High-Performance Prediction of Molten Steel Temperature in Tundish through Gray-Box Model

Abstract

A novel gray-box model is proposed to estimate molten steel temperature in a continuous casting process at a steel making plant by combining a first-principle model and a statistical model. The first-principle model was developed on the basis of computational fluid dynamics (CFD) simulations to simplify the model and to improve estimation accuracy. Since the derived first-principle model was not able to estimate the molten steel temperature in the tundish with sufficient accuracy, statistical models were developed to estimate the estimation errors of the first-principle model through partial least squares (PLS) and random forest (RF). As a result of comparing the three models, i.e., the first-principle model, the PLS-based graybox model, and the RF-based gray-box model, the RF-based gray-box model achieved the best estimation performance. Thus, the molten steel temperature in the tundish can be estimated with accuracy by adding estimates of the first-principle model and those of the statistical RF model. The proposed gray-box model was applied to the real process data and the results demonstrated its advantage over other models.