Shortcut Design of Extractive Distillation Columns

Extractive distillation is a common process for the separation of homogeneous azeotropic mixtures. In this process entrainer feed flowrate and reflux ratio of the extractive column represent the crucial design degrees of freedom which govern feasibility and operating cost. In this work feasibility and operational stability of the extractive column is related to the analysis of the nonlinear dynamics of the extractive section. A fully-automated shortcut design method for the simultaneous determination of minimum entrainer feed flowrate and minimum reflux ratio is presented. The application of this method to a ternary and a quaternary example is shown. INTRODUCTION Extractive distillation is commonly used to separate mixtures which display minimum boiling azeotropes. In the extractive column a heavy boiling entrainer is fed to a tray above the main feed stream. The entrainer facilitates the separation by interacting with the azeotropic mixture and altering the thermodynamic equilibrium in the extractive section of the column. Figure 1 displays the column configuration for a binary process feed. The heavy entrainer E preferably associates to component B and takes it down the column. Therefore a binary mixture of B and E is recovered in the bottom product whereas high purity A is obtained in the distillate product. Separation feasibility and process cost are characterized by two major parameters: entrainer feed flowrate and reflux ratio (or condenser and reboiler heat duties, see shaded degrees of freedom in Figure 1). In addition to minimum reflux, which limits feasibility for all zeotropic and azeotropic separations, there is a maximum reflux above which separation cannot be achieved. These bounds for the reflux ratio depend on the entrainer feed flowrate. The range of feasible choices for the reflux ratio decreases with decreasing entrainer flowrates. Below a minimum entrainer flowrate the extractive effect is no longer sufficient for separation and a feasible reflux policy cannot be found. DESIGN OF EXTRACTIVE DISTILLATION COLUMNS The standard approach for process design of such separation processes usually involves the detailed specification of all relevant design parameters: the number of trays, the location of process and entrainer feed trays and the reflux ratio (or either condenser or reboiler heat duty, see Figure 1). This column configuration is analyzed using a standard process simulator such as Aspen+ [1] or Hysys [2]. The design engineer is left with the difficult task to iteratively change the design variables and repeat this process until all design constraints like required purities etc. are met and convergence is achieved. Such simulation-based design is a good method to specify all relevant design variables, however it is also a rather tedious and time-consuming undertaking. In the early stages of process synthesis a large number of different design alternatives can be formulated. A good example is the purification of a binary alcoholwater mixture which can be achieved by heteroazeotropic distillation using a heterogeneous entrainer [3], extractive distillation using a homogeneous entrainer [4] or by using a hybrid membrane-distillation configuration [5]. For both the homogeneous and the heterogeneous azeotropic distillation processes the choice of the entrainer, usually from a list of potential candidates, defines the structure of the process and is crucial to the economic performance. Then the number of structurally different design alternatives easily grows too large to be effectively tackled using simulation tools. Therefore there is a need for simple and fast algorithms that support process screening and provide an estimate of economic potential. Shortcut methods for the determination of the minimum energy demand of distillation are well suited for this task because they allow fast evaluation of the separation without needing detailed unit information and present valuable insight into the thermodynamic limitations of the mixture. For the separation of ideal mixtures the shortcut method of Underwood [6] has become a standard tool for process design. In the last 20 years several research groups have proposed geometric criteria for the determination of the minimum energy demand of nonideal distillation. Bausa et al. [7] present a critical review of A,B A