ABSTRACT
PID controllers have been widely used in the process industry for more than the past 50 years. The principle of PID controllers seems to have originated in the work of Nicholas Minorsky in connection with the steering of ships in the early 1920s. During the 1940s, Ziegler-Nichols (Z-N) (Ziegler and Nichols, 1942) tuning formulae appeared and have remained in wide use until today in process control. The Z-N tuning rules use some features of the so-called process reaction curve, which is actually the step response curve of the process to be controlled. There are other methods, which use the features of the Nyquist curve. The principle of autotuning is based on the latter, wherein the critical amplitude and frequency at the Nyquist crossover are used in simple tuning rules. Since the early 1980s, commercial autotuners have been designed on the basis of these tuning methods (Smith and Corripio, 1985). The amount of information on the plant dynamics used by these methods being meager, they do not provide good tuning. For example, the heuristic Z-N tuning laws give rise to an oscillatory closed-loop response.
