Energy savings in distillation via identification of useful configurations
Abstract
Recent market and environmental forces require the rapid development of better and cheaper separation process solutions. Especially for multicomponent mixtures, there are several feasible separation process solutions differing significantly in cost and energy consumption in spite of carrying out the same overall process. Therefore a systematic method to identify and design optimal multicomponent separation sequences is needed instead of relying on the inventive activity of a few experienced engineers. Even for a commonly perceived "mature" technology such as distillation, until recently there has been an absence of systematic methods to (i) elucidate all possible separation configurations and to (ii) identify energy efficient candidates. This research aims to address these needs. In this work, we focus on the continuous distillation of non-azeotropic mixtures into n distinct composition final product streams. We develop a computationally efficient and easy-to-use mathematical framework to generate all the basic distillation configurations that use exactly (n-1) distillation columns to carry out this n-component separation. We extend the framework to generate all the additional distillation configurations with thermal coupling. We observe that the search space of distillation configurations grows very rapidly as the number of product streams increases. For instance, for a mixture to be separated into 4 product streams, we can choose from 18 basic configurations and 134 additional configurations with thermal coupling; while for a mixture to be separated into 8 product streams, we can choose from 15,767,207 basic configurations and 29,006,926,681 additional configurations with thermal coupling. The next challenge for a process engineer is to be able to quickly prune the search space to a handful of attractive energy efficient candidates that can be studied in greater detail. To this effect, we develop a quick screening optimization tool that identifies configurations with low heat duties. The heat duty requirements are estimated using the Underwood equations. Application of these frameworks has provided an array of distillation configurations that can potentially have up to 50% lower heat duty than the currently used distillation configuration for petroleum crude distillation. Since petroleum crude distillation is a highly energy-intensive process, these configurations have tremendous potential to improve the energy efficiency of a refinery. Furthermore, we provide evidence to disprove some conventional notions about thermally coupled configurations. We also describe some previously unknown distillation configurations that use less than (n-1) distillation columns for an n-component separation. We demonstrate that these novel distillation configurations have significantly lower heat duties than the currently known distillation configurations with less than (n-1) columns. Since these configurations have one or more distillation columns that produce sidestreams, a quick screening tool based on Underwood's equations cannot be used to estimate the heat duty requirements of these configurations. Therefore, we lay the foundation for equations analogous to Underwood's equations to estimate the minimum heat duty requirements of such distillation configurations.
Degree
Ph.D.
Advisors
Agrawal, Purdue University.
Subject Area
Chemical engineering|Energy
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.