Subspace-based identification integrated to recursive PLS modeling: A preliminary result

Abstract

In this work a new multivariable dynamic recursive estimation approach is presented. The overall meth-odology is based on the classical subspace state-space identification (4SID) theory and integrated with the recursive partial least squares (RPLS) tools. The main theory and concepts are shown and some pre-liminary results are displayed for the classical Shell fractionator process.