Control of cold-rolled strip shape has become an important issue in recent years. But the classical control methods that use the mathmatical equations are not adaptive for this process, because cold-rolling is a nonlinear system and many parameters are involved.
In this study, fuzzy theory is introduced to the controlling of the strip shape in cold rolling. Fuzzy controller is developed based on the production data and the operational knowledges.
The cold rolled products are characterized into several types based on its irregularity. For each type of irregular strip shape, fuzzy controller calculates the change of bender forces of work roll and intermediate roll using fuzzy rules and fuzzy inference. To simulate the continuous shape control, fuzzy controller is linked with emulator which is developed using neural network. The developed fuzzy controller and emulator simulate the cold rolling process until the irregularities converge to the tolerable range to produce uniform cross-sectioned shape.
The results from the various simulations are reasonable for various types of irregular cross-sectioned shapes as well as simple shapes.