Search

Open-source environments such as the Computational Fluid Dynamics software OpenFOAM are very appealing for research groups since they allow for an efficient prototyping of new models or concepts. However, for downstream developments to be sustainable, i.e. reproduci-ble and reusable in the long term, a significant amount of maintenance work must be account-ed for. To allow for growth and extensibility, the maintenance work should be underpinned by a high degree of automation for repetitive tasks such as build tests, code deployment and validation runs, in order to keep the focus on scientific work. Here, an information technology environment referred to as OpenFOAM_RCS is presented that aids the centralized mainte-nance of simulation code and setup files for OpenFOAM developments concerned with reac-tor coolant system safety research. It fosters collaborative developments and review processes. State-of-the-art tools for managing software developments are adapted to meet the require-ments of OpenFOAM. A flexible approach for upgrading the underlying installation is pro-posed, based on snapshots of the OpenFOAM development line rather than yearly version releases, to make new functionality available when needed by associated research projects. The process of upgrading within so-called sprint cycles is accompanied by several checks to ensure compatibility of downstream code and simulation setups. Furthermore, the foundation for building a validation data base from contributed simulation setups is laid, creating a basis for continuous quality assurance. The content of this article was initially presented at the 33rd German CFD Network of Com-petence Meeting, held on March 22-23 2022 at GRS in Garching, Germany.

The Euler-Euler model is widely used in bubbly flow simulations up to industrial dimensions. The standard Euler-Euler model is based on the phase-averaging method. After averaging, the bubble forces in the field equations are functions of the local gas volume fraction. In simulations, when the bubble diameter is larger than the computational cell spacing, the forces can transport the gas belonging to the same bubble in different directions. By contrast, a closure model for the bubble force is typically developed based on the assumption that the force is a resultant force that acts on the bubble's center-of-mass. This inconsistency can lead to a nonphysical gas concentration in the center of a channel or near the channel wall if the bubble diameter is larger than the cell spacing. The purpose of the present contribution is to developed an Euler-Euler model where the bubble force consistency is recovered for two-phase flow simulations where the diameter of the disperse phase can be larger than the cell spacing. Such an Euler-Euler model is developed by combining an existing particle-center-averaged Euler-Euler framework with a Gaussian convolution method. To validate this Euler-Euler approach, a comparison is made with experimental data for the bubbly flows in two different vertical pipes. The results show that the proposed Euler-Euler model recovers the bubble force consistency and alleviates the over-prediction of the gas volume fraction peak near the wall, while its simulation results in the axial gas and liquid velocity and the liquid turbulence kinetic energy are similar to the results of the standard Euler-Euler model.

Pressure relief by blowdown is one of the most important measures to prevent excessive pressures in the primary circuit or containment in severe nuclear accidents. Pool scrubbing can significantly reduce the release of radioactive materials, e.g., aerosols, to the environment during the pressure relief. The decontamination factor indicating the particle retention efficiency depends, among other factors, on the hydrodynamic conditions of the gas-liquid two-phase flow inside the pool. In the present work, the hydrodynamics in two typical pool scrubbing experiments is investigated with the two-fluid model, and the influence of some key factors including bubble diameter, nozzle submergence as well as interaction models are analysed. One case is a rectangular pool and the other is a cylindrical column, and their injection Weber number is around 2×103 and 4×105, respectively. The numerical results show that as the distance from the nozzle exit increases, the void fraction and velocity field expand from the central region, where the nozzle is located, to the whole cross section. The profile and its development depends largely on the bubble size and the interaction force model. It reveals that in the monodisperse simulation, the tuning of bubble diameter is necessary for achieving good agreement, although it is difficult for high velocity gas injection. More information is required to properly describe the bubble size distribution as well as its evolution in pool scrubbing conditions. Furthermore, the experimental data show clear drag reduction in the bubble swarm generated by the gas jet, and the mechanism and model improvement possibilities need to be investigated.

In line with the best practice guidelines for computational fluid dynamics in nuclear reactor safety, Helmholtz-Zentrum Dresden – Rossendorf proposed an Euler-Euler baseline closure concept some years ago. The consideration is that simulations with a fixed set of closures instead of case by case tuning may help to identify model inadequacy and facilitate further improvement. Currently, the baseline model concerns interphase momentum and turbulence exchange as well as bubble coalescence and breakup. It has been tested for a wide range of isothermal applications with different geometrical configurations and material systems. In the present work, the baseline model is extended to non-isothermal flows by including a mechanistic model for the computation of overall interphase heat transfer coefficient. The extended baseline model is validated for both bubble-growing in superheated liquid and -condensing in sub-cooled liquid. The baseline model proposal is independent on the use of a certain CFD code. The presented simulation is carried out with the commercial software ANSYS CFX by employing the best practice guidelines. The simulated liquid temperature, gas volume fraction, vapor bubble size and velocity are compared with the measured ones. The effectivity of the model is demonstrated by a general good agreement.

Eulerian-Eulerian computational fluid dynamic (CFD) models allow the prediction of complex and large-scale industrial multiphase gas–liquid bubbly flows with a relatively limited computational load. However, the interfacial transfer processes are entirely modelled, with closure relations that often dictate the accuracy of the entire model. Numerous sets of closures have been developed, often optimized over few experimental data sets and achieving remarkable accuracy that, however, becomes difficult to replicate outside of the range of the selected data. This makes a reliable comparison of available model capabilities difficult and obstructs their further development. In this paper, the CFD models developed at the University of Leeds and the Helmholtz-Zentrum Dresden-Rossendorf are benchmarked against a large database of bubbly flows in vertical pipes. The research groups adopt a similar modelling strategy, aimed at identifying a single universal set of widely applicable closures. The main focus of the paper is interfacial momentum transfer, which essentially governs the void fraction distribution in the flow, and turbulence modelling closures. To focus on these aspects, the validation database is limited to experiments with a monodispersed bubble diameter distribution. Overall, the models prove to be reliable and robust and can be applied with confidence over the range of parameters tested. Areas are identified where further development is needed, such as the modelling of bubble-induced turbulence and the near-wall region, as well as the best features of both models to be combined in a future harmonized model. A benchmark is also established and is available for the testing of other models. Similar exercises are encouraged to support the confident application of multiphase CFD models, together with the definition of a set of experiments accepted community-wide for model benchmarking.

