Title: Neural network PID control for a water level system

Authors: Xiaoli Li, Longhui Shi, Ji Li

Addresses: Department of Automation, School of Information Engineering, University of Science and Technology Beijing, Beijing 100086, China. ' Department of Automation, School of Information Engineering, University of Science and Technology Beijing, Beijing 100086, China. ' Department of Automation, School of Information Engineering, University of Science and Technology Beijing, Beijing 100086, China

Abstract: Water level system is a classic device for the research of non-linear system control. It is always represented by a First-Order plus Dead Time (FOPDT) model around the equilibrium point. Based on the water level system, different PID strategies (empirical method and Ziegler-Nichols method) are studied for the improvement of control performance. Optimised by gradient descent method, a PID controller based on a radial based function (RBF) neural network is given and applied to an actual A3000 three-tank water level system. From the experiment result, the effectiveness of the proposed method is tested.

Keywords: water level control; PID control; RBF neural networks; nonlinear systems.

DOI: 10.1504/IJMIC.2010.035287

International Journal of Modelling, Identification and Control, 2010 Vol.11 No.1/2, pp.124 - 129

Published online: 20 Sep 2010 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article