CFD modeling of flow,macro-mixing and axial dispersion in a bubble column

The flow pattern in a bubble column depends upon the column diameter, height, sparger design, superficial gas velocity and the nature of gas–liquid system. In this paper, the effect of some of these parameters have been simulated using Computational Fluid Dynamics (CFD). The relationship of these parameters with the interphase force terms has been discussed. A complete energy balance has been established. Using this methodology, the flow patterns reported by Hills (1974), Menzel et al. (1990), Yao et al. (1991) and Yu and Kim (1991) have been simulated. Excellent agreement has been shown between the CFD predictions and the experimental observations. The above model has been extended to homogenization of an inert tracer. In order to confirm this model, mixing experiments were carried out in a 200 mm i.d. bubble column. A radioactive tracer technique was used for the measurement of mixing time. Tc-99m (^99m Tc), in the form of sodium pertechnate salt, was used as the liquid phase tracer. Good agreement has been shown between the predicted and the experimental values of mixing time. The model was further extended for the estimation of axial dispersion coefficient (D_L). Excellent agreement between the simulated and the experimental values of the axial dispersion coefficient confirms the predictive capability of the CFD simulations for the mixing process.     

Influence of high density particles in reduction of axial dispersion in liquid fluidized bed with residence time distribution curve studies

iquid phase RTD curves were investigated in classical fixed and fluidized bed regimes with high density particles. The effect of liquid velocity was studied on bed hydrodynamics. Using an impulse tracer injection technique in a column of 5 cm inner diameter and 1.2 m height, liquid RTD, mean residence time (MRT), axial dispersion coefficient (ADC) and vessel dispersion number (N _D ) were determined. ADC increases with liquid superficial velocity. It varied from 4.63 to 20.7 cm²/s for the particle Reynolds number of 43 to 279, respectively. The experimental results show that the hight density particles cause less ADC than the low density particles at an identical Reynolds number.     

Liquid phase axial mixing in bubble columns operated at high pressures

Liquid phase axial mixing was measured in a 100 mm i.d. bubble column operated in the pressure range of 0.1-0.5 MPa. Water, ethanol and 1-butanol were used as the liquid phase and nitrogen as the gas phase. The temperature and superficial gas velocity were varied in the range of 298-323 K and 0.01-0.21 m/s, respectively. The axial dispersion coefficient increased with an increase in the gas density due to pressure. The temperature had surprisingly a small effect. A CFD model was developed for the prediction of flow pattern in terms of mean velocity and eddy diffusivity profiles. The model was further extended for the prediction of residence time distribution and hence the axial dispersion coefficient (D_L). The predictions of axial dispersion coefficient agree favorably with all the experimental data collected in this work as well as published in the literature. The model was extended for different gas-liquid systems. The predicted values of axial dispersion coefficient were found to agree very well with all the experimental data.     

A study on the flow characteristics in a pulsed doughnut-disc type plate extraction column

The axial dispersion coefficients in the continuous phase and holdup of dispersed phase have been studied in a 4.2 cm inside diameter and 200 cm height pulsed doughnut-disc type plates extraction column. The axial concentration gradient in a continuous extraction column was expressed mathematically in terms of Peclet number by axial dispersion model. Peclet numbers have been calculated from response curves using KC1 solution as an impulse input fracer. Experimental data have been taken for both continuous and dispersed phase with plate spacing, pulsing amplitudes, frequencies, and superficial velocities as system variables. Modified axial dispersion coefficients have been correlated by regression analysis of experimental data, and following equations were obtained. 1. Axial dispersion coefficient (single phase) E_c = 3.5H^-13 A^1.5.1 f + 30.95 U_c 2. Axial dispsion coefficient (two phase) E_c = 2.36 H^{-0 8} A^1.34 f + 20.89 U_c 3. Fractional holdup of the dispersed phase Φ_d = 4 2xl0^-5H^-0.44 Af^1.28U_d ^0.93     

多级鼓泡塔中液相返混系数与交换速度的研究

采用示踪方法对高2 000 mm,内径282 mm多级筛板鼓泡塔内液相返混系数进行测量研究,并通过扩散-返混模型以及RTD曲线给出鼓泡塔内筛板上下二侧液体交换速度,同时考查了表观气速、开孔率等因素对轴向扩散系数与液体交换速度的影响.根据实验得出鼓泡塔内轴向返混系数以及液体交换速度与表观气速、开孔率有很大关系,均随表观气...     