The thermohydraulic conditions in a storage pool for spent nuclear fuel are studied for a loss of cooling accident leading to partially uncovered fuel racks. While under normal operating conditions, the fuel is cooled by means of single-phase natural convection in the water of the pool, the heat transfer rate into the gaseous pool atmosphere for the above scenario involves several mechanisms to a similar amount which are interacting in a complicated manner, such as heat convection, thermal radiation and heat conduction. These mechanisms are analyzed with the aid of Computational Fluid Dynamics using two complementary models that address different length scales, one for the large scales of the pool and the containment geometry, another one for the small scales of the fuel rods. The models emulate the conditions at reactor unit 4 of the former Fukushima Daiichi nuclear power plant at the time of the accident, because this is a case example for a loss of cooling scenario and sufficient material for adequate modeling is available in the literature. The pool scale model serves to analyze the developing flow patterns. It includes the fuel represented as porous medium as well as the pool and reactor building atmosphere. The results show that a characteristic flow field forms in the pool atmosphere that prevails for all studied water levels and decay heat rate distributions. The temperature level, on the other hand, can be reduced significantly by means of a checkerboard storage. However, the boundary conditions in the head region of a fuel assembly are clearly a function of its storage location in the pool. Representative conditions are extracted and applied to the second model that represents a single geometry-resolved fuel assembly with a portion of the atmosphere above it. The corresponding simulations show that the heat exchange rate between the fuel assembly and the atmosphere above is higher for the conditions near the pool wall compared to the conditions in the center of the pool. The study provides important information on how to best arrange the fuel assemblies in a spent fuel pool for optimal heat exchange in case of an accident involving loss of cooling water. Checkerboarding should be combined with a strategy to place fuel with higher decay heat rate near the pool walls.

A unified set of closures have been applied to simulating different configurations and fluids, i.e. pipe flow and bubble column, air/water and air/liquid metal. The simulated velocity, void fraction and turbulence profiles were compared with the measured ones. Starting from the baseline model for poly-disperse flows the present work is intended to prove the performance of a recently published model for bubble-induced turbulence, which was established on the basis of physical analyses and direct numerical simulation data. The model is shown to work well under various conditions without any need of tuning, and significant improvement in the prediction of turbulence parameters in comparison to other models is demonstrated. This is a great step towards developing the baseline closure concept. Finally, a brief discussion on the further development and future work regarding Eulerian closure models was given.

We report on generic experimental studies dealing with the direct contact condensation of steam at a falling sub-cooled water jet at high pressure and temperature. This generic problem concerns some safety-relevant thermal hydraulic scenarios in light water reactors. One of such is the question for heat transfer and mixing when a sub-cooled jet of water in an emergency core-cooling scenario enters a hot pressurized component of the primary reactor circuit. The involved phenomena of heat transfer are complex in this case, reaching from direct condensation of steam via steam bubble entrainment to single phase mixing of hot and cold water. Experiments were performed in the TOPFLOW pressure tank, which is an experimental facility for high-pressure thermal hydraulics experiments in pressure equilibrium. The facility has been designed for studying steam-water two-phase flows at pressures of up to 50 bar. It enables to run experiments in flow domains of complex shape without high difference pressures across the wall. The concept therefore allows us to use thin metal walls and even glass windows to observe flows in complex geometry domains with the help of video camera and to considerably reduce cost and complexity of experimental installations. The experimental results show, that high ambient pressure promotes the direct contact condensation heat transfer of subcooled jet flow heavily.

In this paper, we report on experimental investigations of air/water two-phase flow in a vertical pipe, which is a benchmark experiment for the development and validation of two-phase CFD models and codes. We conducted experiments for co-current upward and downward flow as well as counter-current flow at ambient temperature and pressure conditions. As measurement technique, we employed ultrafast X-ray tomography, which gives two-dimensional cross-sectional images of the gas-liquid distribution in two planes up to 2500 Hz scanning speed. In this paper, we briefly explain this technique, introduce essential data processing steps and discuss the findings of the study.

The paper presents the extension of the GENTOP model for phase transfer and discusses the sub-models used. Boiling flow inside a wall heated vertical pipe is simulated by a multi-field CFD approach. Sub-cooled water enters the pipe from the lower end and heats up first in the near wall region leading to the generation of small bubbles. Further along the pipe larger and larger bubbles are generated by coalescence and evaporation. This leads to transitions of the two-phase flow patterns from bubbly to churn-turbulent and annular flow. The CFD simulation bases on the recently developed GEneralized TwO Phase flow (GENTOP) concept. It is a multi-field model using the Euler-Euler approach. It allows the consideration of different local flow morphologies including transitions between them. Small steam bubbles are handled as dispersed phases while the interface of large gas structures is statistically resolved. The GENTOP sub-models and the Wall Boiling Model need a constant improvement and separate, intensive validation effort using CFD grade experiments.

In this work, CFD simulations using the Euler-Euler approach were performed to model the gas-liquid flow in a swirl-generating device. The computational work was based on experiments, which are conducted in a vertical pipe packed with a static swirl element. Measurements of gas volume fractions at several planes within the swirl element were taken using high-resolution gamma-ray computed tomography (HireCT). The simulations were carried out for the experimental conditions with defined inlet gas volumetric flow rates of 5 and 10%. The profile of several key parameters (e.g. pressure, liquid and gas velocities and gas fraction) are used to understand the flow behavior inside the device. The radial gas phase distribution obtained from the simulations assuming different mono-disperse and bi-modal bubble sizes is compared against the experimental results. The significant influence of the selected bubble sizes on the profile is shown and discussed within this paper. In general, the radial profile of gas fraction is well captured by the CFD simulations except in the transition zone where a significant discrepancy to the experiment is observed.

Steady state counter-current flow limitation (CCFL) experiments were conducted in a 1:3 scaled flat model of the hot leg and part of the steam generator inlet chamber of a German Konvoi pressurized water reactor. The experiments include air-water tests at 1 and 2 bar as well as steam-water tests at 10, 25 and 50 bar. Flooding characteristics are obtained in dependency on the dimensionless gas and liquid flow rates (Wallis-parameter). They showed a slight, but clear dependency on pressure – with increasing pressure more liquid can flow in counter-current to the gas in case of partial CCFL. Also the dimensionless gas flow rate for zero penetration slightly increases with pressure. Beside the flooding characteristics also slug frequencies were analyses basing on pressure measurements along the horizontal part of the test section. Finally, comprehensive video material was obtained which is suited for extracting quantitative data on the local flow structure. This can contribute to the further CFD-code development and validation.

In the frame of the European Union NURESAFE project a benchmark test between NEPTUNE_CFD, CFX and TransAT CFD codes on a reference TOPFLOW-PTS experiment was conducted. The work is a part of the work package on multi-scale and multi-physics simulation of Pressurized Thermal Shock (PTS). The article includes a short description of the TOPFLOW-PTS facility and the reference steam-water experiment. Furthermore the boundary and initial conditions for the CFD simulations are presented. The computational grids that are used for the benchmark simulations and the models used are introduced. Finally, the results of CFD calculations are compared with the experimental data and differences between simulations and experiment are discussed.

The bubble shape is fundamental for every aspect of modeling bubbly flows. The interface is usually highly deformable so that the bubble shape is in general dependent on the surrounding turbulent flow field. Since recent work on this topic addressed almost entirely single-bubbles rising in quiescent flow, the extent of such flow field dependencies as well as swarm effects is rather unknown. This study examines the effect on the bubble shape in bubbly flows with a liquid background flow that is generated by natural convection in the bubbly flow regime when flow properties, i.e. the gas flow rate, sparger setup, and column geometry, are changed by evaluating six different bubble column experiments. The results of this integral approach reveal that the bubble shape of small bubbles is distinctly influenced whereas the shape of large bubbles is unchanged. Averaged over all flow rates, we find that the size-dependent bubble shapes are quite similar for all six experiments. Further studies focusing on single local effects like the shear rate or wake effects are highly desirable to obtain a deeper understanding of the underlying processes; for this purpose, the given results can help to assess the most important effect and in which extend it should be studied.

