Search

The lift force acting on particles, drops, and bubbles in a shear field is a well know effect that was extensively investigated since the 1980s. Experiments with single bubbles in a linear shear field reveal that the lift force coefficient has at a specific bubble size a zero where the coefficient switches its sign from positive to negative. Solving the lateral force balance for a polydisperse bubbly flow with the liquid flow field usually found in a bubble column, the lift force causes a spatial separation of small, which tend to the wall, and big bubbles, which tend to the center. Bubbles with a zero lift force coefficient on the other hand are equally distributed over the cross section. In order to investigate the influence of the flow field on the force balance, simulations with the Euler-Euler two-fluid model were conducted. The simulations are in good agreement with experiments and showed that the liquid flow field found in the used bubble column has a minor influence on the steady state distribution of the bubbles. From evaluating six different bubble column experiments conducted in air/water, the critical diameter at which the lift force coefficient is zero is determined between 5.1 and 5.2 mm. This result is very close to the critical diameter that we were able to determine in previous, single-bubble experiments. Therefore, the conclusion is possible that the lift force coefficients that are determined with single bubble experiments are applicable to the complex flow field found in polydisperse bubbly flows at low void fractions and low turbulence levels. Moreover, a very simple experimental setup is described to determine the critical diameter in complex flow situations, which is the most important point when it comes to modelling the lift force. Such a setup is therefore beneficial when complex substance mixtures used in a specific industrial or environmental application needs to be investigated since lift force correlations are only available for clean, ideal substances.

An Euler–Euler two-fluid approach was used to simulate the behavior of gas bubbles rising in a stagnant liquid metal. A single point injection in the range of moderate gas flow rates results in the formation of bubble chains undergoing distinct oscillations of the bubble trajectories. A set of interfacial closures and a shear stress transport k-ω (SST) turbulence model, namely the baseline model for bubbly flow (Rzehak and Krepper, 2013b) was applied for simulating the transient behavior of the bubble chain. X-ray radiography measurements were conducted to establish an experimental data base for validating the numerical results. The experiments provide a visualization of the two-phase flow in a flat container and allow for determining essential bubble quantities such as the size, shape, trajectory and velocity. The comparison between numerical simulations and experimental data showed a very good qualitative and quantitative agreement with respect to the distribution of the void fraction and the dynamics of the bubble chain. Wrong results were obtained by simulations where the effect of the bubble induced turbulence (BIT) was neglected. Two BIT models were applied within this study, the baseline BIT model and the Sato BIT model. Both models showed a good agreement with the experimental observations, while the results of the baseline model were even closer to the measurements. Thus, the baseline model originally developed for the air-water system has proved to be capable of reproducing the complex transient behavior of oscillating bubble chains in liquid metals.

The shear-induced lift force is known to influence the lateral distribution of gas bubbles in bubbly flows. Although the hydrodynamic behavior of a bubble can be greatly affected by surfactants that are present in the liquid bulk, their influence on the lift force has only been investigated to a limited extent. In our previous work we investigated the influence of impurities on the lift force in air-water flows and could reveal non-negligible changes even without a modification of the bubble drag or shape. To bring further insight on changes caused by higher surfactant concentrations, the lift coefficient of single ellipsoidal bubbles of different sizes, which rise in water with varying degree of contamination are experimentally determined in this work. For this purpose, different amounts of 1-Pentanol as well as Triton X-100 were added to the flow. The results reveal a strong dependency of the lift coefficient on the bubble shape, where different findings in the literature for bubbles with lower Reynolds numbers could also be observed for ellipsoidal bubbles in water.

Scaling of the lift reversal for a deformed bubble in the surface tension-inertial force dominant regime was discussed. Lift data of bubbles in water recently reported in literature were used. The negative lift component was well correlated in terms of the drag coefficient, which, in turn, implies that the vorticity produced at the bubble surface plays a key role in both drag and lift forces as is the case with the viscous force dominant regime. The scaling was confirmed to give good evaluations of the lift coefficients.

Ultrafast X-ray tomography enables non-invasive imaging of gas-liquid flows with high spatial and temporal resolution. While it is relatively straightforward to extract e.g. gas fraction profiles from cross-sectional tomographic images, the extraction of bubble and gas-liquid interface information requires advanced image processing techniques. Thereby it is an important necessity to transform the temporal scale in the scanned sequences into a corresponding length scale for obtaining correct volumetric information. For bubbly flows this means that the velocity of the dispersed phase, e.g. the gas bubbles, has to be determined from dual-plane scans. A common and widely applied method to obtain gas phase velocities is cross-correlating the image sequences of the two scanning planes. This gives an averaged velocity for each position in the cross-section. In the present work, a new method is introduced, which determines the velocity of individual gas bubbles. This new method is termed as “bubble twinning method”, because it tries to identify twin-bubbles in both scanning planes. The developed algorithm compares essential bubble parameters, that is, volume, position and residence time in the slice, by applying a fuzzy-logic based membership function approach. The algorithm was tested for bubbly flow as well as slug flow conditions. Results are compared with established theoretical predictions as well as the cross-correlation method.

