A neuro-fuzzy system for looper tension control in rolling mills

A looper tension control system is common to many rolling processes. Conventional tension controllers for mill actuation systems are based on a rolling model. They therefore cannot deal effectively with unmodeled dynamics and large parameter variations that can lead to scrap runs and machinery damage. In this paper, this problem is tackled by designing a fuzzy controller that possesses different tuning schemes for both off-line and on-line tuning of fuzzy control elements. It is shown that the proposed methods outperform conventional control techniques. Finally, the effects of various design options are discussed and some practical remarks are made.