A Melt Temperature PID Controller Based on RBF Neural Network

Abstract

Traditional melt temperature PID controllers have difficulties in PID parameter tuning. So, they suffer from low accuracy in temperature controlling and the dissatisfaction in high exactitude extrusion processing of the present PID controllers. A new kind of PID controller based on radial basis function (RBF) neural network is proposed. By using a sigmoid function to form the step size function, the proposed controller can not only obtain a higher accuracy in temperature controlling, but also infinitely approach the nonlinear system with lower computations, quicker convergence and more system stability. The simulation results show that the proposed PID controller has a better performance in the melt temperature controlling than other traditional PID controllers.