An advanced model has been developed to track iron and slag levels in the blast furnace hearth. The model is based on measurements of tapped quantities of iron and slag and standard blast furnace measurement variables. The hearth geometry is provided by a previously presented wear model of the hearth refractory, while the floating state of the hearth coke column-the dead man-is estimated from a simplified force balance. The liquid level estimation problem is tackled by an extended Kalman filter, by which the variance of the measurements and parameters can be optimally considered. The results of the model clearly show the dramatic effect of the floating state of the dead man on the tap-cycle evolution Of the liquid levels in the hearth, and therefore point out the importance of applying a proper estimate of the hearth geometry in the model. The findings of the model have been analyzed with respect to the asymptotic limit of the descent below the taphole of the iron-slag interface and the required corrections of the material balances. The model has also been verified by a comparison of its results with the tap-cycle trend of the gas pressure drop over the furnace. (C) 2004 Elsevier Ltd. All rights reserved.