Prediction of Phosphorus Distribution Ratio between Steel and Slag in Converter using Just-In-Time Modeling and Data-Imputation

In converter steelmaking, the quantities of added submaterials are important to produce steel with desired composition at the lowest cost. A system has been developed which optimizes the quantities of the submaterials and provides guidance to operators. Here, the phosphorus distribution ratio between steel and slag has an important role in the optimization. In this paper, we enhance the accuracy of the model by predicting errors. Just-In-Time modeling as well as data imputation is used to make the most of past data. The simulation result with the prediction showed 14.3% improvement in accuracy compared to the result with no imputation.