NCO-Tracking with Changing Set of Active Constraints using Multiple Solution Models

Measurement-based optimization schemes can be used to counteract the effect of uncertainty and to drive the process to optimality. Necessary Conditions of Optimality Tracking (NCO-tracking) is a promising scheme which uses the process measurements to update the inputs directly. The limiting assumption for the application of the NCO-tracking scheme is that the set of active constraints does not change from its nominal value under the influence of the uncertainties. This paper presents a scheme that extends the NCO-tracking approach to handle parametric uncertainties that change the set of active constraints. Several solution structures are implemented to cover the whole range of the uncertain parameters. Based on the identification of the region of the uncertain parameters, the correct solution model is determined and applied to the process. The applicability of the methodology is shown for the benchmark Williams-Otto semi-batch reactor.