Droplet velocity,size, and local holdup measurements in an extraction column by tri‐sensor optical probe

The tri‐sensor optical probe was applied to study the hydrodynamic characteristic in a pulsed sieve plate extraction column. Two immiscible liquids consisting of the dispersed phase (kerosene) and continuous phase (water) were introduced in countercurrent operation. Local parameters such as droplet velocity, drop size, and holdup of the dispersed phase were obtained. It was found that the tri‐sensor optical probe could be used as an efficient and convenient technique for measuring local hydrodynamic parameters inside the pulsed sieve plate extraction column. Furthermore, the results indicated that pulsation intensity imposed more influence on these hydrodynamic parameters than two‐phase superficial flow rates in the investigated ranges. Experimental results were found to be in good agreement with the empirical correlations reported in literature. © 2015 American Institute of Chemical Engineers AIChE J, 61: 3958–3963, 2015     

Comparison of normal phase operation and phase reversal studies in a pulsed sieve plate extraction column

The hydrodynamic characteristics of a pulsed sieve plate extraction column (PSPEC) was studied experimentally using two different liquid phase systems, namely water/kerosene and 30%TBP (tributyl phosphate) in NPH (normal paraffin hydrocarbon)/0.3 M HNO₃. The aqueous phase as the dispersed phase and the organic phase as the continuous phase (phase reversal) and vice versa (normal phase operation) studies in a pulsed sieve plate extraction column 0.076 m in diameter and 1 m height are presented in this paper. The hydrodynamic properties like drop size and holdup are characterized as a function of various operating parameters namely pulse velocity, dispersed phase and continuous phase velocity and duty cycle of pulsing. Flooding in the column was also investigated for the changes involving flow ratio of continuous phase to that of the dispersed phase for both insufficient and excessive pulsing. It has been observed that phase reversal mode of operation is not efficient as compared to normal phase operation for the PSPEC.     

Determining axial dispersion coefficients of pilot-scale annular pulsed disc and doughnut columns

In this study, a computational fluid dynamics(CFD) method was adopted to calculate axial dispersion coefficients of annular pulsed disc and doughnut columns(APDDCs). Passive tracer was uniformly injected by pulse input at the continuous phase inlet, and its concentration governing equation was solved in liquid–liquidtwo-phase flow fields. The residence time distributions(RTDs) were obtained using the surface monitoring technique. The adopted RTD–CFD method was verified by comparing the axial dispersion coefficient between simulation and experimental results in the literature. However, in pilot-scale APDDCs, the axial dispersion coefficients predicted by the CFD–RTD method were approximately three times larger than experimental results determined by the steady-state concentration profile method. This experimental method was demonstrated to be insensitive to the variation of the axial dispersion coefficient. The CFD–RTD method was more recommended to determine the axial dispersion coefficient. It was found that the axial dispersion coefficient increased with an increase in pulsation intensity, column diameter, and plate spacing, but was little affected by the throughput.     

A study on hydrodynamic characteristics in a Φ38 pulsed extraction column by four‐sensor optical fiber probe

Accurate prediction of dispersed phase droplet behavior is crucial to the design and scaling‐up of an extraction column. In this article, the dispersed droplet velocity algorithm and the diameter algorithm in a liquid–liquid two‐phase flow have been developed based on the bubble velocity model in gas–liquid two‐phase flow of Lucas [Measurement Science & Technology. 749, 758(2005)] and Shen [International Journal of Multiphase Flow. 593, 617(2005)]. Hydrodynamic characteristics, including droplet diameter, holdup and droplet velocity, were measured using a self‐made four‐sensor optical fiber probe in a 38 mm‐diameter pulsed sieve‐plate extraction column. Water and kerosene were used as continuous and dispersed phases, respectively. The influences of the pulsed intensity, the continuous and dispersed phase superficial velocities on the hydrodynamic characteristics were investigated. The experimental results show that it is reliable to use a four‐sensor optical probe to measure the hydrodynamic characteristics of a pulsed extraction column. © 2016 American Institute of Chemical Engineers AIChE J, 63: 801–811, 2017     

