: | |
---|---|
Other Authors: | |
Format: | Final Year Project |
Language: | English |
Published: |
Nanyang Technological University
2021
|
Subjects: | |
Online Access: | https://hdl.handle.net/10356/153489 |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Institution: | Nanyang Technological University |
Language: | English |
id |
sg-ntu-dr.10356-153489 |
---|---|
record_format |
dspace |
spelling |
sg-ntu-dr.10356-1534892021-12-03T08:19:19Z Engineer data augmentation practices for machine learning in composites Tan, Jerome Whye Hean Sunil Chandrakant Joshi School of Mechanical and Aerospace Engineering MSCJoshi@ntu.edu.sg Engineering::Mechanical engineering Composites are vital in various sectors, providing materials with properties that can be adjusted according to individual requirements. However, composites are increasingly more advanced and require special steps to manufacture. Due to their various uses, especially in specialized industries such as aerospace and civil engineering, ample testing will be required to ensure the composites are suitable to be used without failure, giving rise to waste of resources given the special steps and materials required to manufacture the composite, hence new ways for prognostic modeling should be explored. In today’s world that is more data-driven than ever, data analysis techniques like machine learning are being used across many industries to improve efficiency and achieve results that are otherwise near impossible for traditional methods. However, to achieve accurate results through data analysis, large data is required to feed and train the model, which is a luxury that not all industry can produce. Data augmentations are used in such sectors to amplify the data within the constraints to allow for use of machine learning. This project report will first explore the newly discovered data augmentation technique, COMPOSITES, a 10-step technique that uses various concepts like mathematics and physics to apply constraints to amplify a dataset reasonably. The COMPOSITES process will be demonstrated on 2 case studies done on inter-ply fracture toughness (ILFT) of carbon nanotube (CNT)-engineered carbon fiber reinforced polymer (CFRP) composite laminates. The case studies will study the effect of varying contents of CNT on (I) Mode I fracture and (II) Mode I and Mode II mixed mode loading fracture. The first case study will serve as a form of validation for the process as well as a platform to delve further into the possible analysis and applications of the process. The second case study will be used to evaluate how well the COMPOSITES process can be applied on a frequently used performance measure in the industry. This report will then go on to discuss on the possible future areas for study to further develop this technique. Bachelor of Engineering (Mechanical Engineering) 2021-12-03T08:19:18Z 2021-12-03T08:19:18Z 2021 Final Year Project (FYP) Tan, J. W. H. (2021). Engineer data augmentation practices for machine learning in composites. Final Year Project (FYP), Nanyang Technological University, Singapore. https://hdl.handle.net/10356/153489 https://hdl.handle.net/10356/153489 en application/pdf Nanyang Technological University |
institution |
Nanyang Technological University |
building |
NTU Library |
continent |
Asia |
country |
Singapore Singapore |
content_provider |
NTU Library |
collection |
DR-NTU |
language |
English |
topic |
Engineering::Mechanical engineering |
spellingShingle |
Engineering::Mechanical engineering Tan, Jerome Whye Hean Engineer data augmentation practices for machine learning in composites |
description |
Composites are vital in various sectors, providing materials with properties that can be adjusted according to individual requirements. However, composites are increasingly more advanced and require special steps to manufacture. Due to their various uses, especially in specialized industries such as aerospace and civil engineering, ample testing will be required to ensure the composites are suitable to be used without failure, giving rise to waste of resources given the special steps and materials required to manufacture the composite, hence new ways for prognostic modeling should be explored.
In today’s world that is more data-driven than ever, data analysis techniques like machine learning are being used across many industries to improve efficiency and achieve results that are otherwise near impossible for traditional methods. However, to achieve accurate results through data analysis, large data is required to feed and train the model, which is a luxury that not all industry can produce. Data augmentations are used in such sectors to amplify the data within the constraints to allow for use of machine learning.
This project report will first explore the newly discovered data augmentation technique, COMPOSITES, a 10-step technique that uses various concepts like mathematics and physics to apply constraints to amplify a dataset reasonably.
The COMPOSITES process will be demonstrated on 2 case studies done on inter-ply fracture toughness (ILFT) of carbon nanotube (CNT)-engineered carbon fiber reinforced polymer (CFRP) composite laminates. The case studies will study the effect of varying contents of CNT on (I) Mode I fracture and (II) Mode I and Mode II mixed mode loading fracture. The first case study will serve as a form of validation for the process as well as a platform to delve further into the possible analysis and applications of the process. The second case study will be used to evaluate how well the COMPOSITES process can be applied on a frequently used performance measure in the industry. This report will then go on to discuss on the possible future areas for study to further develop this technique. |
author2 |
Sunil Chandrakant Joshi |
author_facet |
Sunil Chandrakant Joshi Tan, Jerome Whye Hean |
format |
Final Year Project |
author |
Tan, Jerome Whye Hean |
author_sort |
Tan, Jerome Whye Hean |
title |
Engineer data augmentation practices for machine learning in composites |
title_short |
Engineer data augmentation practices for machine learning in composites |
title_full |
Engineer data augmentation practices for machine learning in composites |
title_fullStr |
Engineer data augmentation practices for machine learning in composites |
title_full_unstemmed |
Engineer data augmentation practices for machine learning in composites |
title_sort |
engineer data augmentation practices for machine learning in composites |
publisher |
Nanyang Technological University |
publishDate |
2021 |
url |
https://hdl.handle.net/10356/153489 |
_version_ |
1718368093753835520 |