Masters Degrees (Chemical Engineering)
Permanent URI for this collection
Browse
Browsing Masters Degrees (Chemical Engineering) by browse.metadata.advisor "Burger, A. J."
Now showing 1 - 20 of 34
Results Per Page
Sort Options
- ItemADM1 Parameter Calibration Method based on partial least squares regression framework for industrial-scale anaerobic digestion modelling(Stellenbosch : Stellenbosch University, 2019-12) Xu, Zhehua; Burger, A. J.; Louw, Tobias M.; Stellenbosch University. Faculty of Engineering. Dept. of Process Engineering.ENGLISH ABSTRACT: Anaerobic Digestion Model 1 (ADM1) is the mainstay modelling tool for Anaerobic Digestion research and development. Its growing popularity is attributed to its sophisticated yet expandable structure. Not only does ADM1 encompass a broad range of biochemical, physicochemical and inhibition reactions, it provides the modeller a structured framework to add or remove reactions per application requirements. Two major challenges that ADM1 faces are the difficulty in translating common quality indicators into ADM1’s 26 state variables, and the complication with calibrating a large number of model parameters – 58 by default. There is currently no consensus with regards to the parameter calibration approach. Researchers utilise various sensitivity analysis techniques to identify sensitive parameters, but the selection of parameters to be calibrated relies largely on the modeller’s discretion. In some cases, decisions are simply made based on prior or expert knowledge. Since the installation, operation and maintenance of advanced instrumentation are often expensive, most industrial digesters are inadequately monitored and thus intentionally over-designed. A model that can be used on-site with acceptable accuracy could serve as a soft sensor to forecast inhibition risks and automate preventive actions. Therefore, this study aimed to develop a standardised way to calibrate parameters when optimising ADM1 models built for industrial-scale digesters. The proposed method, Partial Least Squares (PLS) Method, consists of four steps. In Step 1, a series of Monte Carlo simulations is carried out. For each Monte Carlo run, ADM1 is executed with all its model parameters sampled from independent probability distributions. These probability distributions were obtained by conducting a literature survey across 62 publications and all published parameters compiled into a domain which represents the uncertainty range of each parameter. In Step 2, a multivariate regression technique called PLS Regression (PLSR) is applied to the Monte Carlo results. The motives for employing PLSR are to reduce parameter dimensionality and to identify the underlying relationships between the model parameters and the model outputs. In Step 3, these relationships, which are mathematically described as PLS weights, loadings and latent variables, are utilised to guide parameter calibration. Lastly, the calibrated parameter set is validated against unseen data. This method successfully improved, in the absence of any modeller’s bias, the overall accuracy of a model based on data from an industrial-scale digester. The model is tasked to fit six typical plant measurements: Volatile Fatty Acids (VFA), ammonia, Volatile Suspended Solids (VSS), pH, methane gas flow & carbon dioxide gas flow. A configuration consisting of at least 500 Monte Carlo runs and two latent variables is required to produce a reasonably accurate fit. Although the use of more latent variables could enable PLSR to capture interactions of lesser weighted output variables, the model becomes increasingly prone to overfitting. However, it is envisaged that more latent variables would be necessary if more outputs are modelled. It is recommended to start the PLSR algorithm with one latent variable and only introduce more if necessary. Different parameter calibration methods produce different model outcomes. The PLS Method was benchmarked against two other methods, namely the Group Method and the “Brute Force” Method. In the former method, kinetic parameters were grouped into the three groups of sensitivities (High, Medium, Low) as suggested in the ADM1 Scientific and Technical Report. The three groups are then calibrated sequentially in order of decreasing sensitivity. The “Brute Force” Method involved calibrating all 58 parameters without any particular sequence, prioritisation or expert inputs. Lower and upper limits are, however, set as per the minimum and maximum values identified from the literature. Besides proving to be a suitable method for industrial-scale digester modelling, the PLS Method was found to exhibit several unique traits: • It is the only method that did not show signs of overfitting. • It is the only method that concluded the model optimisation with all calibrated parameter values within the surveyed minimum and maximum range. • It converges on the objective function 30-60% faster than the Group Method and 14 times quicker than the “Brute Force” Method The success is attributed to the fundamentals of PLS regression. Unlike other regression methods where parameters are adjusted independently, PLS enables parameters to be manipulated collectively in a manner that ensures maximum impact on the outputs while considering collinearities among the parameters. This guided approach effectively mitigates the so-called “curse of dimensionality” and, potentially, overfitting and thereby speeds up the calibration process.
- ItemAssessment of the capabilities of two polar sPC-SAFT terms through application to measured ketone-alkane phase equilibria data(Stellenbosch : Stellenbosch University, 2014-04) Cripwell, Jamie Theo; Burger, A. J.; Schwarz, C. E.; Stellenbosch University. Faculty of Engineering. Dept. of Process Engineering.ENGLISH ABSTRACT: Thermodynamic models have been investigated extensively since Johannes van der Waals first devised a mathematical relation capable of predicting both vapour and liquid phases for a mixture at equilibrium. With the advent of modern computing power, these equations of state have gone from their humble empirical beginnings to the comprehensive and fundamentally derived models we have today. One such physically sound model is the Statistical Associating Fluid Theory (SAFT) family of equations, derived from the molecular perturbation theories of the 1980’s. The relative youth of this thermodynamic framework has meant that much work has gone into modification and optimisation of the model recently. The variants of particular interest to this work are the simplified perturbed chain SAFT equations with the Jog & Chapman (sPC-SAFTJC) and Gross & Vrabec (sPC-SAFTGV) polar terms. Each of the polar terms supports one adjustable polar parameter that relates to the quantity of polar segments in the reference fluid but not necessarily its position in the carbon chain. The strength of polar interactions is known to decrease as the functional group moves away from the terminal methyl group and the effects of steric hindrance increase. Thus, in question here is whether the models can account for the change in polar interactions associated with the changing position of the polar group, by only adjusting the values of the existing pure component parameters; that is, in lieu of a position specific parameter. The carbonyl group in ketone molecules is one such polar group, and it is this homologous series that is the focus of this study. The decrease in polar interactions as the carbonyl group in a ketone molecule shifts centrally is apparent from the lower boiling points of the isomers where the polar group is central as compared to those where the functional group is nearer the terminal methyl group. The effect of this functional group shift on binary phase behaviour has not previously been assessed for any system however, as the lack of experimental data attests. Thus, experiments had to be conducted to generate phase equilibrium data for systems comprising each structural isomer of a mid-length ketone with a common second component with no functionality. This limitation was imposed to isolate the cause of experimentally observed phenomena to the shifting polar group alone. The generated data could then be appropriately modelled using the polar sPC-SAFT variants and the capabilities of each model, as outlined above, assessed. To this end, isobaric binary vapour-liquid equilibrium data were measured for 2-, 3- & 4-heptanone with three separate normal alkanes of similar length (n-octane, n-nonane & n-decane) at 40kPa. The apparatus used was a dynamic Gillespie VLE still with temperature and pressure accuracies of 0.03°C and 1.6mbar respectively. Equipment verification was achieved through the reproduction of experimental data for the ethanol/1-butanol system at 1.013bar. The vapour and liquid samples for all nine systems were analysed by gas chromatography with a maximum compositional error of ±0.022 mole fraction. All reported data were found to be thermodynamically consistent using both the L/W and McDermott-Ellis consistency tests. When paired with a common n-alkane, all three structural heptanone isomers displayed similar qualitative trends in phase behaviour. Minimum boiling azeotropes were measured in all nine systems; in the high alkane region for n-octane systems (~98 mole%), the equal concentration region for n-nonane systems (34 mole% to 53 mole%) and in the very dilute n-alkane region for n-decane systems (~3 mole%). The n-nonane systems in particular highlighted the effect of shifting functional group, with completely separate phase envelopes away from the pure alkane composition space evident in a particularly small temperature range. Modelling was performed using in-house developed software, with pure component parameters generated for each system using five different regression procedures. The first was traditional fitting of the segment diameter (σ), segment number (m), segment energy (є/k) and the respective polar parameter (xp, np) to DIPPR correlations of pure component saturated vapour pressure, liquid density and the heat of vaporisation. The latter four procedures included the fixing of the polar parameter according to functional group correlations and the three instances of including the binary VLE data set for each of the three alkanes considered in this work. When applied to the nine binary ketone-alkane systems measured in this work, excellent predictions of the experimental data were in evidence in most cases and only small binary interaction parameters were necessary to correlate the data where pure predictions were poor. The performance of the parameter sets based on the fixing of the polar parameter and the inclusion of VLE data were consistent and of a high quality for both models, with near identical parameters generated in all four cases for each of the nine systems. The parameter sets generated in this fashion were shown to be applicable not only to the systems measured in this work, but also successfully predicted the independently measured experimental data of the n-hexane/4-heptanone system. It was thus concluded that either of these regression alternatives are viable for the generation of accurate component parameters, and the choice of VLE data set included is trivial. The pure predictions of the sPC-SAFTGV model were generally better than its sPC-SAFTJC counterpart, particularly in the case of the traditionally regressed parameter sets. sPC-SAFTGV displayed constant qualitative agreement with the experimental data for each of the heptanone isomers with a given n-alkane. The quality of the predictions of sPC-SAFTJC, however, worsened significantly as the polar interactions diminished from 2- to 4-heptanone, with no predictions even possible for the least polar isomer. This was attributed to the different perturbation theories used in the development of these terms, but a more detailed study would be necessary to confirm this. This work thus shows an apparent inability of the sPC-SAFTJC equation of state to account for the decreasing polar interactions associated with the carbonyl group in a ketone molecule shifting centrally, while sPC-SAFTGV produces qualitatively good fits for all three isomers. These flaws can be overcome through the incorporation of VLE data in the regression procedure if such data is available, or otherwise through the use of group specific correlations for fixing the polar parameter value.
