Continuous-time self-tuning controller using delta model identification

Abstract

This contribution presents the application of self-tuning continuous - time controller for process control modelled by S - models. The process is identified by the regression (ARX) model using the recursive least squares method (RLSM) with applied directional forgetting. The basic RLSM algorithm has been modified for the δ- model structures. Controller synthesis is designed on the basis of the pole placement method (a couple double real poles in s - plane has been chosen in the characteristic polynomial). This continuous - time controller is based on the fact that the δ - parameter estimates have excellent convergence and over a short sampling period converge to their continuous versions. The advantage of this approach is that in the identification part of the self - tuning algorithm is not necessary to use computation the parameter estimates of the differential equation using the state variable filter. Choice of initial parameter estimates does not make problems, too. © 2001 EUCA.