[en] Highlights: • Modelling of boiling sodium flows in a multiphase flow solver. • Rod heated with a constant heat flux in a pipe liquid metal flow. • Sodium boiling flow around a rod heated with a constant heat. • Computations in progress in an assembly constituted of 19 pins equipped with a wrapped wire. - Abstract: In France, Sodium-cooled Fast Reactors (SFR) have recently received a renewed interest. In 2006, the decision was taken by the French Government to initiate research in order to build a first Generation IV prototype (called ASTRID) by 2020. The improvement in the safety of SFR is one of the key points in their conception. Accidental sequences may lead to a significant increase of reactivity. This is for instance the case when the sodium coolant is boiling within the fissile zone. As a consequence, incipient boiling superheat of sodium is an important parameter, as it can influence boiling process which may appear during some postulated accidents as the unexpected loss of flow (ULOF). The problem is that despite the reduction in core power, when boiling conditions are reached, the flow decreases progressively and vapour expands into the heating zone. A crucial investigating way is to optimize the design of the fissile assemblies of the core in order to lead to stable boiling during a ULOF accident, without voiding of the fissile zone. Moreover, in order to evaluate nuclear plant design and safety, a CFD tool has been developed at EDF in the framework of the nuclear industry. Advanced models dedicated to boiling flows have been implemented and validated against experimental data for ten years now including a wall law for boiling flows, wall transfer for nucleate boiling, turbulence and polydispersion model. This paper aims at evaluating the generalization of these models to SFR. At least two main issues are encountered. Firstly, at low Prandtl numbers such as those of liquid metal, classical approaches derived for unity or close to unity fail to accurately predict the heat transfer. In order to evaluate the wall law implemented in the CFD tool, computations have been compared with KALLA experimental results obtained in the case of a rod heated with a constant heat flux which is concentrically embedded in a pipe liquid metal flow (single-phase flow). Secondly, the incipient boiling superheat of sodium is quite different from that of conventional fluids. As a consequence, the nucleate boiling model has been improved and validated against the Charlety’s experiment where a rod heated with a constant heat flux is concentrically embedded in a pipe sodium flow. For different values of the heat flux, the pressure is measured at different locations as function of the mass flow rate. A reasonable agreement has been reached which is very encouraging for further applications. Finally, preliminary computations have been carried out in an assembly constituted of 19 pins equipped with a wrapped wire where partial experimental results are available. Computations have shown a pressure drop at the end of the heated length due to the sudden increase of the hydraulic diameter. Thus, the pressure can drop below the vapour pressure leading to liquid vaporization. This first result supports the assumption of boiling in the upper subassembly zone which could possibly lead to a sodium boiling stabilization.

[en] Highlights: • A new mechanistic model dedicated to DNB has been implemented in the Neptune-CFD code. • The model has been validated against 150 tests. • Neptune-CFD code is a CFD tool dedicated to boiling flows. - Abstract: Extensive efforts have been made in the last five decades to evaluate the boiling heat transfer coefficient and the critical heat flux in particular. Boiling crisis remains a major limiting phenomenon for the analysis of operation and safety of both nuclear reactors and conventional thermal power systems. As a consequence, models dedicated to boiling flows have being improved. For example, Reynolds Stress Transport Model, polydispersion and two-phase flow wall law have been recently implemented. In a previous work, we have evaluated computational fluid dynamics results against single-phase liquid water tests equipped with a mixing vane and against two-phase boiling cases. The objective of this paper is to propose a new mechanistic model in a computational multi-fluid dynamics tool leading to wall temperature excursion and onset of boiling crisis. Critical heat flux is calculated against 150 tests and the mean relative error between calculations and experimental values is equal to 8.3%. The model tested covers a large physics scope in terms of mass flux, pressure, quality and channel diameter. Water and R12 refrigerant fluid are considered. Furthermore, it was found that the sensitivity to the grid refinement was acceptable.

