The paper presents comparative study of several tuning methods for PI controllers aimed at improving disturbance rejection properties. The results showed that Multiple integration method results in the most consistent decay ratios for different types of processes. This result was unexpected, since some methods which are aimed at achieving similar closed-loop responses (e.g. method based on selecting maximum sensitivity) were not so successful.
COBISS.SI-ID: 21338663
The paper describes the usage of predictive controller for low-level control of two-inputs and two-outputs process and comparison with classical two PID control loops. The predictive controller is used for low-level control, where the main emphasis is disturbance rejection performance and robustness. The main results are the following: the closed-loop performance of the classical predictive controller is less efficient than of the PID controller, while it becomes more efficient when using predictive controller with Kalman filter.
COBISS.SI-ID: 20987943
The paper gives modification of standard magnitude optimum criterion for controller tuning in order to increase disturbance rejection performance. The method was applied to PID types of controllers. Calculating of controller parameters requires usage of optimisation methods. Results of experiments showed that disturbance rejection efficiency can be considerably increased, especially for lower-order processes.
COBISS.SI-ID: 23246887
It is known that MOMI tuning method is not suitable for controlling integral processes when using one-degree of freedom (1-DOF) PI controllers, since the magnitude optimum (MO) criteria cannot be met. In this paper it is shown that the MO criteria can be satisfied by using 2-DOF PI controllers. Simplicity and efficiency of the method is demonstrated on several process models and on one hydraulic laboratory plant.
COBISS.SI-ID: 21906727
The paper presents a novel tuning method for different types of controllers. The main advantage of Equalisation Tuning method is that it does not require any additional data from the operator except the measurement of the process steady-state change in an open-loop experiment. The equalisation method is also relatively insensitive to process output noise. Simplicity and efficiency of the method is demonstrated on several process models and on a hydraulic laboratory plant.
COBISS.SI-ID: 21982247