Identification for the control of MIMO industrial processes

A procedure for the identification of industrial processes with the intention of control system design is proposed, discussed, and illustrated by an application to a full-scale production process. The various identification steps are discussed, with a focus on industrial applicability of the procedure. An industrial production process usually consists of various multi-input multi-output (MIMO) model set used is the common denominator form or minimum polynomial form. Parameter estimation is performed in several steps, thus adapting to estimation and control requirements. As an indicative example of practical results obtained, the identification and control of a quartz tube glass process is described