The paper presents an application of the GENTOP model for phase transfer and discusses the sub-models used. Boiling of a heated surface under atmospheric conditions is simulated by the multi-field CFD approach. Sub-cooled water in a generic pool is heated up first in the near wall region leading to the generation of small bubbles. Further away from the bottom wall larger bubbles are generated by coalescence and evaporation. The CFD simulation bases on the recently developed GEneralized TwO Phase flow (GENTOP) concept. It is a multi-field model using the Euler-Euler approach and it allows the consideration of different local flow morphologies including transitions between them. Small steam bubbles are handled as dispersed phases while the interface of large gas structures is statistically resolved. The GENTOP sub-models need a constant improvement and separate, intensive validation effort using CFD grade experiments. Recently, a bubble dynamics model inside the heat partitioning approach has been developed and will be applied in GENTOP in future.

Two-equation turbulence models based on the Boussinesq eddy viscosity hypothesis that have been used in the vast majority of previous simulation studies on bubbly pipe flows contain a term which renders the radial pressure distribution non-constant. In single phase simulations this effect is invariably absorbed in the definition of a modified pressure, from which the real pressure may be recovered if necessary. For bubbly multiphase flows however, this is not possible since the bubbles experience a force which depends, of course, on the real pressure rather than the modified one. As it turns out, most software codes by default rely on the approximation of neglecting the difference between modified and real pressure for bubbly flows. The purpose of the present study is to assess the influence of this approximation on the final simulations results. Fortunately it turns out that at least for the conditions considered in this study, the error is small.

The paper presents Euler–Euler Large Eddy Simulations of dispersed bubbly flow in a rectangular bubble column at a low Reynolds number. The physical models describing the momentum exchange between the phases including drag, lift and wall force were chosen according to previous experiences of the authors. The emphasis of the study is the analysis of bubbly flows concerning the investigation of the influence of the bubble-induced turbulence model. It is found that the presented modeling combination provides fairly good agreement with experimental data for the mean flow. The impact of the modeling on the liquid velocity fluctuations is investigated and the energy spectrum obtained from the resolved velocity is discussed.

In this part of the paper, we introduce and discuss the numerical investigation of two-phase flow and heat transfer during steam condensation inside an inclined tube. For that we developed and employed a three-dimensional two-phase computational fluid dynamics model in ANSYS CFX which comprises a consistent heat transfer model that considers wall condensation and direct contact condensation. Wall condensation is covered by a new subgrid model for the thin liquid film. For modelling of the heat transfer on the steam-liquid interface, three different heat transfer correlations have been implemented to check their performance. Simulation results were compared with experimental data of the COSMEA facility, which has been documented in part I of this paper. Particular focus was given to the development of the liquid film inside the tube and its effects on the wall heat transfer. Moreover, we compared the CFD simulations with system code simulations performed with the ATHLET submodule of AC².

In this paper, experimental investigations on the flow morphology and heat transfer in a single steam condenser tube are presented, which were performed at the thermal hydraulic test facility COSMEA ( CO nden s ation test rig for flow M orphology and h EA t transfer studies). This facility was setup to study the interrelation of condensation heat transfer with two-phase flow in a single condenser tube that is cooled by forced convection. Studies were performed for elevated pressures up to 65 bar at saturation conditions and for inlet steam mass flow of up to 1 kg/s and different inlet steam qualities. The wall heat flux is measured with distributed heat flux probe and global condensation rates were obtained from integral heat and mass balances. As a unique feature the cross-sectional phase distribution was studied via X-ray computed tomography. The data is going to be used for the validation of numerical simulations with 1D ATHLET and 3D CFD codes as presented in the second part of this paper.

In the present paper a set of closure relations for interphase momentum exchange and for bubble-induced turbulence within the Euler-Euler framework is presented and validated against a set of tests performed at the HZDR-facility MT-Loop. The facility was equipped with wire-mesh sensors that allow cross sectional distributions of gas fraction, gas velocity, and bubble sizes to be measured at different distances from the gas injection. The here applied models were derived for the simulation of fully developed turbulent vertical upward bubbly flow in a pipe. The radial gas fraction profile for this kind of flow is the result of the ratio of the radial force components of the so-called non-drag forces and can be used for model validation. However, only the ratio, not the absolute value of the bubble forces can be tested in this way. In the present paper more detailed information from the experiment is exploited by consideration of the evolution of gas fraction distribution particularly after the gas injection region. The change of cross sectional gas volume fraction distribution is the result of the action of non-drag forces. In addition to vertical pipe flow tests further insight is obtained from the investigation of the effect of a slight tube inclination which shifts the gas distribution. Here the disturbance of the cylindrical symmetry of a vertical pipe gives hints on the absolute value of the non-drag force components. The results show that the presented model framework at least is able to describe the phenomena qualitatively. Possible reasons for quantitative deviations are discussed and require further investigations.

Measuring the local liquid velocity and turbulence in large-scale bubble columns with optical methods is complex and usually limited to low gas holdups in thin geometries. Comprehensive datasets in large bubble columns are therefore seldom published. Since the importance of Computational Fluid Dynamics (CFD) is increasing for multiphase applications, such data is also important for validating models for dispersed bubbly flows. In the present work the liquid velocity and turbulence in a pilot-scale bubble column is studied and a CFD validation database is generated by completing previous measurements of the gas void fractions and bubble sizes. For this purpose, we used a Particle Shadowgraph Velocimetry (PSV) technique that was intentionally designed to study the fluid dynamics in large-scale facilities. The measurements were realized in the 5.3 m high and 0.24 m diameter counter-current bubble column at Politecnico di Milano. The superficial gas velocity ranged from 0.37 to 1.85 cm/s, the counter current superficial liquid velocity from 0 to 9.2 cm/s. All operation points are in the so-called pseudo-homogeneous flow regime, in which the integral gas holdup (ranging from 1.02 to 7.55%) increases linear with the superficial gas velocity but an inhomogeneous flow is present. The dominant frequencies of the bubbly flow, shear rates, and turbulence levels are increasing with increasing superficial gas velocity. With increasing superficial liquid velocities, the dominant frequencies are decreasing, the averaged liquid velocities are shifted downwards, but the overall turbulence levels remain constant. In order to investigate the smaller scales at which the bubble-induced turbulence is expected, a filtering process is proposed. As a result, the filtered turbulence levels of all operation points fall on a linear trend line when plotted over the local void fraction, which is the same result obtained in other studies in small, homogenous tabletop columns. The now available database for CFD validation contains the averaged liquid velocities, basic turbulence information, local void fractions, and the bubble sizes at two different heights. The data will be in particular useful to validate the capabilities of models to upscale bubbly flows from tabletop to the pilot-scale bubble columns.

