A large amount of Hydrogen gas is expected to be released within the dry containment of a pressurized water reactor (PWR), shortly after the hypothetical beginning of a severe accident leading to the melting of the core. According to local gas concentrations, the gaseous mixture of hydrogen, air and steam can reach the flammability limit, threatening the containment integrity. In order to prevent mechanical loads resulting from a possible conflagration of the gas mixture, French and German reactor containments are equipped with passive autocatalytic recombiners (PARs) which preventively oxidize hydrogen for concentrations lower than that of the flammability limit. The objective of the paper is to present numerical assessments of the recombiner models implemented in CFD solvers NEPTUNE_CFD and Code_Saturne. Under the EDF/EPRI agreement, CEA has been committed to perform 42 tests of PARs. The experimental program named KALI-H2, consists checking the performance and behaviour of PAR. Unrealistic values for the gas temperature are calculated if the conjugate heat transfer and the wall steam condensation are not taken into account. The combined effects of these models give a good agreement between computational results and experimental data.
During a design-basis accident (DBA) or a severe accident (SA) in a nuclear power plant, certain chemical reactions may produce hydrogen (hydrogen is produced from the oxidation of zirconium sheaths and structures of fuel elements during the phase of core degradation), in such a way that hydrogen and oxygen volumetric concentrations may exceed the lower flammability limits (LFLs). The hydrogen risk in a nuclear power plant may be defined as the risk of hydrogen combustion in the containment building which represents a threat to the integrity of the confinement due to pressure and temperature levels. The nuclear safety is based on the concept called “defence-in-depth” that consists of a hierarchical deployment of different levels of equipment and procedures to maintain the efficiency of the physical barriers.
The transport and distribution of hydrogen inside the containment or the different compartments are critical phenomena to determine the kinetic and the nature of combustion. The prediction of stratification phenomena and location of hydrogen pockets are essential to assess the hydrogen risk and then to optimise the hydrogen hazard mitigation system.
Among the different safety systems for limiting the pressure increase during the course of the accident and the impact of possible combustion (deflagration), French and German PWR reactors have three types of mitigation means. Sprinkler systems: the injected water droplets cool the containment and lower the pressure by condensing steam on the droplets. They also promote mixing of gas and rapidly break possible stratifications of the lightest gases. The walls of the containment building and metal structures also play an important role from a thermal viewpoint. The walls, significantly cooler than the gas, condense the water vapor in the gas mixture and thus limit the pressure increase in the containment. Furthermore, the temperature difference between fluid and walls generates a convection loop, enhancing the mixing of gases of different density. The passive autocatalytic recombiners (PAR): their role is to proactively oxidize hydrogen for preventing its accumulation in the containment. The catalytic recombiners initiate a controlled combustion, which is similar to a slow deflagration.
A catalytic recombiner consists of a vertical channel or stack equipped with a catalyst bed in the lower part. In case of an accident, the catalyst is in contact with the gas mixture of the containment. Hydrogen molecules coming into contact with the catalyst surface are reacted with oxygen to form steam. The heat of the reaction at the catalyst surface causes a buoyancy-induced flow accelerating the inflow rate and thereby feeding the catalyst with a large amount of hydrogen that ensures high efficiency of recombination. The natural air circulation ensures a continuous air supply to the catalytic recombiner. This effect is increased by the height of the chimney, the inlet free area, and the fast heat up of the catalytic plates.
Catalytic recombiners favour the chemical reaction
This paper focuses on numerical assessments of PAR’s modeling implemented in CFD solver Code_Saturne and CMFD solver NEPTUNE_CFD [
The solver belongs to the well-known class of pressure-based methods. It is able to simulate multi-component multiphase flows by solving a set of three balance equations for each field (fluid component and/or phase) [
The main interest of the numerical method is the so-called “volume fraction-pressure-energy cycle” that ensures mass and energy conservation and allows strong interface source term coupling [
Mass balance equations, momentum balance equations and total enthalpy balance equations, are solved for each phase. The gas turbulence is taken into account by the classical
The model of drop-wall interaction which was developed and implemented is written as a symmetric extension of the nucleate boiling model at the wall, and uses as a starting point the model of mass transfer in the core flow. To establish this model, we made the following assumptions: the drops which accumulate on the walls take a hemispherical form; there is no nucleate boiling inside the drops at the wall; the drops which impact the walls successively see a stage of cooling (resp., heating) and a stage of condensation (resp., evaporation); the droplets stick to the wall (no rebound), or slide along the wall.
