Smith Predictor Design in IMC Framework for Integrating Plants

Abstract

PID controllers have been the first and foremost choice of control engineers for their simplicity, ease of application and satisfactory performances. Standard IMC is incapable of taking care of disturbances occurring on the input of plant to overcome this drawback of IMC, a modification is performed on IMC configuration which not only takes care of servo response but also takes care of the regulatory response of the system. Simple methods for tuning this controller is presented in this thesis. The controller structure and the proposed tuning method ensure smooth closed-loop response to servo changes, fast regulatory control and adequate robustness with respect to parametric uncertainty.
In this thesis an IMC controller is modified with the help of Smith-predictor configuration and is used for the control of an integrating plus dead time (IPDT) processes which gives parameters of PI/PID in terms of filter time constant which is adjusted by user and parameters of process model. The proposed method require small computational effort since it requires some prior information which can easily be obtained. Simulation results obtained shows that this method posses presemptuous characteristics compared to IMC-PID & previously known PI/PID controller tuning methods for IPDT processes. To validate the controller thus designed an experiment is performed on a temperature control plant [i.e.IPTD process]. The experimental results thus obtained are compared with simulation results which justifies not only the usefulness but also confirms that the controller can be applied to real time systems too.