Application of adaptive PID based on RBF neural networks in temperature control

Abstract

Electric-heating process has strong nonlinearity and time-varying properties. They are difficult to control accurately using the traditional PID controller with fixed PID parameters. PID parameters are need to be retuned if the control conditions are changed. Combined with traditional PID controller and radial basis function (RBF) neural networks, a PID controller based on RBF neural network is proposed. The parameters of PID controller are tuned on-line using the self-learning ability of RBF neural network. This PID control algorithm is successfully implemented in Matlab software which is integrated with configuration software KingView through their dynamic date exchange (DDE) channel. The PID controller is used in the temperature control of a small electric-heating reactor. The control result shows that the RBF neural network PID has much better control performance than the conventional PID controller.