The total heat flux exchanged between the wall and the flow is split into four terms:
Details can be found in [
In fact, a particular model is developed to reduce the sensitivity to the mesh refinement, for example, the gas temperature and the gas velocity in the nearest cell at the wall depend on the size of the cell with a volume finite method. As a consequence, on one hand, we use the value calculated in the cell and given by the direct resolution of the momentum equation (gas velocity) and the energy equation (gas temperature). On the other hand, we use standard wall functions for the gas velocity and temperature at a nondimensional distance to the wall. The combination of these values provides a weakly dependent value for the gas velocity and temperature near the wall, and these values are the input of the model of heat and mass transfer at the wall.
The motion, the distribution of gases, and heat transfer in containment enclosures can be described by the general momentum, partial masses, and energy conservation equations.
The predominant physical phenomena driving the motion, the distribution, and heat transfer of fluids within containment enclosures are follows.
Mixing and/or segregation of gas whose velocity, density, and temperature are different. “Swelling” of containment: the compressibility of gas is taken into account, even if the flow velocities are low. Laminar and controlled combustion of hydrogen in recombiners, in order to limit the concentration of this gas. Condensation of steam on cold structure surfaces, which has the main effect of limiting the pressure rise.
The general momentum, partial masses, and energy conservation equations describing these phenomena can be simplified, and stiffness due to the presence of physics having very different characteristic length and time scales can be removed or relaxed.
The used turbulence model for containment applications is the standard
The flows are mainly low Mach number flows, whose motion is predominantly driven by free convection. A low Mach number model can be implemented in a pressure correction-based solver usually used for incompressible or steady dilatable flows, as Code_Saturne [
Thanks to the low Mach number approximation, the mechanical pressure is neglected for the computation of density, through the thermal equation of state:
The supplementary unknown
The enthalpy equation of the mixture is quite complex and contains several terms. The body forces, the viscous constraint contributions, and the supplementary terms due to the presence of more than two different species are negligible, when compared to the convective and turbulent transport contributions. For low Mach number flows, the kinetic energy remains small when compared to the thermal energy. On the other hand, the unsteady contribution of the thermodynamic pressure is conserved, as it plays a key role in the pressure rise in the containment.
The Fourier laminar and turbulent conduction term is directly written according to the enthalpy variable through the linearized relation
The enthalpy equation is written in the following form:
Then, in presence of exothermic chemical reactions, due to the combustion of hydrogen by the recombiners, the transformation of formation enthalpy into sensitive enthalpy is taken into account through a source term proportional to the sensitive enthalpy
We recall that the formation enthalpies and reaction heat
In the following sections, calculations performed with Code_Saturne do not take into account the condensation phenomena at wall.
In this study we will use the following rate for the reaction
Moreover
An important point to underline is that the hydrogen depletion rate is a semiempirical relation, based on experimental measures made at the inlet and at the outlet of the recombiner. In other words, the rate given by this expression is a global rate on the whole recombiner.
The reaction
The mass conservation equations are written as below: the global mass equation, containing the sink term of wall condensation: the conservation equations of noncondensable gases, containing the slow combustion sink terms due to the recombiners: the relation for obtaining the condensable gas (steam) from the concentration of the other gases:
H2 depletion rate.
To validate the recombiner model, two H2-PAR testing programs were carried out and the results were compared with experimental data. The experimental device consists of a sealed bag of flexible material to an approximate volume of 7.6 m3, which contains a Siemens FR90/1-150 type recombiner of 0.2 m length, 0.166 m wide and 1.03 m in height. The volume of this pocket is not constant over time, the variation being of the order of one m3. The atmosphere is initially composed of air and water vapor. Hydrogen is introduced at the base of the plant for a specified period. Hydrogen concentrations were measured by gas chromatography every minute at different locations (in particular, at recombiner inlet and outlet). They are then averaged over the whole field.
The computational domain is modelled by a cube and does not take into account the change in volume of the chamber (a flexible pouch; Figures
H2-PAR test case.
H2-PAR test: computational domain.
The operating conditions of test-E2 and E19 are summarized below (Table
Initial conditions—H2-PAR test.
Essai E2-Bis | Essai E19 | |
---|---|---|
O2 molar fraction | 0.0828 | 0.1403 |
N2 molar fraction | 0.3112 | 0.5274 |
H2O molar fraction | 0.6054 | 0.3323 |
H2 molar fraction | 0 | 0 |
Temperature | 85°C | 70°C |
Pressure (Pa) | 101300 | 101300 |
E2-bis test: the mass flowrate is successively
E19 test: the mass flowrate is
Adiabatic conditions.
The differences in density, generated by the combustion of hydrogen, create a natural convection loop that enhances the mixing at the same level and above recombiner. Moreover, cold stratified zone below the recombiner is calculated. The schematic modelling performed here may be representative of what can happen in the real case of the reactor building. Two opposing phenomena are in competition: the recombiner promote the controlled combustion of hydrogen, but according to their position in the enclosure, may also promote the creation of stratified layers of gas, “resistant” to the mix. Figures
E19 test—time evolution of hydrogen molar fraction.
