Preparation and property optimization of FeCrAl-based ODS alloy by machine learning combined with wedge-shaped hot-rolling

Abstract

In this study, oxide dispersion strengthened (ODS) steel with a composition of Fe-12%Cr-4.5%Al-2.0%W-0.3%Y 2 O 3 (wt%) was designed by machine learning (ML) and prepared by hot isostatic pressing (HIP). Wedge-shaped hot-rolling was used to investigate its deformation ability and improve its properties. Microstructure before and after rolling was systematically characterized. The results show that the ODS steel has a bimodal ferrite matrix consisting of coarse equiaxed grains tightly surrounded by fine ones. Coarse M 23 C 6 particles are dispersed in the ferrite matrix. The bimodal equiaxed grains are changed to fibrous ones, and the coarse carbides are refined significantly after hot-rolling. The high temperature of HIP and hot-rolling does not significantly coarsen the Y 2 O 3 particles, and high-density spherical and ellipsoidal nano-sized Y 2 O 3 particles are always dispersed in the ferrite matrix. However, whether it is a HIP sample or rolled sample, it is found that there is a small amount of Y-Al-O composite oxide, and its average size is significantly larger than Y 2 O 3 . The tensile strength and elongation measured at room temperature and 700 °C are very consistent with the predicted values of ML. Moreover, the as-HIPed steel exhibits excellent comprehensive deformability at room and high temperatures, and the performance at room and high temperatures are significantly improved after hot-deformation. • FeCrAl-based ODS steel was designed and prepared by a combination of machine learning, HIP and wedge-shaped hot-rolling. • ODS steel has a bimodal structured ferrite matrix after HIP and changed to fibrous ones after hot-rolling. • High-density nano-sized Y 2 O 3 particles mixed with Y-Al-O are dispersed in the ferrite before and after hot-rolling. • Excellent mechanical properties of the ODS steel at room temperatures and 700 °C can be achieved after hot-rolling.