Prediction model for properties of hot-rolled steel strip: An ICME approach

Abstract

The Integrated Computational Materials Engineering (ICME) approach is implemented to develop an integrated model for the processing of steel at a hot strip mill. The effect of temperature, deformation, and subsequent cooling rate on microstructure and its associated properties of the steel are mapped using analytical, numerical as well as machine learning approaches, all applied in tandem. The modeling approaches include a rigid visco-plastic approach for the flow formulation, finite volume method to calculate temperature distribution profile, artificial neural network (ANN) modeling for predicting CCT diagram, and Mecking–Kock model along with stress mixing law for the structure–property correlation. The composite model for mapping the final microstructure and microstructure-dependent properties of the steel, which depends on the composition, and various parameters of the hot strip mill is validated through a comparison of predicted values with the published results. The work shows that the hybridization of constitutive equations with the data-driven approach in ICME can successfully model a complex system like the hot strip mill.