E2bis test—time evolution of hydrogen molar fraction.
The manufacturer correlation for the hydrogen depletion rate is dedicated to scenario code (like MAAP) with 0D model, and the challenge is to find a method to implement this manufacturer correlation in a CFD code. Figures
The exercise carried out in this paragraph remains modest compared to the real complexity of the operation of a catalytic recombiner. But the goal here is to qualitatively reproduce the local effects induced by the operation of recombiner.
To test the catalytic capability of the SIEMENS recombiner model FR90/1-150, we used the KALI vessel. This 15.6 m3 vessel (4.6 m height and 2.1 m of diameter) is able to withstand a pressure of 12 bar. The recombiner is located close to the wall, at the bottom of the vessel.
The vessel is connected with specific systems: steam injection system, hydrogen injection system,n and cold water system (Figures
KALI-H2 test.
Sketch of the computational domain of the KALI-H2 test.
Sketch of KALI-H2 test.
Sketch of PAR unit.
The vessel is also equipped with a fan to avoid any stratification during the hydrogen injection and also to start with a homogeneous mixture (it is stopped 45 s after the beginning of the hydrogen injection). Since the test facility cannot accommodate a full-size recombiner unit, a small-size segment model was tested. Figure
The recombiner was represented as a box of which the dimensions were the ones from the outside of the recombiner. Plates are not modeled.
In our global approach, source terms are distributed over the meshes representing the active recombiner volume containing the vertical plates. Source terms are calculated in each mesh of the recombiner with the local concentration of hydrogen and oxygen, weighted by the active recombiner volume.
The simulations couple the two-phase flow and the solid heat conduction within the wall vessel. The wall condensation of vapor is taken into account with NEPTUNE_CFD code. In fact, in a real situation, heat transfer to the wall and condensation result in an enhanced mixing of the atmosphere. In order to underline these aspects, Code_Saturne simulations are performed with the artificial assumption of adiabatic walls and without wall condensation.
The initial conditions of the test 008 are the following: pressure: 3.25 bars, temperature: 30°C (303 K), 4% of H2, 96% of air, no steam (water) initially.
Successive stages have to be considered for calculations: mixing of the gases with a fan device occurs every 900 s for 120 s. Moreover, an initial mixing is present during the first 120 s of the test. The successive stages of the simulation are given in Figure
Stages, test number 008.
We assume that the end of the mixing leads to a homogeneous gas mixing in the whole KALI-H2 vessel. Hence, the initial thermal-hydraulics properties of the calculations are taken at this time: homogeneous repartition of gas and temperature. The simulation carries on up to the second phase of mixing (
Thus the mixing stage is simplified, but avoids taking into account the fan device for which few data are available. Then, calculations continue up to a second point of experimental recording (1080 s).
The number of cells is about 112000. The time step corresponds to CFL = 1.
The sensitivity to the turbulence modelling has been investigated with NEPTUNE_CFD code, and calculations with three turbulent models (
Figures
H2 mass fraction below the PAR.
H2 mass fraction at outlet of the PAR.
Figure
Time evolution of gas temperature at outlet of the PAR.
Figure
H2 molar fraction just below the PAR.
Figure
Gas temperature and hydrogen mass fraction fields.
We have presented in this paper the models implemented in NEPTUNE_CFD, a three-dimensional two-fluid code dedicated to nuclear reactor applications and in Code_Saturne, a three-dimensional single-phase code. Thanks to a code-to-experiment benchmark based on the COPAIN and TOSQAN facilities [
During the course of a severe accident in a pressurized water reactor (PWR), spray systems are used in a containment in order to limit overpressure, to enhance the gas mixing in case of the presence of hydrogen, and to drive down the fission products. Hence, vapor condensation on a cooled surface, spray effects, and PAR systems act simultaneously in applications which is made possible with the two-phase flow approach proposed in the paper.
The PAR models used in the paper are based on the manufacturer’s correlation to calculate the hydrogen depletion rate. In future works, it would be interesting to test the impact of the correlation uncertainties on gas temperature and hydrogen mass fraction at the PAR outlet calculated by a CFD code. Comparisons between manufacturer’s correlation and more sophisticated models based on gas phase and surface chemical mechanisms should be investigated.
This paper has been completed in the framework of the PAGODES2 project financially supported by EDF (Electricité de France). The NEPTUNE_CFD code is being developed in the framework of the NEPTUNE project financially supported by CEA (Commissariat à l’Energie Atomique), EDF (Electricité de France), IRSN (Institut de Radioprotection et de Sûreté Nucléaire), and AREVA-NP.