Application of fuzzy logic for automatic shape control in stainless steel rolling process

The shape control of the stainless steel rolling has difficulty in obtaining symmetric shape and stabilizing strip shape despite an abnormal state that shows asymmetric shape pattern with respect to strip lateral direction. The objective of the shape control is to stabilize shape and material flow. The method for the strip shape recognition uses neural network (NN) and least squares method (LSQ). NN extracts symmetric component from the shape error. Moreover, LSQ does curve fitting and classifies the shape error into asymmetry. Fuzzy control method utilizing operator´s knowledge will be proposed in this paper. The experiments are carried out online and then the results show very efficient performance on tracking of the target shape