Search

Gas-liquid flows with wide range of bubble sizes are commonly encountered in many nuclear gas-liquid flow systems. In tracking the changes of gas volume fraction and bubble size distribution under complex flow conditions, numerical studies have been performed to validate predictions of the onegroup and two-group approaches against experimental measurements for upward gas-liquid flows in vertical pipes. These experiments have been strategically chosen because of particular flow conditions yielding specific trends of bubble size evolution, which provided the necessary means of carrying out thorough assessments of bubble coalescence and break-up kernels. Predictions of one-group approach for bubbly flows were in good agreement with experimental data. Predictions of two-group approach for bubbly and cap flows were in reasonable agreement with experimental data; additional insights into the appropriate bubble interaction mechanisms are still required for cap flows. Nevertheless, the encouraging results demonstrated the capability of both approaches in capturing the dynamical changes of bubbles size due to bubble interactions and the transition from “wall peak” to “core peak” gas volume fraction profiles caused by the presence of small and large bubbles.

In consideration of the practical importance regarding the application of gas–liquid flow in a vertical pipe and the quest towards the development of more robust physical models to accurately predict the essential interfacial parameters of the two-phase flow, comprehensive analysis of the characteristics and phase distribution patterns of such a flow have been performed on both experimental measurements and theoretical predictions. In this first part, analysis of experimental data in a large diameter pipe with an inner diameter of 195.3 mm via the wire-mesh senor measuring technique was presented. The experiments were performed at the TOPFLOW facility of Helmholtz-Zentrum Dresden-Rossendorf. In the present paper, measurements of local interfacial parameters which included the void fraction, volume equivalent bubble diameter, bubble size distribution and interfacial velocities were discussed. Test points covering flow regimes from bubbly to cap to slug to churn-turbulent flow were selected to investigate the flow of different bubble shapes and sizes and the significant bubble coalescence and break-up mechanism throughout the large vertical pipe. The radial and axial evolutions of the local flow structure were interpreted in terms of the classifications of different groups of bubbles (Group-1 and Group-2). In addition, the phase distribution patterns were analyzed through the concept of skewness, which essentially identified two fundamental patterns, namely, wall peak and core peak. In general, Group-1 bubbles being smaller spherical bubbles have shown to exhibit a wall peaking distribution while Group-2 bubbles being larger non-spherical bubbles corresponded to a core peaking distribution. The classification of bubbles that have been performed in this present study can be employed for the development of bubble coalescence and break-up mechanistic kernels and other interfacial force closure models for a two bubble group approach in the context of computational fluid d

The basic concept of two-group average bubble number density equations along with three-fluid model has been demonstrated for vertical gas-liquid flow. Specifically, the current study focused on: (i) classification of bubble interaction between spherical bubbles (Group-1) and cap bubbles (Group-2), (ii) preliminary consideration of source and sink terms in the averaged bubble number density equations via the model of Hibiki and Ishii [1] and (iii) assessment by means of experimental data sets at bubbly-to-cap flow transition. Reasonable agreement was achieved between measured and predicted distributions of void fraction, interfacial area concentration (IAC) and volume equivalent bubble diameter.

Considerable attention has been concentrated on describing the temporal and spatial evolution of two-phase geometrical structure caused by the effects of bubble interactions in gas-liquid flows. In the published literature, the transport phenomena of dispersed bubbles in bubbly flow conditions can be regarded in a similar view of the drag and interaction of spherical bubbles, which have brought about the development of most coalescence and break-up mechanisms based primarily on the assumption of interaction between such bubbles. Nevertheless, cap bubbles which are precursors to the formation of slug units in the slug flow regime with increasing volume fraction become ever more prevalent at high gas velocity conditions. It has been shown through many experiments that interaction behaviors between non-spherical bubbles in a liquid flow are remarkably different when compared to those of spherical bubbles. It is therefore imperative additional mechanisms of bubble interactions need to be considered, particularly for cap bubbles, in addition to typical mechanisms that have been established for spherical bubbles. In this work, a two-group modeling of bubbly-cap flows via the transport equations of the average bubble number density has been considered to predict the bubble size distribution of the different bubbles co-flowing with the liquid. Based on the computational fluid dynamics (CFD) framework, a three-fluid model was solved, one set of conservation equations for the liquid phase while two sets of conservation equations for the gas phase with one being Group 1 spherical bubbles and the other depicting Group 2 cap bubbles. The drag and non-drag characteristics of the different sizes and shapes of bubbles were thus accounted via the different momentum equations representing Groups 1 and 2 bubbles. In this initial assessment, the bubble mechanistic models proposed by Hibiki and Ishii (2000) have been adopted to describe the intra-group and inter-group interactions. The

Gas–liquid bubbly flows with wide range of bubble sizes are commonly encountered in many industrial gas–liquid flow systems. To assess the performances of two population balance approaches – Average Bubble Number Density (ABND) and Inhomogeneous MUlti-SIze-Group (MUSIG) models – in tracking the changes of gas volume fraction and bubble size distribution under complex flow conditions, numerical studies have been performed to validate predictions from both models against experimental data of Lucas et al. (2005) and Prasser et al. (2007) measured in the Forschungszentrum Dresden-Rossendorf FZD facility. These experiments have been strategically chosen because of flow conditions yielding opposite trend of bubble size evolution, which provided the means of carrying out a thorough examination of existing bubble coalescence and break-up kernels. In general, predictions of both models were in good agreement with experimental data. The encouraging results demonstrated the capability of both models in capturing the dynamical changes of bubbles size due to bubble interactions and the transition from ‘‘wall peak’’ to ‘‘core peak’’ gas volume fraction profiles caused by the presence of small and large bubbles. Predictions of the inhomogeneous MUSIG model appeared marginally superior to those of ABND model. Nevertheless, through the comparison of axial gas volume fraction and Sauter mean bubble diameter profiles, ABND model may be considered an alternative approach for industrial applications of gas–liquid flow systems.

