Decoupling control of thickness and tension based on DRNN-PID in cold-rolling

Considering the features of coupling, multivariable, and nonlinearity in thickness and tension system for cold-rolling, this paper proposes a dynamic coupling model. Furthermore, a new compound decoupling control algorithm is designed based on diagonal recurrent neural network combined with PID(DRNN-PID) for decoupling control. Simulation results show the proposed algorithm has stronger adaptive tracking, faster response, and better anti-interference than PID. The performances can meet the requirement of practical rolling and effectively enhance the control accuracy of thickness and tension.