Mean Bubble Size Research Articles

Oil emulsion separation with fluidic oscillator generated microbubbles

Surfactants stabilise oil droplets in water, forming a dispersed oil–water emulsion. Treatment of oily effluents is a serious challenge owing to the high stability and colloidal nature of the oil droplets. In many applications, microbubbles are employed for separation purposes due to their buoyancy and increased surface area to volume ratio. This property has been exploited in the water treatment industry for separation in a process known as dissolved air flotation (DAF). Though practically efficient, the process is energy intensive operating at >5 bars and consequently consuming ∼90% of the total energy required in water purification plants. In this study microbubbles were produced by fluidic oscillation via a no-moving part diverter valve to cut down the energy consumption considerably. Microbubbles are applied for the separation of emulsified oil in a process known as microflotation. The mean bubble size generated by fluidic oscillation from the 50μm pore diffuser was ∼100μm, otherwise coarse bubbles were produced under steady flow. The effect of surfactant concentration on oil droplet size was investigated. It was found that oil droplet size varied inversely proportional to surfactant concentration. In addition, it was found that the oil removal efficiency also depends on the surfactant concentration. The maximum oil removal efficiency by Microflotation was found to be 91% under lowest surfactant concentration tested (0.3wt%) whilst at highest surfactant concentration used (10wt%); lowest recovery efficiency (19.4%) was recorded.

Critical parameters to determine mean bubble size of generated foams from a foam generator

ABSTRACTThe filtration performance of filter media fabricated by foam coating technology is highly dependent upon their pore sizes. Therefore, we investigated the effect of the mixing ratio of combined coating components (PTFE emulsion, a foam stabilizer, a foaming agent, and a thickener) and operating parameters (feed rate of coating solution, air flow rate, mixing speed, and cooling temperature) on the bubble size distribution of generated foams in a foam generator. The mean bubble size and bubble size distributions were observed through microscopic observations with an optical microscope. We found that viscosity was the most critical parameter in preparing a coating solution, and the entire series of operating parameters as mentioned above influenced the bubble size in a foamed solution. The bubble size decreased with an increasing ratio of thickening agent up to 1.8%, mixing speed, and with decreasing solution feed rate, and cooling temperature. Furthermore, we had a minimum bubble size for the condition of air flow rate at 200 mL/min. © 2013 Wiley Periodicals, Inc. J. Appl. Polym. Sci. 130: 2062–2067, 2013

Numerical Simulation of Bubble Column Flows in Churn-Turbulent Regime: Comparison of Bubble Size Models

Numerical simulations of cylindrical bubble column operating in the churn-turbulent regime have been simulated using Euler–Euler approach incorporated with the RNG k–ε model for liquid turbulence. Single-size bubble model, double-size bubble model, and the multiple size group model (MUSIG), including the homogeneous and inhomogeneous discrete methods, are employed in the simulations. Mass conserved formulations of breakup and coalescence rates were used in the computation of bubble size distributions. The Schiller–Naumann drag force was used in the single-size model, and the Ishii–Zuber drag force was used for the MUSIG simulations. For the double-size bubble model, an empirical drag formulation was adapted. The simulation results of time-averaged axial velocity and gas holdup obtained with the three models were compared with reported experimental data in the literature. The results showed that only MUSIG models with lift force can reproduce the measured radial distribution of gas holdup in the fully developed flow regime and that the inhomogeneous MUSIG model performs a little better than other models in the prediction of axial liquid velocity. The RNG k–ε model was used in all simulations, and the results confirmed that this version of k–ε model did yield relatively high turbulence dissipation rates and high bubble breakup rates and, thus, resulted in a rational bubble size distribution. The ad hoc manipulation of the breakup rates was avoided. The simulation results indicated profound mutual effects of drag force, mean bubble sizes, and turbulence characteristics. An increase in drag force yielded a decrease in the relative velocity between phases, the later could result in decreases in k and ε. A large Sauter diameter results from a low bubble breakup rate which was directly connected to the dissipation rates of turbulence. The change of Sauter diameter, in turn, influenced the drag force.

