An Expert System Based Process Control System for Silicon Steel Mill Furnace of Rourkela Steel Plant

An intriguing problem faced in Industrial Process Control is unavailability of exact mathematical co-relations, which can define the process behavior. Such cases are most appropriate application areas of Artificial Intelligence and more particularly Expert Systems. A close loop Expert System based heating control system has been developed for Decarburization-Annealing Furnace of Tandem Annealing Line at Silicon Steel Mill of Rourkela Steel Plant. The mill produces electrical grade steel used for various electrical and electromagnetic applications. One of the key factors responsible for achieving desired electromagnetic and mechanical quality of steel is selection of optimum temperature ranges inside the furnace and its accurate control. However exact co-relation, which can determine these temperature values for different steel grades, line speed and other steel sheet parameters are not known. The concepts of Expert System have been used to resolve this indeterminate issue of process control. The Expert System predicts most appropriate set temperatures for different zones of furnace based on steel grade, line speed and other steel coil parameters. The knowledge base and inference rules have been prepared utilizing years of experience and expertise of shop floor operation engineers, researchers and quality control agencies. To make it truly a close loop control, Expert System also configures the control function blocks of a Level-I automation PLC using COM based protocol "OPC". PLC controls the furnace temperature as per set points given by Expert System and communicates the actual process variables back to Expert System. The system has rendered significant improvement in plants performance with respect to improvement in product quality and reduction in energy consumption. It also exemplifies an Industrial demonstration of concepts of Artificial Intelligence for enhancing the efficiency of a plant.