In this paper experiments aimed at identifying the systems parameters of a cascade-controlled servomechanism using both the time domain step response method and frequency response domain Bode plot methods were investigated. In the cascade configuration, the velocity feedback attenuator gain caused the response to become overdamped. Hence, the overdamped response made the step response systems identification method inadequately as it requires at least an overshoot to correctly infer the systems transfer function. Nevertheless, using the frequency response based Bode plot it was possible to parameterise the transfer function of the servomechanism. Furthermore, the obtained model of the servomechanism, which was simulated in MATLAB, depicted the effect of varying the velocity feedback gain from 0.3 to 1 in both open and closed-loop configurations. Finally, the output response of the transfer function for the experimental servomechanism which was obtained via the frequency response method has been compared and validated against the experimental systems output response Also, recent methods for systems identification of transfer function parameters based on genetic algorithm and particle swarm optimisation have all been compared. The output of the transfer function model obtained via the frequency response based method is in good agreement with the output response of the experimental servomechanisms. Hence, this validates the frequency response-based method.
Key words: Systems Identification, Parameter Estimation, Servomechanism, Bode Plot, Frequency Response
|
