A single neuron PID controller for tension control based on RBF NN identification

Abstract

Tension control in FGS (flexible graphite sheet) forming process is crucial to ensure product quality. Because the traditional PID controller is ineffective to regulate the tension when the radius of the unwinding roll is getting smaller, a single neuron adaptive PID controller based on RBFNN (Radial Basis Function neural network) identification is proposed to improve the system performance. RBFNN identifies accurate Jacobian information first and then the single neuron controller adjusts PID parameters is for implementation. The simulation results show that, compared to traditional PID controller, the method possesses the advantages of high precision, quick response and great adaptability and robustness.