Performance rebustness of minimum variance and self-tuning regulators to disturbance model mismatch

✍ Scribed by W.R. Cluett; P.R.C. Nelson

• Publisher: Elsevier Science
• Year: 1991
• Tongue: English
• Weight: 815 KB
• Volume: 46
• Category: Article
• ISSN: 0009-2509
• DOI: 10.1016/0009-2509(91)80137-n

✦ Synopsis

This paper addresses the issue of mismatch between the actual disturbance model of a process and an assumed model for controller design. In current practice, process control engineers often avoid estimation of disturbance model parameters and prefer to use estimation schemes with deterministic models in mind. Most practical processes have several unmeasured disturbances acting on them. These disturb-anceS can often be combined from a modelling point of view into a single random sequence. If the structure and parameters of this disturbance model are not to be or cannot be estimated off-line (for minimum variance control) or on-line (for self-tuning control), a design polynomial ( TI) can be used in its place. This paper will examine the performance. trade-off with minimum variance control when the disturbance model is taken as T,. and the robustness of the self-tuning regulator when T, is used as a filter in the parameter 2.

estimation scheme.