[en] This paper deals with the modelling and the numerical simulation of cavitation phenomena which are of relevant interest in many industrial applications, especially regarding pipes and valves maintenance in nuclear power plant where cavitating flows are responsible for harmful acoustics effects. The cavitation phenomena are calculated by means of a compressible, unsteady, turbulent 3D two-phase flow solver, developed jointly by EDF R and D, CEA, IRSN and AREVA and called NEPTUNE-CFD. The numerical approach is based on a finite volume co-located cell-centred approach and makes use of an original pressure-based multi-field coupling algorithm (Mechitoua, 2003). The cavitation nuclei are pre-existing in the flow. Vapor bubbles generated are advected by the flow and expand in the regions where the local pressure is below the saturation pressure with a tendency to agglomerate into slug bubbles. The interfacial mass transfer is derived to the droplets evaporation theory in drop flow regions and to the Rayleigh-Plesset equation in bubbly flow regions. Results are validated against experimental data of pressure profiles and void fraction visualizations obtained downstream an orifice with the EPOCA facility (EDF R and D). (author)

[en] Condensation heat transfer in the presence of non-condensable gases is a relevant phenomenon in many industrial applications. The present work is focused on the condensation heat transfer that plays a dominant role in many accident scenarios postulated to occur in the containment of nuclear reactors. The aim of the study is to contribute to the understanding of the heat and mass transfer mechanisms involved in the problem. The modelling proposed in the paper assumes that liquid droplets form along the wall at nucleation sites. Vapor condensation on droplets makes them to grow. Once the droplet diameter reaches a critical value, gravitational forces compensate surface tension force and then droplets slide over the wall. Droplets can also join the surrounding droplets and form a film layer. As a consequence of the modelling adopted in the paper, the starting point is the balance of heat and mass transfer between droplets and the gas mixture surrounding the droplet. So, the flow in the simulation domain is modelled as a two-phase flow. This approach allows taking into account simultaneously heat and mass transfer on droplets in the core of the flow and condensation or evaporation phenomena at the wall. Two tests were performed to validate the condensation model against experimental data: the COPAIN experiment (CEA Grenoble) and the TOSQAN ISP47 experiment (IRSN Saclay). Calculated profiles compare favourably with experimental results particularly for the helium and steam volume fraction. Nevertheless the cross-comparison of the gas velocities profiles should be improved in plume-jet configuration. Hence more investigations are needed in turbulence modelling for accurate predictions of heat transfer in the whole containment. (author)

[en] A two-phase CFD modelling approach of the Pressurized Thermal Shock (PTS) problem has been developed and is being validated in the context of PWR life time safety studies. The cold water injection results in strong condensation and complex 3D two-phase phenomena. Direct Contact Condensation (DCC) occurs on the jet and on the free surface of the stratified flow in the leg. These surfaces are much larger than the cells size used in the computational domain, in this sense they can be called large interfaces. DCC depends strongly on the liquid side heat transfer, which is modelled as a function of turbulence, which itself depends on momentum exchange between gas and liquid. A statistical model is used to represent turbulence in each phase. The large interfaces require a special modelling. It has been recently developed and implemented in the NEPTUNE-CFD code which is based on an Eulerian two-fluid model. The present status of this large interface modelling is presented. The validation relies on separate effects experiments such as air-water or steam-water stratified flows and on a more integral experiment, COSI, which represents a cold leg scaled 1/100 for volume and power from a PWR under SBLOCA conditions. The interest of the considered experimental data for PTS CFD is discussed. (authors)

[en] This paper deals with the modeling and the numerical simulation of cavitation phenomena. The cavitation nuclei come from wall nucleation or are pre-existing in the flow. Vapor bubbles generated are carried by the flow and expand in the regions where the local pressure is below the saturation with a tendency to agglomerate into slug bubbles. Compressible, unsteady, turbulent 3D two-phase flow is computed by the NEPTUNE CFD solver, developed jointly by EDF R and D and CEA. The numerical approach is based on a finite volume co-located cell-centered approach and makes use of an original pressure-based multi-field coupling algorithm. The model predictions compared with experimental data on enough selective local variables showed that satisfactory agreement could be obtained without any floating parameter to adjust the data. (authors)

