Department of Chemical Engineering

Permanent URI for this community

https://scholar.sun.ac.za/handle/10019.1/321

Department Process Engineering now has a new name, and will be known from March 2023, as Department of Chemical Engineering.

Browse

Now showing 1 - 11 of 11

Characterisation of liquid distribution and behaviour within randomly packed columns using electric impedance tomography
(Stellenbosch : Stellenbosch University, 2022-04) Lamprecht, Johannes Hendrik; Burger, Andries Jacobus; Stellenbosch University. Faculty of Engineering. Dept. of Process Engineering.
ENGLISH SUMMARY: The optimum design of column internals plays a prominent role in the economic viability of distillation setups, due to such internals’ notable contribution to both operating and capital costs. Progression in both our understanding and characterisation of column internals is therefore paramount. Both hydrodynamic and kinetic characterisation methodologies consider the influence of the vapour-liquid interface, whether directly (effective interfacial area) or indirectly (pressure drop and liquid hold-up). Most of the random packing literature, however, focuses on the evaluation of macro parameters (e.g. pressure drop, holdup, flow rates, packing dimensions and fluid physical properties), with notably less attention to the fluid behaviour at a micro level (e.g. droplet formation, distribution and rivulet formation). This limits the fundamental basis of the available models, introducing numerous regressed empirical constants. In other words, while modern random packing designs are strongly influenced by the optimisation of inter-packing droplet and rivulet formation, the available mathematical models lack predictive capabilities of such micro-behaviour. Against this background, and in pursuit of a better understanding of fluid behaviour and distribution in random packing, an Electrical Impedance Tomography (EIT) measurement system was designed and constructed to visualize and quantify liquid distribution behaviour inside randomly packed columns. The EIT system was preferred to conventional X-Ray tomography, due to a) safety, b) cost-effectiveness, and c) simplicity, while it can be utilised for both conducting and non-conducting liquids. The sensor of the EIT system consisted of a stainless-steel wire matrix, installed at a horizontal plane directly below 3m random packing in a 400mm diameter column. It provided 1369 measuring points, with measuring frequencies of 207 Hz and 21 Hz for conductive and non-conductive liquids, respectively. The data were processed using 2-D and 3-D image processing algorithms to enable quantification of individual liquid elements. The individual elements were evaluated based on their reconstructed volume, surface area and sphericity. The experimental characterisations were used to evaluate the liquid distributions inside two types of industrial random packing, FlexiRing® and Intalox® Ultra, at sizes ranges between 1.5” to 2.5”. The evaluations considered various liquid- and vapour loadings using both water and ethylene glycol to vary the liquid physical properties; water being electrically conductive and ethylene glycol being predominantly non-conductive. The presented results show increased element uniformity in favour of the Intalox® Ultra throughout and illustrated the presence of a force-balance transition in the mechanism of liquid hold-up creation. This indicated the transition from conglomerating inter-packing liquid (IPL) streams, towards droplet-creation. The onset of this transition was found distinctly related to the relative velocity profiles and vapour - liquid shear forces of the respective packings. The contribution of droplets in the inter-packing space to the total vapour-liquid interfacial area was also evaluated. The Intalox® Ultra presented ca 17% and 9.4% increase in total reconstructed surface area for the respective 2” and 1.5” equivalent comparisons with FlexiRing® (for the air-water system). This confirmed the applicability of the EIT characterisation system for both hydrodynamic and kinetic prototyping. Several novel contributions were developed in this work. These are: [1] The development of a characterisation methodology based on EIT for better understanding of inter-packing liquid distributions. [2] Novel experimental inter-packing distribution data for IPL element-volumes and -areas and their relation to: i. packing type, ii. liquid and vapour loadings, and iii. liquid physical properties. [3] Presenting the existence of a packing-specific transitional point, based on liquid and vapour loadings, where the mechanism of liquid hold-up changes. This point marks the cross-over between the conglomeration of inter-packing liquid elements into streams, and their break-up/ redistribution into smaller elements. This alludes to a possible increase in interfacial turbulence (decreasing liquid phase resistance to mass transfer) while adding to the understanding of the pressure drop mechanisms in packed columns. [4] Presenting the total IPL element-surface area as a comparative kinetic characterisation parameter for use in prototyping. This is posed to assist in the design of future packings, in finding the optimum packing area and structure to minimize entrainment and maximize efficiency.
Characterisation of the adsorption and regeneration behaviour of a commercial activated alumina adsorbent when separating 1-hexanol and 1-decanol from n-decane
(Stellenbosch : Stellenbosch University, 2023-03) Khalifa, Muzafar Tagelsir Hamza; Burger, Andries Jacobus; Stellenbosch University. Faculty of Engineering. Dept. of Chemical Engineering.
