Bubble flow mechanisms in trickle beds—an experimental study using image processing

The mechanisms of bubble motion in concurrent gas-liquid down flow through trickle beds are investigated. The laboratory reactor is a structured quasi-two-dimensional porous medium with an average pore diameter close to the values encountered in trickle beds. The accuracy of the reactor design is demonstrated by hydrodynamic investigations on the reactor scale where it is shown that the flow regimes encountered and the experimental pressure drop are comparable to those observed in trickle beds. The investigations on the pore scale are focused on the dispersed bubble flow regime where the liquid flow is continuous and the gas is divided into elongated bubbles. The bubble motion is recorded with the aid of a high-speed video camera and the images are processed and analysed in a quantitative manner. The investigations clearly show that in dispersed bubble flow, the bubbles are frequently pulsing on the pore scale. The mechanism of this flow pattern is discussed.     

Effect of partial wetting on liquid/solid mass transfer in trickle bed reactors

The wetting efficiency of liquid trickle flow over a fixed bed reactor has been measured for a wide range of parameters including operating conditions, bed structure and physico-chemistry of liquid/solid phases. This data bank has been used to develop a new correlation for averaged wetting efficiency based on five different non-dimensional numbers. Finally liquid/solid mass transfer has been determined in partial wetting conditions to analyse what are the respective effects of wetting and liquid/gas flow turbulence. These effects appear to be separated: wetting being acting on liquid/solid interfacial area while the liquid/solid mass transfer coefficient is mainly connected to flow turbulence through the interstitial liquid velocity. A correlation has been proposed for liquid/solid mass transfer coefficient at very low liquid flow rate.     

Prediction of pulsation frequency of pulsing flow in trickle-bed reactors using artificial neural network

Based on an extensive experimental database (946 measurements) set up from the literature published over past 30 years, a new correlation relying on artificial neural network (ANN) was proposed to predict the basic pulsation frequency of pulsing flow in the trickle-bed reactors. Seven dimensionless groups employed in the proposed correlation were liquid and gas Reynolds (Re_L,Re_G), liquid Weber (We_L), gas Froude (Fr_G), gas Stokes (St_G) and liquid Eötvös numbers and a bed correction factor (S_b). The performance comparisons of literature and present correlations showed that ANN correlation is significantly an improvement in predicting pulsation frequency with an AARE of 10% and a standard deviation less than 18%. The effects of the variables including the properties of fluid and bed, and flow rate of liquid and gas on pulsing frequency were investigated by ANN parametric simulations and the trends were compared with exiting experimental results that confirmed the coherence of the proposed method with the previous experiments.     

Induced pulsing in trickle beds—characteristics and attenuation of pulses

Gas/liquid down-flow in packed beds is studied, under periodic liquid feeding (at sufficiently high frequencies to be classified as “fast” mode of pulsing), in a range of mean liquid and gas flow rates within the steady “trickling flow regime”. The aim is to identify periodic feeding conditions resulting in improved fluid-mechanical characteristics (e.g. uniform fluids distribution) and possibly enhanced transport rates in this flow regime, which is common in industrial processes. From instantaneous, cross-sectionally averaged holdup measurements, at various locations along the packed bed, quantitative information is obtained on the axial propagation and attenuation of induced pulses. A phenomenological treatment of the pulse decay process facilitates data interpretation and leads to the determination of a characteristic attenuation factor for the various conditions tested. Key parameters of the process studied include, in addition to dynamic holdup, pressure drop, pulse celerity and intensity, as a function of fluid feed rates (G,L) and liquid cyclic frequency. Under the conditions of these tests, and for fixed mean rates G,L, the time averaged holdup and the pulse celerity are practically constant along the bed; furthermore, these quantities as well as the pressure drop do not seem to be affected by the imposed cyclic liquid feeding frequency. An expression to tentatively correlate pulse celerity data is recommended.The computed attenuation factors indicate that there is a rather narrow band of mean gas and liquid rates (along the so-called “pseudo-transition” boundary to pulsing flow) where pulse decay is at a minimum. Based on these results as well as on pulse intensity vs. bed length data, recommendations are made on preferred conditions for induced pulsing (from the fluid-mechanical standpoint) which would maximize expected benefits.     

