| About this Abstract |
| Meeting |
2025 TMS Annual Meeting & Exhibition
|
| Symposium
|
Fatigue in Materials: Fundamentals, Multiscale Characterizations and Computational Modeling
|
| Presentation Title |
Unsupervised Learning to Cluster Fatigue Life Based on Fatigue Fracture Surfaces |
| Author(s) |
Katelyn Jones, Paul Shade, Reji John, Patrick Golden, Elizabeth Holm, Anthony Rollett |
| On-Site Speaker (Planned) |
Katelyn Jones |
| Abstract Scope |
This work collects ~6000 scanning electron microscope (SEM) images of Ti-6Al-4V fatigue fracture surfaces and applies convolutional neural networks (CNNs) to build a dataset of fatigue fracture surface images and make a connection between them and fatigue life. Secondary electron images of round bar specimens used in a previous study on fatigue life at various stress levels have been imaged at regular intervals and systematically stored. The images capture the crack initiation site, short crack, and steady crack regions; multiple magnifications were utilized to determine which length scale allows the machine learning algorithms to infer physically meaningful information. Pretrained CNNs are fine-tuned and compared against unsupervised machine learning methods to determine if fracture surface alone can be linked to the fatigue lifetime. The images taken, the algorithms used, identified fatigue properties, and fracture characteristics will be presented. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Machine Learning, Mechanical Properties, Other |