An innovative inverse model for predicting the minimum protective bank inside a cylindrical electric arc furnace

Abstract

An innovative fast inverse heat transfer method for predicting the protective phase-change banks inside a cylindrical electric arc smelter is presented. The proposed model takes into account the phase change processes and the thermal effect of the chemical reactions for the smelting of the ore. The Particle Swarm Optimization algorithm is combined with the Levenberg-Marquardt method in order to determine the minimum thickness of the protective bank from non-invasive temperature measurements taken outside the refractory brick walls of the furnace. The required number of temperature sensors and their location is determined using the Principal Component Analysis. Results show that the proposed overall inverse method is computationally faster than other existing inverse methods and, as a result, it may be incorporated to existing on-line control systems of industrial smelters.