Online Parameter Estimation for Adaptive Feedforward Control of the Strip Thickness in a Hot Strip Rolling Mill

Abstract

A new adaptive disturbance feedforward control strategy of the strip thickness in a hot strip rolling mill with online parameter estimation is proposed. The feedforward control strategy makes use of the measured strip temperature and strip entry thickness. To avoid that these disturbances cause a nonuniform strip exit thickness, the Sims’ roll gap model and a linear mill stand deflection model are used to compute control inputs, which compensate for these disturbances. By minimizing the difference between the expected roll force from the model and the measured roll force, uncertain parameters of the model and also errors of the strip tracking are estimated in real time. The estimated parameters are immediately used in the adaptive feedforward controller. Experimental results of the proposed control approach obtained from an industrial hot strip rolling mill show a significant improvement of the strip thickness accuracy compared to the existing standard controllers. The proposed adaptive feedforward control strategy is now in permanent operation at the considered rolling mill.