| About this Abstract |
| Meeting |
MS&T24: Materials Science & Technology
|
| Symposium
|
Integrated Computational Materials Engineering for Physics-Based Machine Learning Models
|
| Presentation Title |
Hybrid Machine Learning Informed Design Guidelines for Structural Gradient Alloys with Enhanced Performances |
| Author(s) |
Jixuan Dong, S. Mohadeseh Taheri-Mousavi |
| On-Site Speaker (Planned) |
Jixuan Dong |
| Abstract Scope |
Gradient composition/microstructure alloys printed by emerging additive manufacturing at voxel size resolution enables a leap in the performance of next-generation structural components. However, defining the configurations of the gradient pattern is challenging. The first step is to define the potential material candidates which meet the requirements at each voxel. Defining these candidates is extremely challenging in experiments due to its time-consuming and expensive nature. We improve this process by developing surrogate models that connect the potential compositional and processing space to microstructural features and properties with hybrid machine learning and CALPHAD based simulations. We can define constraints such as limitation of specific phases, while preparing these surrogate models. We will showcase our analysis for a blisk in rocket engines with gradient composition from Inconel 718 to Monel K-500. Our computational design framework can be applied as guidelines for gradient designs of various structural alloys with different target combinations of properties. |