Adsorptive separations in the production of neutral wine alcohol

Goliath, Elroy Mario (2002-03)

Thesis (PhD)--University of Stellenbosch, 2002.

Thesis

ENGLISH ABSTRACT: This study describes the design, construction, complete industrialisation and operation of a dual bed vacuum swing adsorption (VSA) demonstration plant, which operates at atmospheric and sub-atmospheric conditions. All design objectives as set out initially were met. The plant removes contaminants such as methanol and water from neutral wine spirit. Neutral wine spirit is a key component of various local and international spirituous products which include liqueurs, gin, vodka, fortified wines and brandy. Neutral wine spirit can chemically be described as the azeotropic mixture of ethanol and water, which occurs at an ethanol content of 96.4 vol. %. Methanol is naturally present in all products from the vine. Fermentation and distillation concentrate methanol even more, and due to physical and chemical characteristics, its separation consumes as much as 45 % of total production costs. Neutral wine spirit is produced by the proven technology of continuous atmospheric distillation. Continuous improvement of the distillation process is limited due to the physical constraints of an old facility, but also due to previous design philosophies and approaches. The VSA plant consists of two adsorbers, packed to a total height of 1.71 m and a diameter of 0.4 m. Adsorption took place at 100 °C and regeneration at the same bed temperature with purified nitrogen gas at 170 °C and a vacuum of 17 kPa (abs). Experiments were divided into Group I and Group II experiments. Group I investigated the ability to separate methanol and water from the azeotrope and to which efficiency it occurred. It consisted of 120 adsorption cycles of 5 minutes each and 60 samples were drawn for analyses. Breakthrough was not allowed to occur. The azeotropic feed was consistently dehydrated to a water content < 0.05 wt %, while methanol was reduced to < 4 mg/100mLAA. The type of 3A molecular sieve (MS 564 CS) was specifically selected to ensure analytic as well as organoleptic compliance with the product specification. Molecular sieve 4A was removed due to organoleptic problems with the product. Group" experiments were performed in the format of a sensitivity analysis. The effects of various process parameters on the methanol breakthrough curves were individually assessed. Eighteen experiments were performed over a period of 8 days, with 86 samples drawn. The duration of an adsorption cycle was 30 minutes, allowing methanol breakthrough to occur. Water was preferentially adsorbed. Negative methanol bed loadings during high water loadings confirmed that water was able to displace methanol molecules. In the presence of water, molecular sieve 3A was capable of adsorbing 0.6 mg methanol/100mLAA, while in the absence of water with synthetically dosed methanol, molecular sieve 3A achieved a maximum loading of 12.3 mg methanol/100mLAA. The latter corresponded with a maximum methanol feed content of 1118 mg/100mLAA. In general, quicker breakthrough occurred at higher flow rates and feed concentrations. Continuous breakthrough caused bed contamination and a 24-hour thermal regeneration was performed following experiment 12. The feed flow rate was increased from the theoretical 50 f/hr to 70 f/hr without any additional capital layout. Selected process conditions were found to be effective in continuously separating methanol from ethanol. Depending on the strategy of integration, profitability studies shows a Return on Investment of between 110.1% - 220.8% for the adsorption project. Adsorption is superior to distillation in the separation of methanol. Due to the level of innovation involved, it is recommended that the contents of this study remain confidential and patent protection is to be extended. This dissertation speaks to both the wine making as well as the chemical engineering fraternity. It seeks to provide credibility to both parties, by clarifying the unknown issues fundamental to the respective disciplines.

