Fractionation of tyre derived oil

Date
2017-03
Journal Title
Journal ISSN
Volume Title
Publisher
Stellenbosch : Stellenbosch University
Abstract
ENGLISH SUMMARY: Tyre derived oil (TDO) has shown to be a source of dl-limonene, a valuable chemical compound with a diverse industrial application. There has been an ever increasing interest in the recovery of dl-limonene from TDO. Separation of dl-limonene from TDO by ordinary distillation poses a challenge due to the presence of other chemical compounds with similar properties to dl-limonene. These compounds are impurities and need to be removed. Before dl-limonene can be sold as commercial product it needs to be purified in excess of 90wt%. The main objective of this work is to develop process models using Aspen Plus ® V8.2 to separate dl-limonene from TDO at a sufficient purity, validate the model with experimental data and analyse the economic viability of the developed processes. As distillation cannot achieve the desired separation, enhanced distillation was adopted in this study as it has the ability to separate compounds with similar properties. This thesis investigates the production of high purity (> 90 wt%) dl-limonene from TDO using extractive and azeotropic distillation. A selection of candidate entrainers was identified. The entrainers investigated include diethylene glycol, triethylene glycol, n,n-dimethylformamide, n-methyl-2-pyrrolidone, quinoline, 4-formylmorpholine and tetratethylene glycol dimethyl ether. The thermodynamic models used in the simulation include the non-random two-liquid (NRTL) and the universal quasichemical activity coefficient (UNIQUAC) model. The missing thermodynamic properties are predicted by the universal functional activity coefficient (UNIFAC) model. The process models developed in Aspen Plus® V8.2 using various entrainers in extractive and azeotropic distillation showed that it is possible to obtain dl-limonene recoveries in excess of 95% and at purities as high as 99wt%. Experimental verification of the modelling results was conducted in a batch distillation setup at a pressure of 60 kPa and entrainer to feed ratio of 2, 4 and 6. Aspen Plus ® V8.2 results were compared with experimental data. The ability of the thermodynamic models to predict all sets of experimental data was determined by calculating percentage error. It was found that none of the models could predict experimental data with acceptable accuracy. This shows there is a need for experimental determination of detailed vapour-liquid equilibrium (VLE) and vapour-liquid-liquid equilibrium (VLLE) data for binary and ternary systems (chemical compounds in TDO and entrainers). A techno economic analysis was conducted to investigate the economic viability of different process models developed. From an economic point of view, four out of seven process options developed using different entrainers proved to be profitable. The best process in terms of economic performance was found to be extractive distillation using tetraethylene glycol dimethyl ether with a payback period of 1.23 years and a discounted cash flow rate on return (DCFROR) of 83.21%. The worst process option was found to be azeotropic distillation using n-methyl-2-pyrrolidone with a payback period of greater than 15 years and a DCFROR of 7.72%. As the thermodynamic models do not accurately predict the phase equilibrium data, the economic analysis conducted is not of the desired level of accuracy. Nevertheless, the separation of dl-limonene from TDO using azeotropic/extractive distillation process has shown to be a potential method.
