A novel supervisory control system was developed for a boutique, walking beam type, natural gas fired, industrial reheating furnace for steel slabs. The control system was developed with input from the furnace operators. The system development ideology has been to utilise the considerable and inexpensive computing resources available today, to solve problems in real time in a discrete and digitised manner in place of complex analytical solutions. The system provides a true three-dimensional interface displaying multiple sensor outputs, recommended heating profiles, predicted rougher delivery temperatures and accurately coloured two dimensional (three dimensionally wrapped) slab temperature profiles.

The slab temperatures are calculated using iterative algorithms which employ a Kalman filter developed to ensure that the calculated slab temperatures match the surface temperatures measured by temperature compensated laser pyrometers. The slab modelling system is adaptive; it applies learned coefficients to the absorptivity and emissivity coefficients of slabs as they are processed.

The slab radiation exposure is determined by interpolating between twenty four thermocouples embedded in the furnace surfaces and then calculating the hemispherical equivalent of the radiating surfaces in line of sight to a section of the slab.

The control system utilises an entirely iterative regime to calculate the required heating profiles to ensure that slab delivery temperatures are optimised.

The results are promising with estimated control performance at least as good as is currently manually achieved but with the benefit of additional information being provided to the expert human operator in the control loop.