Modeling and Simulation of Hydraulic Roll Bending System Based on CMAC Neural Network and PID Coupling Control Strategy

Abstract

The hydraulic roll bending control system usually has the dynamic characteristics of nonlinearity, slow time-variance and strong outside interference in the rolling process, so it is difficult to establish a precise mathematical model for control. So, a new method for establishing a hydraulic roll bending control system is put forward by cerebellar model articulation controller (CMAC) neural network and proportional-integral-derivative (PID) coupling control strategy. The non-linear relationship between input and output can be achieved by the concept mapping and the actual mapping of CMAC. The simulation results show that, compared with the conventional PID control algorithm, the parallel control algorithm can overcome the influence of parameter change of roll bending system on the control performance, thus improve the anti-jamming capability of the system greatly, reduce the dependence of control performance on the accuracy of the analytical model, enhance the tracking performance of hydraulic roll bending loop for the hydraulic and roll bending force and increase system response speed. The results indicate that the CMAC-PID coupling control strategy for hydraulic roll bending system is effective.