Mixing in a co‐current upflow bubble column reactors with and without internals

Liquid phase mixing is a phenomenon that results mainly due to convective and turbulent flow fields, which are generated by hydrodynamic interactions between the gas and liquid phases within a continuous co‐current upflow bubble column reactor. The extent of liquid phase mixing is usually quantified through the mixing time, or the axial dispersion coefficient. In the present work, the computational fluid dynamics (CFD) simulations for mixing and RTD in a continuous bubble column (with and without internals) are performed by using OpenFOAM 2.3.1. The superficial gas velocities were 0.014, 0.088, and 0.221 m/s and the superficial liquid velocities were 0.005 and 0.014 m/s. The simulations have been performed for three different configurations of the bubble column, that is, (a) an open bubble column, (b) a column with one vertical central rod of 36 mm diameter, (c) a column with the same central rod and four vertical additional rods of 12 mm diameter. The effects of superficial gas and liquid velocities and column internals were investigated on liquid phase mixing and the axial dispersion coefficient. Comparisons have been made between the experimental measurements and the CFD simulations.

     

CFD Modeling of a Bubble Column Reactor Carrying out a Consecutive A → B → C Reaction

In this paper, we develop a CFD model for describing a bubble column reactor for carrying out a consecutive first‐order reaction sequence A → B → C. Three reactor configurations, all operating in the homogeneous bubbly regime, were investigated: (I) column diameter D_T = 0.1 m, column height H_T = 1.1 m, (II) D_T = 0.1 m, H_T = 2 m, and (III) D_T = 1 m, H_T = 5 m. Eulerian simulations were carried out for superficial gas velocities U_G in the range of 0.005–0.04 m/s, assuming cylindrical axisymmetry. Additionally, for configurations I and III fully three‐dimensional transient simulations were carried out for checking the assumption of cylindrical axisymmetry. For the 0.1 m diameter column (configuration I), 2‐D axisymmetric and 3‐D transient simulations yield nearly the same results for gas holdup ?_G, centerline liquid velocity V_L(0), conversion of A, χ_A, and selectivity to B, S_B. In sharp contrast, for the 1 m diameter column (configuration III), there are significant differences in the CFD predictions of ?_G, V_L(0), χ_A, and S_B using 2‐D and 3‐D simulations; the 2‐D strategies tend to exaggerate V_L(0), and underpredict ?_G, χ_A, and S_B. The transient 3‐D simulation results appear to be more realistic. The CFD simulation results for χ_A and S_B are also compared with a simple analytic model, often employed in practice, in which the gas phase is assumed to be in plug flow and the liquid phase is well mixed. For the smaller diameter columns (configurations I and II) the CFD simulation results for χ_A are in excellent agreement with the analytic model, but for the larger diameter column the analytic model is somewhat optimistic. There are two reasons for this deviation. Firstly, the gas phase is not in perfect plug flow and secondly, the liquid phase is not perfectly mixed. The computational results obtained in this paper demonstrate the power of CFD for predicting the performance of bubble column reactors. Of particular use is the ability of CFD to describe scale effects.     

CFD Modeling of Bubble Column Reactor Including the Influence of Gas Contraction

In this paper a CFD model for a bubble column reactor undergoing a first order reaction A → B is developed. The reactor operates in the homogeneous bubbly regime and has a diameter D_T = 1 m and height H_T = 5 m. The incoming gas stream contains inerts, varying in proportion from 10 % to 90 %. Three‐dimensional transient Eulerian simulations were carried out for an inlet superficial gas velocity U_G = 0.04 m/s. Due to the consumption of A, the gas phase suffers contraction along the height of the reactor and as a consequence there is a significant change in the gas velocity along the column height; this variation in gas velocity is stronger when the incoming gas contains a smaller proportion of inerts. The CFD simulations show that there is a considerable influence of gas contraction on both the bubble column hydrodynamics and on the reactor conversion. None of the conventionally used reactor models is capable of describing the reactor performance in the case of high gas phase contraction.     

Measurement and CFD simulation of single-phase flow in solvent extraction pulsed column

A CFD (computational fluid dynamics) model of a solvent extraction pulsed column has been developed and run with a single water phase. The results are compared with experimental measurements taken on a pilot scale column using PIV (particle image velocimetry).The pulsed column investigated had disk-doughnut internals and was operated under pulsing intensities ranging from 10 to 32.5 mm/s. PIV measurements of velocity were used to validate the CFD model and to characterise the pulsing flow of a single phase through the column. The CFD modelling was performed for the same geometry and operating conditions using a 2D computational grid and a low Reynolds Number k-ε turbulence model. An improved velocity prediction was achieved by adding a gap between the doughnut internal and the pulsed column wall. The combined measurements and predictions give insight into the effect of the geometry internals on the flow hydrodynamics in the pulsed column.     