AFRIKAANSE OPSOMMING: Hierdie studie definieer die ontwerp, vervaardiging en volledige industrialisasie van 'n dubbelbed vakuum adsorpsie demonstrasie aanleg (VSA) wat by atmosferiese en sub-atmosferiese kondisies bedryf word. Alle ontwerpsdoelwitte is bereik. Die aanleg verwyder selektief metanol en water vanuit neutrale wynspiritus. Neutrale wynspiritus is 'n sleutelkomponent van verskeie spiritualieë in die plaaslike en internasionale wyn en spiritus bedryf. Hierdie produkte sluit in likeurs, jenewer, vodka, gefortifiseerde wyne en brandewyn. Chemies, kan neutrale wynspiritus beskryf word as die azeotropiese mengsel van etanol en water teen 96.4 vol. %. Metanol het 'n natuurlike teenwoordigheid in alle produkte vanaf die wynstok. Gisting en distillasie konsentreer metanol tot 'n hoër mate en weens fisiese en chemiese eienskappe word metanol teen hoë koste vanaf die etanol stroom geskei. Die metanol verwyderingskomponent beloop soveel as 45 % van die produksiekoste van die totale proses. Neutrale wynspiritus word deur die gevestigde tegnologie van kontinue atmosferiese distillasie geproduseer. Kontinue verbetering van die proses word beperk deur die fisiese ouderdom en toestand van die fasiliteite, maar is ongelukkig ook die resultaat van vorige ontwerpsfilosofieë en benaderings. Die adsorbeerders is gepak tot 'n hoogte van 1.71 m met 'n deursnit van 0.4 m. Adsorpsie het by 100°C plaasgevind en regenerasie by dieselfde bedtemperatuur met stikstofgas by 170°C en 'n vakuum van 17 kPa (abs). Eksperimentele werk is in Groep I en Groep II eksperimente verdeel. Groep I het die effektiewe prosesvermoë om metanol en water vanuit die azeotroop te verwyder ondersoek. Dit het bestaan uit 120 adsorpsie siklusse van 5 minute elk. Sestig monsters is getrek vir analise. Deurbreek van metanol was nie toe gelaat om plaas te vind nie. Die azeotropiese toevoer is konsekwent tot 'n water inhoud < 0.05 massa % gedehidrateer is. 'n Metanol inhoud < 4 mg/100 mLAA is bereik. Die tipe 3A molekulêre sif (MS 564 CS) was spesifiek vir die toepassing geselekteer om sodoende 'n analities sowel as organolepties aanvaarbare produk te lewer. Molekulêre sif 4A was verwyder weens die vorming van produk wangeure. Groep II eksperimente is in die vorm van 'n sensitiwiteits analise uitgevoer. Die effek van verskeie veranderlikes is individueel op die metanol deurbreekkurwe getoets. Agtien eksperimente is oor 'n tydperk van 8 dae gedoen, met 'n totaal van 86 monsters wat getrek is. 'n Adsorpsie siklus het 30 minute geduur en het deurbreek van metanol toegelaat. Water is by voorkeur geadsorbeer. Negatiewe metanol bed ladings tydens hoë waterteenwoordigheid toon dat water wel metanolmolekule op 'n adsorpsie-posisie kan verplaas. In die teenwoordigheid van water is 'n bedlading van 0.6 mg metanol/100mLAA verkry, met 'n maksimum van 12.3 mg metanol/100mLAA in die afwesigheid van water. Laasgenoemde verteenwoordig 'n toevoer met 'n metanol inhoud van 1118 mg/100mLAA. In die algemeen is gevind dat 'n toename in toevoer vloeitempo en konsentrasie die tyd vir deurbreek verkort. Kontinue deurbreek het kontaminasie van die bed teweeg gebring en 'n termiese regenerasie is vir 24 ure na eksperiment 12 gedoen. Die teoretiese ontwerps vloeitempo was 50 .elhr, maar resultate het getoon dat die aanleg tot 70 .t'/hr kan verwerk sonder addisionele koste. Die geselekteerde proseskondisies was effektief in die versekering van die kontinue skeiding van metanol en etanol. Die lewensvatbaarheidstudie toon, afhangende van die strategie van integrasie, 'n Opbrengs op Belegging van tussen 110.1% - 220.8%. Adsorpsie het 'n beter skeidingsvermoë as konvensionele distillasie vir die verwydering van metanol vanuit etanol. Weens die vlak van innovasie betrokke, word dit voorgestel dat die inhoud van hierdie studie vertroulik gehou word en dat patent beskerming verleng sal word. Hierdie verhandeling spreek tot beide die wynmaak sowel as chemiese ingenieurs dissiplines. Daar is gepoog om geloofwaardigheid vir beide partye te skep deur die onbekende aspekte van albei dissiplines aan te spreek.

Please refer to this item in SUNScholar by using the following persistent URL: http://hdl.handle.net/10019.1/52634
This item appears in the following collections: