Simulation of Neural Network based PID Controller for Pressure Process

Abstract

This paper provides a Neural Network PID controller based on Back Propagation (BP) algorithm applied to pressure control in a tank. The controller has many advantages like that more convenient in parameter regulating, better robust. Neural network is to adjust the parameters of PID controller based on the operational status of the system, to achieve a better performance, making the output of the output neurons corresponding to the three adjustable parameters of a PID controller. Through neural network self-learning and weighting coefficient adjustment, the neural network output will corresponds to the PID controller parameters under a certain optimal control law. The simulation results of pressure control in a tank by using Neuro-PID controller shows that it can get better control.

Keywords: PID controller, back propagation neural network, pressure control