(203ae) On the Attainable Region for Separator Networks

In this work the attainable region (AR) concept for separator networks is introduced for the first time. For separator unit models to which the IDEAS conceptual framework is applicable, it is shown that identification of AR boundary membership is equivalent to feasibility assessment of an infinite linear program (ILP). A number of important AR properties are then theoretically established, including AR convexity for fixed network inlet and outlet flowrate ratios, and representation of the AR in a concentration state space of reduced dimension. Finite dimensional approximations of the aforementioned ILP are then employed in creating increasingly accurate representations of the separator AR. Two case studies will be presented: (1) a 4-Dimensional AR for the vapor-liquid equilibrium based separation of a ternary azeotropic mixture, and (2) a 3-Dimensional AR for binary separation in which the outlet flowrate ratios are allowed to vary.