[en] Based on experience gained at EDF and CEA, a more general and robust 3D multiphase flow solver is being currently developed for over three years. This solver, based on an elliptic oriented fractional step approach, is able to simulate multicomponent/multiphase flows. Discretization follows a 3D full unstructured finite volume approach, with a collocated arrangement of all variables. The non linear behaviour between pressure and volume fractions and a symmetric treatment of all fields are taken into account in the iterative procedure, within the time step. It greatly enforces the realizability of volume fractions (i.e 0<□<1), without artificial numerical needs. Applications to widespread test cases are shown to assess the accuracy and the robustness of the flow solver in different flow conditions, encountered in nuclear reactor pipes

[en] NEPTUNE-CFD is a code based on a 3D transient Eulerian two-fluid model. It is developed within the framework of the NEPTUNE project, financially supported by CEA (Commissariat a l'energie Atomique), EDF, IRSN (Institut de Radioprotection et de Surete Nucleaire) and AREVA-NP. NEPTUNE-CFD is mainly devoted to Nuclear Reactor Safety (NRS) issues. One of the main application targets is the two-phase Pressurized Thermal Shock (PTS), which is related to PWR reactor pressure vessel lifetime safety studies, when sub-cooled water from Emergency Core Cooling (ECC) system is injected into the uncovered cold leg, and penetrates into the RPV downcomer. Following the NEA/CSNI Best Practice Guidelines (BPGs), relevant PTS-scenarios have been identified; a Phenomena Identification and Ranking Table (PIRT) process, the related state of the art of modeling and the existing data basis have been reviewed by a panel of European experts, mainly within the ECORA and NURESIM projects. Consistently, the following five experiments were selected for the NEPTUNE_CFD validation presented in this paper. The first four are useful for separate effects validation. The Fabre et al., 1987, experiment is a co-current smooth and wavy Air Water Stratified (AWST) flow in a rectangular channel with detailed measurements of turbulence and velocities. It allows to validate the dynamic models (turbulence and interfacial friction). The Lim et al., 1984, experiment is a co-current smooth and wavy Steam Water Stratified (SWST) flow in a rectangular channel with measurements of the steam flow rates at six axial positions along the channel. It allows to validate the condensation models. The Bonetto and Lahey, 1993, and the Iguchi et al., 1998, experiments deal with a water jet impingement on a water pool free surface in air environment. In the first one, the void fraction and the mean velocities are measured whereas in the second one, mean and rms velocities are measured. Both allow to validate the dynamic models in the situation of a jet impinging a pool free surface - a challenging case for two-phase CFD - the first one mainly versus gas entrainment phenomena and the second one mainly versus turbulence. Finally, the COSI experiment represents a cold leg scaled 1/100 for volume and power from a 900 MW PWR under LOCA conditions, and therefore can be used for global validation. The measurements include condensation rates and temperature profiles at eight axial positions in the pipe, at various ECC flow rates, inlet steam flow rates and water level in the cold leg. It allows to validate all the models involved in a PTS. The five experiments were calculated with NEPTUNE-CFD 1.0.8 with the same set of models. It includes the Large Interface Method (LIM) and a RANS approach with (k-ε) transport equations in each phase. The available measurement uncertainties are generally smaller than typical calculation / measurement discrepancies. Unfortunately there are often lacks in the available experimental data which stress the need for new ones such as the on-going TOPFLOW-PTS. Following the BPGs, the mesh sensitivity is investigated. The five experiments all deal of course with free surfaces. In this case, the BPGs concede that it is not possible to obtain completely grid-independent results and this is actually what we found. However, some calculations show that the LIM transfer models at the free surface, which are written under the format of wall-functions, allow to better master some mesh size effects, confirming the adequacy of this modeling approach for the industrial application. (authors)