AFRIKAANS OPSOMMING: Dis bevind dat die olie wat vanaf motor bande (MBO) verkry kan word, as ʼn bron dien vir dl limonene wat ʼn waardevolle chemiese verbinding is met verskeie industriële toepassings. Daar is ‘n toenemende belangstelling om dl-limonene van MBO te herwin. Skeiding van dl limonene van MBO deur standaarde distillasie is ‘n uitdaging omdat die ander chemiese verbindings wat teenwoordig is soortgelyke eienskappe as dl-limonene het. Hierdie verbindings is onsuiwerhede en die verwydering daarvan is van kardinale belang. dl-Limonene het slegs kommersiële waarde tensy dit ʼn suiwerheid van 90%(m/m) of meer het. Die hoofdoel van hierdie studie is om proses modelle te ontwikkel wat dl-limonene van MBO skei tot ʼn voldoende suiwerheid, bevestiging van die modelle deur van eksperimentele data gebruik te maak asook om die ekonomiese vatbaarheid van die ontwikkelde proses te analiseer. Standaarde distillasie skeidingstegnieke kan nie alle vloeistof mengsels skei nie. Daarom was gevorderde distillasie gekies omdat dit nie net die kapasiteit het om verbindings te skei met soortgelyke eienskappe, maar ook omdat dit ʼn breë operasionele band het. Hierdie verslag bespreek die skeidings prosedure van hierdie komponente deur ekstraktiewe en azeotropiese distillasie om sodoende die suiwerheid van dl-limonene te verbeter. Ekstraktiewe en azeotropiese distillasie behels die gebruik van ʼn massa skeidingsagent om die skeiding van die vloeistof mengsel te fasiliteer wat nie haalbaar is deur standaard distillasie. ‘n Lys van moontlike skeidingsagente was ondersoek. Hierdie sluit in diëtileen glikol, triëtileen glikol, n,n-dimetielformamiede, n-metiel-2-pyrrolidon, kinolien, 4-formylmorfoline. Aspen Plus® V8.2 was gebruik om die skeidings proses te modelleer. Die termodinamiese modelle wat in die simulasie gebruik was, is die Non-Random Two Liquid (NRTL) model en die Universal quasi chemical activity coefficient (UNIQUAC) model. Die onbekende termodinamiese eienskappe was bepaal deur gebruik te maak van die Universal functional activity coeffiecient (UNIFAC) model. Die proses modelle wat in Aspen Plus® V8.2 ontwikkel was, wys dat dit moontlik is om minstens 95% van die dl-limonene te herwin met ʼn suiwerheid van so hoog as 99% (m/m). Eksperimentele werk was uitgevoer in ‘n enkellade distillasie opstelling wat by ‘n druk van 60 kPa bedryf was met ʼn skeidingsagent tot toevoer verhouding van 2, 4 en 6 onderskeidelik. Aspen Plus® V8.2 resultate was vergelyk met die eksperimentele data. ‘n Persentasie fout was bereken om die akkuraatheid van die gebruik van termodinamiese modelle om eksperimentele data te voorspel, te bepaal. Dit was gevind dat nie een van die modelle gebruik kan word nie om eksperimentele data te voorspel met aanvaarbare akkuraatheid. Hierdie dryf dus die nood vir eksperimentele data vir damp-vloeistof ewewigsdata (DVE) en damp-vloeistof-vloeistof ewewigsdata (DVVE) van binêre en drieledige sisteme (chemiese verbindings in MBO en skeidingsagent). ʼn Tegno-ekonomiese analise was uitgevoer om die ekonomiese vatbaarheid van verskillende proses modelle te ondersoek. Vanuit ʼn ekonomiese standpunt, vier uit sewe proses keuses wat verskillende skeidingsagente gebruik blyk om meer winsgewend te wees. Die beste proses opsie in terme van ekonomiese prestasie was ekstraktiewe distillasie met ʼn terugbetaalde periode van 1.23 jaar en ʼn VKOK van 83.21 %. Die minder gunstige proses opsie was die azeotropiese distillasie wat die skeidingsagent n-metiel-2-pyrrolidon gebruik met ʼn terugbetaal periode van 15 jaar en ʼn VKOK van 7.72%. Dit is bevind dat weens die onakkuraatheid van die termodinamiese modelle vir die voorspelling van fase-ewewig, die verlangde vlak van akkuraatheid vir die ekonomiese ontledings nie geskik is nie. Gevorderde distillasie wat azeotropiese/ekstraktiewe distillasie proses gebruik, het dus die potensiaal om die ekonomiese uitdagings rondom die afval van motorbande deur pirolise tegnologie te verbeter of aan te spreek.
Description
Thesis (MEng)--Stellenbosch University, 2017.
Keywords
Limonene, Tyre derived oil, Enhanced distillation, Entrainer, Extractive distillation, Azeotropic distillation, Tire derived oil, UCTD
Citation