Application of Modified Smith Predictor Controller for SOPDT Systems

Abstract

This dissertation presents a new and modern approach for controller design for delayed time systems. Delay time or dead time appear in all real processes such as in industries and in other fields of work, consisting economical systems and biological systems also. The control of delayed time systems with widely used Proportional Integral Derivative (PID) controllers is not very effective. The simplest and accurate delay or dead time restitution method for stable and well damped processes is the Smith Predictor Controller. In this dissertation a modified version of Smith Predictor Controller is applied to different real systems which are acting like delayed time system. The modified Smith Predictor Controller is a two Degree of Freedom (DOF) controller. It is superior to the simple Smith Predictor controller. It removes the delay part from the characteristic equation and it also separate the disturbance output from the set point output. The set point output and the disturbance output of the closed-loop systems are accommodated by two disjoint parameters 1  and 2  , respectively. Therefore, the modified Smith predictor structure has the strong potential capability to achieve a remarkable improvement in the set point response and disturbance response. In this dissertation work, the proposed controller is applied to different SOPDT systems. It can also be applied to higher order systems by reducing it to first order or second order by model order reduction technique. In this dissertation work, this technique is applied to different real systems like Boost Converter, Solar Powered Boost Converter and Speed Control of DC Servo Motor with Communication Delay for the validation of this control technique. Experiments are performed in MATLAB and Simulink software. Simulation results are found out and the performance of Modified Smith Predictor Control technique is compared with Internal Model Control (IMC) control technique for all systems that are discussed in this dissertation work.