From Feedback Control to Real-Time Business Decision Making in the Process Industrie

Wolfgang Marquardt

The process industries is forced to optimally use existing assets in a flexible and economical manner to cope with the extremely competitive business climate which is due to global market places with largely saturating demand and full cost transparency. The objective of plant management has already been evolving and will continue to evolve from decentralized control at a given nominal operating point to an economically optimal operation at any given point in time. The objective of plant operation is therefore being shifted from set-point tracking and disturbance rejection to model-based real-time optimization to assist business decision making. This talk will briefly introduce current trends in the process industries. A general operations and control problem is formulated. Various decomposition strategies are reviewed. The focus of the presentation will be on recent advances in dynamic real-time optimization including numerical methods as well as strategies to deal with unavoidable model uncertainties. Chemical process applications will serve as illustrating examples.