The two- or multi-fluid approach is frequently used for Nuclear Reactor Safety (NRS)-related simulations of gas-liquid flows. To enable reliable predictions the closure models have to reflect the involved local physical phenomena at the non-resolved scale properly. To consolidate the CFD-modelling in the frame of the multi-fluid approach the so-called baseline model strategy was recently proposed (Lucas et al., 2016). The present technical note discusses a long-term strategy for the baseline model development and ways to obtain or improve closure models. Guidelines for the model development are given by listing requirements for appropriate closure models as well as frequently made mistakes. This is illustrated by examples for recent developments done for HZDR baseline models for poly-disperse bubbly flows.

New closure models for bubble coalescence and breakup proposed in Liao et al. (2011) , Nucl. Eng. Des. 241, 1024, are assessed for the case of condensing steam–water pipe flows. Steady-state CFD calculations are performed employing the commercial CFD solver ANSYS CFX. Predicted evolution of cross-section averaged bubble size and gas volume fraction distribution along the pipe is compared with the measurements provided by the TOPFLOW facility ( Lucas et al., 2010 . CFD4NRS-3, Int. Workshop on Experimental Validation and Application of CFD and CMFD Codes to Nuclear Reactor Safety Issues, Paper 13.1, 14.16.09, Washington D.C., USA.). It is shown that for cases with small initial bubble size and low gas volume fraction, bubble coalescence and breakup can be taken to be nearly negligible and the change of bubble size is primarily due to condensation. Nevertheless, with the increase of initial bubble size or gas volume fraction, bubble coalescence and breakup become more prevalent. Performance of new and standard closure models of bubble coalescence and breakup is investigated It is shown that both models overestimated the breakup rate; in particular, the standard model. The numerical results are also found to be dependent on the inlet liquid temperatures and inter-phase heat transfer models have a significant impact on the results.

A numerical study is presented to examine the effects on countercurrent flow limitation (CCFL) of shape and size of hot leg models with a rectangular cross section. The CCFL was described in terms of Wallis parameters using the channel height H as the characteristic length. Numerical simulations, using the CFD software code FLUENT 6.3.26, were done for the air-water CCFL experiments carried out previously at Helmholtz-Zentrum Dresden-Rossendorf (HZDR) in the 1/3 scale hot leg model with a rectangular channel (HxW = 0.25x0.05 m2), and the results were compared with the air-water CCFL data obtained at Kobe University in the 1/5 scale hot leg model with rectangular cross section (HxW = 0.15x0.01 m2) and the results of simulations. It was found that both the height-to-width ratio and the size of the cross section affected the CCFL characteristics in the Wallis diagram. Comparison of CCFL characteristics in rectangular channels with those in circular channels showed that the hydraulic diameter, Dh, was a major cross section geometry term influencing the CCFL characteristics. CCFL constants of the Wallis correlation were about 0.61 on average for the range of 0.05 m < Dh < 0.75 m but became small for Dh < 0.0254 m, and these tendencies were well reproduced by the numerical simulations.

Small Break Loss of Coolant Accident (SB LOCA) is one of the most severe transients which may lead to Pressurized Thermal Shock (PTS) on the Reactor Pressure Vessel (RPV) wall. During postulated SB LOCA Emergency Core Cooling (ECC) water is injected into the cold leg, where it mixes with the hot coolant. The mixture of cold and hot coolants flows toward the downcomer. Knowledge of transient temperature distribution in the downcomer is necessary to predict thermal gradients in the structural components of the RPV wall. For the prediction of the temperature fields and heat transfer coefficient between the fluid and wall in the cold leg and the downcomer, reliable computational fluid dynamics (CFD) simulations are needed. To validate CFD models for two-phase PTS scenarios numerical simulations of the TOPFLOW-PTS experiments were performed in the framework of the EU NURISP (NUclear Reactor Integrated Simulation Project) project. The paper presents the post-test CFD simulations of a steady-state TOPFLOW-PTS air/water experiment and the pre-test blind simulations of a steady-state TOPFLOW-PTS steam/water case with condensation. CFD simulations were performed with ANSYS FLUENT, ANSYS CFX and NEPTUNE-CFD. The simulations of the air/water test have shown that correct modeling of the ECC jet behavior is essential for the temperature prediction in the cold leg. For modeling these two-phase flows with rather smooth large free surfaces, Reynolds Averaged Navier-Stokes approach seems to be appropriate. The pre-test simulations of steam/water flow predicted a thermal stratification at the entrance of the downcomer. Finally, the simulations of the TOPFLOW-PTS experiments have depicted considerable differences between the codes and the models. © 2013 Elsevier B.V. All rights reserved.

The evolution of steam–water flow with an initial sub-cooling along an 8 m long pipe with an inner diameter of about 200 mm was measured to obtain a database suitable for the qualification of CFD-codes on poly-dispersed flows with phase transfer. Phase transfer has to be considered in Nuclear Reactor Safety Research, e.g. in case of sub-cooled boiling or bubble entrainment caused by Emergency Core Cooling (ECC) injection. Bubble size distributions are important, since the phase transfer rate is proportional to the interfacial area density. To develop and validate closure models for CFD codes experimental data with high resolution in space and time are required. The experiments were conducted at the TOPFLOW facility of the Helmholtz–Zentrum Dresden–Rossendorf. Steam was injected into upwards flowing sub-cooled water via orifices in the pipe wall located at different distances from measuring plane. 1 mm and 4 mm injection orifices are used to vary the initial bubble size distribution. The variation of the distance between the location of the gas injection and the measuring plane allows investigating the evolution of the flow along the pipe. Pressure, steam and water flow rates and the sub-cooling were also varied. Measurements are done using wire-mesh sensors and thermocouples. Data on averaged void fraction, radial gas volume fraction profiles, profiles of the gas velocity and bubble size distributions in dependency of the L/D ratio are available.

We proposed prediction methods for countercurrent flow limitation (CCFL) in horizontal and slightly inclined pipes with one-dimensional (1-D) computations and uncertainty of computed CCFL. In this study, we applied the proposed methods to a full-scale pressurizer surge line [inclination angle theta = 0.6 deg, diameter D = 300 mm, and ratio of the length to the diameter (L/D) = 63] in a specific pressurized water reactor, performed 1-D computations and three-dimensional (3-D) numerical simulations, and found that uncertainties caused by effects of the diameter and fluid properties on CCFL were small. We also applied the proposed methods to experiments for hot-leg and surge line models (theta = 0 and 0.6 deg, D = 0.03 to 0.65 m, and L/D = 4.5 to 63) to generalize them, performed 1-D computations, and found that uncertainties caused by effects of theta and L on CCFL were large due to the setting error for theta and differences among experiments. This shows that a small-scale air-water experiment with the same theta and L/D as those in an actual plant is effective to reduce the uncertainty of CCFL prediction.