- ItemCharacterisation and desalination of typical South African abalone farm effluent sea water(Stellenbosch : Stellenbosch University, 2012-12) Steynberg, Leander Duvan; Burger, A. J.; Stellenbosch University. Faculty of Engineering. Dept. of Process Engineering.ENGLISH ABSTRACT: Nearly all South African abalone farms function on an intensive pump-ashore, flow-through system. Large volumes of sea water that are pumped ashore flow through abalone or kelp harvesting tanks and finally gravitate back to the ocean. If the effluent from an abalone farm can be desalinated without permanent membrane fouling, then sea water reverse osmosis (SWRO) technology can be integrated effectively with established abalone farms without having to increase the farms’ intake system capacities. Without the need to construct and maintain an intake system, the overall cost of desalination can be reduced. Therefore, the aim of this study was to determine the feasibility and viability of integrating a SWRO desalination plant with a typical South African abalone farm. The project focused on four areas of concern, namely: - characterisation of typical South African abalone farm water - SWRO desalination plant pilot study and membrane fouling behaviour - general operation of a typical abalone farm and its implications for desalination - cost estimates and implications for the integration of an SWRO desalination plant with an abalone farm During a nine-month on-site investigation, sea water turbidity was reduced by up to 43% from a mean value of 0.82 NTU in the influent stream to 0.47 NTU in the combined effluent stream from the abalone tanks. Even with spikes in the influent turbidity, the turbidity of the combined effluent from all abalone tanks (excluding tank flush water) remained below 1 NTU. Dissolved organic carbon (DOC) in both the influent and combined effluent remained below 1 mg/litre. Ultrafiltration (UF) was selected as pre-treatment to the reverse osmosis (RO) in order to minimise potential fouling of the RO membranes. Membrane compaction of both the UF and RO membranes contributed significantly to initial flux losses – as much as 18% for the polyethersulfone (PESM) UF membranes and 20% for the thin film composite (TFC) polyamide RO membrane. However, this is comfortably in line with typical compaction values quoted in the literature. Without pre-flocculation, the UF was able to operate at a specific flux between 45 and 55 litre/m2/h (LMH) and recoveries ranging between 60 and 75%. Corresponding trans-membrane pressure (TMP) ranged between 0.59 and 0.76 bar. With ferric chloride pre-flocculation at a concentration of 3 mg/litre (as Fe3+) the UF could be operated at notably lower TMP values between 0.11 and 0.36 bar. These results indicate that provision should definitely be made for pre-flocculation when using UF as pre-treatment, despite the fact that the DOC concentrations and turbidity of the abalone farm effluent are quite low (DOC <1 mg/litre, NTU <1). It furthermore highlights the inability of DOC and turbidity alone to predict the membrane fouling potential of water. A better indicator of membrane fouling potential, albeit not perfect, is the modified fouling index (MFI0.45). This index follows a linear trend with foulant concentration and serves as a good indicator of the filterability of water. On-site measurements showed an increase in mean MFI0.45 values from 29 s/litre2 for the influent to 48 s/litre2 for the effluent from the abalone tanks, thereby confirming the need for pre-flocculation as part of UF pre-treatment. Chemically enhanced backwashing (CEB) of the UF membrane at least every 24 hours was found to be essential for its stable operation. Therefore, UF with pre-flocculation (3 mg/litre Fe3+) and regular CEB can be used effectively as pre-treatment method for the desalination of abalone farm effluent water. An RO ‘feed-and-bleed’ system was used to simulate the typical performance of the last membrane in a full-scale RO membrane bank. This RO membrane performed well with no signs of extreme fouling. The membrane produced a good quality permeate – for the last membrane in a membrane bank – reducing the TDS of the RO feed from 33 493 mg/litre to 969 mg/litre. These results compared well with simulated values by Reverse Osmosis System Analysis (ROSA; an RO simulator by DOW), indicating a TDS reduction from 33 271 mg/litre to 1 409 mg/litre at a feed pressure of 56 bar, and overall recovery of 44%. A steady performance of the RO membrane during the pilot study indicated that it is possible to desalinate abalone effluent water without notable permanent membrane fouling. A stable normalised flux rate of 8 LMH was achieved and RO membrane integrity remained intact with a salt rejection that ranged from 98.0 to 98.5%. No sudden reduction in permeate flux was observed as a result of fouling by unknown constituents present in the UF permeate. DBNPA (a non-oxidising disinfectant) was dosed once per week at a concentration of 10 – 30 mg/litre for 30 minutes. Scaling was controlled effectively by means of an antiscalant dosed at a concentration of 11 – 12 mg/litre in the feed stream. The CIP frequency was not optimised but a CIP frequency of once every 6 – 8 weeks appeared to be more than adequate to prevent permanent membrane fouling Advantages of integrating an SWRO desalination plant with a South African abalone farm include: - no lengthy and costly environmental impact assessment (EIA) is required to build a new intake system - shared capital and operational cost of intake system - dual incentive to keep constant good quality water flowing through the farm - early warnings regarding occurrences such as algal bloom and red bait - shared operational and management cost to keep pipelines clean - electricity saved (pumps for intake system) Disadvantages of integrating an SWRO desalination plant with a South African abalone farm include: - will require diverting of the abalone tank wash water from regular effluent - possible water ‘down-times’ due to maintenance operations on abalone farm Based on information from the literature the fixed capital cost depreciation rate (FCCDR) typically contributes approximately 40% and the operation and maintenance (O&M) cost typically contributes 60% to the unit production cost (UPC) of desalinated water. Furthermore, a SWRO desalination plant’s intake system can contribute between 5% and 33% to the FCCDR, depending on the nature and design of the plant. Consequently, the intake system can contribute between 2% and 13% of the UPC of desalinated sea water. This implies possible cost savings of between R0.15/m3 and R2.37/m3 for the production of fresh water (depending on site-specific design factors) when desalinating sea water effluent from on-shore abalone tanks. Integration of an SWRO desalination plant with a South African abalone farm is feasible and viable, provided that the necessary steps and precautions are taken to ensure a smooth and stable operation of the SWRO desalination plant. Cost savings on the part of all the stakeholders are possible if the correct contract can be negotiated.
- ItemThe chemical manipulation of meta-stable brine super-saturated with gypsum: forcing precipitation by overriding the inhibitory effect of antiscalants on crystal formation.(Stellenbosch : Stellenbosch University, 2011-12) Gerber, Daniel Hendrik; Burger, A. J.; Stellenbosch University. Faculty of Engineering. Dept. of Process Engineering.ENGLISH ABSTRACT: Desalination, by means of reverse osmosis (RO), in combination with other processes, can produce potable water at high recoveries. Antiscalants are generally used to reduce scaling on equipment surfaces and to improve water recovery during RO by slowing down the precipitation kinetics of sparingly soluble salts in the RO feed, thereby allowing concentration levels in the RO brine at several times the solubility limit of these salts. In addition, a fraction of the concentrate may be recycled back to the feed of the RO-membrane to improve the overall recovery, but only after the super saturated salts in the concentrate have been precipitated. The inhibitory character of the antiscalants (which are rejected into the concentrate stream) complicates the removal of salt from the concentrate and therefore prohibits such recycling. The focus of this study is aimed at properly understanding some of the parameters that influence the functionality or effectiveness of antiscalants used in high sulphate waters, with the purpose to override the effect of the antiscalant in the concentrate stream and force precipitation of the super saturated salts in solution. A batch crystallization technique, which considers the precipitation of calcium sulphate dehydrate (gypsum) from a solution of changing super saturation, was used to perform precipitation tests 1) on synthetically prepared solutions, super saturated with gypsum and 2) industrial concentrate, rich in sulphate (produced by concentrating acid mine drainage (AMD) by means of a lab scale RO unit). During batch crystallization, the precipitation process was observed by means of monitoring the depletion of calcium, using a calcium selective electrode (ISE). Deductions concerning the kinetics of precipitation were made from observing two kinetic variables (response variables) e.g. the induction time and the growth rate (tC80 – inferential variable). Two antiscalants have been evaluated in this study: a phosphonate based antiscalant (HYDREX) and a polyacrylate antiscalant (BULAB), at concentrations of 4 mg/l and 12 mg/l. The objective was to chemically and physically manipulate the antiscalant effectiveness, override its effect and force precipitation of gypsum by means of changing parameters in the system, such as the temperature (15°C- 25°C), pH (4-10), ferric chloride concentration (2-10 mg/l) or seeding the solution with gypsum seed at a concentration of 0-2000 mg/l. In addition, lime and a combination of gypsum and lime were also used for seeding at concentrations of 2000 mg/l. The induction time, prior to precipitation, was found to be most strongly affected by the change in seed concentration and pH at a given antiscalant concentration. Seed at a concentration of 2000 mg/l was sufficient in most cases to immediately override the effect of HYDREX and BULAB (at 4-12 mg/l) and produce ~ 0 minutes induction time. A pH of 10 increased the adsorption capacity of HYDREX and BULAB, leading to longer induction times (exceeding 24 hours in some cases). At a pH of 4 the adsorption capacity was very low for both HYDREX and BULAB (lower) leading to shorter induction times (zero to 100 minutes). It was especially in the ‘no-seed’ cases that the effect of pH on the induction time was prominent. The rate of precipitation (crystal growth rate) was increased at a temperature of 25°C, compared to 15°C (the rate increased two fold for an increase in 10°C). The addition of lime-seed, instead of gypsum, (at 2000 mg/l) produced growth rates, two times higher compared to when gypsum was used at the same conditions. In Addition, seeding with lime produced induction times (150 minutes for HYDREX and 50 minutes for BULAB) prior to precipitation, compared to zero induction time when gypsum was used at the same conditions. It was proven that an induction time could be eliminated by adding a combination of gypsum and lime both at a concentration of 2000 mg/l. with the added benefit of the higher growth rate. An increase in the calcium concentration increased the crystal growth rate in the presence of HYDREX. The presence of a high pH, however caused the effect of calcium on the growth (in the presence of BULAB) to be overshadowed. At a higher pH the growth rate of gypsum slowed down as a result of the increase in adsorption capacity of the polymer onto the crystal surface. The interaction of the antiscalant with FeCl3 seemed to be important with regard to crystal growth. Higher ferric concentrations (10 mg/l) were sufficient to limit the inhibitory effect of 12 mg/l antiscalant (HYDREX and BULAB) on the crystal growth rate. Conversely, low ferric concentration resulted in slower growth rates in the presence of an antiscalant. The best conditions (within the scope of the current study), sufficient 1) to override the inhibitory effect of antiscalants (HYDREX and BULAB) and 2) to produce rapid precipitation of gypsum, lie in the use of seeding with gypsum and lime (2000 mg/l), adding ferric chloride (10 mg/l), lowering the pH to 4 or lower (which can only be obtained when lime is not added) and setting the solution temperature to a moderate value of 25°C or higher. These ‘best’ conditions were subsequently applied to a concentrate, produced from concentrating AMD in a RO unit, and proved to be even more successful in overriding the effect of HYDREX and BULAB than in synthetic aqueous solutions. The induction times of precipitation of AMD in all cases were ~ 0 minutes, whereas the growth rate increased threefold compared to the synthetic tests. The presence of additional foreign precipitates of aluminum, calcium and magnesium as well as an increased [SO4ª-] x [Caª+] product of 3.73 (AMD concentrate) vs. 3.46 (synthetic solutions) is thought to be responsible for the increase in precipitation kinetics when only gypsum seed was used. The addition of lime caused an increase in the precipitation potential of the brine by increasing the calcium concentration. Although the addition of lime caused an increase in the pH to 12.3 (at which point the antiscalant was most effective), the increase in pH is likely to cause an increase in the natural carbonate in the water, which would stimulate CaCO3 precipitation. The CaCO3 precipitate would be responsible for the adsorption of antiscalants, reducing their efficiency.
