Applications of the KLV method for the auto-calibration of PID controllers

Abstract

The KLV method (Landau, Voda, 1992) for the auto-calibration of PI and PID controllers is briefly recalled, and an extension of it (for PI controller) is done. The general idea of the proposed approach combines a "frequency experiment" (identification in closed loop of one point of the frequency characteristics of the plant and the steady state gain at the nominal setpoint). With a frequency type "tuning rule", taking into account the closed loop performances and robustness issues. The tuning rules of this indirect design method are inspired from the symmetrical optimum principles, introduced by Kessler (1958). Their advantage is to consider both robustness aspects (phase margin, gain margin, sensitivity, neglected dynamics) and desired closed loop characteristics. They also cover a large domain of current real applications. The proposed method is compared with the Ziegler-Nichols method. Results from industrial experiences are presented.<>