Home > Downloads > Closed Loop Identification for Multivariable Model Predictive Controller on a Catalytic Gasoline Splitter

All Downloads

MatrikonOPC Drivers

Portal Login

Enter your email address to receive updates about new developments.

Closed Loop Identification for Multivariable Model Predictive Controller on a Catalytic Gasoline Splitter

This paper describes the application of closed loop model identification on a Catalytic Gasoline Splitter CGS) unit advanced control application. The closed loop step test was carried out, with the model predictive controller (MPC) running, in an automated fashion with all the variables being perturbed simultaneously. Modeling was carried out in an iterative manner during the step test to adjust the test plan (experiment design) in order to deliver the best possible models. Uncertainty estimates guided these changes. Testing with the MPC running results in models that are more controller relevant and better for closed loop control (Van den Hof and Schrama, 1995). Generalized binary signals (GBN) were used for designing the dither signals and the ASYM (Zhu, 1996) method was used for multivariable closed loop identification.

Get the full Technical Paper Now!

Details:
Technical Paper (946 KB , PDF)
Author: Matrikon Inc.

Description:
Matrikon™ Control Performance Monitor Technical Paper.