Experimental Parametric Identification in Closed-Loop Feedback using Novel Approach to the Linear Control Systems

Abstract

The process of experimental identification refers to a challenging task since most of the closed-loop identification methods depend on some existing knowledge of the controlled object. Although various identification methods have been described and implemented in the past, the common issue refers to the use of nonlinear equipment such as relays that may cause an undesirable error and consequently affect the accuracy of the controlled object. The main aim of this research is to conduct experimental verification of a novel approach to linear system identification in the closed-loop feedback system. The novel approach represents an extension to the Ziegler-Nichols experiment performed in the closed-loop feedback system, but without using any additional equipment. One servomechanism composed of two motors is used as an object of identification where data acquisition is performed using a dSPACE embedded module allowing real-time reading of measurement results. The conducted experiment succeeds in locating a representative number of points in the Nyquist plot bringing only damped oscillations. Finally, the considered motor is identified as a first-order object due to the dynamic of available mechanical as well as electrical subsystems.