A Comparative Design Study for Multivariable Feedback Control System Regulation for pneumatic feed mixing system
The British University in Dubai (BUiD)
In this research, mixing system regulation was investigated. Two control methodologies were studied. Least effort method from the Modern British School, and H-infinity controller from Modern American School. The closed-loop system transients and steady-state response were validated. Controller disturbance rejection and energy consumption were verified, and the results were compared between the two design methodologies. This research is started with a brief history of control systems, major effective millstones of this science and examples of main implementations. After which the main idea about multivariable systems is presented. The system under study is a pneumatic feed mixing system used to simulate an industrial mixing process. It consists of two inputs and two outputs. Research objectives are to evaluate the two controllers under study and compare their performance and energy consumption used to achieve this performance. This system was raised as a control problem used by Dutton (1997) to compare several controllers, and re-used by Whalley and Ebrahimi (2006). Revalidation work is done for least effort controller, MATLAB® code and Simulink® model were designed, and the result is compared with another controller designed based on optimal H-infinity method. After a comparison between the two control techniques, this research concludes that the least effort control method is capable of giving superb solution to the control problem, all the objectives achieved with simple model and perfect energy consumption. H-infinity controller offered perfect response in terms of speed, disturbance rejection, and steady state de-coupling, but it has shown high energy cost with complicated model. This shall promote the use of least effort controller for general industrial multivariable mixing process and similar processes.
multivariable feedback control, energy consumption, control techniques