For the simulation of bubbly flows, knowledge of the lift force as an interaction between gas bubbles and a surrounding shear field is of great importance. The sign of the lift coefficient C L changes with increasing bubble size, i.e. with more pronounced bubble deformation. Beside this, impurities in terms of surface-active components are well-known to change the complete hydrodynamic behavior of a bubble even if the amount is very small. In the present work, the lift coefficient of single ellipsoidal bubbles is determined with a recently developed method, which is suitable to overcome difficulties connected to low viscous systems. In order to investigate the influence of impurities on the lift force, we conducted experiments with single bubbles of different sizes in purified, deionized and tap water. Overall, the determined lift coefficients show no difference between deionized and tap water but reveal differences to results obtained with purified water. As no significant differences in shape and velocity are found between the different water qualities, it remains unclear how the impurities cause the observed differences. For the deionized and tap water results that are more relevant in practice, a new correlation is proposed to account for the observed differences in comparison to data from the literature. It can be used to calculate C L of ellipsoidal bubbles in the investigated size range.

A three-dimensional numerical study of air/water bubbly flow in a cylindrical large-scale bubble column is performed using Euler-Euler approach. The main objective is to investigate the influence of different boundary conditions such as bubble size distribution (BSD), polydisperse effects (mono and bi-dispersed approach), lift force modelling and flow rate distribution at sparger using a computationally affordable approach. In the bi-dispersed approach the population of bubbles are divided into two groups of small and large bubbles and a mean diameter is considered for each group. The division is based on the critical bubble diameter, for which the lift coefficient changes its sign from positive to negative. For air/water system Tomiyama lift coefficient model is widely used and the critical bubble diameter is equal to 5.8 mm. An alternative critical bubble diameter and lift force coefficient correlation is used in this study and compared with the well-known Tomiyama model. The numerical predictions are compared against the experimental data and the effect of different conditions is assessed on basis of comparison of axial gas fraction (local holdup) and global holdup. Better predictions are obtained by taking into account polydisperse flow with new critical bubble diameter and new lift coefficient model. Also, it was found that mass flow-rate distribution at the sparger does not affect numerical results for global and local holdup, however a different flow pattern is observed near the sparger region.

Flashing flow is an important phenomenon in many industrial contexts; however simulation of these flows remains difficult. CFD simulations are able to describe the distribution and evolution of 3D structures in the flow but are dependent on good closure relations for interphase transfer. Nucleation during flashing flow is often neglected in CFD simulation where a minimum starting vapour fraction and a constant bubble number density are given. Models that include nucleation have used wall nucleation terms from 1D system code models, averaged over the domain. In this work, three models for wall nucleation are tested and compared with experimental data from a converging–diverging nozzle. Nucleation is applied at the walls of the domain, and various models are investigated. Good agreement with the critical flow rate and axial profiles are found, but agreement with the radial void fraction data is not satisfactory. Methods of addressing this are explored, and it is found that including a small bulk heterogeneous nucleation term gives the best agreement with the radial profiles, with negligible impact on the axial average properties.

The role of the bubble size dependent lateral lift force on the flow regime is experimentally investigated in a tall bubble column. In the experiments, only the initial bubble size distribution was modified by using a gas sparger with different injection needles and varying the flow through specific injection needles. The gas flow rate was kept unchanged in all experiments. Depending on the bubble size distribution homogeneous, transitional or heterogeneous flow regimes were observed. The flow regime is in good agreement with the predictions from the stability criterion obtained by a previously conducted linear stability analysis for poly-dispersed flow. It was found that as a rule of thumb the following criterion can be used. If the majority of the gas volume is transported by bubbles smaller than the critical diameter at which the lift force changes its sign, the flow is stabilized leading to a homogeneous flow regime. If most of the gas volume is transported by bubbles larger than that diameter the flow is de-stabilized leading a heterogeneous flow regime. If the fraction of gas transported by small and large bubbles is about the same, the initial conditions remain dominant throughout the column height.