- ItemComparing the group-contribution SAFT-γ Mie equation of state with SAFT-VR Mie(Stellenbosch : Stellenbosch University, 2019-12) Hurter, Ruan Martin; Burger, A. J.; Cripwell, Jamie T.; Stellenbosch University. Faculty of Engineering. Dept. of Process Engineering.ENGLISH ABSTRACT: Group-contribution methods (GCMs) allow engineers to reduce time and other resources spent on conducting experiments for parameterisation of thermodynamic models, because GCMs grant users the ability to build new model fluids using previously parameterised functional groups (FGs). GCMs have widely been applied to semi-empirical cubic equations of state (EoS), activity coefficient models, and simpler variants of statistical associating fluid theory (SAFT) EoS, yet these models remain limited in regard to the types of systems and properties they can describe. The rigorous SAFT-VR Mie EoS has a variable-range Mie-potential reference fluid and a complex dispersion term that enable it to accurately model second-derivative dependent properties, and properties in the near-critical region, but its model parameters are specific to components. The latter trait poses a problem when no pure-component data are available. A group-contribution (GC) variant of SAFT-VR Mie, SAFT-γ Mie, was recently developed in an attempt to combine the convenience of a GC model with the holistic predictions of SAFT-VR Mie. However, this model is relatively new and prior to this investigation, effects of the GC approach in the SAFT-VR Mie framework had not been evaluated in detail. The purpose of this project was to investigate whether assumptions made by the GC approach benefit or deteriorate different applications of this complex SAFT formulation. A general comparison between SAFT-γ Mie and SAFT-VR Mie was done to identify characteristics posing a unique challenge to SAFT-γ Mie by modelling components of increasing complexity: nonpolar, non-associating n-alkanes and 1-alkenes; polar, non-self-associating n-alkyl acetates; and polar, self-associating 1-alcohols. It was found that SAFT-γ Mie is able tomodel alkanes, alkenes, and acetates accurately, but it failed to produce equally accurate results for 1-alcohols, suggesting that the modelling of small polar molecules poses a problem for SAFT-γ Mie. This notion cannot be verified without doing a comparison involving a larger sample of polar, non-self-associating components. Ketones and esters were modelled to evaluate the performance of SAFT-γ Mie for polar components. This part of the study also provided the opportunity to evaluate the performance of the pseudo-association approach used to account for dipolar interactions, as well as the consistency in modelling accuracy between linear isomers. SAFT-VR Mie with the Gross & Vrabec (GV) polar term was used as a benchmark. SAFT-γ Mie cannot distinguish between structural isomers using exactly the same functional groups; therefore, new groups were defined for 2-ketones, 3-ketones, and n-alkyl propanoates. Results mirrored the 1-alcohol results, indicating that the modelling of smaller, more polar molecules poses a challenge for SAFT-γ Mie likely due to the disregarding of proximity effects — a change in functional group characteristics based on its environment, i.e. its surrounding groups or atoms. One likely solution to the problem is to introduce second-order group contributions to act as adjustments to first-order contributions, but this empirical adjustment would reduce the model’s fundamental predictive capability. Besides proximity effects, SAFT-γ Mie also disregards structural considerations such as steric hindrance and the order of intramolecular bond formation; this is also expected to have an impact on SAFT-γ Mie’s performance. It was found that these structural considerations are vital for accurate modelling of branched alkanes, and that significant differences can be observed in properties of branched alkane isomers. While SAFT-VR Mie models all of the considered branched alkanes accurately, SAFT-γ Mie does not. It was found that the homosegmented approach followed in SAFT-γ Mie’s chain term prevents the model from making any distinctions based on a molecule’s layout. A new heterosegmented chain term was proposed: Bonding contributions would be calculated between segments of unique groups instead of between approximated molecular-average segments. Different methods for weighing the contributions of intra- and intergroup bonds were discussed. Although preliminary results suggest that a heterosegmented chain term would allow the model to distinguish between isomers, further investigation is required to evaluate the consequences of the proposed changes.
- ItemControl of biofouling on reverse osmosis membranes using DBNPA(Stellenbosch : Stellenbosch University, 2016-03) Ras, Gabriel Retief; Burger, A. J.; Stellenbosch University. Faculty of Engineering. Dept. of Process Engineering.ENGLISH ABSTRACT: Reverse Osmosis (RO) is used throughout the world for water desalination and it has gained wide popularity due to its efficient energy consumption and the safe operating process. Fouling (of which biological fouling is the most problematic) of the membranes is, however, an inevitable process that cannot be avoided, only managed. Biological fouling is the growth of microorganisms in the membrane system, causing undesirable effects. The correct pre-treatment can reduce (but not necessarily prevent) biofouling. This is because microorganisms have the ability to reproduce and form secondary populations throughout the membrane system, even if 99.99% of the microorganisms are removed in the pre-treatment process. Most modern RO plants are equipped with thin film composite polyamide (TFC PA) membranes. However, biological control on such membranes is restricted, since oxidising biocides like chlorine degrade the membrane material, thereby increasing salt passage and reducing membrane life. Therefore, this study investigated the use of a common non-oxidising biocide, i.e. 2,2-dibromo-3- propionamide (DBNPA) to manage biological growth on TFC PA membranes. The primary aim was to demonstrate the influence of three DBNPA dosing variables on the control of biofouling on the RO membranes. These variables were dosage (10 ppm to 200 ppm), dosing frequency (twice daily to once every 2nd day) and dosing duration (30 min to 2 hours). The work also strongly relied on the characterisation of biological fouling through determination of biomass parameters (protein concentration, polysaccharide concentration, total cell count and colony-forming units) and linking it to flux decline. Tests were conducted in lab-scale RO membrane blocks, housing flat-sheet TFC PA membranes with appropriate flow spacers typically found in commercial spiral-wound membrane cartridges. Since clean municipal water was used as feed water, nutrients (sodium acetate, sodium nitrate and sodium dihydrogen orthophosphate, in the ratio of 100:20:10 to give a final carbon concentration of 100 μg/ℓ) were supplemented to stimulate sufficient microbial growth, thereby enabling a sensible study on the effect of DBNPA dosing. During the removal of the biofilm from the membrane, no combination of the removal and homogenisation techniques (e.g. scraping the biofilm from the membrane, ultrasonic bath and ultrasonic probe treatment) yielded significantly higher colony forming unit (CFU) counts. R2A agar, however, produced significantly higher CFU counts compared to nutrient agar. Therefore, the agar used during plate counts appears to have been of greater significance on cell enumeration than the combination of biofilm removal and homogenisation techniques, which had little effect on cell counts, irrespective of agar used. DBNPA dosing reduced the amount of biofouling, regardless of the dosing strategy used. However, within the scope of this study, biofouling was best controlled with a DBNPA dosage of 100 ppm for two hours once per day. Applying the same dosing strategy every second day, was not as effective in limiting flux decline, but still produced better results than the remaining dosing strategies. This supports the notion of a sufficiently high dosage for an optimal time, rather than high concentration shock-dosages for a short period. A significant increase in biomass parameters (cell count, colony forming units, and protein- and polysaccharide concentration) was observed when nutrients were added to the feed water. Protein concentration (p=4.29 x 10-5, R2=0.71) and polysaccharide concentrations (p=0.0053, R2=0.58) on the membrane had a strong and significant relationship with the flux decline, making it suitable parameters for biofouling quantification. CFU showed a significant, but not strong, (p=0.0011, R2=0.54) relationship to the flux decline, whereas total cell count did not provide a significant (p=0.14) relationship. Protein- and polysaccharide concentrations could therefore be used for the quantification of the biofouling. A destructive study should, however, be performed to determine these parameters. A practical tool is therefore still necessary for the early diagnosis of biofouling. For future studies, it is recommended that larger ranges of cross-flow velocities and pressures be investigated, together with the effect of DBNPA dosing. Ideally, the work should be performed on a membrane that is packed in a spiral-wound format to simulate real-life situations.
