About this Abstract |
| Meeting |
2026 TMS Annual Meeting & Exhibition
|
| Symposium
|
Fracture and Deformation Across Length Scales: Celebrating the Legacy of William Gerberich
|
| Presentation Title |
Deep Learning in Characterization of Nanoindentation Induced Acoustic Events |
| Author(s) |
Antanas Daugela |
| On-Site Speaker (Planned) |
Antanas Daugela |
| Abstract Scope |
Acoustic emission monitoring of nanoindentation induced high frequency events have been explored in the past resulting in phase transformations, plasticity yield point and thin film fracture phenomena characterizations. This technique provides additional information in registering ultra-fast nanoscale phenomena where in some cases such as phase transformations quasi-static nanoindentation loading-unloading curves are not informative. Various acoustic system configurations and signal processing methodologies have been explored through the years. Conventional acoustic signal RMS or integrated energy values were found to be very consistent but not capable of differentiating similar events. Thus, the advanced signal processing techniques such as wavelet based acoustic signal decomposition and Joint Time-Frequency domain representation methodologies can play a critical role in “fingerprinting” undergoing high speed phenomena. In order to complete and automate the process of signal characterization a Deep Learning has been implemented into the nanoindentation induced acoustic signature classification. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Characterization, Machine Learning, Thin Films and Interfaces |