Three-dimensional numerical simulations are presented considering the impact of a steady magnetic field on a bubble-driven liquid metal flow inside a cylinder. The injection of moderate gas flow rates through a single orifice at the bottom of the fluid vessel results in the formation of a bubble plume. The magnetic field is applied in either vertical or horizontal direction. The calculations were performed by means of the commercial software package CFX using the Euler–Euler multiphase model and the RANS–SST turbulence model. The non-isotropic nature of MHD turbulence was taken into account by specific modifications of the turbulence model. The numerical models are validated with recent experimental results. (Zhang, C., Eckert, S., Gerbeth, G., 2007. The flow structure of a bubble-driven liquid–metal jet in a horizontal magnetic field, J. Fluid Mech. 575, 57–82.) The comparison between the numerical simulations and the experimental findings shows a good agreement. The calculations are able to reproduce a striking feature of a horizontal magnetic field found in the range of moderate Hartmann numbers revealing that such a steady transverse magnetic field may destabilize the flow and cause distinct oscillations of the liquid velocity.

This paper presents a concept for the CFD-modelling of multiphase flows where both segregated and dispersed flow structures occur simultaneously. Transitions between such morphologies, characterized by different scales of interfacial structures, are investigated and a new multi-field two-fluid strategy for a generalized two-phase flow (GENTOP) is presented. The GENTOP-approach extends the inhomogeneous Multiple Size Group (MUSIG)-model by adding an additional continuous gas phase. Within the MUSIG-framework, mass transfers between different bubble size groups due to coalescence and breakup as well as gas–liquid transfers are described. By modelling an additional mass transfer between the polydispersed and continuous gas phase, transitions between the different gas morphologies can be considered dependent on the flow situation. The continuous gas phase summarizes gas structures which are large enough to be resolved within the computed mesh. Therefore a free surface detection and generalized formulations for interfacial transfer models are introduced. The appearance of one particular gas phase due to mass transfer from another gaseous morphology is demonstrated by means of two demonstration cases: the impingement of a liquid jet on a free surface with an associated entrainment of dispersed bubbles as well as an evolving vertical bubble column showing a wide spectrum of bubble sizes.

Void fraction, Taylor bubble and liquid slug lengths, and slug frequency are parameters essential to any description of the structure of slug flow. In the present study, these parameters were extracted from the time series of cross-sectionally averaged void fraction obtained from two vertical facilities having similar internal pipe diameter but significantly different axial lengths; 51.2 mm/3.5 m and 52.3 mm/9 m. In order to study slug flow, the flow rates for which it occurred were first identified. To investigate the effect of flow development on slug characteristics measurements were carried out at several axial locations from the mixing section for both facilities. For slug frequency, a new correlation including the effect of the axial length has been proposed and assessed using previously published data.

Although, most of the work reported on two-phase flows are limited to small pipe diameters, two-phase flow in large risers are increasingly being encountered in the petroleum and nuclear industries. In the present work, a wire mesh sensor was employed to obtain void fraction and bubble size distribution data and visualizations of steam/water flow in a large vertical pipe (194mm in diameter) at 46bar. For comparison purposes, measurements were made at similar phase velocities and physical properties to the data of Omebere-Iyari et al. (2006), which is for the flow of a nitrogen/naphtha mixture in a similar-sized riser. There exist significant differences between both sets of data. Churn-turbulent flow is observed in the present work instead of slug flow which differs from the intermittent and semi-annular flow patterns reported by Omebere-Iyari et al. (2006). The mean void fraction of the nitrogen/naphtha data is higher that of the present steam/water data. On examination of radial void fraction profiles, core peak distributions are observed for the present work in contrast to the wall peak profiles predicted for the data of Omebere-Iyari et al. (2006) using a power law relationship.

Measurements of the cross-sectional distribution of the gas fraction and bubble size distributions were conducted in a vertical pipe with an inner diameter of 51.2 mm and a length of about 3 m for air/water bubbly and slug flow regimes. The use of a wire-mesh sensor obtained a high resolution of the gas fraction data in space as well as in time. From this data, time averaged values for the two-dimensional gas fraction profiles were decomposed into a large number of bubble size classes. This allowed the extraction of the radial gas fraction profiles for a given range of bubble sizes as well as data for local bubble size distributions. The structure of the flow can be characterized by such data. The measurements were performed for up to 10 different inlet lengths and for about 100 combinations of gas and liquid volume flow rates. The data is very useful for the development and validation of meso-scale models to account for the forces acting on a bubble in a shear liquid flow and models for bubble coalescence and break-up. Such models are necessary for the validation of CFD codes for the simulation of bubbly flows.