- ItemA critical evaluation of the mass transfer and fouling behaviour in forward osmosis with integrated flow-reversal(Stellenbosch : Stellenbosch University, 2019-04) Hurter, Marielle; Burger, A. J.; Stellenbosch University. Faculty of Engineering. Dept. of Process Engineering.ENGLISH ABSTRACT: Forward osmosis (FO) is a membrane-based technology that can be operated at relatively low mechanical pressures and may be utilised in processes where water needs to be added or removed from process streams. Options for its potential application are diverse and it could, for example, be used in the regulation of water content in fruit juices, or in the augmentation of clean water to high-TDS cooling water circuits. Similar to reverse osmosis (RO) processes, scale formation by sparingly soluble salts can limit the maximum allowed recovery of water, while flux profiles, salt rejection characteristics and cross-flow velocity (CFV) play key roles in the overall behaviour of the system. However, FO systems are more amenable to the utilisation of osmotic backwashing than RO systems. Therefore, this study endeavoured to critically evaluate the mass transfer and fouling behaviour of FO membranes at different operating conditions, including the intermittent switching of the flow path (i.e. intermittently reversing the flux). To support this study, a bench scale FO setup was designed, constructed and commissioned. Subsequent laboratory work entailed: - Evaluate and assess the bench scale setup by comparing the theoretical and measured recovery, based on the measured water flux. - Evaluate the effects of changes in the CFV on the mass transfer of water and solutes over the membrane, while using a feed solution with TDS well below 100 mg·L-1. - Determine the effects of the operational configuration on the mass transfer over the membrane. - Investigate the process realities and limitations of intermittent flow path switching on reducing scale formation. Two operational modes were considered, viz. with the membrane active layer (1) facing the feed solution (AL-FS) or (2) facing the draw solution (AL-DS), with CFVs ranging from 13 cm.s-1 to 52 cm·s-1.Within this CFV range, the water fluxes attained in the AL-FS configuration were on average 40%lower than those in the AL-DS configuration. In the AL-FS configuration, the flux increased from 11.2 L·m-2·h-1 to 20 L·m-2·h-1 when the CFV was increased from 13 cm.s-1 to 37 cm·s-1. However, a further increase in CFV above 37 cm·s-1 did not result in higher fluxes and the limiting flux of 20 L·m-2·h-1 was reached. This is ascribed to the potential increase in dilutive internal concentration polarisation in the support layer of the membrane, thereby limiting the effective driving force (effective osmotic pressure difference) over the membrane. In the AL-DS configuration, this limiting flux was not reached within the defined CFV range. However, it was found that operation in the AL-DS configuration tended to a limiting flux of 20 L·m-2·h-1 when operating at draw solution concentrations above 50 000 mg·L-1 TDS. This is considered to be partly the result of an increased reverse solute flux (RSF) along with dilutive external concentration polarisation on the active layer side of the membrane. During operation with intermittent flow path switching when recovering water from a 1.9 super-saturated gypsum feed solution, ca 15 minutes were required to purge the flow channels of the respective residual solutions in the specific laboratory system under investigation. Operation at a CFV of ca 28 cm·s-1 then proved to enable the most rapid alleviation of internal concentration polarisation (ICP) in the AL-FS configuration (or mostly RSF in the AL-DS configuration). Under the most stable conditions in the AL-FS configuration, the operational flux dropped from 12 L·m-2·h-1 to ca 9 L·m-2·h-1 over a period of only 12 hours. In other words, flux declines of ca 38% were observed over a period of 12 hours when operating in the AL-DS configuration at 15-minute switch-intervals every two hours. This indicated the formation of gypsum scale in the support layer and highlighted the detrimental effects of the support layer in a scaling environment.
- ItemThe development of simplistic and cost-effective methods for the evaluation of tray and packed column efficiencies(Stellenbosch : Stellenbosch University, 2017-03) Lamprecht, Johannes Hendrik; Burger, A. J.; Stellenbosch University. Faculty of Engineering. Dept. of Process Engineering.ENGLISH SUMMARY: Phase contacting column internals are manufactured through a series of punching, die moulding and bending. Although cost-effective on a large scale, this process is considered unfavourable for prototyping, as it is both time-consuming and expensive. This limits designer creativity and introduces extended waiting periods between the design, fabrication and evaluation phases. This translates into development timelines in excess of two years. During the evaluation stage, efficiencies are conventionally measured using hydrocarbons system at total reflux. This introduces notable constraints on the prototype packing and tray material, due to the high temperatures and pressures required for these tests. Therefore, this research project focused on the development and experimental validation of two simplistic and cost effective methods that can be used to quantify column internal efficiency. The ADIBAA (aqueous desorption of isobutyl acetate in air) and HA (Humidification of air) methods are hereby proposed for efficiency measurements in packed and tray columns respectively. For validation of both methods, two separate pilot plant facilities were designed and constructed, one at Stellenbosch University and one at an industrial research laboratory. The ADIBAA-method involved using a liquid phase limiting system to isolate the performance parameters in the liquid phase. The combination of the method and experimental setup offered rapid quantification, while remaining cost-effective and environmentally friendly. The ADIBBA-method was experimentally validated in a 400mm diameter stainless steel packed column, with a bed height of 1.1 metre. Such validation entailed (a) experimental measurement of isobutyl acetate concentrations, (b) calculation of volumetric liquid phase mass transfer coefficients and (c) comparison of these calculated coefficients with predictions by four independent correlations from literature. Agreement between the literature correlations and the newly-determined experimental data was found to be within 10%. The applicability of the ADIBAA-method, in evaluating column internal efficiencies, was confirmed through comparison of 1.5’ FlexiRings® and the equivalent Intalox® Ultra™. A quantifiable improvement of 15% was recorded in the preloading regime, in favour of the Intalox® Ultra™. Further justification of the ADIBAA-method was presented in the evaluation of 2.5’ Intalox® Ultra™ packing. The HA-method, proposed to use for tray columns, focused on the evaluation of Murphree tray efficiencies. This method offered large improvements over the constant reflux method in terms of environmental and safety considerations, while also reducing the experimental time by an order of magnitude. A rectangular tray column with respective weir and flow path lengths of 762mm and 870mm was used in the experimental evaluation. The HA-method was found to adequately quantify hydrodynamic variations in both weeping and vapour bypass. The comparative ability of the method was experimentally verified by relating a 12% open area sieve tray with two separate prototypes. The method enabled rapid evaluation and quantitatively illustrated difference in efficiencies between the prototypes. From this research, it follows that the ADIBAA- and HA- methods can indeed be used as cost-effective, simple and time-efficient methods to evaluate prospective designs of random packing and trays.
- ItemThe effect of feed pH and in-situ pH adjustment on the behaviour of an anaerobic sequencing batch reactor treating synthetic winery wastewater(Stellenbosch : Stellenbosch University, 2017-03) Smit, Jason; Burger, A. J.; Sigge, G. O.; Stellenbosch University. Faculty of Engineering. Dept. of Process Engineering.ENGLISH SUMMARY:Legal requirements and high production costs have resulted in wineries having to seek alternative methods of reducing operating costs with regards to the reduction of fresh water intake as well as treating the wastewater they produce. Winery wastewater typically contains varying concentrations of monosaccharides, volatile fatty acids and ethanol. To legally dispose of winery wastewater, its chemical oxygen demand (COD) needs to be reduced to below 400 mg/ℓ when disposing between 50 and 500 m3 daily. An anaerobic sequencing batch reactor (ASBR) has been recommended as a possible method to treat winery wastewater due to several benefits over typical aerobic systems. Anaerobic digestion systems produce useful biogas and, compared to aerobic digestion systems, generate low volumes of sludge. These advantages, together with the simplicity and relatively low installation costs of sequencing batch reactor (SBR) systems, make ASBR technology an attractive option for treatment of winery wastewater. Winery wastewaters have varying pH (3.5 – 8.0) due to the various sources from the plant. As such, the main objective of this study was to determine whether an ASBR can be operated with in-situ pH control and without adjusting the feed alkalinity. Exploratory simulations were performed with the Anaerobic Digestion Model no. 1 (ADM1) to understand where the potential problems could occur with the experimental ASBR. Sludge retention in the ASBR was simulated through the incorporation of a clarification model in the ADM1. The simulation results indicated that winery wastewater with high monosaccharide concentrations would cause a sudden drop in pH early in the ASBR process due to rapid production of volatile fatty acids. It therefore followed that in-situ pH control would be required. The ADM1 was found to be unstable when poor initial guesses of the soluble and sludge component concentrations were used in the simulations. With the ADM1 simulation, the pH was identified as the variable which would easily indicate instability in the model. Following the ADM1 simulations, a 14-litre laboratory scale ASBR was used to treat different synthetic winery wastewaters while operating with in-situ pH control. Two artificial feed solutions were prepared, the first with a high ammonium sulphate concentration and the second without ammonium sulphate. Both solutions contained high concentrations of glucose and fructose. The ASBR could handle the ammonium sulphate between organic loading rates (OLRs) 1.1 and 2.1 g-CODfeed.ℓ-1ASBR.day-1. Under these conditions, the ASBR achieved a COD reduction of at least 60 %. In the absence of ammonium sulphate, the ASBR achieved a COD reduction of at least 80 %. Biogas containing methane, carbon dioxide and nitrogen was produced. Theoretically excluding the nitrogen from the biogas resulted in a methane fraction in excess of 80 mol%, with the balance being carbon dioxide. 𝐾𝑂𝐻 was dosed as a nutrient. Correcting the feed pH to 7.4, allows for an approximate saving of 8 – 12% on the total amount of 𝐾𝑂𝐻 required for feed substrate dosing and in-situ pH control. In-situ pH control was deemed to be the most important during the first five hours of a batch. After this, the methanogens generally consumed acetic acid fast enough to counter the effect that volatile fatty acid formation has on the pH of the system.