ENGLISH ABSTRACT: Paraffins produced in industry are usually associated with impurities such as oxygenates, and adsorption is commonly used to remove these oxygenates (e.g. alcohols). However, the recyclability of such adsorbents has rarely been studied. In this study, the reusability of activated alumina adsorbent used to adsorb 1-alcohols from a paraffin solvent was investigated. The objectives of this study included: (I) the measurement and investigation of single and binary-component adsorption data; (II) the modelling of equilibrium adsorption isotherms and adsorption kinetics; and (III) the measurement and investigation of the efficiency of adsorbent throughout consecutive adsorption/thermal regeneration cycles. One type of commercialized activated alumina adsorbent was used in this study. 1-hexanol and 1-decanol were used as adsorbates, while n-decane was used as the solvent. The effect of initial adsorbate concentration and alcohol carbon chain length on the adsorption and desorption were also briefly investigated. A bench-scale batch adsorption set-up was used for adsorption experiments. Schott bottles containing 1-alcohols + n-alkane solutions (at various initial concentrations of adsorbates) and adsorbents (10g of fresh adsorbent at first experiment) were immersed in a water bath, where the water temperature was maintained at 65 °C. The regeneration experiments were carried out at 205 °C, 185 °C and 165 °C using thermal regeneration columns. Heated purging carrier gas (nitrogen) was used to sweep through the adsorbent for the full duration of regeneration. For both single-component and binary systems, the total equilibrium adsorbent loading was found to increase when increasing the adsorbates’ initial concentration. For single-component solutions, the total adsorption capacity increased systematically as the initial concentration of adsorbates increased up to 2.5 mass%, but at even higher initial concentrations the capacity remained relatively constant. For the binary system, the equilibrium adsorbent loading began to plateau at initial concentrations higher than 2.0 mass%. The effect of carbon chain length exhibited different behaviour for single-component and binary systems. The adsorbent proved a slightly higher affinity to adsorb more of the 1-decanol in the single-component system, whereas in the binary system it was inclined to adsorb more of the 1-hexanol. At similar 1-alcohol initial concentrations in single and binary systems, it was found that the adsorption of a specific 1-alcohol from the binary systems was remarkably poorer than the adsorption of the same 1-alcohol from the corresponding singlecomponent systems. Equilibrium studies showed that the Freundlich and Sips isotherms are suitable for the single 1-hexanol and single 1-decanol adsorption systems, respectively. However, for the binary system, poor correlation was exhibited between the date and the Extended Langmuir, Extended Freundlich and Extended Sips models. Both single-component and binary systems can be described by Elovich kinetics and the Intra-Particle Diffusional Model. As expected, higher efficiencies were achieved for the respective single-component systems than for the binary system. As far as regeneration behaviour is concerned, the activated alumina adsorbent exhibited an adverse response to the highest regeneration temperature (205 °C). For both single-component and binary systems, the adsorbent offered a better efficiency at the regeneration temperature of 185 °C than at 205 °C and 165 °C. BET analysis showed a notable reduction in the total surface area of the regenerated adsorbent at 165 °C, indicating incomplete regeneration at such low temperature, was also confirmed by a systematic drop in adsorption efficiency down to 40% at the final regeneration cycle.
Contributions to theoretical developments and practical exploitation of mass transfer principles in separation technologies
(Stellenbosch : Stellenbosch University, 2023-12) Nieuwoudt, Izak; Burger, Andries Jacobus; Stellenbosch University. Faculty of Engineering. Dept. of Chemical Engineering. Process Engineering.
