About this Abstract |
Meeting |
MS&T25: Materials Science & Technology
|
Symposium
|
Applications of Uncertainty Quantification (UQ) in Science and Engineering
|
Presentation Title |
Representative microstructure for macro-scale property prediction using multi-scale models |
Author(s) |
Arulmurugan Senthilnathan, Pranav Karve, Sankaran Mahadevan |
On-Site Speaker (Planned) |
Arulmurugan Senthilnathan |
Abstract Scope |
Micromechanical models require a representative microstructure of the meso-scale coupon to predict macro-scale properties. However, microstructures typically obtained using the Electron Back Scattered Diffraction (EBSD) technique may not be representative of the meso-scale coupon due to spatial variability of microstructural features that influence macro-scale properties. This work develops a novel comprehensive methodology to determine the representative microstructure (RM) size of a meso-scale coupon for predicting macro-scale properties using micromechanical model. RM size is determined by using a dissimilarity metric that quantifies the spatial variability of the microstructure features and a reliability metric that estimates the accuracy of mechanical property prediction. The developed methodology is illustrated on synthetically generated conventionally forged Titanium-7%(wt) Aluminum (Ti-7Al) microstructure and the RM size of a Ti-7Al is determined. Developing a methodology to determine RM size paves the way for accurate macro-scale property prediction that is used for material design, uncertainty quantification, and optimum process design. |