- ItemThe effect of forward flushing, with permeate, on gypsum scale formation during reverse osmosis treatment of CaSO4-rich water in the absence of anti-scalant(Stellenbosch : Stellenbosch University, 2014-12) Otto, Dietmar Norman; Burger, A. J.; Stellenbosch University. Faculty of Engineering. Dept. of Process Engineering.ENGLISH ABSTRACT: When desalinating brackish water by reverse osmosis (RO) or other techniques, high overall water recoveries are essential to minimize brine production and the associated disposal costs thereof. As the overall water recovery increases, concentrations of sparingly soluble salts (e.g. barium sulphate, calcium sulphate) reach levels above saturation, especially near the membrane surface, drastically increasing the scaling propensity. Antiscalants are typically dosed into the feed water to prevent such scaling during RO desalination. However, the carry-over of antiscalant into the concentrate stream can complicate subsequent salt precipitation processes that may be used to increase overall water recovery. These precipitation techniques are sometimes used to reduce the levels of super-saturation in the RO concentrate prior to a subsequent RO desalination step. The purpose of this study was to assess the feasibility of reducing calcium sulphate scaling on RO membranes, by using periodic permeate flushing when feeding a lab-scale RO unit with a supersaturated calcium sulphate solution in the absence of anti-scalant. The overall water recovery was increased by recycling the concentrate, after an intermediate de-supersaturation step. This simulated a multiple-stage RO system, typical of processes used in high-recovery acid mine drainage (AMD) treatment plants. De-supersaturation of the concentrate intermediate was achieved with direct seeded gypsum precipitation, in the absence of any antiscalant. On the membrane surface inside the membrane unit, calcium sulphate concentrations greatly exceeded saturation levels – a combined consequence of the normal concentration process and the well-known surface-based concentration polarisation phenomenon. Therefore, periodic forward-flushing of the supersaturated solution from the membrane unit was performed with permeate. In theory, the periodic flushing removes the highly concentrated layer at the membrane surface during every flush, before scaling can occur. Various flushing regimes were evaluated to assess the effectiveness of the process. A lab-scale desalination unit with a 0.106 m2 flat sheet polyamide RO membrane was designed and constructed. The unit could operate at a feed rate of 12-14 L/h and at permeate fluxes of 12-24 LMH. Super-saturated feed solutions were prepared by mixing sodium sulphate and calcium chloride dihydrate salts with demineralised water, with an initial salinity of ± 5300 mg/L TDS, corresponding to a gypsum saturation index (SIg) of 1.2 for most experiments. The total production time, net permeate production and flux decline were used to compare the flushing efficiency in different experimental runs. Initial tests showed that scaling could be prevented (when operating the unit in full recycle mode, i.e. where both concentrate and permeate were recycled to feed), at flushing frequencies between 12 and 2.4 h-1, when the membrane feed and concentrate were slightly under-saturated (SIg = 0.9) and slightly super-saturated (SIg = 1.1) respectively. However, when switching the same system to non-flushing mode after 24 hours of operation, membrane scaling occurred within 2-3 hours, as indicated by a strong decline in flux. However, when operating the system in concentrate recycle mode (i.e. permeate is withdrawn) with super-saturated feed solutions (e.g. SIg = 1.2), and thus a notably more super-saturated solution in the membrane concentrate, scaling could not be prevented (albeit delayed for some time) with intermittent permeate flushing. A fractional 25-1 factorial design was used to determine which factors had the most significant effect on total production time and permeate production rate, testing five factors: 1) flushing frequency, 2) flushing volume, 3) permeate soak time, 4) permeate flux and 5) instantaneous recovery. The ANOVA analysis showed that total production times were, not surprisingly, primarily affected by the permeate flux, where operation at 24 LMH resulted in a lower net permeate production between 3.0 - 4.2 L, compared to 7.6 - 9.7 L at 12 LMH. Higher permeate fluxes clearly resulted in higher levels of concentration polarisation at the membrane surface, thus increasing the propensity for membrane scaling. Flushing frequency and instantaneous recovery also affected the net permeate production, where 6 h-1 and 10 % were the optimal values respectively within the range of test conditions. The lowest permeate production rate resulted in the highest net permeate volume production (i.e. also longest total production time), confirmed by a least squares regression. In summary: This study showed that periodic permeate flushing could delay the membrane scaling process. However, it failed to prevent membrane scaling completely when operating the system with supersaturated calcium sulphate solutions in the absence of antiscalants. The flushing technique effectively delayed the onset of precipitation, but scaling eventually occurred if the lab-scale RO system was operated in concentrate recycle mode with oversaturated feed solutions (SIg = 1.2). Additional experiments at different cross-flow velocities during permeate flushing, while using an optimised RO test cell flow channel design, are recommended for future studies.
- ItemThe effect of humic substances on the crystallisation of gypsum(Stellenbosch : Stellenbosch University, 2017-03) Bock, Heinrich; Burger, A. J.; Stellenbosch University. Faculty of Engineering. Dept. of Process Engineering.ENGLISH SUMMARY: High recovery mine water treatment plants generate brine streams that are highly supersaturated with inorganic salts. Intermediate crystallisation of gypsum is required for further treatment of these brine streams. The crystallisation of gypsum is influenced by various factors such astemperature, supersaturation, additives orimpurities, pH and seeding. The presence of natural organic matter (NOM), consisting of humic substances (HS), can prevent the onset of crystallisation. These substances, mainly composed of humic (HA) and fulvic acids (FA), are considered to be weak polyelectrolytes due to their carboxylic and phenolic functional groups. The aim of this study is to investigate the effect of HS on the crystallisation of gypsum to understand the mechanisms (nucleation and crystal growth) of crystallisation better. This knowledge can be used to improve the sizing and operation of crystallisers. The effect of HS was investigated at supersaturation (SS) of 2-4, pH of 4.5–9.5 and seed loading of 200, 1000 and 2000 mg/l through a batch crystallisation process. An increase in HS concentration resulted in an increase in induction times due to the increased inhibitory effects of HA and FA through their functional groups. Induction times increased from 25 to 295 minutes with an increase in HA concentration from 0 to 15 mg/l at SS3 (0.0419 mol/l). At a HA concentration of 15 mg/l, an increase to SS4 (0.0566 mol/l) resulted ina decrease of induction times from 295 to 15 minutes, indicating the driving force of supersaturation. Increase in initial pH enhanced the inhibitory abilities of both HA and FA. Induction time increased from 115 to 415 min with an initial pH increase from4.5to9.5 in the presence of 15 mg/l HA at SS3 (0.0419mol/l). The effect of FA was far greater than HA, with crystallisation completely inhibited for a period of 2 days at a FA concentration of 5 and 15 mg/l in the absence of any seed crystals. At FA concentrations of 1.0 and 2.5 mg/l, induction times were 185 and 480 minutes, respectively. The greater effect of FA is attributed to an increase inthe number of functional groupswith a decrease in molecular weight. Seeding the crystallisation process successfully overcame the inhibitory effects of HS (both HA and FA) at concentrations of 1000 and 2000 mg/l gypsum seed crystals. With a seed concentration of 200 mg/l, an induction period of 50 min was observed in the presence of HA at 15 mg/l. With FA at 10mg/l and a gypsum seeding of 200 mg/l, no crystallisation was induced. This again illustrated the enhanced effect of FA to block active growth sites successfully. In the presence of seed crystals pH has no effect, suggesting that only surface interaction is taking place. With HA, an increase in seed crystals resulted in an increased growth rate (from 0.50 to 4.91 litre.mol-1.min-1) due to an increase in available growth sites. The inhibiting and retarding effect of HS on crystallisation is significant. Increasing supersaturation can override the inhibitory abilities of HS, while the presence of sufficient seed material will completely override the inhibitory abilities of HS and minimise the effects of these substances on crystallisation.
- ItemEffect of liquid and gas physical properties on the hydrodynamics of packed columns(Stellenbosch : Stellenbosch University, 2017-12) Minne, Ulrich Ludkin; Schwarz, C. E.; Burger, A. J.; Stellenbosch University. Faculty of Engineering. Dept. of Process Engineering.ENGLISH SUMMARY: To fully utilise the higher capacity and separation efficiency of modern random packings, models are required that are capable of accurately predicting the capacity and separation efficiency of these packings. Since the capacity and separation efficiency in packed columns are influenced by the physical properties of liquids and gases, experimental data with varied liquid and gas physical properties are required for both the development and validation of these models. The aim of this project was to investigate the effect of liquid and gas physical properties on the hydrodynamic behaviour of Intalox Ultra A and O random packing with nominal sizes of 1.5” and 2.5”, respectively. This was achieved experimentally by measuring the pressure drop, liquid hold-up and liquid entrainment for four liquids with different viscosities, densities and surface tensions, and two gasses with different densities, at different liquid flow rates over the entire hydrodynamic range. The pilot plant in which the experimental work was performed had a column diameter of 400 mm and a packed bed height of 3 m. Liquid flow rates of 6, 37, 73, 98 and 122 m3/(m2.h) were investigated. The small open area of the existing pan type distributor restricted the accurate measurement of liquid entrainment. Therefore, it was replaced with a channel-type distributor, doubling the available open area for gas flow to 60 %. A method was developed to identify the loading and flooding points when presented with the experimental pressure drop data, based on the statistical approach of prediction intervals of regressed curves. Overall, this method provides relatively accurate identification of the loading and flooding points at high liquid flow rates, while the use of entrainment data in identifying the flooding point was found to over-predict the flooding point at high liquid flow rates. The high viscosity of silicone oil and ethylene glycol resulted in these liquids having higher pre-loading liquid hold-up compared to that of water and Isopar G. For silicone oil and ethylene glycol, the ratio of viscous forces over the gravitational forces is much larger than for water and Isopar G. Despite the difference in their physical properties, Isopar G and silicone oil produced very similar flooding velocities, which were lower than that of water and ethylene glycol. While the high viscosity of silicone oil resists the flow of liquid down the column, resulting in low capacity, the low density of Isopar G allows the upward drag force of the gas to counteract the gravitational force on the liquid at a lower gas flow rate than the other liquids with higher densities. At the same superficial gas velocity, carbon dioxide, with a higher gas density, produced a higher pressure drop than air/nitrogen, as well as a higher liquid hold-up above the loading point. This is a result of the higher gas kinetic energy of carbon dioxide compared to air/nitrogen at the same superficial gas velocity. The general effect of increasing the packing size from 1.5” Intalox Ultra A to 2.5” Intalox Ultra O is a reduction in both pressure drop and liquid hold-up, as well as an increase in capacity of approximately 11 % at a liquid flow rate of 6 m3/(m2.h), increasing to approximately 37 % at a liquid flow rate of 122 m3/(m2.h). A comparison of the Billet and Schultes, Maćkowiak and Stichlmair model predictions for the experimental data showed that in general all three models predict the pre-loading pressure drop for both packings with all four liquids poorly. Overall, the Maćkowiak model predicts flooding points better than the other two models. The Maćkowiak model is the only model based on the droplet entrainment modelling approach, whereas the Billet and Schultes and Stichlmair models are based on the liquid film modelling approach. The Maćkowiak model is the only model that takes into account the surface tension of the liquid. The experimental pressure drop was also compared to the pressure drop predicted with KG-Tower. The predicted and experimental pressure drop were in good agreement in the pre-loading region, with some deviation at higher gas flow rates. Overall, KG-Tower predicted the pressure drop better than the Billet and Schultes, Maćkowiak and Stichlmair models. The project created an extensive data set of experimental hydrodynamic data, for a range of physical fluid properties, which can be used in both the verification of existing hydrodynamic models and the development of new models.