Experimental study of a two-phase surface jet

Results of an experimental study of a two-phase jet are presented, with the jet issued near and below a free surface, parallel to it. The jet under study is isothermal and in fresh water, with air injectors that allow variation of the inlet air volume fraction between 0 and 13 %. Measurements of water velocity have been performed using LDV, and the jet exit conditions measured with PIV. Air volume fraction, bubble velocity and chord length distributions were measured with sapphire optical local phase detection probes. The mean free surface elevation and RMS fluctuations were obtained using local phase detection probes as well. Visualization was performed with laser-induced fluorescence. Measurements reveal that the mean free surface elevation and turbulent fluctuations significantly increase with the injection of air. The water normal Reynolds stresses are damped by the presence of bubbles in the bulk of the liquid, but very close to the free surface the effect is reversed and the normal Reynolds stresses increase slightly for the bubbly flow. The Reynolds shear stresses $$\left\langle {u^{\prime } w^{\prime } } \right\rangle$$ decrease when bubbles are injected, indicating turbulence attenuation, and are negative at deeper locations, as turbulent eddies shed downward carry high axial momentum deeper into the flow. Flow visualization reveals that the two-phase jet is lifted with the presence of bubbles and reaches the free surface sooner. Significant bubble coalescence is observed, leading to an increase in mean bubble size as the jet develops. The coalescence near the free surface is particularly strong, due to the time it takes the bubbles to pierce the free surface, resulting in a considerable increase in the local air volume fraction. In addition to first explore a bubbly surface jet, the comprehensive dataset reported herein can be used to validate two-phase flow models and computational tools.

English

  The dispersion of gas into liquid is a complex part and it is depending on bubble size and its distribution including coalescence and breakup. The study of flow pattern, bubble size and its distribution and factors affecting the bubble size was reviewed by several authors. Various measurement techniques of bubble size are also reported in literature. In this paper we have made an attempt to show the literature survey on bubble dynamics as it plays very crucial role in mass transfer characteristics.   Key words: Computational fluid dynamics (CFD), multiphase flow gas-liquid two-phase flow, bubbly flow, bubble size distribution, Sauter-mean bubble diameter, mean bubble size, bubble breakup and coalescence, free surface flow, slug flow.

Controlling the Size Distribution of Lipid-Coated Bubbles via Fluidity Regulation

Lipid-coated bubbles exhibit oscillation responses capable of enhancing the sensitivity of ultrasound imaging by improving contrast. Further improvements in performance enhancement require control of the size distribution of bubbles to promote correspondence between their resonance frequency and the frequency of the ultrasound. Here we describe a size-controlling technique that can shift the size distribution using a currently available agitation method. This technique is based on regulating the membrane dynamic fluidity of lipid mixtures and provides a general size-controlling variable that could also be applied in other fabrication methods. Three materials (1,2-dihexadecanoyl-sn-glycero-3-phosphocholine, 1,2-dioctadecanoyl-sn-glycero-3-phospho-(1'-rac-glycerol) and polyethylene glycol 40 stearate) with distinct initial fluidities and phase behaviors were used to demonstrate the use of fluidity regulation to control bubble sizes. Bubble size distributions of different formulations were determined by electrical impedance sensing, and bubble volumes and surface areas were calculated. To confirm the relationship between regulated fluidity and mean bubble size, the membrane fluidity of each composition was determined by fluorescence anisotropy, with the results indicating linear relations in the compositions with similar main transition temperatures. Compositions with a higher molar proportion of polyethylene glycol 40 stearate showed higher fluidities and larger bubbles. B-mode ultrasound imaging was performed to investigate the echogenicity and lifetime of the fabricated bubbles, with the results indicating that co-mixing a high-transition-temperature charged lipid (i.e., 1,2-dioctadecanoyl-sn-glycero-3-phospho-(1'-rac-glycerol)) extends the tailoring range of this fluidity regulation technique, allowing refined and continuous changes in mean bubble size (from 0.93 to 2.86 μm in steps of ∼0.5 μm), and also prolongs bubble lifetime. The polydispersity of each composition was also determined to evaluate practicality in particular applications. Our study demonstrates a feasible approach to naturally controling bubble size distribution and provides a practical reference for other fabrication systems and ultrasound imaging applications.

