Multivariable closed-loop system identification of plants under model predictive control

Abstract

This paper discusses a simulation where a multi variable plant under Modelbased Predictive Control was identified from closed-loop data. A motivation for closedloop system identification in this context is given and an identification methodology is proposed. To evaluate the consistency of the methodology, the plant was identified for different controller settings and different added disturbances. Different methods to ensure identifiability were also investigated.