BYU

Abstract by Brayden Bekker

Personal Infomation


Presenter's Name

Brayden Bekker

Degree Level

Undergraduate

Co-Authors

Chandramouli Nyshadham
Hayden Oliver

Abstract Infomation


Department

Physics and Astronomy

Faculty Advisor

Gus Hart

Title

How to Find a Superalloy

Abstract

The search for new materials has accompanied human history since its origin. In recent centuries many have turned to chemistry and physics with the goal to develop methods for efficient materials discovery. A superalloy is a material which, among other properties, exhibits great mechanical strength at high temperatures. A new superalloy has the potential to for example increase the efficieny of jet engines and increase the lifespan of marine and aerospace turbines. We present a machine learning based approach to search for potential superalloys. We pick three elements and create a model which predicts the properties of a material composed of some composition of those three elements. The accuracy and efficiency of this machine learning approach allows us to study in detail the predicted properties of a new material as the concentration of different elements in the material are varied. Machine learning as a tool for computational materials discovery helps hone the search for the superalloys that will shape the future.