Establishing a facility to measure the efficiency of structured packing under total reflux

Paquet, Emil Friedrich (2011-03)

Thesis (MScEng (Process Engineering))--University of Stellenbosch, 2011.

Thesis

ENGLISH ABSTRACT: Structured packing is often the preferred choice for column internals because of its low pressure drop and high efficiencies compared to that of trays and random packing. However, the mass transfer phenomena in these gas-liquid contacting devices is still not well understood, even though it is widely used in industry. A contributing factor to this is the lack of understanding and availability of experimental data in the open literature. These shortcomings complicate the design of a distillation column and make practical experience essential. There is thus a need for more experimental data, especially for packings where only limited information is available. The focus of this study was to establish a testing facility that can be used to measure the efficiency of structured packing under total reflux, and not to measure vast quantities of experimental data; the latter will be done in future. The facilities available at Stellenbosch University limited the internal diameter of the column to 0.2 m, which is sufficient to test higher surface area structured packings (≥350 m2/m3). The column is used with a thermosyphon reboiler that uses steam as the heating source and is equipped with a total condenser. Two sections are used for the packed bed that allow for a total packed height of 3.78 m (2x1.89 m). The column is set up to operate under total reflux and was designed to operate at pressures ranging from 0.3 to 1 bar abs, vapour flow rates of 0.73 – 3.65 (m/s) (kg/m3)0.5 and liquid flow rates of 5 – 25 m3/(m2.h). It was found that the 2-butanol/iso-butanol and the p-xylene/o-xylene systems are suitable test mixtures for this pilot plant setup. The VLE data from Kutsarov et al. (1993) and Zong et al. (1983) for p-xylene/o-xylene and 2-butanol/iso-butanol are thermodynamic consistent and was validated by VLE experiments done in this study. It was found that the experimentally obtained efficiency (HETP) and pressure drop data for Mellapak 350Y compared well with published results of Spiegel and Meier (1987). With regard to the predictive models, it was found that i) the SRP model predicted the HETP of Mellapak 350Y structured packing accurately in the pre-loading region and slightly over predicted the HETP in the loading region, whereas ii) the Delft model over predicted HETP and iii) the Billet and Schultes model under-predicted HETP under the entire tested range (i.e. over-predict efficiency). With regard to the pressure drop data i) the Billet model iv accurately predicted the pressure drop over the entire tested range, whereas ii) the SRP model accurately predicted the pressure drop in the pre-loading region and slightly over predicted the pressure drop in the loading region and iii) the Delft model over predicted the pressure drop over the entire range and followed an almost parallel trend to the results from the SRP model. It was also found that information in the field of mass transfer in a packed column is far from saturated, and there is a need for more experimental data and better understanding of the mass transfer phenomena in packed columns.

AFRIKAANSE OPSOMMING: Gestruktureerde pakking het ʼn laer drukval en ʼn hoër effektiwiteit in vergelyking met willekeurige pakkings en plate, en is daarom dikwels die voorkeur keuse vir pakkings materiaal in ʼn distilleer kolom. Die massa-oordrags verskynsels in hierdie gas-vloeistof kontaktors word egter nog nie goed verstaan nie, ten spyte van die grootskaalse aanwending in die nywerheid. ʼn Bydraende faktor is die tekort aan eksperimentele data in die ope literatuur. Die tekortkomings bemoeilik die ontwerp van distilleerkolomme en maak praktiese ervaring ʼn noodsaaklike vereiste. Daar is dus ʼn behoefte aan meer eksperimentele data veral vir pakkings waar daar min of geen data beskikbaar is nie. Die fokus van die studie was om ʼn toetsfasiliteit op te rig wat gebruik kan word om die effektiwiteit van gestruktureerde pakking onder totale terugvloei kondisies te bepaal, en dus nie om ‘n groot hoeveelheid data te genereer nie; laasgenoemde sal wel deel uitmaak van toekomstige studies. Die fasiliteite beskikbaar by die Universiteit van Stellenbosch het die binne diameter van die kolom beperk tot 0.2 m. Die diameter is voldoende om gestruktureerde pakkings met ʼn hoë oppervlakarea te toets byvoorbeeld pakkings met areas 350 m2/m3 en hoër. Die kolom gebruik ‘n verdamper (met stoom as energie bron) om die vloeistof te verdamp en ‘n totale kondensator (verkoel met verkoelingswater) om die damp te laat kondenseer. Twee seksies van 1.89 m elk word gebruik vir die gepakte bed en die kolom het dus ‘n totale pakkingshoogte van 3.78 m. Die kolom is opgestel vir totale terugvloei en is ontwerp om bedryf te word by drukke tussen 0.3 en 1 bar abs, damp snelhede van 0.73 tot 3.65 (m/s) (kg/m3)0.5 en vloeistof vloeitempo’s tussen 5 en 25 m3/(m2.h). 2-butanol/iso-butanol en p-xylene/o-xylene is gevind om geskik te wees as mengsels vir die toetsopstelling. Die damp-vloeistof fase-ewewig data van Kutsarov et al. (1993) en Zong et al. (1983) vir p-xylene/o-xylene and 2-butanol/iso-butanol is termodinamies konsistent en is gevalideer deur damp-vloeistof fase ewewig toetse in die studie. Daar is gevind dat die eksperimenteel bepaalde effektiwiteit en drukval data vir Mellapak 350Y pakking goed vergelyk met gepubliseerde data van Spiegel and Meier (1987). Die eksperimenteel bepaalde effektiwiteit data is met waardes van beskikbare modelle model vergelyk en daar is gevind dat: i) die SRP voorspel die effektiwiteit van Mellapak 350Y pakking akkuraat in die ‘pre-loading’ gebied maar toon afwykings van die eksperimentele data in die ‘loading’ gebied, ii) die Delft model voorspel ‘n hoër hoogte ekwivalent aan ‘n teoreties plaat (HETP) oor die hele gebied terwyl iii) die Billet en Schultes model weer ‘n laer HETP voorspel oor die hele gebied. Met betrekking toe die drukval data i) voorspel die Billet model die drukval akkuraat oor die hele gebied, ii) die SRP model voorspel die drukval korrek in die ‘pre-loading’ gebied maar begin afwyk van die eksperimentele data in die ‘loading’ gebied en iii) die Delft model voorspel groter waardes vir drukval oor die hele gebied en volg amper ʼn parallelle tendens met die SRP model. In die studie is daar gevind dat daarin die gebied van massa-oordrag nog ʼn tekort is aan eksperimentele data en daar baie navorsings geleenthede is.

Please refer to this item in SUNScholar by using the following persistent URL: http://hdl.handle.net/10019.1/6834
This item appears in the following collections: