Tensile property prediction by feature engineering guided machine learning in reduced activation ferritic/martensitic steels

Abstract The accurate prediction of tensile properties has great importance for the service life assessment and alloy design of RAFM steels. In order to overcome the limitation of traditional physical metallurgical models, machine learning algorithm was used to establish universal models for the prediction of RAFM steels’ yield strength and total elongation. A database with a wide range of compositions and treatment processes of RAFM steels was first established. Then, feature engineering methods were used to select the highly correlated features. With the reasonable selection of machine learning algorithm and test/training set partitioning strategy, random forests regressors were trained by the selected features. The prediction results proved that, compared with traditional physical metallurgical models, the feature engineering guided random forests regressors had advantages of accuracy and universality for the prediction of RAFM steels’ yield strength and total elongation. And the calculated process window for the balance of strength and plasticity could provide guidance for the further design and development of RAFM steels.