High-resolution variations in size, number and arrangement of air bubbles in the EPICA DML (Antarctica) ice core

AbstractWe investigated the large-scale (10–1000 m) and small-scale (mm–cm) variations in size, number and arrangement of air bubbles in the EPICA Dronning Maud Land (EDML) (Antarctica) ice core, down to the end of the bubble/hydrate transition (BHT) zone. On the large scale, the bubble number density shows a general correlation with the palaeo-temperature proxy, δ18O, and the dust concentration, which means that in Holocene ice there are fewer bubbles than in glacial ice. Small-scale variations in bubble number and size were identified and compared. Above the BHT zone there exists a strong anticorrelation between bubble number density and mean bubble size. In glacial ice, layers of high number density and small bubble size are linked with layers with high impurity content, identified as cloudy bands. Therefore, we regard impurities as a controlling factor for the formation and distribution of bubbles in glacial ice. The anticorrelation inverts in the middle of the BHT zone. In the lower part of the BHT zone, bubble-free layers exist that are also associated with cloudy bands. The high contrast in bubble number density in glacial ice, induced by the impurities, indicates a much more pronounced layering in glacial firn than in modern firn.

Bubble size, gas holdup and bubble velocity profile of some alcohols and commercial frothers

Frothers are used in flotation to reduce bubble size and as a consequence increase gas holdup. For 1-alcohols, C 4 to C 7, and three commercial frothers, MIBC, DF250 and F150, it is shown that at equivalent Sauter mean bubble size gas holdup depends on frother type. The paper explains the gas holdup result by showing the effect of frother type on single bubble rise velocity profile, i.e., local velocity vs. distance (up to 350cm). For example, a 1.45mm bubble in 0.4mmol/L pentanol hardly slows relative to water only while in 0.4mmol/L hexanol it reaches minimum velocity. This corresponds to the higher gas holdup in 1-hexanol compared to 1-pentanol. To compare frothers we introduce velocity at 300cm vs. concentration and the concentration to reach minimum velocity, CMV. The CMV illustrates the strong frother effect: for example, CMV for the alcohols decreases by about one order of magnitude for every 1 C increase in chain length. The decrease in CMV with frother type correlates with the increase in gas holdup. It is shown that all frothers tested follow the same minimum velocity-size trend modeled reasonably by the relationship proposed by Karamanev for terminal velocity.

Surface and foam properties of SLES + CAPB + fatty acid mixtures: Effect of pH for C12–C16 acids

