A New Image-Based Temperature Prediction Method

Abstract

The hearth temperature is an important index to measure the condition of blast furnace (BF). A good and reasonable hearth temperature can maintain the stable production of BF, which is a direct guarantee to realize the longevity and efficiency of BF. Therefore, an image-based multiple rank matrix regression (MRMR) temperature prediction method is proposed to study the change of tuyere raceway temperature. This method takes matrix as the input of the model and considers the spatial position information of the matrix elements. According to the intrinsic property of the projection vectors, the relevant constraints are applied to avoid the overfitting problem. Finally, the prediction performance of the model is validated by using BF data.