Innovation in industrial model predictive control

The industrial application of model predictive control (MPC) has been largely confined to the petrochemical arena. Crossover into the general process industries has been limited. Among the reasons that may be proposed for this can be highlighted: a lack of precedent applications; large engineering costs; inappropriate control technologies. This paper describes initiatives to address such barriers so that the benefits of MPC can be more widely available: demonstration projects on mainstream process applications; software tools to minimise engineering costs; embedding MPC within the plant regulatory control system; a flexible approach to control technology so that solutions match the problem encountered rather than the other way round