About this Abstract |
| Meeting |
2026 TMS Annual Meeting & Exhibition
|
| Symposium
|
Advancing the Frontier of Powder Materials Processing and Sintering: A MPMD/EPD Symposium in Honor of Eugene Olevsky
|
| Presentation Title |
Machine-learning for Efficient Design of Ceramic Microstructures – Challenges and Opportunities |
| Author(s) |
Rajendra K. Bordia, Xiao Geng, Fei Peng, Hai Xiao |
| On-Site Speaker (Planned) |
Rajendra K. Bordia |
| Abstract Scope |
Although there is significant interest in machine learning (ML) for design of ceramics, success has been limited in using ML to design microstructures and predict microstructure sensitive properties. The primary reason for this is the lack of large, correlated datasets. Laser sintering allows ultra-fast sintering of ceramics close to full density within a few tens of seconds. We developed a machine learning (ML) algorithm to predict the microstructure under arbitrary laser power. Next, using ultra-fast laser sintering, we fabricated, in a single laser scan, an alumina sample array that contains around one hundred individual sample units of varying density and microstructure over a broad range. A microstructure-sensitive property, hardness, of the units was measured using micro-indentation. Using the correlated results of microstructure and hardness we developed an ML algorithm to correlate the microstructure to the hardness with less than 5% error. |
| Proceedings Inclusion? |
Planned: |