Thesis (MEng) -- Stellenbosch University, 2014. / ENGLISH ABSTRACT: Supercritical fluids are enjoying ever increasing popularity as a solvent medium for extraction, stripping and absorption processes. Being readily tuneable and able to achieve sharp, highly efficient separations, supercritical fluids present an attractive alternative to traditional solvents, while using less intrinsically harmful compounds. Although the potential of supercritical fluids as solvents have been known for more than a century, there are still several areas of uncertainty, one being the hydrodynamics of extraction columns operating under supercritical conditions. This shortcoming can be attributed to the satisfactory performance of modified standard hydrodynamics to approximate column design, along with a predominant culture of overdesign in process engineering. Even though modified subcritical hydrodynamic models provide a good approximation they do not successfully predict the effect of changes in density, viscosity and surface tension of a supercritical fluid, leading to inaccuracies in column design. This study investigates the state of hydrodynamics under supercritical conditions in counter current packed columns discussed in literature, identifies shortcomings in existing literature and devises a way of addressing the said shortcomings. The primary objective of this study is to establish a multipurpose supercritical pilot plant capable of measuring hydrodynamics under supercritical conditions, followed by the secondary objective of measuring preliminary hydrodynamic data to prove the plant can deliver on its design requirements in measuring reliable hydrodynamic data. During a survey of available literature it was found that very little experimental work has been performed on hydrodynamics under supercritical conditions and especially on random packings. Further it is found that the systems investigated in literature were conducted under conditions of significant mass transfer. As mass transfer directly affects flow rates and fluid properties of the fluids in the column, it is vital to use systems with very little to no mass transfer. This ensures the most accurate approach possible when investigating fundamental hydrodynamic behaviour. Finally it was found that there are no well-defined correlations available for a wide range of packings, fluid properties and hydrodynamic phenomena for columns under supercritical conditions. To remedy the shortcomings in hydrodynamic data it was decided that more pilot plant work is required. It was found that no pilot plants available can measure hydrodynamic data. An investigation was performed into retrofitting available pilot plants, plants used by other research groups and commercially available plants. It was concluded that the best option was to salvage the major parts of an existing old pilot plant and use them to construct a new, customized pilot plant. This provides the opportunity of constructing a custom, multipurpose pilot plant capable of use in future research. After an initial concept design a full design of the new pilot plant was performed. The plant consists of two columns of 17 mm and 38 mm inside diameter and 3.5 m and 1.5 m packed height, respectively, and is capable of pressures and temperatures of up to 300 bar and 200°C. Furthermore the pilot plant can measure liquid hold-up, pressure drop, flooding and entrainment in accordance with the objective of measuring supercritical hydrodynamic data. Liquid hold-up was determined by stopping the process and allowing the column to drain, after which the volume drained was measured. To measure the pressure drop an Endress+Hauser Deltabar S PMD75 DP cell was used. Flooding was determined using the measured pressure drop and volumetric rate of column overheads, from where a hydrodynamically inoperable state is defined. Overall entrainment, although unlikely due to the presence of a demister in the column, was investigated by comparing the column overheads to literature phase equilibria. Preliminary hydrodynamic testing was performed using the 38mm diameter column packed with 1/4” Dixon rings. Testing is performed with at 120 bar and 40°C with a CO2 supercritical phase and polyethylene glycol liquid phase with an average molar mass of 400 (PEG 400). The hydrodynamic data gathered showed expected trends, but showed discrepancy with literature due to differences in liquids used, column packing and experimental system between the respective studies. / AFRIKAANSE OPSOMMING: Superkritiese vloeistowwe is besig om toenemende gewildheid as 'n oplosmiddel vir ekstraksie, stroping en absorpsie prosesse te geniet. Hierdie gewildheid is as gevolg van ʼn vermoë om skerp, hoogs effektiewe skeidings te bewerkstellig