The dividing wall distillation column has a larger number of design variables than a conventional column. For design of the column, it will be desirable to define a priori the feasible space over which all the designs lie. An attempt has been made in this paper to address this problem through a graphical representation of all the possible dividing wall column (DWC) designs for a specified separation of a ternary feed. The development of the theory is based on splitting the dividing wall column into three simple columns (a prefractionator and two downstream columns) and applying the shortcut methods of Fenske, Underwood, Gilliland and Kirkbride. For specified terminal product compositions, it is shown that the design space can be constructed on a 3-dimensional plot, the axes being the flow rates of two of the components in the ’net distillate’ from the prefractionator (dividing wall column being representable as a Petlyuk system), and the effective reflux ratio of the prefractionator. For ease of graphical representation, the designs will be projected on to a 2 dimensional space of prefractionator output flow rate variables for a fixed prefractionator reflux ratio. Constraints related to the availability of feed components to downstream columns, infeasible reflux ratio and imbalance in plate assignment on either side of the wall are also placed on the 2 dimensional design space to generate a feasible design space. On this design space, enveloped by various constraints, various equi-parameter curves are drawn depicting locus of points on which the chosen parameter has a constant value. The parameter chosen can be either the total number of column plates or the number of plates above/below the dividing wall, reboiler duty, or the cost. The design space proposed in this paper, even though it uses the shortcut methods, provides the designer with a broad view of what all designs are available, out of which some attractive options may be explored further. The location of equi-cost or equi-energy curves assist the designer in identifying design changes which could lead to either decreased cost or decreased energy.
Read full abstract