PID Control, Design, Extension, Application

Sammanfattning: This thesis considers the design of PID controllers, the extensions of these controllers to improve their performance, and the applications of these design methods to industrial processes. New tuning methods for PI and PID controllers have been presented. These methods uses a model of the process to be controlled, given as a transfer function. The methods captures essential requirements of a control system, such as: load disturbance response, robustness with respect to model uncertainties, measurement noise response, set point response. The primary design goal of the proposed design methods are to achieve good rejection of a load disturbance with the constraint on robustness to guarantee the stability of the closed loop system. The use of PID control have been extended to bridge the gap between it and the theoretical H-infinity control. It is shown how the robustness constraint of the proposed design methods for PI and PID controllers should be chosen to guarantee that the weighted H-infinity norm of the transfer function from load and measurement disturbance to process inputs and outputs is less than a specified value gamma. A new way to determine for which class of systems a PID controller will be stabilizing is also presented. Furthermore, the use of PID control have been extended to handle processes with undamped modes. A modular approach has been taken, where an active control system has been designed, which consists of an allpass filter and a bandpass filter. To determine the parameters of these two filters the only information needed is a few characteristics of the process frequency response. The proposed design methods for PI and PID controllers have been evaluated in a benchmark for control of steam generator water level in a power plant, and at the pulp and paper company Modo Paper, in Husum, Sweden.