Model-Free Controller Tuning Based on DFT Processing: Application to Induction Motor Drives

Abstract

In this paper, we present a new approach based on discrete Fourier transform (DFT) analysis for controller tuning of the closed-loop system with unknown plant. The DFT analysis is used to process the closed-loop measurements collected online to derive the frequency response of an initial closed-loop system that does not provide a good performance. Based on the closed-loop frequency response data, we propose two methods for tuning PID controller parameters according to some desired closed-loop performance specifications. The proposed approach can be applied online because the closed-loop system does not need to be stopped for data collection. The tuning problem of rotor speed controllers of electric drives, is chosen as an example to experimentally validate our proposed method. Specifically, we are interested here in the control of an induction motor. The availability of high-performance computational and storage facilities greatly simplifies the collection of measured data used for controller tuning. The experimental results presented in this paper demonstrate the efficacy and usefulness of the proposed control design method in many industrial applications.