About this Abstract |
Meeting |
2025 TMS Annual Meeting & Exhibition
|
Symposium
|
Additive Manufacturing Modeling, Simulation and Machine Learning
|
Presentation Title |
Quantifying the Characteristics of Pore Features using Gaussian Process Machine Learning in LPBF Process Parameter Space |
Author(s) |
Tasrif Ul Anwar, Maher Alghalayini, Nadia Kouraytem |
On-Site Speaker (Planned) |
Tasrif Ul Anwar |
Abstract Scope |
Laser powder bed fusion (LPBF) additive manufacturing provides intricate materials with varying microstructural properties depending on process parameters like laser power and scan speed. LPBF is widely used for applications that require engineered porous materials, such as in the biomedical and energy fields. However, due to the complex physics in the melt pool, additional stochastic pores are introduced to the material’s structure. Quantifying the pore structures’ features is crucial for better performance quantification. Here, we combine optical imaging and Gaussian process machine learning to quantify pore features as a function of key process parameters. Pore properties, such as size, shape, and distribution, are extracted from printed material images at specific points in the process parameter space. The Gaussian process then predicts the characteristics of those pore features in the continuous parameter space. This study opens new avenues for quality control using distribution analysis in LPBF. |
Proceedings Inclusion? |
Planned: |
Keywords |
Additive Manufacturing, Machine Learning, |