An investigation of liquid maldistribution in trickle beds

The influence of liquid maldistribution at the top of the packing on flow characteristics in packed beds of gas and liquid cocurrent downflow (trickle beds) is experimentally investigated. Particular attention is paid to the effect of gas and liquid flow rates on flow development. Tests are made in the trickling and pulsing flow regimes. A uniform, a half-blocked and a quarter-blocked liquid distributor is tested. Packings of various sizes and shapes are employed. Data are presented on pressure drop and liquid holdup as well as trickling to pulsing flow transition. Diagnosis of radial and axial liquid distribution is made by means of conductance probes. The effects of liquid foaming, bed pre-wetting, top-bed material, and blockage midway the bed on liquid distribution are also examined. Overall, liquid waves in the pulsing flow regime have a beneficial effect, promoting uniform liquid distribution in the bed cross section.     

Phenomenological approach to interpret the effect of liquid flow modulation in trickle bed reactors at the particle scale

This work analyzes the influence of liquid flow modulation on the behavior of a reaction occurring in a spherical porous particle within a trickle bed reactor. A single first-order reaction between a gaseous reactant and a non-volatile liquid reactant is considered. Non-steady-state mass balances for gas and liquid reactants are formulated and solved under isothermal conditions in order to focus the analysis on the mass transport effects. Dynamic reactant profiles inside the catalytic particle are obtained for different cycling and system conditions. The enhancement factor (ε) due to periodic operation is defined to evaluate the impact of induced liquid flow modulation on reaction rate. Influence of cycling and system parameters on the enhancement factor is also reported for a wide range of conditions. Experimental trends observed by several authors can be explained with this approach.     

Cyclic operation strategy for extending cycle life of trickle beds under gas-liquid filtration

The hydrodynamics of a trickle-bed reactor subjected to fines-contaminated feed flows was studied under cyclic operation to assess whether or not periodic flows are able to reduce fines deposition and hence to extend reactor operational life under filtration conditions. The bed was subjected to single-pass kaolin+kerosene suspension and air flows to study liquid-, gas- and alternating liquid-gas cyclic operation policies. It was found that fast- and slow-mode liquid cyclic operation policies were not be able to decrease either the pressure drop or the specific deposit; however, with adjusting carefully the parameters of liquid cyclic operation it would be possible to prolong the trickle-bed cycle life. The alternating gas-liquid cyclic feed was also found useful to reduce the levels of pressure drop and bed specific deposit and exhibited the same efficiency for both short and tall beds.     

Three-dimensional analysis of trickle flow hydrodynamics: Computed tomography image acquisition and processing

Computed tomography (CT) is a powerful technique that can be used to image multi-phase flow in three dimensions. In this study, X-ray CT is used to image trickle flow in a stationary packed bed with a spatial resolution of . Errors introduced during the radiograph acquisition and tomographic reconstruction of the volume image make it difficult to identify the three phases (gas, liquid and solid) and in particular the interfacial areas. A novel post-processing strategy based on the matrix convolution operation and a priori knowledge of the shape of the particles is developed that makes it possible to accurately identify the phase interfaces in an unbiased way. The result is a ternary three-dimensional image where each voxel is one of gas, liquid or solid. From this, the gas-liquid, gas-solid and liquid-solid interfacial areas can be calculated. The proposed procedure yields images that are superior to those obtained from the usually employed thresholding operation.     

Cyclic variation of the liquid flow residence time in periodically operated trickle-bed reactors

The cyclic variation of the mean residence time of the liquid phase is investigated in a trickle-bed reactor operated with a liquid feed rate modulated in a periodic square wave pattern. Experiments made using a salt tracer method are compared to a residence time model, based on the concept of continuity shock waves. The model predicts accurately the mean residence times and their cyclic variation in case of a non-zero base liquid flow rate. A particular application of the model is the adjustment of the feed cycle parameters in order to obtain a constant residence time of the liquid, no matter the moment at which it enters the bed. This particular cycle duration depends, among others, on the feed rates, but also on the bed length.     

