- 12:30 PM

Dynamics and Control of Vapor Recompression Distillation

Sujit S. Jogwar1, Michael Baldea2, and Prodromos Daoutidis1. (1) Chemical Engineering and Materials Science, University of Minnesota, 421 Washington Ave SE, 151 Amundson Hall, Minneapolis, MN 55455, (2) Praxair, Inc., 175 East Park Drive, Tonawanda, NY 14150

Distillation is one of the most energy consuming units in a chemical plant, motivating the need for energy integration. Vapor recompression distillation is one such energy integrated distillation configuration, wherein the vapor coming from the top of the distillation column is used to reboil the bottoms stream. Vapor recompression distillation is favored for the separations involving close-boiling liquids. Such separations result into large reflux ratios and small compressor duty is needed to facilitate heat transfer in the reboiler-condenser.

Most research on vapor recompression distillation has emphasized steady state economics focusing on capital costs, operating costs and optimal steady state operating conditions. Control objectives in these systems are the control of exit concentrations, pressure and holdups of various units. These integrated systems show strong control-loop interactions, because of the thermal coupling between units. However, most control studies have been within a multi-loop linear control framework. In this talk, we propose a comprehensive modeling analysis and control framework for such columns. Specifically, we illustrate that the discrepancies in material and energy flows in the column lead to a multi-time scale behavior. Large material recycle owing to large reflux ratio leads to time scale separation in the material balance dynamics. Energy recycle through the reboiler-condenser is dominated by latent heat effects, whose contribution to enthalpy is larger compared to sensible heat effects. This additional stiffness leads to further time scale separation in the energy dynamics, leading to overall three-time scale dynamic behavior.

Using singular perturbation arguments, reduced order non-stiff models valid in each time scale are obtained, which can be used for controller design. This analysis also leads to a classification of manipulated variables available in each time scale. The theoretical results will be illustrated via a simulation case study on a propane-propylene system.