Integrated predictive optimiser and constraint supervisor for processes with basic feedback control

Two important problems in supervisory control are: constraint handling and economic optimisation, especially, if unconstrained direct feedback controllers (e.g. PIDs) are used. The new solution proposed in the paper is an integrated economic set-point optimiser and constraint supervisor which provides the basic feedback controllers with set-points calculated for both optimality and constraint handling. A model-based predictive optimiser is proposed which takes into account input and output constraints and models of both the process and basic feedback controllers. To get the formulation of the nonlinear optimiser suitable for on-line applications, successive linearization of steady-state nonlinear process model used for the economic optimisation is further applied. This leads to formulation of the integrated optimiser's problem as a fast and reliable quadratic programming problem. Example simulation results proving efficacy of the proposed approach for a MIMO nonlinear process are provided.