Dynamic modeling and optimization of an electric arc furnace

A dynamic model of the electric arc furnace (EAF) based on fundamental principles has been developed, validated, and optimized. The model includes material and energy balances over three control volumes, an equilibrium chemistry model, and a scrap melting model. The model dynamically predicts the temperature and composition in the liquid steel bath, slag, and off-gas. Furnace operations such as charging, carbon injection, oxygen lancing, oxy-fuel burner operation, and excess stoichiometric burner oxygen addition are inputs to the model and can be varied at will.; Using off-gas and operating data, eight model parameters were adjusted to minimize the sum-squared error between measurements and model predictions. Two different furnace operating practices were matched satisfactorily for off-gas composition using a nonlinear least squares algorithm in Matlab.; Optimization of the two different practices was carried out using iterative dynamic programming (IDP). The optimization performance measure consisted of maximizing chemical energy utilization, maximizing raw material yield, and minimizing process time. The IDP algorithm manipulated carbon injection, oxygen lancing, and excess stoichiometric burner oxygen simultaneously and individually to minimize the performance measure. IDP results show that both unique furnace operations may be dramatically improved by manipulating any or all of the controls. Chemical energy utilization and raw material yield were improved with almost any control combination, but process time improved only for control combinations which provide significant chemical energy.; The modeling and optimization efforts demonstrate the tremendous potential for the improvement of EAF operation through off-line modeling and optimization.