Nonlinear model predictive control method for carbon fiber angle link weaving machine tension based on neural network

In this paper, a neural network-based nonlinear model predictive control strategy for carbon fiber angle link weaving machine tension is proposed. Firstly, the tension nonlinear model considering the opening disturbance is established. Secondly, considering the existence of nonlinear terms in the system, a radial basis function (RBF) neural network is proposed to approximate the nonlinear terms online to improve the control accuracy of the system. A tension nonlinear model predictive control(NMPC) is designed to achieve constant tension control under the driving torque constraint. And compared with the conventional nonlinear model predictive controller, the RBFNMPC improves the time to reach the steady state in the system state variation curve by 0.2s, reduces the overshoot of the three states in the state response curve by 19.1%, 6.1% and 12.3%, respectively, and effectively reduces the time to reach the steady state.