ENGLISH ABSTRACT: Our standard of living and quality of life rely on producing consumer products from chemicals. These chemicals typically have to be purified, and distillation, extraction and absorption are the primary technologies for affecting these separations. Such technologies are energy intensive, and further development and optimization can reduce their impact on greenhouse gas emissions. Furthermore, reduction in the capital consumed by these processes can significantly improve value creation. My contributions to these fields over the past 36 years are covered in detail in this dissertation under the sections highlighted below. It is conservatively estimated that these contributions generated more than two billion dollars of value for the companies who commercialized and use these technologies. In the process of making contributions to this field, I have been blessed to collaborate with brilliant people and their contributions to my endeavors are gratefully acknowledged. Improved separation processes My research into separation processes with reduced energy and capital consumption was focused on solvent-driven separations. Computer Aided Molecular Design (CAMD) methodologies for solventdriven separation processes were developed and, from this, improved extractive distillation, azeotropic distillation and liquid extraction processes were conceived. The energy and capital consumption of these processes were significantly lower than that of competing technologies, which also translates into lower greenhouse gas emissions. These processes were commercialized and create significant value for the chemical producers and the companies who produce consumer products. Improved separation tower internals My research into separation tower internals was focused on creating novel equipment that had higher separation efficiency and higher hydraulic capacity than the best equipment available at the time. The novel INTALOX® ULTRATM random packing exhibited higher efficiency and capacity than other random packing. This allowed chemical companies and refiners to debottleneck towers for capacity and/or separation efficiency. This reduced energy consumption and corresponding greenhouse gas emissions. The size of new towers could also be reduced. The PROFLUX® severe service packing allowed refiners to increase capacity, increase run lengths and increase the product yield in vacuum towers. I developed several tray configurations that improved the capacity and/or efficiency of separation towers. Other developments included improved liquid distributors and droplet separators. These new, improved separation tower products generated significant value for both the companies that commercialized it and the end-users who installed it in their separation towers. Separation technology education Although the separation processes and separation tower internals discussed in this dissertation generated significant revenue for the companies who commercialized and used these technologies, the lasting value lies in the underlying methodologies and knowledge that were developed. It is important that this knowledge be passed on to the next generation of engineers. To this end, I have developed university courses, the Koch-Glitsch Mass Transfer School and the FRI Distillation Academy. Under my guidance 14 Masters and 7 PhD students graduated in the field of separation technology. Almost 1000 students attended these courses and many commented that they have received unique knowledge that equipped them for the future. The details of my non-confidential contributions to the field of separation technology have been summarized in 410 patents (46 patent families), 48 papers and 80 conference contributions. At the 2022 Spring Meeting the Separations Division of the American Institute of Chemical Engineers (AIChE) gave special recognition to my “many outstanding achievements in distillation, extraction, 5 absorption, and troubleshooting” by dedicating an honors session to celebrate my “lifetime of contributions as an engineer, educator, inventor, and R&D leader in separations technology and engineering”.
Developing the s-SAFT-γ Mie equation of state toward nonaqueous alkanolamine-based carbon capture systems
(Stellenbosch : Stellenbosch University, 2024-03) Schulze-Hulbe, Alexander; Cripwell, Jamie Theo; Burger, Andries Jacobus; Stellenbosch University. Faculty of Engineering. Dept. of Chemical Engineering. Process Engineering.
ENGLISH ABSTRACT: Decarbonizing industrial processes is imperative for mitigating the harmful effects of climate change. A promising route to decarbonization lies in developing nonaqueous alkanolamine-based carbon capture processes. However, there is a very wide range of nonaqueous formulations to choose from, and little available thermodynamic data. Accordingly, an apt starting point for assessment of nonaqueous alkanolamine-based carbon capture is the development of a predictive thermodynamic modeling tool which captures the salient phenomena of these systems. The Statistical Associating Fluid Theory (SAFT) equations of state (EoSs) present a fundamental approach to thermodynamic modeling. Combining these EoSs with the group-contribution (GC) approach provides these rigorous models with considerable predictive capabilities. This renders GC-approach SAFT EoSs particularly useful in the data-scarce context of nonaqueous alkanolamine-based carbon capture. Accordingly, the main aim of this work was to develop structural SAFT-γ Mie (“s-SAFT-γ Mie”), a stateof- the-art GC-approach SAFT EoS, toward a description of alkanolamine solvent/CO2/organic cosolvent systems. This presents the first instance in which the predictive capabilities of a GC-approach EoS are extended to nonaqueous alkanolamine-based carbon capture systems. However, myriad approaches can be followed in developing parameters for GC-approach EoSs. This renders parameterization challenging, thus presenting an obstacle to industrial implementation of these models. To facilitate use of GC-approach EoSs, a further aim of this work was to illustrate how GCapproach EoSs can be parameterized for nonaqueous alkanolamine-based carbon capture systems using a systematic and consistent approach. Transferable s-SAFT-γ Mie group interaction parameters were developed from the ground up for primary and secondary alcohols, as well as primary amines. The model exhibited robust capabilities in modelling these components as well as their mixtures with n-alkanes. However, results for linear alkanolamines indicate that s-SAFT-γ Mie’s generalizability comes at the expense of quantitative accuracy. In the process of developing these parameters, a novel and generalizable approach was devised to account for the effect of changing hydroxyl group position in secondary alcohols. This further developed s-SAFT- γ Mie’s capabilities in distinguishing between the properties of isomers, an important characteristic for solvent/cosolvent screening purposes. s-SAFT-γ Mie further provided qualitatively accurate descriptions for a wide range of organic cosolvents with a single parameter set. This broadly generalizable modeling approach can be extended to components for which little or no reliable data are available, highlighting its value to carbon capture process designers. The parameters thus developed were transferred to CO2-containing mixtures. Pertinently, s-SAFT-γ Mie provided qualitatively accurate descriptions of CO2 solubility in polyethylene glycols, which are important components for nonaqueous carbon capture. Regarding alkanolamine solvent/CO2/organic cosolvent systems, s-SAFT-γ Mie was capable of qualitatively reproducing the effects of temperature, liquid-phase composition as well as organic cosolvent chain length on CO2 solubility. This holds for lower pressures, where CO2 solubility is driven by chemical absorption, as well as higher pressures, where CO2 is dissolved by physical absorption. These robust predictive capabilities render s-SAFT-γ Mie well-suited to comparing CO2 solubility in several alkanolamine solvent/organic cosolvent formulations, highlighting its potential future use within the context of a solvent/cosolvent screening tool.
