Chemical Engineering and Processing 48 (2009) 1199–1211 Contents lists available at ScienceDirect Chemical Engineering and Processing: Process Intensification journal homepage: www.elsevier.com/locate/cep Modeling of PSA separation process including friction pressure drop in adsorbent bed Krzysztof Kupiec a,∗ , Jan Rakoczy a , Erwin Lalik b a Cracow University of Technology, Faculty of Chemical Engineering and Technology, ul. Warszawska 24, 31-155 Kraków, Poland b Institute of Catalysis and Surface Chemistry, Polish Academy of Science, ul. Niezapominajek 8, 30-239 Kraków, Poland article info Article history: Received 19 February 2009 Accepted 14 April 2009 Available online 3 May 2009 Keywords: Friction pressure drop Ethanol dehydration Pressure swing adsorption abstract A sequential process of pressure swing adsorption–desorption has been investigated experimentally and numerically. Two mathematical models have been developed. The exact model includes the momentum balance, apart from the mass balances, whereas the lumped model assumes that both individual steps of the process are isobaric. Experimental separation of a gaseous mixture of water and ethanol over a zeolitic adsorbent has been carried out. Regeneration of the adsorbent was performed under lowered pressure by purging the bed with unhydrated ethanol. The predictions of both models were compared with the experimental results and a satisfactory agreement was found only in case of the exact model. The lumped model tended to overestimate the efficiency of regeneration of bed during the purge step. It was concluded that, under conditions applied in measurements, a model adequately representing the process should include the momentum balance because a pressure drop arising over the purge step is significant in relation to the total pressure. © 2009 Elsevier B.V. All rights reserved. 1. Introduction Adsorption is often a method of choice for the industrial gas separation because it combines low operation costs with high effec- tiveness, as well as offering a wide variety of sorbents that can ensure highly selective performance for a number of gaseous mix- tures that are in use within the industry. The process is sequential: following the adsorption, the sorbent has to be regenerated by removal of the retained component. The regeneration is carried out under a pressure lower than that of adsorption; afterwards the sorbent is ready for the next adsorption–desorption cycle. Analysis of the processes based on adsorption–desorption cycles usually involves numerical simulation using mathematical models that differ widely in their degrees of complexity. In compromising between simplicity of the easily solvable models, and complexity of the very time-consuming ones, one has to select simplifying assumptions very carefully. In doing so it is necessary to account for individual conditions under which a process is being run. It is obviously worthwhile to apply every approximation that may be possible under given circumstances. However it should be borne in mind that an oversimplification can cause the model to lose its relevance to industrial reality and to yield useless results as a consequence. ∗ Corresponding author. E-mail address: kkupiec@chemia.pk.edu.pl (K. Kupiec). The equations applied in chemical engineering usually include the mass balances as well as the heat balances, if the thermal effects accompanying the process are significant. The momentum balance equations however are rarely involved because it is often assumed that a process is isobaric. This may often be justifiable; however in cases of low pressure processes that involve a gas flowing through a pellet bed, the constant pressure assumption is rather problem- atic. A criterion of whether a process can be considered isobaric is in these cases the ratio of the pressure drop in the gas phase (over the entire bed) to the pressure of the gas at the inlet to the bed. If this ratio is close to zero, the process can be considered to be isobaric. Conversely, if the pressure drop is comparable to the inlet pressure, the isobaric assumption is unacceptable, and the model should be extended by including the momentum bal- ance. Technically, it causes additional complications in modeling due to including an extra variable. This may be worth the effort, however, because departure from isobaricity appears to affect both the adsorption and, to greater extent, the desorption stage of the process. Generally, the decrease of the total pressure (caused by the pressure drop over the bed) results in a decrease of the partial pressure of the adsorbed component. For the adsorption stage, it leads to a decrease of the equilibrium content of this component within pellets, which amounts to a decrease of the adsorption driv- ing force, and thereby results in a slow down of its intensity. For desorption, on the contrary, the decrease of equilibrium content in pellets actually enhances the process. The problem of accounting for the pressure drop of gas as it flows through the pellet bed has been a subject of several studies, most of 0255-2701/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.cep.2009.04.009