Pseudo-derivative feedback-based identification of unstable processes with application to bioreactors

Abstract

The use of a pseudo-derivative feedback structure for closed-loop parameter identification of unstable first-order plus dead-time processes is investigated in this paper. Two alternative identification methods are presented. The proposed identification methods require small computational effort and they are particularly useful for on-line applications. The new methods and the simpler structure of the pseudo-derivative feedback controller used ensure considerably smoother response (i.e., less overshoot), than other well-known identification methods based on standard PI/PID controllers, and better accuracy along with simplicity in implementation. The synergism of intelligent tools, which can be based on functional and linguistic rules for advanced decisions at the supervisory level, with low-level process PD-0F controllers and identifiers, can provide adaptive systems. These can optimize for both the long-term objectives and the short-term process dynamics.