Adaptive neural torsional vibration suppression of the rolling mill main drive system subject to state and input constraints with sensor errors

Abstract

Torsional vibration often occurs in rolling mill drive system, which seriously affects the product quality accuracy and the service life of transmission equipment. This paper studies the adaptive neural torsional vibration suppression control problem for the rolling mill main drive system with state and input constraints subject to unknown measurement sensitivities. Firstly, considering the nonlinear friction between the work roll and strip, nonlinear damping at the motor and the load and unknown uncertainties on system parameters, a new torsional vibration model of the main drive system of rolling mill is established. Then, by selecting the proper asymmetric tangent barrier Lyapunov function, the motor torque control law is proposed based on backstepping algorithm. The adaptive neural networks are introduced to solve the unknown uncertainties and the unknown measurement errors and a continuous differentiable Gaussian error function is employed to deal with actuator saturation. It is strictly proved that the designed main drive torsional vibration system is stable and the performances of the transformed states are preserved. Finally, simulation shows the validity and the advantages of the proposed algorithm.