A temperature control of steel strip using neural network in continuous annealing process

In the iron making process, most of control systems include mathematical models which play an important role in calculating preset values and adapting parameters of the control system. So it is continuously required to adjust the mathematical model for maintaining and improving an accuracy of a control system. For this, there must be an abundant experience and knowledge about the special field. Therefore it is difficult to get a great result without experts. In Kwangyang works, we have improved the control accuracy of some control systems by substituting neural networks for the functions of mathematical model. We present an application example that improves the accuracy of some control system by replacing the mathematical model with a neural network in continuous annealing process