Research On Excitation Signal And Frequency Domain Identification Of The Servo Control System

Motion control systems have been applied into almost every field in industry and they play an important role in the development of a country's economy. System identification is the basis for controller design, especially for high speed and high precision motion control system design. Proper identification provides the most basic knowledge to improve the control system performance.The content of this thesis mainly contains several aspects as follow:(1) Details of the modeling by mechanism of the servo control plant are given. In the gray box identification, the control system usually works below the cut-off frequency. For the purpose of simplifying the controller design, the system model is approximated by ignoring high-frequency model from the perspective of the excitation signal design. The frequency domain characteristics of the electromechanical system, such as the cut-off frequency, and the 3dB fallen frequency are identified.(2) The conditions that the excitation signal should meet are analyzed. Since the spectrum of the pseudo random binary signal is broadband, the signal cause the excitation of high-frequency uncertainties. From the simulation, we get the conclusion that the PRBS is not suitable as the input signal for ignoring high-frequency model.(3) Analysis about the spectrum bandwidth and the amplitude of the input signal, the amplitude of its differential signal is done. The suitable swept-frequency sinusoidal signal is applied to the mechatronic system. The estimated parameters oscillate periodically since the instantaneous frequency of this particular excitation signal is explicitly dependent on time and the estimated results are not convergent.(4) A particular class of chaotic signal is proposed as the excitation signal for the mechatronic system identification, because chaotic signal typically has stationary, continuous, and band-limited power spectra. According to the frequency domain response experiment, the suitable chaotic signal is applied to the practical mechatronic system. The estimated parameters are convergent, and the frequency domain character of the system is obtained. The verification experiments show that this type of chaotic signal can be safely applied to mechatronic system, and the identification model can approximate the model of the practical system.