Boiling flow inside a wall heated vertical pipe is simulated by a multi-field CFD approach. Sub-cooled water enters the pipe from the lower end and heats up first in the near wall region leading to the generation of small bubbles. Further along the pipe larger and larger bubbles are generated by coalescence and evaporation. This leads to transitions of the two-phase flow patterns from bubbly to churn-turbulent and annular flow. The CFD simulation bases on the recently developed GEneralized TwO Phase flow (GENTOP) concept. It is a multi-field model using the Euler-Euler approach. It allows the consideration of different local flow morphologies including transitions between them. Small steam bubbles are handled as dispersed phases while the interface of large gas structures is statistically resolved. The paper presents the extension of the GENTOP model for phase transfer and discusses the sub-models used. Finally the above mentioned boiling pipe is considered as demonstration case.

The method for predicting countercurrent flow limitation (CCFL) and its uncertainty in an actual pressurizer surge line of a pressurized water reactor (PWR) using 1/10-scale air–water experimental data, one-dimensional (1D) computations, and three-dimensional (3D) numerical simulations was proposed. As one step of the prediction method, 3D numerical simulations were carried out for countercurrent air–water flows in a 1/10-scale model of the pressurizer surge line to evaluate capability of the 3D simulation method and decide uncertainty of CCFL characteristics evaluated for the 1/10-scale model. The model consisted of a vertical pipe, a vertical elbow, and a slightly inclined pipe with elbows. In the actual 1/10-scale experiment, air supplied into the lower tank flowed upward to the upper tank and water supplied into the upper tank gravitationally flowed downward to the lower tank through the pressurizer surge line. In the 3D simulation, however, water was supplied from the wall surface of the vertical pipe to avoid effects of flooding at the upper end (the 3D simulation largely underestimated falling water flow rates at the upper end). Then, the flow pattern in the slightly inclined pipe was successfully reproduced, and the simulated CCFL values for the inclination angle of θ=0.6 deg (slope of 1/100) agreed well with the experimental CCFL data. The uncertainty among air–water experiments, 1D computations, and 3D simulations for the 1/10-scale model was dC=±0.015 for the CCFL constant of C=0.50. The effects of θ (θ=0,1.0 deg) on CCFL characteristics were simulated and discussed.

In reflux condensation, steam generated in the reactor core and water condensed in a steam generator (SG) form countercurrent flow in a hot leg, which consists of a horizontal pipe, an elbow and an inclined pipe. Both countercurrent air–water and steam–water tests were previously carried out at Helmholtz-Zentrum Dresden-Rossendorf (HZDR) using the 1/3 scale rectangular channel simulating a PWR hot leg installed in the pressure chamber of the TOPFLOW facility. In this paper, in order to evaluate the effects of fluid properties, the steam–water CCFL (countercurrent flow limitation) tests at HZDR were simulated using the CFD software, FLUENT 6.3.26. The computational domain included the reactor vessel simulator, hot leg and SG inlet chamber in order to avoid uncertainties of boundary conditions at both ends of the hot leg. The VOF (volume of fluid) method and two-fluid (2F) model were used. In the 2F model, the combination of three correlations on the interfacial friction coefficients, which had been validated for the 1/15 and 1/5 scale tests at Kobe University, was used as a function of local void fractions. The CCFL characteristics predicted by the 2F and VOF agreed relatively well with the steam–water CCFL data at HZDR but overestimated the effects of fluid properties on CCFL. The VOF simulations were better able to reproduce the fluid properties than the 2F simulations.

Numerical simulations have been done to evaluate CCFL (countercurrent flow limitation) in a PWR hot leg under reflux condensation by using a VOF (volume of fluid) method implemented in the CFD software, FLUENT6.3.26. The calculated CCFL characteristics have been verified and agreed well with known values including the UPTF data at 1.5 MPa. Therefore, in this paper, parameter calculations using the VOF method were done for system pressures up to 8 MPa under PWR full-scale conditions with the diameter of 750 mm. As a result, calculated CCFL characteristics in the Wallis diagram were slightly mitigated from 0.1 MPa to 1.5 MPa with increasing system pressure, but they did not change from 1.5 MPa to 8 MPa. The CCFL database calculated in this study and values measured under air-water and steam-water conditions were used to derive a CCFL correlation and its uncertainty, where the CCFL constant was . Most of the CCFL data and the current correlation predictions were within the uncertainty of +-0.03.

In order to investigate the two-phase flow behaviour in a complex reactor-typical geometry and to supply suitable data for computational fluid dynamics (CFD) code validation, a model of the hot leg of a pressurised water reactor was built at Helmholtz-Zentrum Dresden Rossendorf (HZDR). The hot leg model is devoted to optical measurement techniques, therefore, a flat test section design was chosen and equipped with large windows. In order to enable the operation at high pressures, the test section is installed in the pressure chamber of the TOPFLOW ( T ransient tw O P hase FLOW ) test facility of HZDR, which is used to perform the experiments under pressure equilibrium with the inside atmosphere. Counter-current flow limitation (CCFL) experiments were performed, simulating the reflux-condenser cooling mode appearing in small break loss-of-coolant-accident (LOCA) scenarios. The fluids used were air and water at room temperature and pressures of up to 3.0 bar, as well as steam and water at pressures of up to 50 bar and the corresponding saturation temperature of 264 °C. One selected 50 bar experiment is presented in detail: the observed behaviour is analysed and illustrated by typical high-speed camera images of the flow. Furthermore, the flooding characteristics obtained from the different experimental runs are presented in terms of the Wallis parameter and Kutateladze number, which are commonly used in the literature. However, a discrepancy was first observed between the air/water and steam/water series. Further investigations show that the steam was probably wet due to heat losses and liquid entrainment from the heater circuit. Consequently, a correction of the steam measurements was required. The amount of parasitic water was evaluated indirectly over the shift of the zero liquid penetration noticed in the CCFL diagram. Finally, the experimental results confirm that the Wallis similarity is appropriate to scale flooding in the hot leg of a pressurised water reactor over a wide range of pressure and temperature conditions.

Within the current study, experimental investigations and computational fluid dynamics (CFD) simulations were performed to investigate the flow field structure developed under a turbulent liquid jet plunging into a quiescent pool of water. This topic is still challenging for CFD codes. Indeed, the study of turbulence in two-phase bubbly flows is one domain where experimental, numerical, and theoretical work is being extensively done nowadays. A correct description of closure laws for drag, lift, and interfacial forces is of great importance in numerical simulations. Most critical with respect to CFD is the impact region between the jet and free surface of the liquid pool. Here, a complex interaction between surface waves and turbulence leads to the entrainment of air. These phenomena occur on very small scales. Up to now, it is not possible to resolve all relevant scales in one simulation due to limited computational resources. Therefore in this work, all phenomena above the pool surface and the impact region are excluded and the focus is set on the development of the flow field below the pool surface. The jet is modeled as a two-phase bubbly flow injecting into the pool. For this purpose, the Particle Image Velocimetry (PIV) was utilized as measuring technique. Velocity fields for both impinging region and recirculation zone developed in the tank below the free surface were quantified and instantaneous and time-averaged velocity fields were obtained. For test cases where air entrainment occurred, only the recirculation region situated outside the bubble plume was quantified. On the other hand, the numerical simulations were performed using ANSYS-CFX 12.0, a commercial CFD package that solves the Navier–Stokes equations via a finite volume method and a coupled solver. The 2D as well as 3D simulation results are presented and compared with experimental results. Comparisons with the experimental data reveal satisfactory predictions of mean flow quantities, obtained by applying proper models of inter-phase momentum transfer.

