The Neural Network Proportion Integral Differential Controller and the Application on Mill Hydraulic Pressure Automatic Gauge Control System

Abstract

This paper presents a neural network proportion integral differential (PID) controller for automatic gauge control (AGC) System of rolling mill, it is an high non-linear and time-varying system. The traditional PID controller has the invariable parameters. However in the actual factory, the environment of the controlled object is often changed. If the three parameters of PID controller can’t adjusted adaptively, the controller will have a badly control effect. The neural network can adjust the three parameter based on the control error. If the control error becomes zero, the parameter didn’t adjust too. The simulation shows that neural network PID controller has good dynamic quality. The control system has short response time, small over modulation, highly steady state behavior and robustness comparing with the traditional PID controller.