Conceptual Modeling and Referential Control Applied To Batch Distillations

Abstract

This paper proposes the combined use of a conceptual modeling method and a referential control strategy to operate batch distillation columns. Using the pinch theory, the modeling method, which is valid for multicomponent mixtures, allows the derivation of a quasi-optimal recipe to guide de operation properly. The referential control implies the empirical determination of reduced-order models for designing and tuning feedback controllers dedicated to tracking non-stationary conditions in batch processes. While defining a feasible recipe implies adjusting the reference trajectory until the desired product purity and recovery is achieved, defining the feedback control system implies first the selection of an appropriate tray temperature evolution, and then the isolation of the dynamics associated to the manipulated variable from the main time-variable behavior desired for the operation. The effectiveness of this combined procedure is illustrated through rigorous simulations where the light component is recovered from binary and ternary mixtures of alcohols.