Performance analysis of various Smith predictors for integrating processes with longer dead-time

Abstract

Processes that contain a large transport lag[exp (?τDS)] can be difficult to control because a disturbance in set point or load does not reach the output of the process until τD units of time have elapsed. The control strategy known as dead-time compensating controller is also referred to as a Smith predictor. The control algorithm in a Smith predictor is normally a PI controller. This paper describes the advantage of modified Smith predictors which has the advantage over the dead-time compensating controllers. Although it contains five parameters, only three are adjusted by the operator namely the gain, the integral time and the dead-time. The controller is also suited for processes with varying longer dead-times and can be tuned manually in the same way as a PID controller and has the advantage over the other Smith predictor based controllers.