Juergen Hahn, Rensselaer Polytechnic Institute
While it is the general purpose of this project to develop a fundamental understanding of the dynamics exhibited by reactive distillation columns where competing reactions are taking place, one has to look at a specific system in order to use simulators for investigating a process. The process under investigation deals with the hydrogenation of benzene in a reformate stream. This operation is one of the key steps for the production of gasoline as benzene removal from the product stream is an EPA requirement. One approach which has found acceptance in practice is to hydrogenate benzene to cyclohexane. However, a problem arises as the catalyst used for the reaction is not selective for benzene, and toluene, which is present in the reformate stream in considerable quantities, will also be hydrogenated. However, toluene hydrogenation is undesirable as toluene has a high octane rating and should be retained in the final product. The solution to this problem is to use catalytic distillation and capitalize on the differing volatilities and resulting column concentration profiles of benzene and toluene. The column can be designed to have a reactive zone at a location where benzene is commonly found but where the concentration of toluene is negligible. As a result benzene will be hydrogenated while toluene remains intact, as it does not enter the reactive zone, thus meeting the environmental and economic objectives with a single unit operation.
Continuous reactive distillation columns are commonly designed based on specifications for steady state operation of the process. The dynamic aspects of the column are rarely taken into account for the design, and control configurations are determined from design principles for regular distillation columns. While this approach has proven successful for set point changes for many processes, it can lead to problems for disturbance rejection. The behavior exhibited by reactive distillation columns can be extremely sensitive to disturbances in the feed concentration, whereas regular distillation columns exhibit a lesser degree of sensitivity to this type of disturbance. This difference is due to the complexity introduced by the reaction phenomena in addition to the separation. This is a crucial aspect for benzene hydrogenation as the concentration of some of the main components in the feed can vary by 50% or more due to disturbances upstream from the column. The importance of addressing these disturbances is increased by the fact that changes in the feed are happening on a daily basis and a column operating under a feedback plus ratio control scheme, will take about 2-3 hours to rectify the upset conditions. This situation clearly presents a need for a detailed investigation into the dynamic behavior exhibited by reactive distillation columns for benzene hydrogenation and methodologies used for influencing the dynamics.
The work that has so far been conducted as part of this grant makes three contributions: (1) a detailed literature review has been conducted, (2) a rigorous model has been constructed which is able to validate the observed effects that disturbances have on column operations, and (3) different control schemes, with a particular emphasis on disturbance rejection have been investigated.
Four students have contributed to the project so far. Vishal Mahindrakar is a third year PhD student in my group and is the main person in charge of the modeling effort. He is supported by Arjun Bhadouria, a second year PhD student in my group. Wei (David) Dai, another third year PhD student in my group, has contributed his expertise on modeling and parameter estimation. Lastly, Dunie Navarro, an undergraduate student supported by a NSF-REU grant, has contributed to this effort. Dunie won the 1st poster price at the undergraduate research competition at Texas A&M for his poster “Dynamics and Control of Benzene Hydrogenation via Reactive Distillation”.
As far as impact of the grant is concerned, the topic of this work forms the core of Vishal Mahindrakar’s PhD thesis. The modeling efforts that we undertake has allowed my group to build up additional expertise for modeling separation systems. One indicator of this is that a post-doctoral research, supported by a CONACYT grant, decided to join my research group in October 2012 for a project involving modeling of distillation columns for separation of complex mixtures.
Summarizing, the grant PRF# 50978-ND9 will partially supported three PhD students (including the no-cost extension period), includes contributions from a post-doctoral researcher and an undergraduate student support by a NSF-REU grant. Furthermore, ACS-PRF support or partial support has so far been acknowledged in five journal papers and one conference paper.