(627b) Optimal Design and Control of Batch Distillation Columns

Batch distillation is maybe the oldest separation process used for separate liquid mixtures. It is used for the output in small and big scale of chemical products with high quality. The flexibility of the production arrangements can also cope with the fluctuations or rapid changes in demand. This operation is widely used for the production of fine chemicals and specialized products such as alcoholic beverages, essential oils, perfumes, pharmaceutical and petroleum products. [1]

The main advantages of batch distillation over a continuous distillation lie in the use of a single column as opposed to multiple columns and its flexible operation. Moreover, only one batch distillation column and only one sequence of operation are necessary to separate all the components in a mixture. [1]

Given the transient nature of this operation, to describe its behavior, a dynamic model is taken into account according to rigorous model (MESH) that includes energy and mass balance, equilibrium relationships and summation of composition. On the other hand, there are conventional and unconventional configurations, between the last ones are the inverted batch distillation column, the middle vessel batch distillation column, the multi-vessel batch distillation column, but for this work a conventional configuration is selected.

After implementation of the dynamic simulation without binaries variables done in GAMS, profiles obtained are compared with profiles extracted by simulation results in Aspen Batch Distillation®, in order to check the model used.

The optimization of batch distillation has been intensively studied in the literature, where overall objective is to determine the optimal strategy based on a given objective function and satisfy several constraints.  In particular, the goal is to determine the optimal reflux policy to obtain a specified quality of product. In the literature are three categories, the problem of maximize distillate, the problem of minimize time and the problem of maximize profit. [2] In addition, an optimal design had been integrated with the optimal operation; where those issues are optimized simultaneously.  

In this work a methodology of optimal design and control in batch distillation column is presented in order to maximize an economic objective function, which binaries variables are taken into account. That is done using a mixed integer dynamic optimization strategy [3], in special a MINLP solver in GAMSis used, which all continuous variables are discretized in order to obtain a simultaneous approach categorized as large-scale MINLP problem. [4] In addition, orthogonal collocation is implemented as discretization method, based on the work done by Hendro et al. [5] where they allude about the stability and accuracy of results so as improve to computational efficiency significantly. 

Once the optimal policies are obtained, these are tested with post-optimal analysis which basically consists of a sensitivity analysis when profiles are manipulated. Finally, a case of study of hydrocarbons mixture is presented in order to illustrate the methodology presented in this work.

References:

[1] I. M. Mujtaba, Batch Distillation: Design and Operation, London, UK.: Imperial College Press, 2004.

[2] A. Arias Barreto, I. Rodriguez-Donnis, V. Gerbaud y X. Joulia, «Optimization of Heterogeneous Batch Extractive Distillation,» Industrial & Engineering Chemistry Research, vol. 50, 2011.

[3] B. Chachuat, A. B. Singer y P. I. Barto, «Global Mixed-Integer Dynamic Optimization,» AIChE JOURNAL, vol. 51, p. 2235–2253, 3 June 2005.

[4] L. T. Biegler, «An overview of simultaneous strategies for dynamic optimization,» Chemical Engineering & Processing, vol. 46, pp. 1043-1053, 2007.

[5] P. L. Hendro, P. Hoo y G. Wozny, «Efficient simulation of Batch distillation Process by Using Orthogonal Collocation,» Chem. Eng. Technol., vol. 21, nº 11, pp. 853-862, 1998.