[en] This paper focuses on the modeling and the numerical simulation with the NEPTUNE-CFD code of cavitation phenomena and boiling bubbly flows. Compressible, unsteady, turbulent 3D two-phase flow is computed by the NEPTUNE-CFD solver, developed jointly by EDF R and D and CEA. The numerical approach is based on a finite volume co-located cell-centered approach and makes use of an original pressure-based multi-field coupling algorithm [15]. The cavitation nuclei come from wall nucleation or are pre-existing in the flow. Generated vapor bubbles are advected by the flow and expand in the regions where the local pressure is below the saturation with a tendency to agglomerate into slug bubbles. The model predictions compared with experimental data on enough selective local variables showed that satisfactory agreement could be obtained without any floating parameter to fit the data. After cavitation flows, the second part of the paper deals with boiling bubbly flow through a mixing device representing the effect of a fuel assembly spacer grid equipped with mixing blades (DEBORA-mixing experiment, CEA, Grenoble). Local measurements of the void fraction are provided downstream the mixing enhancer. The computations compare favourably with the experimental results, in particular the global effect of the mixing blades was observed. A modification of the classical nucleate boiling model is proposed to overcome the strong model sensitivity with respect to near wall grid refinement. (authors)

[en] The NEPTUNE-CFD code, which is based on an Eulerian two-fluid model, is developed within the framework of the NEPTUNE project, financially supported by CEA (Commissariat a l'Energie Atomique), EDF, IRSN (Institut de Radioprotection et de Surete Nucleaire) and AREVA-NP. NEPTUNE-CFD is mainly focused on Nuclear Reactor Safety applications involving two-phase flows, like two-phase Pressurized Thermal Shock (PTS) and Departure from Nucleate Boiling (DNB). Since the maturity of two-phase CFD has not reached yet the same level as single phase CFD, an important work of model development and thorough validation is needed, as stated for example in NEA/CSNI Writing Group dedicated to the 'Extension of CFD Codes to Two-Phase Flow Safety Problems' (draft6c, 2009). Many of these applications involve bubbly and boiling flows, and therefore it is essential to validate the software on such configurations. In particular, this is crucial for applications to flow in PWR fuel assemblies, including studies related to DNB. This work aims at presenting the present status of NEPTUNE-CFD validation in this area, as a step in an iterative process of improvement. To this end, this paper presents NEPTUNE-CFD code validation against four test cases based on experimental results. These data have been selected to allow separate effects validation. The adequacy of the measured quantities and the corresponding basic model of the CFD code to validate is underlined in each case. The selected test cases are the following. The Liu and Bankhoff experiment (1993) is an adiabatic air-water bubbly flow inside a vertical pipe. It allows to validate forces applied to the bubbles. The Bel F'Dhila and Simonin (1992) experiment is an adiabatic bubbly air-water flow inside a sudden pipe expansion. It allows to validate the dynamic models and turbulence. The DEBORA (CEA, 2002) and the ASU (Arizona State University, Hassan 1990) facilities provide results for boiling flows inside a vertical pipe. The working fluid is refrigerant R12 for DEBORA and R113 for ASU. Both allow to validate the nucleation modeling on a heated wall, and ASU allows also the validation of the two-phase wall function (Mimouni, 2009). A key feature of this work is that all these calculations were performed with a single standard version (1.0.8) of NEPTUNE-CFD, and with a single and consistent set of models, avoiding case-dependent 'tuning' of the modeling: a RANS approach with a Reynolds Stress Model for the turbulence of the continuous phase; the drag force from Ishii (1990), the added mass from Zuber (1964), the lift force from Tomiyama (1998) and a turbulent dispersion force are chosen for the dispersed phase. The NEA/CSNI Best Practice Guidelines were followed as much as possible, especially in the mesh generation process by keeping acceptable quality for the grids, by exploring the grid convergence, and also by assessing the numerical convergence. Comparisons with experimental data show that NEPTUNE-CFD has captured experimental profiles with reasonable accuracy for dynamical quantities and void fraction. Improvement must be done for the prediction of the bubbles size distribution. The need of new experiments will also be addressed to validate other specific models, like those used for bubble condensation in subcooled convective flow, which is the goal of the new TESS program. A companion paper presenting validation computations against these very recently obtained data is also submitted to the workshop. (authors)