Hydrodynamics of trickle bed reactors at high pressure: Two-phase flow model for pressure drop and liquid holdup, formulation and experimental validation

A large experimental database has been established at IFP on the same experimental setup to measure simultaneously pressure drop and liquid holdup in packed bed reactor operated in trickle for a large range of operating conditions. The varying parameters are liquid viscosity and density, gas density, bed particle shape and size. The range for gas density range is particularly large (from 1.3 to ), thanks to the use of dense gas to simulate very high pressure conditions. This data bank has been first used to compare the prediction accuracy of the different models from the literature. Finally, the mechanistic model proposed by Attou et al. [1999. Modelling of the hydrodynamics of the cocurrent gas-liquid trickle flow through a trickle-bed reactor. Chemical Engineering Science 54, 785-802] has been improved by adding a new formulation for liquid film tortuosity in two-phase flow conditions. This model has been validated over the whole data range and the accuracy has been checked with data external to the data bank. The prediction accuracy is significantly increased when compared with the best available models for pressure drop and liquid retention in trickle flow reactors.     

Investigation of the unsteady two-phase flow with small bubbles in a model bubble column using phase-Doppler anemometry

The unsteady two-phase flow of water laden with small air bubbles in a model bubble column is investigated experimentally. Phase-Doppler anemometry (PDA) is used for measuring the velocities of water and bubbles. The measured sizes of reflecting tracers in the water and of the air bubbles are used to discriminate between water and bubble data. The investigations are focussed on the unsteady behaviour of the flow and on the interaction between the two phases. The measurement of relative (slip) velocities between bubbles and water reveals information about the dynamic behaviour of the two-phase system under the action of buoyancy on the disperse phase. The evaluation of time series of bubble velocities yields insight into typical frequencies at which the flow fluctuates. It is shown that, at all locations in the flow field, the velocity probability density functions of bubbles and liquid can be described by two superimposed Gaussian functions. The bubbles belonging to the two Gaussians exhibit different slip velocities. The probability for the occurrence of bubble collisions is quantified on the basis of the PDA data.     

Flow regimes in trickle beds using magnetic emulation of micro/macrogravity

Strong inhomogeneous magnetic fields in atmospheric-bore superconducting solenoid magnets were used to investigate the hydrodynamics in bore-fitted trickle beds which undergo emulated earth-bound artificial micro/macrogravity. This environment was able to modify the apparent gravity for both diamagnetic and paramagnetic materials by means of magnetization body force densities. Body force vectors can be co-linear or antiparallel to the cocurrent two-phase downflow in trickle beds depending on material magnetic susceptibilities, magnetic field gradient and direction of magnetic field. Trickle-to-pulse flow transition was experimentally studied in microgravity, macrogravity and beyond-levitation conditions for the air-water and the phenylacetylene-kerosene/hydrogen systems. Magnetic fields were found to displace the transition boundary from trickle to pulse flow. This was rationalized in terms of an equivalent artificial gravity effect by formally commuting magnetization forces into an equivalent gravitational acceleration. A theoretical analysis, using a modified Grosser et al. [1988. Onset of pulsing in two-phase cocurrent downflow through a packed bed. A.I.Ch.E. Journal 34, 1850-1860] “artificial gravity” transition model, was carried out and model predictions were found to follow qualitatively the experimental findings.     

Liquid flow texture analysis in trickle bed reactors using high-resolution gamma ray tomography

Trickle bed reactor performance and safety may suffer from radial and axial liquid maldistribution and thus from non-uniform utilization of the catalyst packing. Therefore, experimental analysis and fluid dynamic simulation of liquid–gas flow in trickle bed reactors is an important topic in chemical engineering. In the present study for the first time a truly high-resolution gamma ray tomography technique was applied to the quantitative analysis of the liquid flow texture in a laboratory cold flow trickle bed reactor of 90 mm diameter. The objective of this study was to present the comparative analysis of the liquid flow dynamics for two different initial liquid distributions and two different types of reactor configurations. Thus, the hydrodynamic behavior of a glass bead packing was compared to a porous Al₂O₃ catalyst particle packing using inlet flow from a commercial spray nozzle (uniform initial liquid distribution) and inlet flow from a central point source (strongly non-uniform initial liquid distribution), respectively. The column was operated in downflow mode at a gas flow rate of 180 L h⁻¹ and at liquid flow rates of 15 and 25 L h⁻¹.     