The effect of hydraulic conditions on the performance of an upflow anaerobic filter at ambient temperatures treating synthetic winery wastewater
(Stellenbosch : Stellenbosch University, 2024-03) Fleming, Warren Francis; Burger, Andries Jacobus; Stellenbosch University. Faculty of Engineering. Dept. of Electrical and Electronic Engineering.
ENGLISH ABSTRACT: Wineries produce wastewater throughout the year, although peak production is experienced during harvest and bottling periods. Winery wastewater (WWW) also varies in strength and composition throughout the year, with COD concentrations ranging from 400 mg/ℓ to 50 000 mg/ℓ. Major components of the COD in WWW include monosaccharides, ethanol, and acetic acid. Wineries seeking to mitigate their demand for fresh irrigation water while adhering to legal wastewater treatment standards are presented with various treatment options. Among these, anaerobic wastewater treatment provides an environmentally friendly and effective option. However, anaerobic wastewater treatment systems can be expensive to implement and operate. They are often complex to control, operate best at elevated temperatures, and can be extremely sensitive to environmental conditions and wastewater compositions, generally making them inaccessible to small-scale wineries. The upflow anaerobic filter (AF) is an implementation of high-rate anaerobic wastewater treatment that makes use of packing media within the digester to foster biofilm development. Upflow AFs are able to operate over a larger range of flow conditions to comparable upflow anaerobic sludge blankets (UASB) and require smaller digester volumes than suspended sludge continuously stirred tank reactor (CSTR) counterparts. Therefore, this research focused on the evaluation of an uncomplicated packed bed anaerobic system for the potential treatment of winery effluent, with only pH and upflow control while operating at ambient temperatures (set-point temperature of 22 °C). The experimental AF was a 3 m tall PVC column, with a diameter of 110 mm. It had a working volume of 25.7 ℓ and was packed with 25 mm plastic pall rings. A synthetic WWW (5 213 ± 376 mgCOD/ℓ) was used as typical feed to the AF. The pH of the solution entering the column was controlled at 7.2, while two hydraulic flow parameters were controlled, viz the hydraulic residence time (HRT) of fresh feed in the system and the upflow velocity through the column. These hydraulic parameters were manipulated by a fresh feed pump and a recirculation pump, respectively. The system pH, temperature, and biogas production rate were monitored through a bespoke Arduino data logging system. The biogas was produced at rates between 0.7 mℓ/min and 2.8 mℓ/min, with yields between 0.3 mℓ/mgCODconsumed and 1.1 mℓ/mgCODconsumed, and methane concentrations ranging between 46% and 72% (gas readings taken at ambient temperature and pressure). The system showed unstable and poor performance when operating at too low upflow velocities. Although, it consistently reduced the COD of the feed by more than 55% when operating at upflow velocities between 30 – 48 m/day and HRTs between 24 – 60 hours. The highest COD reduction of 69.5% was at an upflow velocity of 31 m/day and 42 hours HRT. In other words, when treating a synthetic winery effluent with ca 5 000 mg/ℓ COD, this simple system can comfortably reduce the COD levels to around 1 900 mg/ℓ. Since the system was able to operate acceptably at ambient temperatures over various hydrodynamic conditions (showing resilience to changes in selected operating parameters), it could perhaps be considered as a viable anaerobic treatment method for small-scale and boutique wineries. The Anaerobic Digestion Model No.1 (ADM1) was applied – with selected regressed parameters – to predict the COD reduction in a system of similar process design, but at different COD feed concentrations. https://scholar.sun.ac.za iii The final calibrated model was regressed against experimental data at and HRT of 42 hours and upflow velocity of 39 m/day. Assuming CSTR conditions, the calibrated model showed good effluent COD predictions (errors below 5.6%) of the optimal experimental conditions and offers a reasonable means of predicting the performance of the system.
Evaluation of structural (s)-SAFT-γ Mie using newly measured binary VLE data of alkanes mixed with branched alcohols
(Stellenbosch : Stellenbosch University, 2024-03) Slabbert, Robert Malcolm; Burger, Andries Jacobus; Stellenbosch University. Faculty of Engineering. Dept. of Chemical Engineering. Process Engineering.
