Estimation of Temperature Profiles within Slabs of a Simulated Reheat Furnace using the Extended Kalman Filter

THE control of a reheat furnace is complicated by modern furnace configurations. Slabs of various thicknesses, width and length are charged to the furnace. The slabs are heated to a specified average temperature for rolling. The furnace operator must exercise considerable judgement, based on experience, to achieve properly heated slabs; there is no means of measuring the temperature gradient across the slabs in the furnace. It is desirable therefore to reconstruct the unmeasured average slab temperature from the measured furnace signals, e.g., zone temperatures. The framework for combining furnace measurements and process knowledge is the extended Kalman filter. The filter, however, requires the definition of a dynamic furnace model.