About this Abstract |
Meeting |
2021 TMS Annual Meeting & Exhibition
|
Symposium
|
Materials for High Temperature Applications: Next Generation Superalloys and Beyond
|
Presentation Title |
Inverse Design of Chemistry of High Temperature Ni-base Superalloys Using CALPHAD and Machine Learning |
Author(s) |
Rajesh Jha, George S. Dulikravich |
On-Site Speaker (Planned) |
Rajesh Jha |
Abstract Scope |
Using 120 experimentally verified Nickel base superalloys we performed calculations to estimate the critical phases responsible for high temperature applications (FCC_A1 (γ) and FCC_L12 (γ’)) and one detrimental TCP phase that needs to be avoided (µ-phase). Phase stability calculations were performed at desired temperature of application to determine critical phases, followed by a set of machine learning algorithms on the data obtained from CALPHAD approach. The purpose was to find the composition range so that alloys manufactured through a similar thermal treatment protocol will have maximized critical phases responsible for high temperature properties, while minimizing the detrimental phases. We determined correlations between alloying elements and critical phases; critical phases and properties of interest, and strengthening (γ’) phase and detrimental (µ) phase. Machine learning was used to inversely determine the amount of critical phases and chemistry/composition that would yield this amount of critical phases for given stress-to-rupture and time-to-rupture values. |
Proceedings Inclusion? |
Planned: |
Keywords |
Computational Materials Science & Engineering, Computational Materials Science & Engineering, Machine Learning |