Closed-loop identification of fractional-order models using FOMCON toolbox for MATLAB

Abstract

Closed-loop identification methods attract significant interest because they grant the possibility to obtain industrial process models without the need to perform separate open-loop experiments thus avoiding potential production losses. In this paper, we study the problem of identifying a fractional-order process model of a plant working in a closed control loop. We consider both the direct and the indirect identification approaches. Identification is performed offline by means of minimizing the model output error and is based on data collected in the time domain. We provide the description of the software implementation of the proposed identification procedure. All necessary computations are performed in FOMCON toolbox for MATLAB. An exemplary closed-loop system is studied. It comprises a real-life fractional-order PID controller prototype, and a simulated model of a plant running on a personal computer based real-time prototyping platform. Experimental results are provided with relevant discussion.