|About this Abstract
||2019 TMS Annual Meeting & Exhibition
||Computational Approaches for Big Data, Artificial Intelligence and Uncertainty Quantification in Computational Materials Science
||Machine Learning to Predict Continuous Cooling Phase Transformations in Steels
||Peter Hedström, Moshiour Rahaman, Wangzhong Mu, Joakim Odqvist
|On-Site Speaker (Planned)
The microstructure of steels is governing their mechanical properties and thus a key aim of metallurgists is to optimize the microstructure by alloying and heat treatment. A common tool to assist this work is continuous cooling transformation (CCT) diagrams, which are experimentally determined. CCT diagrams are challenging to predict computationally, but the reward of such capability would be significant in the development of steel alloys and their heat treatments. The recent progress in machine learning (ML) and the availability of open materials data spur us to make an attempt at predicting CCT diagrams using ML. A large database is collected and we make a systematic comparison of the predictive power of different ML ensemble learners and Artificial Neural Networks (ANN).
||Planned: Supplemental Proceedings volume