deur gebruik te maak van ʼn maklik aanpasbare oplosmiddel wat minder intrinsiek skadelik is as tradisionele oplosmiddels. Hierdie voordele lei daartoe dat superkritiese vloeiers as ʼn aantreklike alternatief tot tradisionele oplosmiddels gesien kan word. Alhoewel die potensiaal van superkritiese vloeistowwe as oplosmiddels al vir meer as ʼn eeu bekend is, is nog weinig eksperimentele werk al gedoen oor die hidrodinamiese gedrag van superkritiese gepakte kolomme. Hierdie tekortkoming kan toegeskryf word tot die bevredigende prestasie van aangepaste standaard hidrodinamiese korrelasies gedurende superkritiese kolomontwerp en ʼn oorheersende kultuur van oorontwerp in proses-ingenieurswese. Alhoewel aangepaste standaard hidrodinamiese korrelasies ʼn aanvaarbare benadering bied, beeld dit nie die effek van die veranderde digtheid, viskositeit en oppervlakspanning van ʼn superkritiese vloeistof uit nie, wat lei tot foute in kolomontwerp. Hierdie studie ondersoek die stand van superkritiese hidrodinamika in literatuur, spesifiek in teenstroom gepakte kolomme. Tekortkominge in die bestaande literatuur is geïdentifiseer en 'n metode om die genoemde tekortkominge reg te stel is bedink. Die primêre doel van hierdie studie is om 'n veeldoelige superkritiese loodsaanleg te bou wat tot staat is om superkritiese hidrodinamika te meet, gevolg deur die sekondêre doelwit wat die meet van voorlopige hidrodinamiese data behels, wat sal bewys dat die loodsaanleg voldoen aan ontwerpsvereistes. Tydens 'n opname van beskikbare literatuur was daar gevind dat weinig eksperimentele werk al gedoen is in die veld van superkritiese hidrodinamika, en nog minder oor sogenoemde ongeordende of ‘random’ kolompakkings. Verder is daar gevind dat eksperimente uitgevoer in literatuur slegs bestaan uit stelsels waar beduidende massa-oordrag plaasvind. Aangesien massa-oordrag die vloeitempo en fisiese eienskappe van die vloeiers in ʼn kolom direk beïnvloed, is dit noodsaaklik om gebruik te maak van stelsels met baie min of geen massaoordrag. Dit verseker ʼn akkurate benadering tot die meet van fundamentele hidrodinamiese gedrag. Laastens is gevind dat daar geen hidrodinamiese korrelasies beskikbaar is wat ʼn wye verskeidenheid van kolompakkings, vloeier eienskappe en hidrodinamiese verskynsels onder superkritiese toestande dek nie. Om die tekortkominge in superkritiese hidrodinamika in literatuur aan te spreek, word meer eksperimentele loodsaanlegwerk vereis. Daar is gevind dat geen van die beskikbare loodsaanlegte hidrodinamiese data kan meet nie. Ondersoek is ingestel tot die ombouing van bestaande loodsaanlegte, aanlegte wat gebruik is deur ander navorsingsgroepe en kommersieel beskikbare aanlegte. Daar is tot die gevolgtrekking gekom dat die beste opsie is om ʼn nuwe loodsaanleg self te bou en gebruik te maak van parte uit een van die ou bestaande aanlegte om kostes laag te hou. Sodoende kan ʼn veeldoelige, pasgemaakte loodsaanleg gebou word wat ook vir toekomstige navorsing gebruik kan word. Na ʼn aanvanklike konsep ontwerp vir die nuwe loodsaanleg, is ʼn volledige ontwerp gedoen. Die aanleg bestaan uit twee kolomme van onderskeidelik 17 mm en 38 mm binnedeursnee en 3,5 m en 1,5 m gepakte hoogte, en is in staat om eksperimente by ʼn maksimum druk en temperatuur van tot 300 bar en 200°C uit te voer. Verder is die loodsaanleg in staat daartoe hidrodinamiese data te meet, naamlik die vloeistofophoud in die kolom, drukval oor die kolompakking, kolomvloed en druppel meesleuring. Die vloeistofophoud in die kolom is bepaal deur alle voer tot die kolom te stop en tyd toe te laat vir die vloeistof om te dreineer, waarna die gedreineerde afgetap en gemeet is. Om die drukval te meet word ʼn Endress+Hauser Deltabar S PMD75 DP sel gebruik. Kolomvloed is bepaal met behulp van die drukval oor die kolom en die vloeitempo van die kolom se boonste produkstroom, van waar ʼn hidrodinamies onbruikbare toestand gedefinieer word. Algehele druppel meesleuring, alhoewel onwaarskynlik as gevolg van die teenwoordigheid ʼn ontwasemer in die kolom, is wel ondersoek deur die vloeistofinhoud in die kolom se boonste produkstroom te vergelyk met fase ewewigsdata in literatuur. Voorlopige hidrodinamiese eksperimente is uitgevoer met behulp van die 38mm deursnee kolom gepak met 1/4 " Dixon ringe. Eksperimente is uitgevoer by 120 bar en 40 ° C met 'n CO2 kritiese fase en 'n poliëtileenglikol vloeistof fase met ʼn gemiddelde molêre massa van 400 (PEG 400). Die hidrodinamiese data het verwagte tendense getoon, maar diskrepansies met literatuur waardes. Die verskille tussen die eksperimentele en literatuur data word geregverdig deur die verskille tussen die vloeistowwe, pakking en eksperimentele stelsels wat in die onderskeie studies gebruik is.