Data-based Tuning of PI Controllers for Vertical Three-Tank Systems

Abstract

This paper suggests the application of Iterative Feedback Tuning (IFT) as a data-based control technique to parameter tuning of PI controllers dedicated to vertical three-tank systems. The level control in the first two tanks is carried out using a multivariable control system structure which consists of two control loops, one for each level. The two PI controllers in these control loops are first tuned in terms of the Modulus Optimum method. New IFT algorithms are proposed in order to ensure the performance improvement of level control systems by means of six steps assisted by experiments. The experimental results show the strong performance improvement obtained after few iterations of IFT algorithms in terms of a model reference tracking optimization problem.