Machine learning for prediction of retained austenite fraction and optimization of processing in quenched and partitioned steels

The metastable retained austenite (RA) plays a significant role in the excellent mechanical performance of quenching and partitioning (Q&P) steels, while the volume fraction of RA (V-RA) is challengeable to directly predict due to the complicated relationships between the chemical composition and process (like quenching temperature (Q(T))). A Gaussian process regression model in machine learning was developed to predict V-RA, and the model accuracy was further improved by introducing a metallurgical parameter of martensite fraction alpha(f alpha ') to accurately predict V-RA in Q&P steels. The developed machine learning model combined with Bayesian global optimization can serve as another selection strategy for the quenching temperature, and this strategy is very efficient as it found the optimum Q(T) with the maximum V-RA using only seven consecutive iterations. The benchmark experiment also reveals that the developed machine learning model predicts V-RA more accurately than the popular constrained carbon equilibrium thermodynamic model, even better than a thermo-kinetic quenching-partitioning-tempering-local equilibrium model.

Automated summary

This study developed a Gaussian process regression model in machine learning to predict the volume fraction of metastable retained austenite (V-RA) in quenching and partitioning (Q&P) steels. The accuracy of the model was further improved by introducing a metallurgical parameter and combining it with a Bayesian global optimization method for selecting the quenching temperature. The developed machine learning model showed better accuracy in predicting V-RA compared to other popular models.