Rationalising MRI, conductance and pressure drop measurements of the trickle-to-pulse transition in trickle beds

Recent MRI data have shown that the transition from trickle to pulsing flow in trickle-bed reactors occurs over a range of liquid velocities at constant gas velocity. The transition is initiated by isolated local pulsing events, which increase in number with increase in liquid velocity until a maximum number exists. Above this liquid velocity, which we have termed the transition point, the individual pulses merge until a single macro-scale pulse is formed and the whole bed demonstrates pulsing flow. In this paper we compare the characterisation of the transition obtained using conductance and pressure drop measurements with that obtained using MRI. Using the insights gained from the 3-D MRI measurements, recorded with a data acquisition time of 280 ms, it is shown that the conductance and pressure drop measurements are sensitive to different stages of the evolution of the hydrodynamic transition, a factor important when using these different measurements in the development and validation of numerical and theoretical models. Conductance measurements identify unambiguously only the onset of the single macro-scale pulse regime, consistent with a determination of the transition point made by visual observation. In contrast, pressure drop measurements are sensitive to both the onset of formation of local pulses and the liquid velocity at which the maximum number of liquid pulses occurs. We also show how a combination of conductance and pressure drop measurements can be used to fully characterise the transition, thereby enabling translation of the insights gained by MRI into a robust measurement strategy for use on larger-scale reactors. Data are reported for a cylindrical column of length 70 cm and inner diameter 43 mm, packed with cylindrical porous γ-Al₂O₃ packing elements of length and diameter 3 mm. The bed was operated under conditions of co-current downflow of air and water, at ambient temperature and a pressure of 2 barg. Gas and liquid superficial velocities were in the range 25-300 and 0.9-, respectively.     

Application of Magnetic Resonance Imaging (MRI) for investigation of fluid dynamics in trickle bed reactors and of droplet separation kinetics in packed beds

Magnetic Resonance Imaging technique was used to investigate the fluid dynamics of multiphase flow in two different applications: (a) stationary two-phase flow in trickle beds, and (b) time-dependent droplet separation in granular bed filters. The experiments were carried out with different gas/liquid systems at either atmospheric conditions or at elevated pressure and temperature. Two-dimensional and three-dimensional image data were then evaluated to quantify the porosity profiles and gas/liquid distribution in packed beds. The results compare well with data from integral measurements.     

Identification of flow regime transitions in an annulus sparged internal loop airlift reactor based on higher order statistics and Winger trispectrum

Higher order statistics and Wigner higher order moment spectra were used to extract useful flow regime characteristics from wall pressure fluctuation signals in an annulus sparged internal loop airlift reactor. It is found that the pressure fluctuation in the airlift reactor is a typical nonlinear and non-stationary process, which exhibits different frequency characteristics depending on flow regimes. Analysis methods based on bispectrum and Wigner trispectrum are powerful tools to reveal frequency characteristics of pressure signals. To identify flow regime transitions in the reactor, two new characteristic quantities, namely average bispectrum and generalized average frequency, are defined from bispectrum and Wigner trispectrum of the pressure signal, respectively. Two flow regime transition points corresponding to three flow regimes in the reactor are successfully detected by using these two characteristic quantities.     

Origin of pressure fluctuations in an internal-loop airlift reactor and its application in flow regime detection

The origin of pressure fluctuations in an internal-loop airlift reactor (ILALR) and its application in the flow transition detection are investigated. It is found that pressure fluctuations can be characterized as global pressure fluctuations and local pressure fluctuations by frequency domain analysis and wavelet analysis. The global pressure fluctuations generated by gas compression in the gas plenum and flow fluctuations in the gas-supply system have almost a linear attenuation in the downcomer and almost no attenuation in the riser, especially in heterogeneous flow regime. However, it is found that the pressure wave from bubble eruption at bed surface has little impact on the wall pressure fluctuations. The global pressure fluctuations may be explained by Sasic's model. The local bubble-induced pressure fluctuations generated by bubble passage, coalescence and breakage can be determined by bubble passage frequency bandwidth and lower coherence. After extracting the local bubble-induced pressure fluctuations from the origin wall pressure fluctuations, it is shown that the Hurst exponent of the local pressure fluctuations increases faster in the homogeneous flow regime than in the heterogeneous flow regime, which can be employed to indicate the flow regime transition.     

