Computational Alloy Design and Discovery

People

Advisor

Research areas

Description

Engineering alloys are a blend of elements, often a combination of more than 10 deliberate alloying additions. As a consequence, the number of possible alloys that can be produced is (empirically speaking) nearly infinite……..    For example, if one takes 30 elements, and uses factorial analysis, that means a possible 2.6525286 x 1032 alloys are possible……..  that is…. a very large number.   Adding to this, the performance of any metallic alloy is also influenced by numerous factors, including thermomechanical processing and heat treatments.

Needless to say, materials engineers have barely scratched the surface of the possible alloys (and processing conditions) that may be produced. This means a future full of wonderful materials – lies in the waiting. 

The project herein therefore relies on the use of machine learning, to assist in the development of A.I. predicted alloy compositions that are potentially useful for future metallic alloys.  Future metallic alloys will need to have at least one property that is enhanced (i.e. strength), but other properties may also be sought to be enhanced, such as corrosion resistance, conductivity, or toughness.  The project will involve some coding, and the use of machine learning / artificial intelligence tools (such as Tensorflow). Students eager to explore machine learning, data-mining (and who are not afraid to learn to code if they cannot code already) are encouraged to apply. 

Requirements

Some experience in coding and an interest in machine learning is essential.

Updated:  1 June 2019/Responsible Officer:  Dean, CECS/Page Contact:  CECS Marketing