ROBUST STEADY-STATE TARGET CALCULATION FOR MODEL PREDICTIVE CONTROL

摘要

A method and apparatus for steady-state target calculation that explicitlyaccountsfor model uncertainty is disclosed. In accordance with one aspect of theinvention, whenmodel uncertainty is incorporated, the linear program associated with thesteady-statetarget calculation can be recast as a highly structured nonlinear program. Inaccordancewith another aspect of the invention, primal-dual interior point methods canbe applied totake advantage of the resulting special structure. For a system havingcharacteristic gainparameters G having a known uncertainty description, the present inventionprovides amethod and apparatus for selecting steady-state targets for said system-manipulatedvariables such that all system-controlled variables will remain feasible atsteady-state forall possible values of the parameters G within the known uncertaintydescription. Anominal estimate ~ of the system parameters G is made, and in accordance withanother aspect of the invention, the steady-state targets are selected suchthat when~ = G, the system is driven to an operational steady-state in which theobjective functionis extremized.