| About this Abstract |
| Meeting |
2021 TMS Annual Meeting & Exhibition
|
| Symposium
|
Friction Stir Welding and Processing XI
|
| Presentation Title |
Application of Machine Learning for Prediction of Microstructure and Mechanical Performances in Solid-state Joining Processes |
| Author(s) |
Benjamin Klusemann, Frederic E. Bock, Uceu F.H. Suhuddin, Lucian A. Blaga, Jorge F. dos Santos |
| On-Site Speaker (Planned) |
Jorge F. dos Santos |
| Abstract Scope |
Data-driven machine learning models exhibit strong capabilities to identify and quantify relationships along the process-microstructure-property chain. Via regression analysis, highly non-linear correlations along these chain domains can be established and used for predictions. The integration of microstructural feature classification and segmentation, enables strengthening and refinement of those correlations. Predictions of mechanical properties based on particular process parameters but also inverse determination of required process parameters for desired properties can be accomplished.
In this contribution, results from the application of various machine-learning-models to correlate process parameters and microstructure characteristics with desired joint properties in solid-state joining techniques will be presented. Two exemplary processes of Refill Friction Stir Spot Welding and Friction Riveting will be analyzed: Experimental data is generated through Designs of Experiments with addition of many experiments that enrich the input data. The predicting capabilities and learning effects of such machine learning models for solid-state joining techniques will be discussed. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Machine Learning, Process Technology, Mechanical Properties |