Dynamic Modeling of the BOF for Endpoint Prediction Using EFSOP~? Technology Results and Implementation at Riva Taranto

摘要

Reliable endpoint prediction is an invaluable tool in the operation of a BOF. It provides the steelmaker with increased productivity and yield while reducing operating costs. Technicians have traditionally relied on static charge models for endpoint prediction of temperature & carbon. These models are limited in their ability to accurately predict endpoint because they do not account for process dynamics & are adversely affected by uncertainties in the initial conditions. With the EFSOP~? strategy, a rigorous, non-linear dynamic model was used to predict the mass, temperature & compositions of the hot metal, slag and gas phases with closed-loop control and continuous tuning from the feedback of real-time off-gas composition measurements. To enhance endpoint prediction with a Sublance System in place, a supplemental regression model was incorporated. This model, using either an inblow measurement from the Sublance or the predictions of Carbon and temperature from the dynamic model, calculates the endpoints of Carbon and temperature in the final stages of the blow.