Design and optimization of HIDiC columns using a constrained Boltzmann-based estimation of distribution algorithm-evaluating the effect of relative volatility

In this paper we study the design and optimization of heat-integrated distillation columns (HIDiC) using a constrained Boltzmann-based estimation of distribution algorithm. The total annual cost was defined as the fitness function of the problem, and the effect of the relative volatility on the performance of the HIDiC configurations was evaluated. We analyzed a set of equimolar mixtures with relative volatility varying from 1.12 to 2.4. The designoptimization strategy implemented was executed through Matlab-Excel-Aspen Plus interface. The results showed that the best interrelation between energy savings and TAC of the HIDiC columns were obtained in the separation of mixtures with relative volatilities between 1.12 and 1.5, particularly for mixtures of isomers. As a result, energy savings from 50 to 87% were obtained for this class of mixture. At the same time, we obtained HIDiC sequences with TACs lower than that TAC of conventional columns (2–16%). Nevertheless, we found that the TAC of some HIDiC schemes resulted higher than the TAC of conventional columns (6–17%). In addition, carbon dioxide emissions released into the atmosphere were reduced in the same proportion as savings in energy required to fuel the reboilers. ã 2016 Elsevier B.V. All rights reserved.