A theoretical analysis of Smith and analytical predictors

A theoretical analysis has been performed to elucidate the relationship between two popular time delay compensation techniques, the Smith predictor and the analytical predictor. In general, use of these two techniques results in different closed-loop responses. It is only for the special case where the process model is perfect and set point changes occur, that the two techniques are equivalent. A modification of the analytical predictor has been proposed that allows the technique to be used with any feedback controller, instead of the special PI controller employed by Moore et al. (1970).