ENGLISH ABSTRACT: The application of group contribution (GC) methods to an equation of state (EoS) increases the flexibility of the EoS and has shown improvement in the predictive capabilities of EoSs. The application of GC methods to EoSs requires parameterisation of several functional groups, and it then allows for the prediction of physical properties of mixtures (for which no experimental data exist) constructed out of such pre-parameterised functional groups. The application of GC methods to statistical associating fluid theory (SAFT) EoSs has been more naturally adopted due to the spherical chain segments present in the reference fluid, with one such EoS being SAFT-γ Mie. Although its predictive capabilities are well-established, SAFT-γ Mie cannot distinguish between structural isomers. A recent modification to the chain term, structural (s)-SAFT-γ Mie, seeks to rectify this flaw by allowing SAFT-γ Mie to consider both the inter- and intragroup bonds without new fitting parameters. Assessing the accuracy of thermodynamic models is primarily done using reliable experimental data; however, gaps in the available literature data hinder the improvement of these models. To evaluate (s)-SAFT-γ Mie, novel isobaric vapour-liquid equilibrium (VLE) data for methyl butanol isomers/alkanes (C6 – C8) were measured. The verification of the experimental methodology was achieved using vapour pressure curves correlated to DIPPR correlations and the replication of binary VLE data for the system hexane/1-butanol at 101.3 kPa. New isobaric VLE data for twelve binary systems were then produced at 101.32 kPa. The maximum uncertainties calculated were determined to be 0.531 kPa, 0.431 K, 0.0135 mole/mole and 0.0126 mole/mole for the pressure, temperature, and liquid and vapour composition, respectively. With the exception of systems containing 2-methyl-1-butanol, all systems were found thermodynamically consistent according to the L-W area test and McDermott-Ellis test. It is quite possible that the heat of vaporisation correlation used for 2-methyl-1-butanol mixtures caused these systems to fail the L-W test. Both OH and first order hydroxyl groups (CHnOH) were used for modelling of branched alcohols. Pure component property predictions (vapour pressure, liquid density and heat of vaporisation) of the methyl butanol isomers were generated using SAFT-γ Mie and s-SAFT-γ Mie. SAFT-γ Mie matched the predictions of (s)-SAFT-γ Mie for pure component properties. For binary phase predictions, UNIFAC, acting as a reference for activity coefficient models (ACMs), matched those of SAFT EoSs. SAFT-γ Mie matched or outperformed the predictions of (s)-SAFT-γ Mie for binary phase predictions. This was deemed to be a result of the imported parameters used for s-SAFT-γ Mie, as these parameters were previously regressed using linear alcohol data. The modification to the chain term of s-SAFT-γ Mie allows for the distinction between the structural isomers, but often resulted in larger deviation from experimental data. This is most likely due to the inherent limitations of SAFT, with non-neighbouring interactions not considered and only limited steric hinderance effects being accounted for. Therefore, it is recommended that, while s-SAFT-γ Mie has notable advantagesfor certain applications, its use should be limited to primary branched chain alcohols, branched chain molecules using first order functional groups, branched chain molecules that are non-associating, and branched molecules with short chain branches.
Investigating the potential of near-infrared spectroscopy for the detection of low-concentration CECs in water
(Stellenbosch : Stellenbosch University, 2023-03) Lourens, Cordilia; Burger, Andries Jacobus; Archer, Edward; Swartz, CD; Stellenbosch University. Faculty of Engineering. Dept. of Chemical Engineering.
ENGLISH ABSTRACT: The presence of so-called contaminants of emerging concern (CECs) in water is a significant issue in many countries. Their low concentrations, complex matrices, and wide range of compositions (with varying physical-chemical properties) pose many challenges to their rapid identification and quantification. Consequently, there is a need for the development of reliable, fast, and low-cost monitoring techniques that warn of any irregular changes in CEC-related water quality, which can then prompt detailed analysis by standard methods. Near-infrared (NIR) spectroscopy, in conjunction with chemometrics, is one such technique that could potentially be utilised for near-real-time monitoring and detection of selected CECs. Therefore, this study endeavoured to develop a better understanding of the potential of NIR spectroscopy for detection of changes in the water spectrum that may indicate a CEC-related deterioration in water quality. The applicability of the method was considered by performing multivariate analysis on the NIR spectral data of deionised water spiked with surrogate CECs typically associated with anthropogenic pollution. Standard deviations of less than 0.16% and less than 0.71%, during the repeatability and reproducibility tests, respectively, suggested a high precision of the NIR method. The ratio of the standard deviation over the mean were less than 0.2% for the repeatability tests, and less than 1.16% for the reproducibility tests. This allowed for the development of a CEC classification model. A pre-feasibility study was firstly conducted, considering the following five CECs at concentrations varying between 0.001 mg/L and 1000 mg/L: carbamazepine, caffeine, efavirenz, sulfamethoxazole and trimethoprim. Positive pre-feasibility results – indicating that NIR spectroscopy indeed has potential as a rapid screening method – prompted further detailed experiments with three typical surrogate CECs, viz acetaminophen, benzotriazole and caffeine. Exploratory data analysis using principal component analysis (PCA) revealed that two spectral regions, 1300–1600 nm and 1600–2200 nm, where the appropriate wavelength regions in which to perform