Mold level control in continuous caster by neural network model

In continuous billet casting, keeping the mold level steady is one of the m ost important technologies for maintaining steel quality. Using conventiona l methods, it is difficult to attain precise control of the mold level beca use of the nonlinear characteristics of the process. We have developed a co ntrol system using a neural network model to overcome this problem. In this paper, control problems of a continuous easter are introduced first. Next, the structure of the control system is proposed. In our proposed system, t he neural network model recognizes the temporal patterns of inlet flow and controls the stopper stroke for a main control loop with a PI controller. T he problems involved in construction of a valid neural network model that h as good generalization and robust properties, are discussed from the viewpo int of optimizing the number of hidden layer units by the information crite rion. Finally some results of its application are described.