Intelligent Temperature Control in an Industrial Furnace

The paper presents an original solution for the intelligent control of the temperature in a rotary hearth furnace. The controlled heating process is approached as a distributed parameter one. The process dynamics in relation to time is identified based on experimental data obtained from the real plant. For the appropriate temperature control in the furnace, an advanced control system based on intelligent control is designed. The neural networks are used both to adapt the parameters of the main controller and to estimate the values of the disturbances, respectively of the faults which occur in the system. The designed control system can reject the effect of the faults which occur in the furnace operation, being a fault tolerant one. The estimation of the fault value is made by neural networks, based on the model of the spatial temperature distribution between the furnace sectors. The efficiency of the proposed control system is proved through simulation.