Recently we showed that triple surfactant mixtures, comprising the anionic surfactant SLES, zwitterionic surfactant CAPB and long-chain fatty acid (FAc), possess very high surface modulus which can be used to modify the dynamic properties of foams (Golemanov et al. Langmuir 2008, 24, 9956). In the current paper we perform a systematic experimental study of the effects of several factors on the surface tension, surface dilatational modulus, viscous friction inside sheared foam, foam-wall friction, and mean bubble size in sheared foams, for such triple surfactant mixtures. The factors studied are: (1) chain length of fatty acid; (2) fatty acid concentration; (3) pH; (4) presence of glycerol in the aqueous solution. Lauric (C12Ac), myristic (C14Ac) and palmitic (C16Ac) acids are studied. The concentrations of the main surfactants were fixed, 10mM for SLES and 5mM for CAPB. Varying the myristic acid (C14Ac) concentration, CC14Ac, we observed two ranges: (1) At CC14Ac<0.06mM, the acid increases the solution surface modulus up to 50mN/m, however, without affecting significantly any of the other properties studied; (2) At CC14Ac>0.11mM, C14Ac leads to significant decrease of surface tension (from 29 to 22mN/m), increase of surface modulus (from 50 to 400mN/m), decrease of mean bubble size in sheared foams (from 300 to 150μm), and strong increase of the inside-foam friction and foam-wall friction. Qualitatively similar results are obtained with the other FAc studied. Mass-balance estimates show that low surface tension and high surface modulus are observed, when the molar fraction of FAc in the adsorption layer approaches ca. 30%. The main role of FAc is to induce a surface phase transition, leading to formation of surface condensed phase in the mixed adsorption layer. For all systems studied, the increase of pH above a certain transitional value leads to a sharp increase of surface tension and decrease of surface dilatational modulus, which is accompanied with a decrease in the inside-foam and foam-wall viscous friction. The transitional pH value varies between 8 and 11, and increases with the fatty acid chain-length. The increase of pH causes ionization of the FAc molecules – the latter start to behave as usual anionic surfactant without forming surface condensed phase. The main effect of glycerol is to decrease the transitional pH (by 1 to 1.5 units for 40wt% glycerol), without affecting significantly the system properties away from the transitional pH. These results clarify the compositional domain, in which the fatty acids can be used for control of surface and foam properties in such surfactant mixtures.

Effect of bulk viscosity and surface tension kinetics on structure of foam generated at the pilot scale

The texture properties and the stability of foam are highly dependent on foam structure after processing. The Weber Number, widely used to predict mean bubble size, takes in account the shearing rate during process, the bulk viscosity, and the surface tension at equilibrium. The objective of this study was to illustrate the importance of surface tension kinetics on structure of food foam. Since surfactant adsorption is a complex process, it is essential to decide whether equilibrium or surface tensions at short times have to be considered for the size prediction of bubbles generated at pilot scale. To do so, we formulated model systems with controlled bulk viscosity and interfacial properties: Newtonian fluids at different viscosity levels and various surface tension kinetics. Characteristic parameters were obtained by adjusting the dynamic surface tension measurements to kinetic models found in the literature. Structure of processed foam was accessed by Sauter Diameter and bubble size distribution. Whipping conditions were the same for all solutions. The experimental design showed a clear effect of viscosity and surfactant concentration on adsorption kinetics. These had in turn an impact on foam structure: slow adsorption kinetic led to bigger bubbles with larger size distribution whereas fast adsorbing conditions produced smaller and more homogeneous bubbles.

Radial Bubble Distribution in a Fluidized Bed with Vertical Tubes

The radial distribution of bubble characteristics such as the bubble frequency, the bubble size, and the bubble rise velocity, is a key to better understanding the flow patterns of gas (and solids) in bubbling fluidized beds. Therefore, the radial distribution of bubble characteristics is investigated in a fluidized bed with vertical tubes by means of optical probes. Measurements are taken for Geldart B particles at two bed heights, three gas velocities, and three vertical tube bank configurations. While the bubble frequency shows a strong dependence on the radial position in the bed and the orientation of the vertical tubes, the bubble size and the bubble rise velocity, in particular in the lower portions of the bed, are more constant. Typical bubble pathways are shown to move from the column wall to the center axis of the bed with increased bed height. Therefore, in order to have a representative mean bubble size and mean bubble rise velocity, which accounts for the whole cross-section of the bed, an area-number weighted mean is introduced.

Numerical Prediction of Effect of Oxide Characteristics on Heterogeneous Nucleation of Bubbles in Aluminium Foaming

Effects of oxide wettability, oxide surface cavities and cavity surface area-to-volume ratio on heterogeneous nucleation of bubbles in hydrogen supersaturated liquid aluminium foam precursor were studied. Increase in cavity surface area-to-volume ratio, oxide contact angle and cavity sharpness led to increase in cavity nucleation rate of bubbles. Oxides with lower contact angles required sharper cavities for nucleation to occur. A characteristic cavity angle exists for given oxide contact angle to make bubble nucleation feasible. A suitable combination of oxide contact angle and cavity angle were selected to estimate heterogeneous nucleation rates for different cavity surface areas and combined with bubble growth analysis to arrive at final bubble size distribution. The mean bubble size predicted numerically compared favourably with experimental observation on cell size of the foam.

Prediction of air bubble dispersion in a viscous fluid in a twin-screw continuous mixer using FEM simulations of dispersive mixing

Bubble break-up was studied during mixing of a viscous Newtonian fluid in a co-rotating twin screw continuous mixer using experimental observations and finite element method (FEM) simulation of dispersive mixing. Bubble break-up was greatest with two paddle element configurations that had no stagger and a reverse stagger respectively while a third configuration with a forward element stagger showed the least break-up. The decrease in bubble break-up was found to be due to the decrease in the elongation flow density and an increase in local forward flow in the forward angle stagger configuration. Maximum stable bubble diameters and effective shear rate for break-up calculated at different locations in the mixer using the fundamental capillary number theory for drop and bubble break-up correlated well with the measured local mean bubble sizes and the mean local shear rates calculated using FEM simulations of the flow.

Simulation of transient fluid flow in mold region during steel continuous casting

A system of models has been developed to study transient flow during continuous casting and applied to simulate an event of multiple stopper-rod movements. It includes four sub-models to incorporate different aspects in this transient event. A three-dimensional (3-D) porous-flow model of the nozzle wall calculates the rate argon gas flow into the liquid steel, and the initial mean bubble size is estimated. Transient CFD models simulate multiphase flow of steel and gas bubbles in the Submerged Entry Nozzle (SEN) and mold and have been validated with experimental data from both nail dipping and Sub-meniscus Velocity Control (SVC) measurements. To obtain the transient inlet boundary conditions for the simulation, two semi-empirical models, a stopper-rod-position based model and a metal-level-based model, predict the liquid steel flow rate through the SEN based on recorded plant data. Finally the model system was applied to study the effects of stopper rod movements on SEN/mold flow patterns. Meniscus level fluctuations were calculated using a simple pressure method and compared well with plant measurements. Insights were gained from the simulation results to explain the cause of meniscus level fluctuations and the formation of sliver defects during stopper rod movements.

Improved Modeling of Bubble Column Reactors by Considering the Bubble Size Distribution

Bioconversion of glucose to gluconic acid with glucose oxidase as enzyme was considered to occur in a bubble column reactor at 40 °C, atmospheric pressure, and pH 5.5 while oxygen gas velocity was varied in the homogeneous and transition regime in the range of 0.0014–0.0112 m s–1. It was found that the superficial gas velocity has a positive effect on gas holdup, mean bubble size, and volumetric mass transfer coefficient, while the glucose concentration has a negative influence on these hydrodynamic parameters. The modeling of bubble column by considering the complete bubble size distribution in the governing equations was carried out in this work. The model was also solved by considering the single and double sizes of bubbles instead of complete bubble size distribution. It was shown that the results of the model, considering the complete bubble size distribution, is in a good agreement with the experimental data, while the considered single and double bubble sizes cannot predict the experimental data of oxygen concentration in the reactor properly. It was also found that the superficial gas velocity has a positive influence on oxygen concentration while glucose concentration has negative influence at low superficial gas velocity and positive influence at high gas velocity.

