Liquid velocity and dispersion coefficient in an airlift reactor with inverse internal loop

Kochbeck, Birgit; Hempel, D. C.

doi:10.1002/ceat.270170607

Cited by 11 publications

(3 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Over the ranges of packing matrices, packing height, bed top, and bottom bed clearances used in this investigation, these features seem to be influence the liquid axial dispersion coefficient in the fibrous‐bed bioreactor. The dependence of liquid axial dispersion coefficient on the liquid circulation velocity in the ILALFBB proposed in this study could be accurately modelled by a modified empirical correlation previously suggested by Kochbeck and Hempel (1994) for unpacked ILAL‐bioreactor which describes the dependence of axial dispersion coefficient on the liquid circulation velocity ( U Lr /U L,I ) and the area ratio ( A r / A d ). In the current study, this model was based on 107 experimental data in each case.…”

Section: Correlationsmentioning

confidence: 78%

Mass transfer and mixing characteristics in an airlift‐driven fibrous‐bed bioreactor

2010

View full text Add to dashboard Cite

An internal loop airlift-driven fibrous bed bioreactor (ILALFBB) was designed and developed with a high degree of flexibility to handle genetically engineered and fragile shear-sensitive cells. The mixing and oxygen mass transfer characteristics have been investigated. A cotton fibre was set in the downcomer of the ILALB to represent the fibrous packed bed and the outcome results were compared with those of the polyurethane foam (PUF) packed and unpacked ILALB systems. The effects of fibre, packing height, bed top and bottom clearances, spacing between adjacent fibre surfaces, and superficial gas velocity were investigated. The liquid phase mixing output variables included the liquid circulation velocity (U Lc ), circulation time (t Lc ), mixing time (t Lm ), Bodenstein number (Bo L ), and axial dispersion coefficient (E L ), whereas the mass transfer out variable was the K L a. Bo L and E L in the riser and downcomer regions of all packed systems increased with increasing in packing height, packing top clearance, and superficial gas velocity, except the overall Bo L was independent of gas velocity at low gas velocities. The Bo L was found highest in the riser of the large cotton and small PUF packed system with large spacing and the E L in the downcomer of PUF packed systems with smaller spacing between fibre surfaces. Increased amounts of packing in the ILALB, whether in the form of cotton or PUF decreased the U Lc in the bioreactor because of the increased frictional resistance and tortuosity. The reduction in U Lc was significant for large packing with smaller spacing between fibre surfaces and increased bottom clearances of the cotton packed system. High circulation times (t Lc ) and shorter mixing times (t Lm ) were achieved using small PUF packing with large top clearance. Relatively high K L a values were obtained using large packing with large top clearances and spacing between fibre surfaces. The boost in K L a was associated with increased gas holdup and/or interfacial area, due to bubble breakage by the shearing action of the fibrous-bed. Empirical correlation proposed for E L , Bo L , and K L a gave a good fit of the experimental data.Un bioréacteurà boucle interne opérant par air-lift età lit fibreux (ILALFBB) aété conçu et développé avec un niveauélevé de flexibilité pour la manipulation des cellules génétiquement modifiées, fragiles et sensibleà l'effet de cisaillement. Les caractéristiques de mélange et de transfert de masse de l'oxygène ontétéétudiées. Les effets de la structure de la garniture de remplissage ontétéétudiés. BoL et E L dans les régions de canalisation montante et de la canalisation descendante de tous les systèmes garnis ont augmenté avec l'accroissement de la hauteur de remplissage, du dégagement supérieur de la garniture de remplissage et de la vitesse superficielle de gaz, sauf que le BoL globalétait indépendant de la vitesse de gazà faible vitesses de gaz. Le BoL le plusélevé aété observé dans la canalisation montante du système constitué de larges garnitures de coton, des pe...

show abstract

Section: Correlationsmentioning

confidence: 78%

Mass transfer and mixing characteristics in an airlift‐driven fibrous‐bed bioreactor

