| Abstract Scope |
Recently, many applications of process data-based machine learning (ML) have been reported in the steel industry. In particular, in order to apply machine learning based on big data in steel factories, micro data from various manufacturing processes such as steelmaking/hot rolling processes must be linked. In this study, mechanical properties such as yield strength, ultimate tensile strength, and elongation of hot rolled steel sheets for automobile parts were predicted. In this study, the model used more than 80 valuable data, including chemical composition, cooling history at the run-out-table, size of the hot rolled coil during winding, cooling history at the yard, and various hot rolling process parameters. By predicting the material of the entire length and width, the material properties deviation of the entire coil can be estimated.
Keywords: steel; hot rolling process; hot rolled steel coil; mechanical property; machine learning |