This paper considers the use of a cascade compensator embedded in a Smith predictor in the forward path of a system being designed, which is an effective way to solve two problems: providing the desired process performance in the system and placing the dead-time element outside the closed loop so that it only increases the transient time by the value of the time delay and does not affect the overshoot value. It is shown that the Smith predictor should be used with caution because, being based on an additive signal compensation scheme, it may be not robust and may require accurate knowledge of the time-delay value for robust implementation. The problem of sensitivity analysis of the system to dead-time variation is stated and solved.
No takes yet. Share an insight, caveat, or question.
Vunder et al. (2016) studied this question.
Synapse has enriched 5 related papers. Consider them for comparative context: