Model Predictive Control of a Dynamic System with Fast and Slow Dynamics: Implementation Using PLC

Abstract

The article presents Model Predictive Control (MPC) of a multivariable laboratory process using Programmable Logic Controller (PLC). The Dynamic Matrix Control (DMC) MPC algorithm is used in which a step-response model of the process is used for prediction. Two main practical issues are discussed. Firstly, it is shown how to deal with DMC control of a dynamical systems in which some variables react quickly and some slowly. In order to deal with fast and slow dynamics, a decomposed control structure based on the DMC algorithm is used. Secondly, the decomposed DMC control structure is implemented using the PLC controller, taking into account the PLC resources. Results of real laboratory experiments are present to show effectiveness of the discussed method.