the investigation. The PCA score plots of the surrogate CECs showed good separation between the NIR spectra of the test chemicals and the deionised water. However, for acetaminophen and benzotriazole good separation was only possible at high concentration ranges (10–1000 mg/L) – notably higher than typical concentrations at which they occur in polluted surface water. On the other hand, caffeine samples showed separation at all the tested concentrations (0.00001–1000 mg/L), and thereby suggesting the potential application of NIR spectroscopy for the detection of caffeine at typical concentrations found in the environment. Partial least squares discriminant analysis (PLS-DA) showed that it was possible to differentiate between acetaminophen and the deionised water at a minimum concentration of 10 mg/L using the PLS-DA model with a 92.6% classification accuracy. For caffeine, differentiation was possible at a minimum concentration of 0.01 mg/L, with model classification accuracies of 87%. These results corroborated the results from the exploratory analysis with PCA. Differentiation between benzotriazole and the deionised water using PLSDA was successful to a minimum concentration of 0.1 mg/L, coinciding with findings from previous studies. The results of some preliminary evaluation tests performed on river water samples were unsatisfactory. As expected, it showed that NIR measurements in water containing salts and organics (potentially also several CECs) would be influenced adversely by such dissolved components, thereby complicating potential realtime application of these methods. Further in-depth research is therefore required, where a wider variety of matrices (such as surface water, groundwater, and wastewater), are included in the experimental design. NIR spectroscopy combined with multivariate data analysis has shown promising potential for detecting contaminants of emerging concern at mg/L and even high µg/L levels. However, implementation of the method as a (near)-real time early warning system for detecting CECs at environmental concentrations (at ng/L levels) is not yet a practical possibility.
The measurement and modelling of vapour-liquid equilibria in ternary systems containing polar, cross-associating, and associating compounds
(Stellenbosch : Stellenbosch University, 2020-12) Thompson, Andrew Cameron; Burger, Andries Jacobus; Cripwell, Jamie Theo; Stellenbosch University. Faculty of Engineering. Dept. of Process Engineering.
ENGLISH ABSTRACT: The accurate production of simple binary vapour liquid equilibria (VLE) data sets is time consuming and expensive, and the measurement of similar data sets for multi component systems is even more laborious and tricky. Therefore, generic thermodynamic models that can predict VLE behaviour accurately are required for the optimisation and design of many processes within industry. This is especially true for complex systems such as polar, associating, and cross-associating mixtures.For the majority of complex VLE mixtures, thermodynamic models incorporating binary interaction parameters (BIPs) are required for an accurate representation.Models in the perturbation theory equations of state (EoS) use pure component data to predict phase equilibria. In doing so, these models predict rather than correlate VLE and phase equilibria. Perturbation theory models such as CPA-GV, SAFT-VR-Mie-GV, and sPC-SAFT-GV have been proved accurate for the prediction of binary systems and assumed to be accurate for multicomponent systems. However, there is a need to systematically test the assumption that these models can accurately predict complex multicomponent mixtures with a similar accuracy that they do for binary systems. This study therefore produced carefully measured new equilibrium data on the ternary VLE of four systems, as well as the nine binary systems within the ternary systems. Each system contained 1-propanol and 2-butanone, with the third component being either one of three different C4 esters, or2-propanol. These data sets were then used to test and evaluate the applicability and accuracyof the aforementioned EoS.The data of the nine binary and the four-ternary isobaric VLE systems were produced at 101.3kPa with the use of a Gillespie type still. This still had experimental uncertainties in temperature of ± 0.62 K and pressure of ± 0.046 bar. The liquid and vapour samples were analysed using a GC and had experimental uncertainties of ±0.016 mole fraction. The experimental procedure with the Gillespie still and the GC analysis was verified using vapour pressure of the binary systems that occur in all four ternary systems, 1-propanol/2-butanone, and existing literature data. All VLE data were thermodynamically consistent, passing both the McDermott-Ellis and the L/W consistency tests. The deviations from the ideal state in the ternary systems were found to correspond to similar deviations in the binary mixtures. This was further evidenced with distillation boundaries when two azeotropes existed (methyl propionate and propyl formate systems) with the 2-propanol system displaying the largest deviations from the ideal state, and the ethyl acetate system behaving the most ideal. Accurate results were found for the binary modelling, with both the SAFT-VR-Mie-GV and sPC-SAFT-GV models being able to cope with the cross-association and the non-ideal nature of the systems, having AAD values for the nine binary systems of 0.614 K and 0.011 mole fraction and 0.589 Kand 0.011mole fraction, respectively. In the ternary systems, the two models that accurately predicted the binary systems also predicted the ternary systems accurately, with comparable results with NRTL and each other. The AAD values of the four ternary systems were 0.820 K and 0.015 mole fraction, and 0.720K and 0.017 mole fraction, respectively. This work therefore shows that the SAFT type models display strong potential to predict the vapour-liquid equilibria in ternary systems containing polar, non-associating and associating compounds.