Research into gas–liquid countercurrent two-phase flow in a model of pressurized water reactor (PWR) hot leg has been carried out over the last several decades. An extensive experimental data base has been accumulated from these studies, leading to the development of phenomenological correlations and scaling parameters of the countercurrent flow limitation (CCFL). However, most of the proposed correlations apply under a relatively narrow range of conditions, generally limited to the test section conditions and/or geometry. Moreover the development of mechanistic models based on the underlying physical processes has been limited. In contrast to this mechanistic form of modelling, the implementation of computational fluid dynamics (CFD) techniques has also been pursued, but the considerable robust three-dimensional (3D) closure relations for this application remain an unachieved goal due to lack of detailed phenomenological knowledge and consequent application of empirical one-dimensional experimental correlations to the multidimensional problem. This paper presents a comprehensive review of research work on countercurrent gas–liquid two-phase flow in a PWR hot leg and provides direction regarding future research on this topic. In the introductory section, the problems facing current research are described. In the following sections, recent experimental as well as theoretical research achievements are overviewed. In the last section, the problems that remain unsolved are discussed, along with some concluding remarks. It was found that only limited theoretical developments exist in the literature, however highly reliable experimental data are needed to support this effort. Additional work, both analytical and experimental, needs to be carried out on the effects of mass transfer on countercurrent flow limitation to improve the existing correlations and analytical models.

This paper presents different CFD-simulations on flows which are relevant for nuclear reactor safety using a new modeling approach for the interfacial drag at free surfaces. The developed drag coefficient model was implemented together with the Algebraic Interfacial Area Density (AIAD) model ( Hohne, 2009 ) into the three-dimensional (3-D) computational fluid dynamics (CFD) code ANSYS-CFX. The applications considered include the prediction of counter-current flow limitations (CCFL) in a PWR hot leg, the development of hydraulic jump during the air–water co-current flow in a horizontal channel, and pressurized thermal shock (PTS) phenomena in a PWR cold leg and downcomer. For the modeling of these tasks, an Euler–Euler approach was used. This approach allows the use of different models depending on the local morphology. In the frame of an Euler–Euler simulation, the local morphology of the phases has to be considered in the drag model. To demonstrate the feasibility of the present approach, the computed main parameters of each case were compared with experimental data. It is shown that the CFD calculations agree well with the experimental data. This indicates that the AIAD model combined with new drag force modeling is a promising way to simulate the phenomena in frame of the Euler–Euler approach. Moreover the further validation of the model by including mass transfer effects should be carried out.

In order to improve the understanding of counter-current two-phase flow and to validate new physical models, CFD simulations of a 1/3rd scale model of the hot leg of a German Konvoi pressurized water reactor (PWR) with rectangular cross section were performed. Selected counter-current flow limitation (CCFL) experiments conducted at Helmholtz-Zentrum Dresden-Rossendorf (HZDR) were calculated with ANSYS CFX using the multi-fluid Euler–Euler modelling approach. The transient calculations were carried out using a gas/liquid inhomogeneous multiphase flow model coupled with a shear stress transport (SST) turbulence model. In the simulation, the drag law was approached by a newly developed correlation of the drag coefficient ( Hohne and Vallee, 2010 ) in the Algebraic Interfacial Area Density (AIAD) model. The model can distinguish the bubbles, droplets and the free surface using the local liquid phase volume fraction value. A comparison with the high-speed video observations shows a good qualitative agreement. The results indicate also a quantitative agreement between calculations and experimental data for the CCFL characteristics and the water level inside the hot leg channel.

Bubble condensation plays an important role, e.g. in sub-cooled boiling or steam injection into pools. Since the condensation rate is proportional to the interfacial area density, bubble size distributions have to be considered in an adequate modeling of the condensation process. The effect of bubble sizes was clearly shown in experimental investigations done previously at the TOPFLOW facility of FZD. Steam bubbles were injected into a sub-cooled upward pipe flow via orifices in the pipe wall located at different distances from measuring plane. 1 mm and 4 mm injection orifices were used to vary the initial bubble size distribution. Measurements were done using a wire-mesh sensor. Condensation is clearly faster in case of the injection via the smaller orifices, i.e. in case of smaller bubble sizes. Recently the Inhomogeneous MUSIG model was implemented into the CFD code CFX from ANSYS enabling the simulation of poly-dispersed flows including the effects of separation of small and large bubbles due to bubble size dependent lift force inversion. It allows to divide the dispersed phase into size classes regarding the mass as well as regarding the momentum balance. Up to now transfers between the classes in the mass balance can be considered only by bubble coalescence and breakup (population balance). Here an extension of the model is proposed to include the effects due to phase transfer. The paper focuses on the derivation of equations for the extension of the Inhomogeneous MUSIG model and presents some first results for verification and validation.

The current paper presents the prediction results of a bubbly flow under plunging jet conditions using multiphase mono- and poly-dispersed approaches. The models consider interfacial momentum transfer terms arising from drag, lift, and turbulent dispersion force for the different bubble sizes. The turbulence is modeled by an extended k – ɛ model which accounts for bubble induced turbulence. Furthermore in case of a poly-dispersed air–water flow the bubble size distribution, bubble break-up and coalescence processes as well as different gas velocities in dependency on the bubble diameter are taken into account using the Inhomogeneous MUSIG model. This model is a generalized inhomogeneous multiple size group model based on the Eulerian modeling framework which was developed in the framework of a cooperative work between ANSYS-CFX and Forschungszentrum Dresden-Rossendorf (FZD). The latter is now implemented into the CFD code CFX. According to the correlation on the lateral lift force obtained by Tomiyama (1998) ; this force changes its sign in dependence on the bubble size. Consequently the entrained small bubbles are trapped below the jet. They can escape from the bubble plume only by turbulent fluctuations or by coalescence. If the size of the bubbles generated by coalescence exceeds the size at which the lift force changes its sign these large bubbles go out from the plume and rise to the surface. A turbulent model based on an additional source term for turbulence kinetic energy and turbulence eddy dissipation equation is compared to the common concept for modeling the turbulence quantities proposed by Sato et al. (1981) . It has been found that the large bubble distribution is slightly affected by the turbulence modeling which affects particularly the bubble coalescence and break-up process.

