A new prediction model for liquid level in blast furnaces based on time series analysis

The operations of Blast Furnace in the steel industry involves high levels of uncertainty since the liquid level in the hearth can not be directly measured. This paper presents the methodology to obtain the model of the health level by measuring the Electromotive Force (EMF) on the blast furnace shell, which is strongly correlated with the hearth level. It describes the results achieved with the installation of a measurement system of the EMF signal, and details all procedures for identification of a new mathematical model using time series analysis. The model is used to forecast the hearth level behavior in up to 30 minutes in a Steel Shop located in Brazil.