Pervaporation of ethanol-water solutions using commercially available pervaporation membranes
(Stellenbosch : Stellenbosch University, 2024-03) Akoodi, Aaisha; Pillay, Visvanathan Lingamvrthi; Burger, Andries Jacobus; Stellenbosch University. Faculty of Engineering. Dept. of Chemical Engineering. Process Engineering.
ENGLISH ABSTRACT: Pervaporation can be used for the recovery of bioethanol from fermentation broths. Most previous studies focused on laboratory-synthesized membranes and the OFAT approach has generally been used to investigate the influence of operating variables. In this study, pervaporation of ethanol-water solutions is investigated using commercial membranes (Pervap 4060 and Pervatech) and key operating parameters and their effects on pervaporation performance are identified. First, a laboratory-scale pervaporation unit was set up and validated in terms of repeatability of results and comparison with results reported. From repeated runs, coefficient of variance values for total and ethanol flux were 2.8% and 8%, respectively, indicating that the results were reproducible. Furthermore, the results obtained were comparable to those reported by the membrane manufacturer and in literature. The second phase of this study involved investigating the effects of key operating parameters on pervaporation performance. The parameters were feed temperature (40 - 60℃), feed flow (7 - 15 litre/min), and feed concentration (5 - 15 mass%). An increase in feed temperature and concentration caused an increase in flux, while feed flow did not have a notable effect. Generally, ethanol selectivity decreased with an increase in feed temperature and concentration and increased slightly with an increase in feed flow rate. In the third stage, an empirical model was developed using a 3-level central composite design was performed. Based on experimental data, a second-order model was developed using Statistica software. The developed flux model had a good accuracy; however, the selectivity model did not fit the experimental data well. Nevertheless, suggested optimum values were identified as: 60℃ feed temperature and 15 mass% EtOH feed concentration, with a predicted total flux of 3.52 kg/(m2h) and ethanol flux of 1.63 kg/(m2h). Possible limitations of pervaporation performance at high feed ethanol concentrations were investigated during the fourth phase of this study. Experiments were performed at increasing feed concentrations up to 75 mass% EtOH. A notable increase in ethanol and total flux were observed with an increase in feed concentration from 1 to 75 mass% EtOH. Conversely, a considerable decrease in the selectivity was observed. This indicates that the Pervap 4060 membrane is ethanol-selective at low ethanol concentrations in the feed. In the fifth stage of this study, the effects of membrane type on pervaporation performance were investigated. The pervaporation performance of commercial membranes, Pervap 4060 and Pervatech PDMS, were compared over a range of feed https://scholar.sun.ac.za iii concentrations (5 - 75 mass% EtOH). The pervaporation separation index (PSI) combines flux and selectivity data and was used to compare overall efficiency. The average PSI values of the Pervap 4060 and Pervatech PDMS membranes were 12 and 25 kg/(m2h), respectively. PSI values of the Pervatech membrane were higher than that of the Pervap 4060 membrane, indicating that the Pervatech membrane may be more efficient for ethanol recovery
Seeding as a mitigation strategy for gypsum scaling in membrane distillation desalination
(Stellenbosch : Stellenbosch University, 2022-12) Bührmann, Erika Isabel; Burger, Andries Jacobus; Stellenbosch University. Faculty of Engineering. Dept. of Process Engineering.
ENGLISH ABSTRACT: Membrane distillation (MD) is a temperature driven separation technology that can desalinate water at higher salinities than can be treated by means of reverse osmosis (RO), and can operate at lower temperatures than multistage flash distillation (MSF) or multi-effect evaporation (MEE), while having a smaller physical footprint. MD presents an alternative to conventional desalination technologies for the treatment of highly saline water especially where solar or low-grade waste heat is available or where a small footprint is desirable. As is typical of membrane processes, MD is susceptible to scaling, especially by calcium scalants, such as gypsum, which has a low solubility in water at elevated temperatures. Gypsum scaling can block the membrane pores and restrict the flux, as well as penetrate the pores and cause the membrane to become wetted, thereby reducing its separation efficiency (SE). It was hypothesised that gypsum scaling in MD can be mitigated by adding gypsum crystals (seeds) to the supersaturated calcium sulphate solution. This procedure (crystal seeding) can restrict crystallisation on the membrane in favour of growth of suspended crystals in the bulk solution. This study endeavoured to demonstrate that seeding can retard the flux decline and delay the onset of pore wetting in an MD process. Furthermore, this study demonstrated the effect of feed agitation and seed loading on the effectiveness of scaling mitigation by seeding. Finally this study investigated the reversibility of gypsum scaling in seeded membrane distillation by rinsing with water. In an experimental investigation a saturated calcium sulphate solution was desalinated using a lab-scale MD process consisting of a membrane fitted to the base of a stirred feed vessel and exposed to cold water that was pumped over the permeate side of the membrane. The MD system was run at 61 °C with a trans-membrane vapour pressure of 16 kPa and a bulk permeability coefficient of 2.3x10-7 s/m. It was demonstrated that the addition of 5 g/L seed crystals agitated to a G-value of 630 s-1 resulted in the flux declining ten times slower than when no scaling mitigation was implemented. However, when crystals were only agitated to a G-value of 210 s-1, seeding accelerated the rate of scaling ostensibly via crystal deposition and secondary nucleation on the membrane. It was demonstrated that increasing the level of agitation (from 210 to 420 s-1 and from 420 to 630 s-1) resulted in a retardation in flux decline and a delayed risk of pore wetting as the crystal growth rate on the membrane and the crystal deposition rate onto the membrane were inferred to be decelerated. Similarly, increasing the seed loading (from 1 to 3 g/L and from 3 to 5 g/L) was also inferred to decelerate the crystal growth rate on the membrane but to accelerate the crystal deposition rate. Therefore a greater seed loading delayed the risk of pore wetting, but did not necessarily retard the flux decline. The scaling that occurred during seeded membrane distillation appeared to be completely reversed by rinsing with and soaking in water. However, only rinsing with water was not sufficient in removing the scale and restoring the membrane properties. It was therefore surmised that while seeding is a feasible gypsum scaling mitigation strategy in MD, it should be paired with a more thorough cleaning sequence than merely periodically rinsing with water.