Bubble condensation in sub-cooled water is a complex process, to which various phenomena contribute. Since the condensation rate depends on the interfacial area density, bubble size distribution changes caused by breakup and coalescence play a crucial role. Experiments on steam bubble condensation in vertical co-current steam/water flows have been carried out in an 8 m long vertical DN200 pipe. Steam is injected into the pipe and the development of the bubbly flow is measured at different distances to the injection using a pair of wire mesh sensors. By varying the steam nozzle diameter the initial bubble size can be influenced. Larger bubbles come along with a lower interfacial area density and therefore condensate slower. Steam pressures between 1 and 6.5 MPa and sub-cooling temperatures from 2 to 12 K were applied. Due to the pressure drop along the pipe, the saturation temperature falls towards the upper pipe end. This affects the sub-cooling temperature and can even cause re-evaporation in the upper part of the test section. The experimental configurations are simulated with the CFD code CFX using an extended MUSIG approach, which includes the bubble shrinking or growth due to condensation or re-evaporation. The development of the vapour phase along the pipe with respect to vapour void fractions and bubble sizes is qualitatively well reproduced in the simulations. For a better quantitative reproduction, reliable models for the heat transfer at high Reynolds number as well as for bubble breakup and coalescence are needed.

In the present study we propose new coalescence and breakup closures for the inhomogeneous MUltiple bubble Size Group (MUSIG) model. The major purpose is to consider bubble coalescence and breakup due to different mechanisms in turbulent gas–liquid mixtures and to develop a generally applicable constitutive model for CFD applications. For bubble coalescence the new model includes coalescence due to turbulence, laminar shear, wake-entrainment and eddy-capture. Bubble breakup mechanisms encompass turbulent fluctuation, laminar shear and interfacial slip velocity. The new model was implemented in the commercial CFD software (CFX-12.0) and applied to the case of turbulent air–water mixtures in a large vertical pipe (DN 200). Simulation results for the evolution of bubble size distributions were compared to the standard closure model of Luo and Svendsen (1996) and Prince and Blanch (1990) as well as the experimental data of TOPFLOW test facility. Better prediction is accomplished by the proposed new model.

A population balance model (the Inhomogeneous MUSIG model) has recently been developed in close cooperation between ANSYS-CFX and Forschungszentrum Dresden-Rossendorf and implemented into the CFD-Code CFX [Krepper, E., Lucas, D., Prasser, H.-M, 2005. On the modelling of bubbly flow in vertical pipes. Nucl. Eng. Des. 235, 597–611; Frank, T., Zwart, P.J., Shi, J.-M., Krepper, E., Rohde, U., 2005. Inhomogeneous MUSIG Model—a population balance approach for polydispersed bubbly flows, International Conference “Nuclear Energy for New Europe 2005”, Bled, Slovenia, September 5–8, 2005; Krepper, E., Beyer, M., Frank, Th., Lucas, D., Prasser, H.-M., 2007. Application of a population balance approach for polydispersed bubbly flows, 6th Int. Conf. on Multiphase Flow Leipzig 2007, (paper 378)]. The current paper presents a brief description of the model principles. The capabilities of this model are discussed via the example of a bubbly flow around a half-moon shaped obstacle arranged in a 200 mm pipe. In applying the Inhomogeneous MUSIG approach, a deeper understanding of the flow structures is possible and the model allows effects of polydispersion to be investigated. For the complex flow around the obstacle, the general structure of the flow was well reproduced in the simulations. This test case demonstrates the complicated interplay between size dependent bubble migration and the effects of bubble coalescence and breakup on real flows. The closure models that characterize the bubble forces responsible for the simulation of bubble migration show agreement with the experimental observations. However, clear deviations occur for bubble coalescence and fragmentation. The models applied here, which describe bubble fragmentation and coalescence could be proved as a weakness in the validity of numerous CFD analyses of vertical upward two-phase pipe flow. Further work on this topic is under way.

A 1D test-solver was developed in recent years for modeling of two phase bubbly flows in pipe geometry. The solver considers a number of bubble classes and calculates bubble-size resolved void fraction profiles in the radial direction. A successful implementation was achieved regarding bubble forces models (non-drag forces). Discrepancies appeared when coalescence and breakup rates were significant. These rates depend upon local turbulence quantities, which are possible reason for discrepancies. Originally the test-solver is equipped by Sato model (Sato, Y., Sadatomi, M., Sekoguchi, K., 1981. Momentum and heat transfer in two-phase bubble flow. I. International Journal Multiphase Flow 7, 167–177 .) which accounts for turbulence via shear- and bubble-induced viscosities calculated out of empirical correlations. One equation for the turbulent kinetic energy was solved, while the dissipation rate was calculated out of a correlation. In order to improve calculation of the local turbulence parameters, a two-phase k – ɛ turbulence model was adopted instead. The account for the bubble-induced turbulence was made via a source term taken out of literature. Comparisons between new and old turbulence modeling against experimental data showed better agreement for the new model. The experiments covered a wide range of water and air superficial velocities for upward bubbly flow in two pipe's diameters: 50 and 200 mm. The main feature of the new model is providing more reliable values of turbulence parameters for application in coalescence and breakup models. A comparison with CFX 5.7 calculations in a 50 mm pipe showed better calculation results when the source term was considered in the k – ɛ equations. An implementation into CFX is planned.

The NURESIM Integrated Project of the 6th European Framework Program was envisaged to initiate the development of the next-generation common European Standard Software Platform for simulating nuclear reactors. The overall objective of NURESIM Thermalhydraulic sub-project is to improve the understanding and the predictive capabilities of the simulation tools for key two-phase flow thermal-hydraulic processes such as the Critical Heat Flux (CHF). A multi-scale analysis of reactor thermalhydraulics is envisaged and two-phase CFD is developed to allow some zoom on local processes when the resolution of system codes is not sufficient. Current industrial methods for CHF mainly use the sub-channel analysis and empirical CHF correlations based on large scale experiments having the real geometry of the reactor assembly. The NURESIM-TH activities regarding CHF aim at using two-phase CFD as a tool for understanding boiling flow processes, in order to subsequently help new fuel assembly design and to develop better CHF predictions in both PWR and BWR. A “Local Predictive Approach” may be envisaged for the long term where CHF empirical correlations would be based on local T/H parameters provided by CFD. This paper presents a review of existing experimental data bases which can be used for validation of the two-phase CFD application to Critical Heat Flux (CHF) investigations with respect to nuclear reactors. The phenomenology of DNB and Dry-Out are detailed identifying all basic flow processes which require a specific modeling in CFD tool. The resulting program of work is given and the current state of the art of the modeling is presented.

This contribution presents different approaches for the modeling of gas entrainment under water by a plunging jet. Since the generation of bubbles happens on a scale which is smaller than the bubbles, this process cannot be resolved in meso-scale simulations, which include the full length of the jet and its environment. This is why the gas entrainment has to be modeled in meso-scale simulations. In the frame of a Euler-Euler simulation, the local morphology of the phases has to be considered in the drag model. For example, the gas is a continuous phase above the water level but bubbly below the water level. Various drag models are tested and their influence on the gas void fraction below the water level is discussed. The algebraic interface area density (AIAD) model applies a drag coefficient for bubbles and a different drag coefficient for the free surface. If the AIAD model is used for the simulation of impinging jets, the gas entrainment depends on the free parameters included in this model. The calculated gas entrainment can be adapted via these parameters. Therefore, an advanced AIAD approach could be used in future for the implementation of models (e.g., correlations) for the gas entrainment.