2010

View full text Add to dashboard Cite

show abstract

“…The differences derive also from the methods used for the estimation of D ax . Most contributions are indeed based on the Blenke method (see, e.g., Blenke, 1979;Fields and Slater, 1983;Kochbeck and Hempel, 1994;Tudose, 1996, 1997;Merchuk et al, 1998) that considers an axial dispersion coefficient for the whole reactor. This method is probably suitable for internal-loop ALRs, for which deaeration is seldom complete, but this point is more doubtful in external-loop ALRs, as shown by Verlaan et al (1989), Lu et al (1994) and Dhaouadi et al (1997).…”

Section: Introductionmentioning

confidence: 99%

Experimental and theoretical analysis of axial dispersion in the liquid phase in external-loop airlift reactors

Vial¹,

Poncin²,

Wild³

et al. 2005

Chemical Engineering Science

View full text Add to dashboard Cite

“…Further, if a strong white noise was added to the measured RTD data, the calculated parameter deviated from the real value and could not represent the real flow behavior in the reactor. Although many researchers have studied the reliability of the data processing methods mentioned above [3][4][5][6], few of them have discussed the anti-disturbance ability of these methods. Therefore, it is important to find a robust method to calculate the fluid velocity and dispersion coefficient from raw experimental RTD data.…”

Section: Introductionmentioning

confidence: 99%

Application of Residence Time Distribution for Measuring the Fluid Velocity and Dispersion Coefficient

Zhang

Wang

2006

Chem Eng & Technol

View full text Add to dashboard Cite

Most studies on residence time distribution (RTD) have focused on the tail of the RTD curve, and very little attention has been paid to the effect of white noise on the measured results. The aim of this work is to study the effect of white noise on the calculated parameters with different data processing methods. The anti-disturbance abilities of the moment method and the least squares method are compared. The results show that the anti-disturbance ability of the least squares method was better than that of the moment method. As a result of peak overlapping in the RTD curve of a loop reactor, the moment method cannot be used to calculate the fluid velocity and dispersion coefficient. Experiments show that the least squares method is still applicable in a loop reactor. IntroductionThe concept of residence time distribution (RTD) is an important issue for reactor design and scale-up. Several models, such as the dispersion model, cross-flow model, gamma distribution time delay model, and modified exponentially distributed time delay model, have been proposed in recent decades to describe the measured RTD data. Among these models, the dispersion model is the most commonly used to describe the measured RTD data in a reactor. The parameters in these models are usually estimated based experimental RTD data [1].The tracer injection and response method was widely used to study the mixing characteristics of a reactor. The moment method and curve-fitting method are the most commonly used to analyze the tracer response signals. Fahim and Wakao suggested that the time domain analysis was the best method for the evaluation of mixing parameters [2]. In the time domain analysis reported in the literature, the measured tracer signal was expressed by a Fourier series for curve fitting. This introduced numerical errors resulting in inaccuracies in the calculated value. Further, if a strong white noise was added to the measured RTD data, the calculated parameter deviated from the real value and could not represent the real flow behavior in the reactor. Although many researchers have studied the reliability of the data processing methods mentioned above [3][4][5][6], few of them have discussed the anti-disturbance ability of these methods. Therefore, it is important to find a robust method to calculate the fluid velocity and dispersion coefficient from raw experimental RTD data.In a loop reactor, the RTD is wavelike because of the fluid circulation [7]. At low fluid velocities, the peaks are independent of each other. In such cases, the moment method can be applied to a single peak and the fluid velocity and dispersion coefficient are obtained. At high fluid velocities, the adjacent peaks overlap with each other, and therefore, the moment method is inapplicable. This work aims to present a general method that can be used at different fluid velocities and discusses the anti-disturbance ability of this method. Data Processing MethodsThe tracer input-response technique is based on the axial dispersion equation. A tracer is injected int...

show abstract

Liquid velocity and dispersion coefficient in an airlift reactor with inverse internal loop

Cited by 11 publications

References 6 publications

Mass transfer and mixing characteristics in an airlift‐driven fibrous‐bed bioreactor

Mass transfer and mixing characteristics in an airlift‐driven fibrous‐bed bioreactor

Experimental and theoretical analysis of axial dispersion in the liquid phase in external-loop airlift reactors

Application of Residence Time Distribution for Measuring the Fluid Velocity and Dispersion Coefficient

Contact Info

Product

Resources

About