Spectroscopic measurement of associating and solvating binary mixtures for the determination of monomer fraction data towards thermodynamic model improvement
(Stellenbosch : Stellenbosch University, 2024-03) Burgess, Aiden Edward; Cripwell, Jamie Theo; Burger, Andries Jacobus; Stellenbosch University. Faculty of Engineering. Dept. of Chemical Engineering. Process Engineering.
ENGLISH ABSTRACT: Amidst an energy transition within the broader chemical engineering sector, there is a strong drive towards making processes more sustainable by enhancing energy efficiency by adopting cleaner alternatives like green solvents and biodiesels. However, the presence of hydrogen bonding phenomena in such chemicals poses a longstanding challenge for thermodynamic modelling and process design. The monomer fraction, indicative of the extent of hydrogen bonding, can be measured using Fourier Transform Infrared (FTIR) spectroscopy. Yet surprisingly, its practical application in testing thermodynamic model performance remains unexplored, partly due to limited data and discrepancies in measurement techniques. Thus this was the overall focus of this work, to propose a rigorous alternative for monomer fraction quantification in associating and solvating systems and thoroughly assess its use in testing thermodynamic model performance. A detailed examination of the uncertainty associated with the single-point calibration in an ethanol/nhexane system revealed that there was a significant dependence of the ethanolic monomer fraction on the choice of the calibration method. With this in mind, establishing a definitive calibration choice proved to be challenging, prompting an alternative methodology which was based on multivariate curve resolution with alternating least squares (MCR-ALS). The approach leveraged purely physical constraints and resulted in physically meaningful profiles. The rotational ambiguity uncertainty in each of the resolved profiles was completely suppressed and the resulting lack of fit (LOF) was determined to be 2.55%. The spectral profiles exhibited features that allowed the assignment of α, β, γ, and δ, bonds where the α-bond type was used to determine the true monomer fraction while the fraction of free hydroxyls (α+β) compared well to previously reported data. Probing for ethanolic monomer features in solution with C4 esters was performed using the hydroxyl band. In each of the three binary systems, no features could be attributed to free ethanol. Thus, it was concluded that ethanol was not present in a free monomeric form and the monomer fraction of ethanol in all C4 ester systems was zero. Conversely, C4 ester monomer fractions, in solution with ethanol, were determined by probing the alkoxyl ester band as this allowed spectra across the entire compositional space to be studied. It was found that the monomer fraction of methyl propionate and ethyl formate had similar gradual decreasing trends while the monomer fraction trend of propyl formate decreased significantly as a function of increasing ethanol composition. The monomer fractions of C4 esters were thermodynamically modelled using the SAFT-VR Mie + GV equation of state (EoS) under non-solvating and solvating conditions. Using the N scheme, improved monomer fraction predictions were seen but were still qualitatively poor. Although regression of new association parameters, against C4 ester monomer fractions, brought improved monomer fraction correlations, significant deterioration in VLE predictions was observed. Marginal improvements were determined when using parameters obtained from a combined monomer fraction and VLE discretised sensitivity analysis. The inclusion of monomer fraction data in thermodynamic model parameterisation highlighted the model’s inability to accurately describe monomer fraction data. This suggests a possible shortcoming of Wertheim’s association theory, indicating the need for further development of the theory.

Browse

Browsing Department of Chemical Engineering by browse.metadata.advisor "Burger, Andries Jacobus"

Results Per Page

Sort Options