| Kristiansson's1 controller design and evaluation method characterizes a control system by dividing its frequency response into four regions. Within each frequency region a criterion is defined as a measure of one of the system's properties---performance, robustness, and control activity. Constraining three of the four criteria at desired levels, controller parameters that optimize the fourth criterion are calculated so that an optimal balance of the control system's properties is attained.; This thesis presents the application of Kristiansson's technique to the design, evaluation, simulation and experimental implementation of optimal PI/PID-based control systems for real and hypothetical processes. The experimental evaluation is carried out on two pilot-scale processes and an industrial control loop.; The salient points made from the application of the evaluation technique are: (1) With the use of either first or second order low-pass filtering, derivative (D) control can be safely incorporated into a PI control system, hence a PID controller. Compared with the traditional PI controller, the PID controller can significantly improve output performance without excessive control activity; (2) For processes having significant time-delay dominance, there's substantial improvement in closed-loop performance when a PID controller is utilized, instead of the PI controller; and (3) For processes with larger time delays, the PID controller can perform better or equal to the Smith-augmented PI controller.; 1Kristiansson, B. (2000). Evaluation and Tuning of PID Controllers. Licentiate Thesis, Dept of Signals and Systems, Chalmers University of Technology, Goteborg, Sweden. |
