Improved Pi Control via Dynamic Inversion

Abstract: In this paper we present a new design methodology to achieve high performances in the setpoint following task by means of a standard Proportional-Integral (PI) controller. In particular, a dynamic inversion procedure is adopted to synthesize a suitable command input to the closed-loop control system. In this way, the PI parameters can be selected for a load rejection purpose, as the designed command input is capable to ensure in any case that a desired output transition is accomplished with a low rise time and a small overshoot. A salient feature of the technique is that it does not require a complex identification phase, as a simple first-order plus time delay (FOPTD) model of the process, which can be easily obtained by means of an open-loop step response, suffices. A Padé approximation is then adopted to deal with time delays. Illustrative examples show the effectiveness of the method, which preserves all the basic characteristics of the PI controllers and it is therefore suitable to be applied in an industrial environment. Copyright c 2002 IFAC