- ItemEntrainment in an air/water system inside a sieve tray column(Stellenbosch : University of Stellenbosch, 2010-03) Uys, Ehbenezer Chris; Knoetze, J. H.; Burger, A. J.; University of Stellenbosch. Faculty of Engineering. Dept. of Process Engineering.ENGLISH ABSTRACT: Mass transfer efficiency in distillation, absorption and stripping depends on both thermodynamic efficiency and hydrodynamic behaviour. Thermodynamic efficiency is dependent on the system kinetics while hydrodynamics is the study of fluid flow behaviour. The focus of this thesis is the hydrodynamic behaviour in tray columns, which affects entrainment. In order to isolate hydrodynamic behaviour from the thermodynamic behaviour that occurs inside sieve tray columns, investigations are conducted under conditions of zero mass transfer. When the gas velocity is sufficiently high to transport liquid droplets to the tray above, entrainment occurs. The onset of entrainment is one of the operating limits that determines the design of the column and thus impacts on the capital cost. By improving the understanding of the parameters that affect entrainment, the design of the tray and column can be improved which will ultimately increase the operability and capacity while reducing capital costs. Existing correlations predicting entrainment in sieve tray columns are based on data generated mainly from an air/water system. Previous publications recommend that more testing should be performed over larger ranges of gas and liquid physical properties. An experimental setup was therefore designed and constructed to test the influence of the following parameters on entrainment: 1. gas and liquid physical properties 2. gas and liquid flow rates 3. tray spacing The experimental setup can also measure weeping rates for a continuation of this project. The hydrodynamic performance of a sieve tray was tested with air and water over a wide range of gas and liquid flow rates and at different downcomer escape areas. It was found that the downcomer escape area should be sized so that the liquid escaping the downcomer always exceeds a velocity of approximately 0.23 m/s in order to create a sufficient liquid seal in the downcomer. For liquid velocities between 0.23 and 0.6 m/s the area of escape did not have an effect on the percentage of liquid entrained. It was also established that entrainment increases with increasing gas velocity. The rate at which entrainment increases as the gas velocity increase depends on the liquid flow rate. As soon as the liquid flow rate exceeded 74 m3/(h.m) a significant increase in entrainment was noted and the gas velocity had to be reduced to maintain a constant entrainment rate. This is because the increased liquid load requires a longer flow path length for the froth to fully develop. The undeveloped froth, caused by the short (455 mm) flow path, then creates a non-uniform froth that is pushed up against the column wall above the downcomer. Consequently, the froth layer is closer to the tray above resulting in most of the droplets ejected from the froth reaching the tray above and increasing entrainment. By reducing the gas velocity, the froth height and ejecting droplet velocity is reduced, resulting in a decrease in entrainment. The results from the experiments followed similar trends to most of the entrainment prediction correlations found in literature, except for the change noted in liquid flow rates above 74 m3/(h.m). There was, however, a significant difference between the experimental results and the correlations developed by Hunt et al. (1955) and Kister and Haas (1988). Although the gas velocities used during the air/water experiments were beyond the suggested range of application developed by Bennett et al. (1995) their air/water correlation followed the results very well. The entrainment prediction correlation developed by Bennett et al. (1995) for non-air/water systems was compared with the experimental air/water results to test for system uniformity. A significant difference was noted between their non-air/water prediction correlation and the air/water results, which motivates the need for a general entrainment prediction correlation over a wider range of gas and liquid physical properties. Based on the shortcomings found in the literature and the observations made during the experiments it is suggested that the influence of liquid flow path length should be investigated so that the effect on entrainment can be quantified. No single correlation was found in the literature, which accurately predicts entrainment for a large range of liquid loads (17 – 112 m3/(h.m)), high superficial gas velocities (3 – 4.6 m/s) and different gas and liquid physical properties. It is therefore recommended that more work be done, as an extension of this project, to investigate the influence of gas and liquid physical properties on entrainment (under zero mass transfer conditions) for a large range of liquid (5 – 74 m3/(h.m)) and gas (2 – 4.6 m/s) flow rates. In order to understand the effect of droplet drag on entrainment, tray spacing should be varied and increased to the extent where droplet ejection velocity is no longer the mechanism for entrainment and droplet drag is responsible for droplet transport to the tray above. Since it is difficult and in most cases impossible to measure exact gas and liquid loads in commercial columns, another method is required to measure or determine entrainment. Since liquid hold-up was found to be directly related to the entrainment rate (Hunt et al. (1955), Payne and Prince (1977) and Van Sinderen et al. (2003) to name but a few), it is suggested that a correlation should be developed between the dynamic pressure drop (liquid hold-up) and entrainment. This will contribute significantly to commercial column operation from a hydrodynamic point of view.
- ItemEstablishing a facility to measure packed column hydrodynamics(Stellenbosch : University of Stellenbosch, 2010-12) Lamprecht, Sarel Marais; Knoetze, J. H.; Burger, A. J.; University of Stellenbosch. Faculty of Engineering. Dept. of Process Engineering.ENGLISH ABSTRACT: Distillation continues to be the most widely used method of separation in the processing industry, in spite of its inherently low thermodynamic efficiency. Two of the critical distillation research needs that arose from the US-Initiative Vision 2020 were to develop a better understanding of the physical phenomena as well as developing better predictive models. Also, characterisation of modern packing materials is required to assist in the CO2 capture optimisation. This thesis deals with both these aspects by establishing a facility that can accurately measure the hydraulic capacity of packed columns. This setup eliminates mass transfer and specific attention can be given to the hydrodynamic behaviour of packed columns. Two phenomena that have a large impact on the mass transfer efficiency of packing materials are the loading and flooding point. The loading point is signified by the following: a.) where the packed column hold-up increases, b.) higher increase in pressure drop, and c.) a decrease in Height Equivalent to a Theoretical Plate (HETP). The onset of flooding is where the shear forces between the gas and liquid become so large (relative to the gravitational forces) that a net upwards movement of liquid occurs, resulting in liquid droplets being heavily entrained. This is normally accompanied by a sharp increase in HETP, pressure drop and liquid hold-up. The prediction of these operating limits is of great value but, despite the many contributions that were made from 1960 to 2010, there is still room for improvement. The operating region of particular interest is between the loading and flooding point, especially for fluids with physical properties significantly different from that of water. In the past, this operating region was not of great importance, but industries are constantly striving to increase their production with minimal capital expenditure. Thus, packed columns are being pushed to their limits and a good understanding of the phenomena occurring near these operational limits is now required. A 400 mm diameter glass packed bed setup (with a bed height of 3000 mm) was designed and constructed to test the effect of the following parameters on packed bed pressure drop and liquid hold-up: · Gas and liquid physical properties · Gas and liquid rates · Type of packing (either random or structured) The experimental setup has been designed so that in the future the influences of the above mentioned parameters on entrainment can also be measured. Initially, hydrodynamic tests on random packing materials (1.5” Pall® Rings, 1.5” IMTP®, 1.5” Intalox® Ultra™) were conducted over a liquid range of 6 - 122 m3/(m2·h). Through a thorough literature study it was found that the most likely semi-theoretical model, that would be able to predict the pressure drop and the liquid hold-up over most of the random packing test range, was the model developed by Billet [1991; 1993; 1995; 1999]. The other models found throughout the literature had at least one of the following deficiencies: · Limited to only the pre-loading region. · Tested (and thus applicable) only over a very select group of packing materials with no attempt to generalise. · Lacked the proper validation of significantly variable fluid properties over multitudes of liquid and gas rates especially, at higher gas and liquid rates. The experimental setup was successfully commissioned, noting the following maximum experimental errors: Vapour flow factor - 2.6 %; liquid rate - 0.75 %; packed bed pressure drop - 0.75 %; liquid hold-up - 1.25 % and entrainment - 1.05 %. Significant deviations were observed between the experimental hold-up and the hold-up from the predictive model of Billet (using Pall® Rings). Careful inspection revealed that this predictive model potentially uses two definitions for hold-up at flooding, one which has a theoretical basis and the other purely empirical. Upon substituting the theoretical value with the empirical value, a significant improvement was observed between the measured and predicted results. Deviations were still observed near the flooding point and were attributed to the difficulty of obtaining reliable flooding data. The range of liquid hold-up prediction by Billet was only verified up to a liquid rate of 82 m3/(m2·h) and the pressure drop prediction only verified up to a liquid rate of 60 m3/(m2·h). This reinforces the need for high liquid, high gas rate data. Due to the empirical nature of the liquid hold-up at flooding prediction, and since pressure drop prediction is directly linked to liquid hold-up, another model was used to compare the experimental pressure drop data. The KG-TOWER® simulator was used to predict IMTP® data and compare it to the experimentally measured values. It was found that the experimental IMTP® data followed the same trends as those from KG-TOWER® within the operating limits of the program. Thus, since the experimental data follows similar trends as models found in the literature, as well as falling within their reliable limits, the experimental setup can correctly measure the parameters in question. The experimental data from the different random packings were compared to one another by using a statistical method to determine the loading point and onset of flooding. This method uses prediction confidence intervals by fitting empirical curves to each operating region and was found to be useful in determining these critical points from experimental hydraulic data (in the absence of HETP data).The only useful comparison was between IMTP® and Intalox® Ultra™ as they both have roughly the same density, size and void fraction. It was found that, on average, the pressure drop of Intalox® Ultra™ is 20 % lower than that of IMTP® over the entire operating range. The hydraulic operating range of Intalox® Ultra™ was found to be on average 16 % larger than that of IMTP®. It is recommended that further testing should be done to investigate the influence of fluid properties (specifically liquid viscosity and to a lesser extent surface tension) on the hydraulic capacity of packed columns. Also, high gas and high liquid rate data should be generated to assist current modelling techniques. Lastly, a comparative characterisation between Intalox® Ultra™ and Raschig Super-Rings would serve as a benchmark for fourth generation random packings.
