Innovation in industrial model predictive control

摘要

The industrial application of model predictive control (MPC) hasbeen largely confined to the petrochemical arena. Crossover into thegeneral process industries has been limited. Among the reasons that maybe proposed for this can be highlighted: (i) a lack of precedentapplications, (ii) large engineering costs, (iii) inappropriate controltechnologies. This paper describes initiatives to address such barriersso that the benefits of MPC can be more widely available: (i)demonstration projects on mainstream process applications, (ii) softwaretools to minimise engineering costs, (iii) embedding MPC within theplant regulatory control system, (iv) a flexible approach to controltechnology so that solutions match the problem encountered rather thanthe other way round