To model the macroscopic bubble interactions (e.g. coalescence and breakage), the two-fluid model in conjunction with the population balance equation (PBE) approach has been considered as a practical and accurate formulation of handling bubbly flow systems. Recently, the MUltiple SIze Group (MUSIG) model appears to be one of the most direct solution methods which solves the PBE with discrete class approach and fuses seamlessly with the computational fluid dynamics (CFD) framework. Nonetheless, for complex bubbly flow structures with wide range of bubble size, large number of classes must be used to attain sufficient resolution for the bubble size distribution (BSD). This poses severe limitations on both computational time and resources. This paper focuses on introducing an alternative direct quadrature method of moments (DQMOM) (Marchisio and Fox, 2005) where the BSD is tracked through its moments by integrating out the internal coordinate. The main advantage of DQMOM is that the number of scalars to be solved is very small (i.e. usually 4-6). To assess the performance of DQMOM in measure-up with the MUSIG model, predictions of both models are validated against two experimental data by Hibiki et al. (2001) and Lucas et al. (2005). In general, the model predictions compared very well against the measured data. Associated numerical issues and drawbacks for the DQMOM model are also discussed.

Polydispersed bubbly flows with wide range of bubble size are commonly encountered in many industrial fields. The use of population balance models coupled with the two-fluid model presents the most viable way of handling such complex flow structures. The main focus of this paper is to access the capabilities of two population balance models – namely Average Bubble Number Density (ABND) and Inhomogeneous MUlti-SIze-Group (MUSIG) model; in resolving the dynamical changes of void fraction and bubble size distribution under polydispersed flow conditions. Numerical predictions are validated against two polydispersed flow measurements. Special attentions are directed towards the performance of the two models in capturing the behavioral transition of “wall peak” to “core peak” void fraction profile. Applicability and drawbacks of the two population balance models for industrial applications are also discussed.

Practicing engineers are constantly confronted with the prospect of solving complex gas-liquid bubbly flow problems in real industrial systems. The use of population balance models coupled with the two-fluid model presents the most viable way of handling such flows. The homogeneous MUltiple-SIze-Group (MUSIG) model has recently become a widely adopted population balance approach whereby the continuous bubbles size range can be represented by a series of discrete classes. The improved inhomogeneous MUSIG model extends the capability of accounting different bubble shapes and travelling gas velocities. Conversely, the Average Bubble Number Density (ABND) model represents another simpler approach in handling bubble interactions in complex gas-liquid bubbly flow. The capability of these three population balance models is assessed. Particular emphasis is directed towards the possible handling of bubbly-to-slug transition flow conditions. Numerical predictions are compared against experimental data obtained from Lucas et al. [1] and Hibiki et al. [2]. Shortcomings and applicability of these models for industrial applications are also discussed.

For increasing the predictability of equipment and improving efficiency of production, there is a high demand to develop a compact and efficient mathematical model capable of modelling the complex bubbly flow structures which frequently occur in large-scale industrial engineering systems. A generalized Average Bubble Number Density (ABND) transport equation model in conjunction with three forms of bubble coalescence and breakage kernels was implemented and incorporated into commercial software ANSYS CFX 11. The main focus of this paper is to assess the overall performance of the ABND model and the three different bubble mechanism kernels under a large-scale gas-liquid bubbly flow system. Based on the high-quality TOPFLOW database for airwater two-phase flows in a large vertical pipe with nominal diameter of 195.3mm, experimental data were strategically selected for model validation. To examine the relative merits and drawbacks of three forms of coalescence and breakage kernels, model predictions of local radial distributions of bubble rise velocity, volume fraction and bubble size were compared against experimental results. The capabilities in predicting the “core peak” volume fraction profiles and evolution process of bubble rise of different kernels were discussed.

Practicing engineers are constantly confronted with the prospect of solving complex gas-liquid bubbly flow problems in real industrial systems. The use of population balance models coupled with the two-fluid model presents the most viable way of handling such flows. The homogeneous MUltiple-SIze-Group (MUSIG) model has recently become a widely adopted population balance approach whereby the continuous bubbles size range can be represented by a series of discrete classes. The improved inhomogeneous MUSIG model extends the capability of accounting different bubble shapes and travelling gas velocities. Conversely, the Average Bubble Number Density (ABND) model represents another simpler approach in handling bubble interactions in complex gas-liquid bubbly flow. The capability of these three population balance models is assessed. Particular emphasis is directed towards the possible handling of bubbly-to-slug transition flow conditions. Numerical predictions are compared against experimental data obtained from Lucas et al. [1] and Hibiki et al. [2]. Shortcomings and applicability of these models for industrial applications are also discussed.

Transient boiling is examined for conditions related to the hypothetical heating of liquids in a storage tank by an external fire, with the potential for evaporation of the liquids and the release of toxic gases into the environment. Temperature and void fraction distribution data were obtained from an experiment on water heated in a tank. Numerical simulations were also performed. These generally agreed reasonably well with measurements.