Model-free iPI controller design and application to a thermal plant

Abstract

The automatic controllers with multiple integral components represent a special category. Although they are not so widely used for standard applications, there is a class of processes that require this type of controller. Thus, in this paper, the intelligent PI (iPI) controllers designed in the Model-Free Control paradigm are presented and compared with a PI controller. Two tuning methods are proposed for the iPI controllers, one based on a pole-placement technique and the other one based on the Iterative Feedback Tuning (IFT) strategy. The IFT algorithm hinges on an iterative data-driven approach that starts from a predesigned set of parameters for the controller and leads to an optimal set of parameters that should improve the system’s performance. To test the capabilities of the proposed design methods, a comparison between the three controllers was performed on a thermal plant used to heat-treat Mg-Li-Al based alloys, and their performances were evaluated based on a numerical analysis.