Monte Carlo simulation of the bubble size distribution in a fluidized bed with intrusive probes

Intrusive probes such as optical probes are commonly used to measure the bubble size distribution in a fluidized bed. However, usually only a chord length distribution is measured which is typically smaller than the actual centerline bubble size distribution of the pierced bubbles. Moreover, since small bubbles are less likely hit by the probe than large bubbles, the effective bubble size distribution in the entire bed is generally hidden to an intrusive probe. In order to elucidate the bubble size distribution in a fluidized bed measured by an intrusive probe, a Monte Carlo (MC) model is established. MC simulations are conducted with varying sample distributions (gamma and Rayleigh), varying probe positions, and varying spatial distributions of bubbles in the cross-section. Provided the bubble shape is ellipsoidal, it is shown that for all of these variations, the mean chord length can be taken as a representative measure of the mean bubble size in the bed. Furthermore, the applicability of statistical backward transforms (analytical, non-parametrical, and maximum entropy approach) to convert the chord length distribution to the overall bubble size distribution in the bed is assessed. None of these backward transforms outperforms the simple and straightforward approach to just take the mean chord length as the representative mean bubble size in the bed.

Effects of Perfluorocarbon Gases on the Size and Stability Characteristics of Phospholipid-Coated Microbubbles: Osmotic Effect versus Interfacial Film Stabilization

Micrometer-sized bubbles coated with phospholipids are used as contrast agents for ultrasound imaging and have potential for oxygen, drug, and gene delivery and as therapeutic devices. An internal perfluorocarbon (FC) gas is generally used to stabilize them osmotically. We report here on the effects of three relatively heavy FCs, perfluorohexane (F-hexane), perfluorodiglyme (F-diglyme ), and perfluorotriglyme (F-triglyme), on the size and stability characteristics of microbubbles coated with a soft shell of dimyristoylphosphatidylcholine (DMPC) and on the surface tension and compressibility of DMPC monolayers. Monomodal populations of small bubbles (~1.3 ± 0.2 μm in radius, polydispersivity index ~8%) were prepared by sonication, followed by centrifugal fractionation. The mean microbubble size, size distribution, and stability were determined by acoustical attenuation measurements, static light scattering, and optical microscopy. The half-lives of F-hexane- and F-diglyme-stabilized bubbles (149 ± 8 and 134 ± 3 min, respectively) were about 2 times longer than with the heavier F-triglyme (76 ± 7 min) and 4-5 times longer than with air (34 ± 3 min). Remarkably, the bubbles are smaller than the minimal size values calculated assuming that the bubbles are stabilized osmotically by the insoluble FC gases. Particularly striking is that bubbles 2 orders of magnitude smaller than the calculated collapse radius can be prepared with F-triglyme, while its very low vapor pressure prohibits any osmotic effect. The interface between an aqueous DMPC dispersion and air, or air (or N(2)) saturated with the FCs, was investigated by tensiometry and by Langmuir monolayer compressions. Remarkably, after 3 h, the tensions at the interface between an aqueous DMPC dispersion (0.5 mmol L(-1)) and air were lowered from ~50 ± 1 to ~37 ± 1 mN m(-1) when F-hexane and F-diglyme were present and to ~40 ± 1 mN m(-1) for F-triglyme. Also noteworthy, the adsorption kinetics of DMPC at the interface, as obtained by dynamic tensiometry, were accelerated up to 3-fold when the FC gases were present. The compression isotherms show that all these FC gases significantly increase the surface pressure (from ~0 to ~10 mN m(-1)) at large molecular areas (70 Å(2)), implying their incorporation into the DMPC monolayer. All three FC gases increase the monolayer's collapse pressures significantly (~61 ± 2 mN m(-1)) as compared to air (~54 ± 2 mN m(-1)), providing for interfacial tensions as low as ~11 mN m(-1) (vs ~18 mN m(-1) in their absence). The FC gases increase the compressibility of the DMPC monolayer by 20-50%. These results establish that, besides their osmotic effect, FC gases contribute to bubble stabilization by decreasing the DMPC interfacial tension, hence reducing the Laplace pressure. This contribution, although significant, still does not suffice to explain the large discrepancy observed between calculated and experimental bubble half-lives. The case of F-triglyme, which has no osmotic effect, indicates that its effects on the DMPC shell (increased collapse pressure, decreased interfacial tension, and increased compressibility) contribute to bubble stabilization. F-hexane and F-diglyme provided both the smallest mean bubble sizes and the longest bubble half-lives.

