About this Abstract |
| Meeting |
2022 TMS Annual Meeting & Exhibition
|
| Symposium
|
Additive Manufacturing: Nano/Micro-mechanics and Length-scale Phenomena
|
| Presentation Title |
Multi-scale Dynamic Strain Measurement and Machine Learning Optimization to Uncover Solidification Dynamics of Ti-5553 L-PBF Melt Pools |
| Author(s) |
Caleb Andrews, Maria Strantza, Nicholas P Calta, Tae Wook Heo, Saad A Khairallah, Rongpei Shi, Manyalibo J Matthews, Mitra L. Taheri |
| On-Site Speaker (Planned) |
Caleb Andrews |
| Abstract Scope |
Laser powder bed additive manufacturing (L-PBF AM) is a technology which enables digital near-net fabrication of complex parts. However, the rapid solidification inherent to the L-PBF process generates residual strains across scales. Understanding how strain spans the micro-millimeter scale as a function of laser processing parameters can elucidate how to form strain free parts by optimizing the L-PBF process, or, create strain anisotropy at the microscale to tune mechanical properties. We demonstrate a high-resolution electron back scattering diffraction (HR-EBSD) framework to measure the absolute strain of Ti-5553 meltpools across their entire length scale with sub-micron resolution. Monte Carlo diffraction simulation is utilized to create a zero-strain point of reference, from which collected Kikuchi patterns can be correlated against, while a deep learning denoising autoencoder is employed to reduce noise from experimental data. This demonstrates a method to obtain the absolute strain tensor at the sub-grain scale without utilizing beamline techniques. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Additive Manufacturing, Machine Learning, Mechanical Properties |