Mass-balance Based Multivariate Modelling of Basic Oxygen Furnace Used in Steel Industry

摘要

A static model for a basic oxygen furnace process in SSAB steel plant is described in this paper. The model aims for predicting the so called distribution factor describing how the chemical elements fed into the process are distributed into the outlet streams. A multi-model approach is used where sub-models use regression models with interaction and quadratic terms with carefully selected input variables. The predictions for manganese, phosphorus and sulphur work well, while the model for iron needs further improvement. Different modelling techniques or clustering may be applied.