摘要：AbstractThis work presents an analysis of the effect of measurement noise on the closed-loop performance for three anti-windup strategies, used together with a PID controller. The study is done both analytically and experimentally and considered stable, integrating and unstable processes with dead time subjected to saturation of the actuator. The PID tuning rule, used for all the presented case studies, is based in a low-order approximation of the filtered Smith predictor. The analysis shows that the error recalculation anti-windup technique gives better reference tracking performance when compared to the incremental algorithm and back-calculation techniques, being able to reduce the effects of noisy measurements on the calculation of the control action, thus resulting in lower control and process variable variability. In addition, it is shown that when the process operating point is near a saturation limit, noise can cause an offset between the process variable and the reference and it is also proven that the error recalculation anti-windup strategy can significantly attenuate this behavior.