LIQUID PHASE AXIAL BACKMIXING IN AN AIRLIFT LOOP BIOREACTOR WITH NON-NEWTONIAN FLUIDS

Liquid phase axial backmixing in the riser and downcomer sections of an airlift loop reactor with non-Newtonian fluids was investigated and determined by dynamic response technique with pulsed tracer input, dual probe detection and computer on-line analysis system under different superficial gas velocity conditions. This method was used to obtain the dispersion coefficient D_z for the individual sections of the reactor.

Kolmogoroff's theory of isotropic turbulence was applied to analyse the results of dispersion coefficient. The results show that the axial dispersion coefficient in the riser or downcomer section increases with increasing of superficial gas velocity and apparent viscosity of the fluid. The degree of mixing in the downcomer is higher than that in the riser under the experimental conditions.     

RTD studies in trickle bed reactors packed with porous particles

The residence time distribution (RTD) for liquid phase in a trickle bed reactor (TBR) has been experimentally studied for air-water system. Experiments were performed in a 15.2 cm diameter column using commerical alumina extrudates with D/d_p ratio equal to 75 to eliminate the radial flow differences. The range of liquid and gas flow rates covered was 3.76 < Re_L < 9.3 and 0 < Re_G < 2.92. The axial dispersion model was used to compute axial dispersion coefficient. The effect of liquid and gas flow rates on total liquid holdup and axial dispersion was investigated. The total liquid holdup has been correlated to liquid and gas flow rates.     

Axial Dispersion and Phase Holdup Characteristics in Reciprocating Plate Extraction Columns

Abstract

Axial dispersion and phase holdup characteristics have been determined in a 0.102-m i.d. × 3.5 m high QVF glass column. The axial dispersion coefficient decreases with increasing reciprocating frequency (f) and amplitude (A) in the inhomogeneous dispersed phase flow regime, whereas it increases in the emulsion flow regime. The axial dispersion coefficient with a perforated plate increases with continuous and dispersed phase velocities. However, the effect of phase velocities on axial dispersion is less pronounced with the fan plate. The axial dispersion coefficient can be correlated with A ² f, fluid velocities, and the free fractional opening area of the plates. The dispersed phase holdup increases with an increase in agitation intensity. Af, and decreases with the free opening area of the plate.     

Comparison of the Axial Dispersion Performance of Pulsed Solvent Extraction Columns with Tenova Pulsed Column–Kinetics Internals and Standard Disc and Doughnut Internals

Axial dispersion performance of a 2-m high 76-mm diameter pilot-scale pulsed solvent extraction column has been studied using two liquid–liquid systems, Alamine 336/isodecanol/Shellsol 2046 (continuous)–tap water (dispersed) and LIX 84/Shellsol 2046 (continuous)–tap water (dispersed). The pulsed column was operated with standard disc and doughnut internals and Tenova pulsed column–kinetics internals using pulsation intensities from 0.005 m/s to 0.025 m/s with polyvinylidene fluoride internal plates of 22.4% open area. The effect of pulsation intensity, dispersed phase velocity, and continuous phase velocity on axial dispersion coefficient have been investigated and compared with the two different column internals, and the experimental data has been correlated with empirical relationships.     

新型脉冲萃取塔存留分数与特性速度

为测试锥形穿流塔板的性能,在内径为75mm的脉冲萃取塔中,以煤油-水和10%磷酸三丁酯/煤油-水为实验体系,在无传质条件下,研究脉冲强度与两相表观流速对分散相存留分数和特性速度的影响。结果表明,在实验范围内,存留分数与分散相流速近似成正比,与连续相流速无关。而随着脉冲强度的增大,存留分数先减小,当脉冲强度达到临界值(Af)_t后,存留分数迅速增大。将此临界值与脉冲筛板塔临界值进行对比,两种体系分别减小约9.7%和41.4%,此外,特性速度随着脉冲强度的增大而减小,且界面张力较低的体系减小幅度更大。在实验结果分析的基础上,利用量纲分析方法得到了存留分数与特性速度的工程经验关联式,预测值与实验值符合较好,相对误差均小于20%,可以应用于脉冲萃取塔的设计计算。