Applying dynamic matrix control in the process industries

Abstract

Advanced multivairiable control system designs had little impact in the process industries until the development of model-based predictive control (MBPC) systems. This paper records some aspects of the application of one particular MBPC technique, Dynamic Matrix Control (DMC), to three different plants in the process industries. The paper begins by recalling the essential principles of MBPC systems, by briefly outlining the origins of their use in the process industries, and by recording some of the theoretical issues associated with these systems, that are currently being examined. It proceeds with a particular focus on DMC. Its use in the control of process plants is discussed by reference to a simulated paper machine. Subject to the constraints imposed by commercial confidentiality, studies are then presented of application to a fluidised catalytic cracker in the petrochemical industry and to paper fibre recycle plants in the paper industry. The paper closes with a discussion of the benefit plant-wide optimisation systems can offer whole process plants, once MBPC systems have been implemented on the individual units making up the process.