Accelerating nano-bainite transformation based on a new constructed microstructural predicting model

Abstract

Shortening the bainite transformation time is one crucial issue on the application of high carbon nano-bainite steel. In this paper, an accurate black-box model that could predict the plate thickness of bainitic ferrite was constructed based on gradient boosting decision tree model, and a polynomial model was constructed to make predicting process easier. According to the constructed model, a two-steps process was proposed, which shortened the transformation time from over 60 h to nearly 25 h. After the new process treatment, the size distribution of bainitic ferrite in microstructure became more uniform, and the strength and toughness of steel were also improved simultaneously.