- ItemEvaluation and comparison of the ability of three industrially relevant adsorbents to remove alcohol contaminants from an alkane solvent(Stellenbosch : Stellenbosch University, 2019-04) Groenewald, Jomare; Schwarz, C. E.; Burger, A. J.; Stellenbosch University. Faculty of Engineering. Dept. of Process Engineering.ENGLISH ABSTRACT: Adsorption has been used successfully in many industries and has replaced the use of processes such as distillation in some cases. The adsorption process is influenced by many factors and its efficiency is largely dependent on the regeneration of the spent adsorbent. Adsorption is system specific and requires experimental adsorption and regeneration data for the system to be separated. A better understanding of adsorbent regeneration and its effects on the adsorption process also aids in improving the overall efficiency of the process. The aims of this project were to investigate the adsorption of alcohol solutes from an alkane solvent using three industrially used adsorbents and to evaluate the effects of regeneration on each adsorbent. This was achieved through the following objectives: (1) designing and constructing batch adsorption and regeneration experimental setups; (2) performing equilibrium and kinetic adsorption tests to evaluate the effect of the type of adsorbent, type of adsorbate and temperature on the adsorption process, (3) modelling this data with existing equilibrium and kinetic models; (4) Perform cyclic adsorption-regeneration batch tests. Objective one was met by designing and constructing two batch experimental setups. The adsorption setup consisted of a heated water bath housing 10 tall form beakers in which the adsorption experiments were run. The column type regeneration system included three regeneration columns, a condenser, a liquid trap and activated carbon trap. Nitrogen gas was used as the carrier gas in the system and electric heat tracing on the column inlet tubing and regeneration columns was used to heat the system to the required regeneration temperatures. Adsorption tests investigating the alcohol adsorbing capabilities were performed using 3 types of adsorbates (1-hexanol, 1-octanol and 1-decanol), 3 types of adsorbents (activated alumina F-220, Selexsorb® CD and Selexsorb® CDx) and 3 temperatures (25, 30 and 35 °C). Overall, the activated alumina F-220 adsorbent performed better than the other two adsorbents. Generally, the 1-hexanol showed higher adsorbent loadings compared to those of 1-octanol and 1-decanol. The performance of the adsorption systems appeared to favour the slightly higher adsorption temperatures, showing the largest adsorbent loadings at 35 °C. The third objective was met by applying three isotherm and three kinetic models to the equilibrium and kinetic adsorption data respectively. The isotherm modelling confirmed that all adsorption systems exhibited favourable adsorption with physical bonds formed between adsorbate and adsorbent. The kinetic modelling gave insight into the rate-limiting step of each kinetic system, indicating that the rate-limiting step of each adsorption system could not be solely defined as intra-particle diffusion or adsorption reaction, but was rather a combination of the two. Lastly, the effects of two regeneration temperatures (185 and 205 °C) were tested on the three adsorbents in 8 adsorption-regeneration cycles. All three adsorbents remained thermally stable for all 8 cycles and the activated alumina F-220 adsorbent showed the lowest decline in alcohol removal efficiency for both regeneration temperatures. At a regeneration temperature of 185 °C the adsorbents showed slightly higher initial adsorbent loadings, but a greater decline in adsorbent loading over the 8 cycles than at the 205 °C regeneration temperature. Recommended future work suggestions included investigating higher adsorption temperatures to find the optimal temperature and investigating the effect of regeneration duration on the regeneration process.
- ItemEvaluation of entrainers for the dehydration of C2 and C3 alcohols via azeotropic distillation(Stellenbosch : Stellenbosch University, 2012-03) Pienaar, Cornelia; Burger, A. J.; Knoetze, J. H.; Stellenbosch University. Faculty of Engineering. Dept. of Process Engineering.ENGLISH ABSTRACT: Distillation is the most widely used separation technique in the chemical process industry and typically accounts for approximately one-third of the total capital cost and more than half of the total energy consumption of a typical petrochemical-chemical plant. Therefore, the design and optimization of the distillation sequence are of critical importance to the economics of the entire process. Azeotropic mixtures cannot be separated into their pure components via normal distillation. Enhanced distillation techniques such as heterogeneous azeotropic distillation should be considered for these mixtures. Isobaric vapour-liquid-liquid equilibrium (VLLE) data are highly important for the design and analysis of heterogeneous distillation columns. However, limited VLLE data are available in literature due to the difficulties involved with measuring such data. The objective of this work was to systematically evaluate and compare the performance of selected entrainers (including benzene, DIPE and cyclohexane) for the dehydration of C2 and C3 alcohols. To meet this objective, phase equilibrium data had to be measured. Isobaric VLLE at standard atmospheric conditions were measured with a dynamic Guillespie unit equipped with an ultrasonic homogenizer, which prevented liquid-liquid separation. Vapour-liquid equilibrium (VLE) and VLLE data were measured for ethanol/water/di-isopropyl ether (DIPE), n-propanol/water/DIPE and n-propanol/water/isooctane. The VLE data were found to be thermodynamically consistent according to the L-W (Wisniak 1993) and McDermott-Ellis consistency tests. No thermodynamic consistency test, specifically for VLLE data, could be found in literature, but the LLE part of the data followed a regular profile according to the Othmer-Tobias correlation. The binary DIPE/water and isooctane/water azeotropes, as well as ternary ethanol/DIPE/water and n-propanol/isooctane/water azeotropes, as measured in this work, agree well with those found in literature. Regressed parameters for the NRTL, UNIQUAC, and UNIFAC models, generally improved the model predictions compared with built-in Aspen parameters. This confirmed the importance of having actual measured VLLE data available for evaluation and improvement of estimations by thermodynamic models. NRTL predicted the ethanol/DIPE/water and n-propanol/DIPE/water VLLE most accurately. Despite the improved regressed parameters for n-propanol/isooctane/water predictions, the models are still considered unsuitable for accurate prediction of the VLLE behaviour of this system. Separation sequences were simulated in Aspen with built-in and regressed parameters, respectively, to illustrate the significant effect such inaccurate parameters have on these simulations. Phase diagram (VLLE data) evaluation of ethanol and isopropanol (IPA) with various entrainers, as found in literature, indicated that DIPE might be a good entrainer for the dehydration of these alcohols. Benzene and cyclohexane are generally used as entrainers in industry for these processes. Benzene is however carcinogenic and therefore an alternative has to be found (United States Department of Labour - Occupational Safety & Health Administration 2011). Separation sequences were simulated for ethanol dehydration with benzene and DIPE as entrainers, respectively. Taking cost and safety into account, DIPE can be considered an acceptable replacement for benzene as entrainer for ethanol dehydration. A separation sequence was also simulated for the dehydration of IPA with DIPE as entrainer and compared to a simulation with cyclohexane (Arifin, Chien 2007) as entrainer. DIPE was found to be a reasonable alternative to cyclohexane as entrainer for IPA dehydration. Two other separation sequences were simulated as practical examples where DIPE could be used as entrainer for the recovery of ethanol or n-propanol from aqueous Fischer Tropsch waste streams. DIPE is therefore found to be a feasible alternative entrainer to benzene and cyclohexane for the dehydration of ethanol and IPA via heterogeneous azeotropic distillation, based on pre-liminary cost considerations, separation ability and safety. Better entrainers than DIPE may exist, but from the data available in literature and the measurements made in this work DIPE appears to be superior to benzene, cyclohexane and isooctane.
- ItemFlotation of a UG2 ore in a novel pneumo-mechanical laboratory cell(Stellenbosch : Stellebosch University, 2018-03) Uys, Nardus; Burger, A. J.; Stellenbosch University. Faculty of Engineering. Dept. of Processing Engineering.ENGLISH ABSTRACT: Flotation is regarded as one of the most significant primary separation techniques in the minerals processing industry and most platinum-bearing minerals are separated from their related gangue minerals through froth flotation. The selectivity of the flotation process is influenced by many factors, amongst others, the degree of liberation of the valuable minerals and the size of the particles in the feed slurry. For a specific flotation system – and with all else being equal – the selectivity of the process drops notably below a certain particle size (e.g. 15 μm) due to decreasing efficiencies of particle-bubble attachment and increasing levels of gangue entrainment. Against this background, conventional mechanically-agitated flotation machines are generally used in the PGM industry, where they are operated to deliver relatively high energy input and high aeration rates. In search of more efficient flotation devices for this industry, a unique flotation machine was designed, built and evaluated in collaboration with a consultant from the PGM industry. In an attempt to achieve both higher valuable mineral recovery and higher concentrate grade, especially in the finer size ranges, the specific design combined the following principles: • high energy intensity at consecutive (in series) particle-bubble attachment points; • release of loaded bubbles to a non-agitated calming zone for froth separation. As such, the flotation unit contained a down-comer (similar to a Jameson CellTM), which was fitted with three internal impellers and subsequent exit ports for the sequential release of aerated slurry. This agitated down-comer represented the high-energy intensity zone for proper particle-bubble attachment, while the outer calming zone provided a non-agitated froth separation environment. The idea behind the sequential release of intensely air-contacted slurry was to establish a concentration gradient within the outer cell in an attempt to improve the concentrate grade at minimal loss of recovery. The first objective of this study was to utilize the extensive knowledge and know-how captured in literature to design and build a pneumatic flotation cell that enables: a) multi-stage energy input during aeration for improved overall particle-bubble attachment, in an attempt to maximise recovery; b) systematic, multi-stage release of the aerated pulp into a non-agitated calming zone for froth separation, in an attempt to maximise grade while minimising recovery losses. The second objective of this study was to perform a preliminary evaluation of the flotation device, with PGM ore as feed, by characterising the cell’s performance in terms of the most prominent hydrodynamic parameters, such as superficial air velocity, energy input and cell superficial slurry velocity. These objectives were achieved by answering the following key questions: • What are the key features of a flotation cell that could potentially produce PGM concentrates with higher grades and similar recoveries than the mechanical cells currently in use? • How is the flotation efficiency of the newly designed cell (as constructed for this project) typically affected by prominent hydrodynamic parameters such as aeration rate, energy input and recycling rate? • How does this newly designed cell typically perform with regards to recovery, grade and Cr2O3 rejection, compared to mechanical flotation machines used in the PGM industry? The cell was operated in batch-like mode, with tailings being recycled to simulate multiples of these flotation cells in series. This also enabled some comparison of the test results with flotation data generated in a conventional, mechanically-agitated, laboratory batch flotation cell at an operating PGM flotation plant. Chemical dosing (i.e. activator, collector and frother) and related conditioning times were in accordance to prescriptions from the consultant and similar to recipes used on the PGM plant. A notable difference from the operation at the PGM plant was that low-TDS reverse osmosis permeate was used to prepare the flotation slurry for test work. The operating conditions were: cell superficial slurry velocity (1.8-2.6 cm/s), superficial air velocity (0.8-1.2 cm/s) and mixing intensity (680-1277 rpm in a 36 mm down-comer). The froth depth was maintained at 10 mm and scraping occurred at 8 second intervals. Within this range of operating conditions and under continuous recycle, promising results could be obtained, similar to batch flotation results from the PGM plant. In a typical run with a high overall 4-element (4E) recovery of 80.8%, the cell produced an overall 4E grade of 76.4 g/t. When the system was operated to improve grade, an overall 4E recovery of 72.6% was achieved at a very high 4E grade of 153.6 g/t. Both these runs produced good chromite rejection.