Trickle flow hydrodynamic multiplicity: Experimental observations and pore-scale capillary mechanism

Trickle bed reactors are encountered throughout the process industry. Considerable attention has been given to the study of the hydrodynamics of this type reactor. It has been identified that, in the trickle flow regime, the hydrodynamic parameters (e.g. pressure drop and liquid holdup) are not unique functions of the operating and system conditions, but depend on the flow history. This study reviews the experimental trends identified in literature on the basis of a limiting cases framework and then evaluates the three-dimensional pore-scale liquid distribution using computed tomography (CT) data. This leads to the identification of 20 phenomenological trends that characterize hydrodynamic multiplicity, including hydrodynamic flow hysteresis as well as the effects of pre-wetting. The CT study yields additional experimental insight into the role of capillary pressure and ultimately leads to the proposal of a capillary gate mechanism based on contact angle hysteresis as the root cause of multiplicity. The mechanism is incorporated into a simple pore-network model. It is shown that the qualitative performance of the model corresponds closely to the majority of phenomenological trends and is capable of explaining the observed experimental behaviour.     

Bubble formation from a free-standing tube in microgravity

The bubble characteristics and the bubble detachment mechanisms during injection of air from a free-standing capillary tube submerged in water were studied in microgravity. The experiments were conducted in the 2.2-s drop tower at the NASA Glenn Research Center. A tube, 0.51 mm in diameter and 150 mm long, in a rectangular test section ( long) served as the injector. Images of the bubbles during the drops were acquired using a high-speed camera for various gas flow rates. Bubble detachment was observed for all the Weber numbers tested (0.28-31.12). This observation was different from previous studies using plate orifices in microgravity when bubble detachment was observed only for Weber numbers larger than 8. In order to resolve these differences, experiments were carried out using plate orifices. It was found that the bubbles detached from the orifice for all Weber numbers and that the bubbles formed were larger than those formed with the tube injector, particularly at low gas flow rates. The availability of a large area for the bubble to anchor itself and the presence of the chamber underneath the orifice could cause these differences. The effects of the chamber volume on the unsteadiness of bubble formation in plate-orifices at low gas flow rates are discussed.     

An experimental study of air-water Taylor flow and mass transfer inside square microchannels

Flow and mass transfer properties under air-water Taylor flow have been investigated in two square microchannels with hydraulic diameters of 400 and 200 μm. Experimental data on Taylor bubble velocity, pressure drop and liquid side volumetric mass transfer coefficient (k_La) have been presented. It was shown that the measured Taylor bubble velocity in square microchannels could be well interpreted based upon an approximate measurement of the liquid film profile therein. Then, the obtained two-phase frictional pressure drop values in both microchannels were found to be significantly higher than the predictions of the correlation proposed by Kreutzer et al. [2005b. Inertial and interfacial effects on pressure drop of Taylor flow in capillaries. A.I.Ch.E. Journal 51, 2428-2440] when the liquid slug was very short, which can be explained by the inadequacy of their correlation to describe the excess pressure drop caused by the strong inner circulation in such short liquid slugs. An appropriate modification has been made to this correlation in order to improve its applicability in microchannels. Finally, the experimental (k_La) values in the microchannel with hydraulic diameter of 400 μm were found to be in poor agreement with those predicted by the existing correlations proposed for capillaries with diameters of several millimeters. The observed deviation was mainly due to the fact that mass transfer experiments in this microchannel actually corresponded to the case of short film contact time and rather poor mixing between the liquid film and the liquid slug, which was not in accordance with mass transfer assumptions associated with these correlations. A new empirical correlation has been proposed to describe mass transfer data in this microchannel.