Izvorni znanstveni članak

Design of GPC's in State Space*

José Vicente Salcedo
Miguel Martínez
Javier Sanchis
Xavier Blasco

Sažetak

This paper introduces a methodology for the original design of generalised predictive controllers (GPC's) based on the use of a state space CARIMA model to carry out those predictions. The CARIMA model presented is equivalent to the CARIMA model commonly used in the input/output (I/O) formulation of the GPC's. A connection is settled among the stochastic part of this model and the filter polynomials Ti(z-1), making possible the design of the controller once any one of them is known. It is also remarkable that for the estimation of non-measurable states, a full rank observer is proposed, and the fact that its poles are equal to the roots of the filter polynomials Ti(z-1) can also be appreciated.

Ključne riječi

CARIMA model; optimization; predictive control; state space design

Hrčak ID:

6620

URI

https://hrcak.srce.hr/6620

Datum izdavanja:

21.12.2001.

Podaci na drugim jezicima: hrvatski

Posjeta: 4.848 *