- ItemFlux enhancement in a spiral wrap ultrafiltration element by using backpulsing(Stellenbosch : University of Stellenbosch, 2009-12) Elarbi, Abdulghader; Burger, A. J.; Sanderson, R. D.; University of Stellenbosch. Faculty of Engineering. Dept. of Process Engineering.AFRIKAANSE OPSOMMING: Die effek van teenpolsing op die aanvuiling van 'n 2.5-duim spiraal kruisvloei ultrafiltrasie element is eksperimenteel ondersoek. Teenpolseksperimente met 'n organiese (dekstraan) oplossing en 'n anorganiese (kaolien) suspensie is uitgevoer deur gebruik te maak van 'n polipropileenmembraan (100 000 molekulêre massa snypunt). Die konsentrasie van die dekstraanoplossing was tussen 250 en 750 mg/L en die konsentrasie van die kaolien oplossing was tussen 100 en 300 mg/L. Teenpolsing behels die aanwending van drukpolse van tussen 100 en 150 kPa aan die kant van die produk (permeaat). Die polstussenposes het gewissel tussen 1 en 15 s en die duur van die polse tussen 0.1 en 0.5 s. Die vloeitempo was tussen 500 en 1500 L/h, en die toegepaste druk was 100 kPa. Eksperimentele resultate het getoon dat terugpols effektief was vir die vermindering van membraanaanvuiling, en die verbetering van vloei deur die membraan. Met aanhoudende terugpolsing het die netto vloei toegeneem met toenemende terugpolsdruk. Daar was 'n effense toename met 'n toename in kruisvloeitempo en 'n sterk afname met toenemende voeroplossingkonsentrasie. Die beste terugpols parameters vir die twee verskillende aanvuilingsmateriale was soos volg: 0.2 s polsduur, 3 s polstussenpose en 150 kPa terugopolsdruk vir die dekstraanoplossing; en 0.2 s polsduur, 5 s polstussenpose en 150 kPa terugopolsdruk vir die kaoliensuspensie. Die beste resulate behaal vir vloei onder hierdie kondisies was 3-maal en 1.5-maal hoër as die vloei behaal sonder polsing, vir dekstraan en kaolien, onderskeidelik. Nadat die membraan aan aanvuiling, gevolg deur terugpolsing, blootgestel is, is dit skoongemaak deur skoon water met terugpolsing te gebruik. Die vloei van die skoon membrane wat voorheen met dekstraan en kaolien aangevuil is was 62% en 71% van die oorspronklike vloei, onderskeidelik. Die Taguchi metode met 'n L9 ortagonale reeks is gebruik om die belangrike terugpolsfaktore te bepaal wat 'n maksimum permeaatvloei tot gevolg gehad het. Die polsdruk het die grootste effek op die membraanvloei gehad. Polstussenpose en polsduur het 'n onbeduidende effek en die dwarsvloeitempo het 'n swak effek op membaanvloei gehad. Daar moet egter opgelet word dat hierdie waarnemings slegs van toepassing is binne die eksperimentele grense soos bepaal in die inleidende ondersoek van hierdie studie.
- ItemFTIR measurement of monomer fractions in dilute alcohol-acetone systems for the evaluation of the sPC-SAFT EoS(Stellenbosch : Stellenbosch University, 2013-12) Kruger, Francois Johan; Burger, A. J.; Stellenbosch University. Faculty of Engineering. Dept. of Process Engineering.ENGLISH ABSTRACT: The simplified Perturbed-Chain Statistical Associating Fluid Theory (sPC-SAFT) is characterised by the dual advantages of decreased computational intensity, while remaining accurate for a variety of systems. Vapour-liquid equilibrium data are used to generate equation of state parameters. However, incorporating monomer fraction data into the parameter regression has long been advocated as a good, or even preferred, practice. Therefore, the monomer fraction data of dilute alcohol-acetone systems were analysed in this study. A small stainless steel sample vessel was constructed with temperature control, manual pressure control and a mechanism for liquid phase analysis via infrared spectroscopy. The performance of the spectrometer was verified by comparison with the ethanol – n-hexane data of von Solms et al. (2007), after which new monomer fraction data were obtained for dilute solutions (between 0.01 and 1.5 mol%) of methanol, ethanol, 1-propanol and 2-propanol in acetone near 23 °C. For dilute alcohol-acetone systems it was found that the propanols had the highest monomer fractions, and methanol the lowest. With increasing alcohol concentration, the monomer fractions decreased exponentially to values of 0.4 and 0.1 for methanol and the other alcohols respectively. The excess availability of hydrogen bond acceptors in the mixtures explains the equivalency observed for ethanol, 1-propanol and 2-propanol. For dilute acetone-alcohol systems it was found that, especially for methanol and ethanol, there was a pronounced trend towards acetone monomer fractions of 1 at infinite dilution. For the acetone – 2-propanol system, a previously unrecorded monomer peak was observed and quantified. Acetone monomer fractions tended to decrease as alcohol chain-length increased, showing that acetone could more easily penetrate the hydrogen bond network of the solvent when the solvent-solvent bonds were weaker. Monomer fraction data were compared to predictions for the sPC-SAFT scheme and parameters combinations published in the literature. The experimental data were accurately modelled using modified association parameters such that the solute associates strongly (εAB≈103 κ≈1), while the solvent parameters were decreased (εAB≈102 κ≈10˗3) to give a weakened solvent association effect. The difficulty for the dilute solute in penetrating the solvent bonding network appeared to be similar to the hydrophobic effect. Two new association schemes were proposed for acetone, assigning a single (N) or two (2N) negative association sites to represent the oxygen valence electron pairs. These schemes showed relative success in modelling acetone as the solvent in the mixture, while not being able to predict acetone monomer fractions when acetone was the solute. For dilute acetone-alcohol systems, the data were best described using the 2B model for acetone, while the best choice of scheme for the alcohol varied from system to system. For dilute alcohol-acetone mixtures it was generally found that a 2B-N model (with modified association parameters) provided the best fit to those experimental data. Accurate modelling below 0.1 mol% was difficult to attain with average errors decreasing to the order of 10% when this area was excluded. In this highly dilute region, not one of the models could describe the rapid change in (monomer fraction) gradient sufficiently while simultaneously offering accurate predictions over the entire experimental range.
- ItemHydrodynamic evaluation of the effects of fluid physical properties and sieve tray geometry on entrainment and weeping(Stellenbosch : Stellenbosch University, 2014-12) Moses, Royston Kyle; Burger, A. J.; Knoetze, J. H.; Stellenbosch University. Faculty of Engineering. Dept. of Process Engineering.ENGLISH ABSTRACT: Distillation is one of the most widely used processes for the separation of fluids with different volatilities. Due to the popularity of this process it is often assumed that the hydrodynamic behaviour inside distillation columns is well-defined. However, this is not always the case and this study therefore endeavoured to provide additional insight into the topic through a systematic investigation into the hydrodynamics and the capacity limitations of a sieve tray distillation column. The objective of the study was to measure and evaluate the effects of the following variables on entrainment and weeping: - Fluid flow rate (gas and liquid). - Plate geometry (i.e. hole diameter and fractional hole area). - Liquid properties (i.e. surface tension, viscosity and density). - Gas properties (i.e. viscosity and density). The hydrodynamic effects were evaluated at zero mass transfer in a pilot-scale tray column, by passing pure liquids and gases in counter current configuration. The pilot column was rectangular in shape with internal dimensions of 175 mm by 635 mm. A chimney tray was used to capture the weeping liquid, while a de-entrainment tray was used in combination with a mist eliminator pad to capture the entrained liquid. The fractional hole areas for the sieve trays under investigation were 7%, 11% and 15% and the hole diameters were 3.2 mm (⅛ in.), 6.4 mm (¼ in.) and 12.7 mm (½ in.). The experimental liquids were ethylene glycol, butanol, water and silicone oil, while the gases were air and carbon dioxide (CO2). These experimental measurements produced over 10 000 data points for entrainment and over 7 000 data points of weeping. The results were repeatable and the entrainment values compared reasonably well with previous data produced by Nutter (1971) and Uys (2012). The differences between entrainment for the different liquids were more significant in the spray regime than in the froth regime, and butanol was entrained more readily than silicone oil, ethylene glycol and water. Fluids that caused a larger spray layer in the dispersion zone produced more entrainment. Entrainment increased with decreasing liquid density, decreasing liquid surface tension and decreasing liquid viscosity. The more unstable the dispersion layer, the higher the entrainment. The liquid density strongly influenced weeping, i.e. weeping increased with increasing liquid density. On the other hand, gases with higher densities – and thus with a higher mass flow rates at similar volumetric flow rates through the sieve tray – displayed less weeping and more entrainment than less dense gases, because of an increased upward drag force on the fluids. When considering tray geometry and when operating in the spray regime, the magnitude of entrainment increased with decreasing fractional hole area, while the dependency of entrainment on fractional hole area was more prominent at lower fractional hole areas. When operating in the froth regime – typically above 23 m3/(h.m) – the fractional hole area had a relatively small influence on the magnitude of entrainment, while the cross-flowing liquid rate dominated related effects. In the spray regime, i.e. typically below 23 m3/(h.m), the entrainment increased with increasing sieve tray hole diameter, while hole diameter had a relatively small influence on entrainment at higher liquid flow rates between 23 and 60 m3/(h.m). However, at even higher liquid flow rates in the froth regime, i.e. above 60 m3/(h.m), the effect of hole diameter on the entrainment became more prominent again, with increased entrainment for smaller hole diameters. The effect of hole diameter on weeping differed with changing fluid combinations and the 12.7 mm hole size caused notably less weeping than the 3.2 mm and 6.4 mm trays at higher liquid flow rates. It is believed that weeping occurred preferentially at so-called localised high pressure zones on the sieve tray. At high gas and liquid flow rates, the resultant extended dispersion layer allows minimal intimate contact between the plate and the liquid (minimising such localized high-pressure zones). In effect, the liquid ‘jumps’ over the entire flow path length in the test rig, thus resulting in low weeping rates at high gas and liquid rates. The effects of fractional hole area and hole diameter on entrainment and weeping can be correlated with combinations of well-known hydrodynamic dimensionless numbers, such as the Weber number (We), Froude number (Fr) and Reynolds number (Re). Within the limitations of this study, the flow-Froude number was shown to be the most useful dimensionless number, since it displayed a monotonic relationship with magnitude of entrainment for different combinations of fluid systems and tray configurations. Furthermore, both the construction number and fluid density ratio could be used in a sensible manner to correlate some of the effects of tray geometry on entrainment.