Polidocanol Concentration and Time Affect the Properties of Foam Used for Sclerotherapy

How foam is created for sclerotherapy varies and is not standardized. Moreover, the effect of several factors on the quality of the foam is not well studied. To investigate the effects of different parameters on foam stability and bubble size. As a measure of foam stability, foam half time (FHT) and bubble size distribution were determined for various parameters (polidocanol (POL) concentration, freshness of the POL, syringe size, liquid-to-air ratio, number of pump cycles, and needle size) in the foam creation process. FHT was 115 to 157 seconds for POL 1% and 143 to 192 seconds for POL 3%. The other parameters had a limited effect on FHT. POL foam 1% (t=0 seconds) had a mean bubble size of 71 ± 9 μm that increased when the foam was maintained horizontally in the syringe: 102 ± 12 μm) at 30 seconds and 121 ± 20 μm at 60 seconds. The other parameters had no significant influence on bubble size distribution. Higher concentration of POL and rapid injection optimize foam stability and bubble size distribution, but other important foam characteristics are largely independent of differences in the generation and injection of foam. The authors have indicated no significant interest with commercial supporters.

CO2 Mass Transfer Induced through an Airlift Loop by a Microbubble Cloud Generated by Fluidic Oscillation

Carbon dioxide is one of the most common gases produced from biological processes. Removal of carbon dioxide from these processes can influence the direction of biological reactions as well as the pH of the medium, which affects bacterial metabolism. Kinetics of carbon dioxide transfer mechanisms are investigated by sparging with conventional fine bubbles and microbubbles. The estimate of the concentrations of CO2(aq), H2CO3, HCO3 –, and CO32– from pH measurement in an airlift loop sparged mixer is derived. The canonical estimate of overall mass transfer coefficient of CO2 has been estimated as 0.092 min–1 for a microbubble size of 550 m compared with 0.0712 min–1 for a fine bubble (mean bubble size of 1.3 mm) sparging. It is observed that the efficiency of CO2 removal has increased up to 29% by microbubble sparging compared with fine bubble sparging. Laminar bubbly flow modeling of the airlift loop configuration correctly predicts the trend of the change in overall mass transfer in both gas stripping with nitrogen and gas scrubbing for CO2 exchange, while demonstrating the expected separated flow structure. The models indicate that the macroscale flow structure is transient and pseudoperiodic. This latter feature should be tested by flow visualization, as preferential frequencies in the flow can be exploited for enhanced mixing.

CFD simulations to portray the bubble distribution and the hydrodynamics in an annulus sparged air‐lift bioreactor

AbstractAir‐lift bioreactor (ABR) are applied in the pharmaceutical industry for those processes of low‐oxygen demand. The characteristics of the air bubbles in an ABR are important since they influence the overall performance of the bioreactor. In this article, two‐dimensional CFD simulations are carried out to portray the flow characteristics in an annulus‐sparged ABR with a low‐aeration rate, which is a scaled geometry of a 200 m3 annulus‐sparged ABR for the L‐ascorbic acid manufacturing. VOF model is applied firstly to capture the flow behaviour of the bubbles in the bioreactor, and the distributions of the bubbles and local mean bubble sizes are approximately obtained. With the local mean bubble size distribution, Eulerian model is adopted to investigate the performance of the bioreactor. The liquid velocity and the gas holdup are further discussed. The results show the existence of the local back mixing as well as the mal‐distributions of the velocity and the gas holdup along the radial direction in the annulus‐sparged ABR.