A generalized inhomogeneous multiple size group (MUSIG) model based on the Eulerian modeling framework was developed in close cooperation of ANSYS-CFX and Forschungszentrum Dresden-Rossendorf and implemented into the CFD code CFX. The model enables the subdivision of the dispersed phase into a number of size groups regarding the mass balance as well as regarding the momentum balance. In this work, the special case of polydispersed bubbly flow is considered. By simulating such flows, the mass exchanged between bubble size classes by bubble coalescence and bubble fragmentation, as well as the momentum transfer between the bubbles and the surrounding liquid due to bubble size dependent interfacial forces have to be considered. Particularly the lift force has been proven to play an important role in establishing a certain bubble size distribution dependent flow regime. In a previous study [Krepper, E., Lucas, D., Prasser, H.-M., 2005. On the modeling of bubbly flow in vertical pipes. Nucl. Eng. Des. 235, 597–611] the application of such effects were considered and justified and a general outline of such a model concept was given. In this paper the model and its validation for several vertical pipe flow situations is presented. The experimental data were obtained from the TOPFLOW test facility at the Forschungszentrum Dresden-Rossendorf (FZD). The wire-mesh technology measuring local gas volume fractions, bubble size distributions and velocities of gas and liquid phases were employed. The inhomogeneous MUSIG model approach was shown as capable of describing bubbly flows with higher gas content. Particularly the separation phenomenon of small and large bubbles is well described. This separation has been proven as a key phenomenon in the establishment of the corresponding flow regime. Weaknesses in this approach can be attributed to the characterization of bubble coalescence and bubble fragmentation, which must be further investigated.

Many flow regimes in nuclear reactor safety (NRS) research and nuclear reactor engineering are characterized by water–steam two-phase flows and require for their accurate CFD simulation the development and validation of appropriate multiphase flow models. This paper is presenting multiphase flow models for the description of mono- and polydisperse bubbly flows in the Eulerian framework of multiphase flow modeling in ANSYS CFX. The models are taking into account interphase momentum transfer due to governing drag and non-drag forces. Furthermore in case of a polydisperse air–water or water–steam bubbly flow the bubble size distribution, bubble break-up and coalescence processes as well as different gas velocities in dependency on the bubble diameter are taken into account in the newly developed inhomogeneous MUSIG model. Both multiphase flow models for mono- and polydisperse bubbly flows have thoroughly been validated and compared to experiments of MT-Loop and TOPFLOW test facilities (FZ Dresden-Rossendorf, Germany). Best practice guidelines (Menter, 2002) have been applied in order to allow for a systematic error quantification and thoroughly assessment of model formulations.

The influence of the grid size on the rise velocity of a single bubble simulated with an Eulerian two-fluid method is investigated. This study is part of the development of an elaborated Eulerian two-fluid framework, which is able to predict complex flow phenomena as arising in nuclear reactor safety research issues. Such flow phenomena cover a wide range of interfacial length scales. An important aspect of the simulation method is the distinction into small flow structures, which are modeled, and large structures, which are resolved. To investigate the requirements on the numerical grid for the simulation of such resolved structures the velocity of rising gas bubbles is a good example since theoretical values are available. It is well known that the rise velocity of resolved bubbles is clearly underestimated in a one-fluid approach if they span over only few numerical cells. In the present paper it is shown that in the case of the two-fluid model the bubble rise velocity depends only slightly on the grid size. This is explained with the use of models for the gas–liquid interfacial forces. Good approximations of the rise velocity and the bubble shape are obtained with only few grid points per bubble diameter. This result justifies the resolved treatment of flow structures, which cover only few grid cells. Thus, a limit for the distinction into resolved and modeled structures in the two-fluid context may be established.

Wire mesh sensors (WMS) are state of the art devices that allow high resolution (in space and time) measurement of 2D void fraction distribution over a wide range of two-phase flow regimes, from bubbly to annular. Data using WMS have been recorded at the Helmholtz-Zentrum Dresden-Rossendorf (HZDR) (Lucas et al., 2010; Beyer et al., 2008; Prasser et al., 2003) for a wide combination of superficial gas and liquid velocities, providing an excellent database for advances in two-phase flow modeling. In two-phase flow, the interfacial area plays an integral role in coupling the mass, momentum and energy transport equations of the liquid and gas phase. While current models used in best-estimate thermal-hydraulic codes (e.g. RELAP5, TRACE, TRACG, etc.) are still based on algebraic correlations for the estimation of the interfacial area in different flow regimes, interfacial area transport equations (IATE) have been proposed to predict the dynamic propagation in space and time of interfacial area (Ishii and Hibiki, 2010). IATE models are still under development and the HZDR WMS experimental data provide an excellent basis for the validation and further advance of these models. The current paper is focused on the uncertainty analysis of algorithms used to reconstruct interfacial area densities from the void-fraction voxel data measured using WMS and their application towards validation efforts of two-group IATE models. In previous research efforts, a surface triangularization algorithm has been developed in order to estimate the surface area of individual bubbles recorded with the WMS, and estimate the interfacial area in the given flow condition. In the present paper, synthetically generated bubbles are used to assess the algorithm’s accuracy. As the interfacial area of the synthetic bubbles are defined by user inputs, the error introduced by the algorithm can be quantitatively obtained. The accuracy of interfacial area measurements is characterized for different bubbles sizes and shapes, and for different WMS acquisition frequencies. It is found that while convex shapes are successfully analyzed by the reconstruction algorithm, some difficulties are faced with bubbles presenting internal cavities. Utilizing the experimental HZDR database, the performance of the current two-group IATE model is evaluated. While the qualitative propagation of interfacial area is predicted sufficiently well, there is a discrepancy in magnitude between the model’s prediction and the experimental results. Overall, the study suggests that differences exist in the incidence of interaction mechanisms between small and large diameter pipes and further efforts are needed in order to extend the range of validity of current IATE models.

The evolution of the structure of a gas–liquid flow in a large vertical pipe of 195 mm inner diameter was investigated at the TOPFLOW test facility in Rossendorf. Wire-mesh sensors were used to measure sequences of two-dimensional distributions of local instantaneous gas fraction within the complete pipe cross-section. The sensors own a resolution of 3 mm at a frequency of 2500 Hz. Superficial velocities were varied in a range covering flow regimes from bubbly to churn-turbulent flow. The distance between the gas injection and the sensor position was changed using a so-called variable gas injection system. It consists of six gas injection units, each equipped with three rings of injection orifices in the pipe wall (orifice diameter: 1 and 4 mm), which are fed from ring chambers. The gas flow towards these distributor chambers is individually controlled by valves. Measured bubble-size resolved radial gas fraction profiles reveal differences in the lateral migration of bubbles of different size starting from the injection at the wall. The evolution of bubble-size distributions allows to study bubble coalescence and break-up. The influence of the physical properties of the fluid was studied by comparing cold air–water experiments with steam–water tests at 65 bar.