Dr. Dirk Lucas

Head Computational Fluid Dynamics
Phone: +49 351 260 2047

Modelling and simulation of flashing boiling flows in modules with large vertical dimensions


Production of steam from water boiling is a familiar phenomenon under normal operations of light water nuclear reactors, for example, in the core of Boiling Water Reactors (BWRs) and in the Steam Generator (SG) of the Pressurized Water Reactors (PWRs). In these normal situations, water is heated to saturation temperature by a hot structure, i.e. the fuel rods in the core or the U-tubes in the SG. However, phase change from water to steam can also be triggered under adiabatic conditions by depressurization. In this case water temperature remains almost constant but the saturation temperature drops. This kind of phase change phenomenon is called cavitation or flashing in the literature. According to Laurien (2005), the main difference between the two terms of “cavitation” and “flashing” is that cavitation takes place usually under low-temperature conditions while flashing is relevant to the liquid flow near saturation temperature.

Actually there are many situations in nuclear power plants, where flashing is a relevant issue. The representative example is that in the modern design of reactors natural circulation is frequently adopted to cool down the core or to remove the residual heat for the purpose of reliance and safety. During such a circulation, warm water coming from the heat source goes up through a long adiabatic rising pipe. It is cooled by some kind of heat exchanger at the top of the circuit and then goes down back to the heat source again. If the water temperature entering the adiabatic rising pipe is high enough flashing boiling could take place there due to a reduction of hydrostatic pressure. The formation of steam leads to an increase of circulating flow rate and causes a subsequent decrease of the water temperature. As a result the process of flashing may eventually stop and the flow rate will be low again, so that the water temperature will increase leading to a new flashing cycle. In this way a self-sustained flow oscillation will be generated (Manera et al., 2005), which may be a source of flow instability or water hammer. For the analysis of heat removal capacity of such a circuit as well as the instability problem, it is of significant meaning to investigate the flashing phenomenon. Furthermore, flashing is often encountered in abnormal or accident scenarios of nuclear power plants, for example the operation of safety-relief valves and the Loss Of Coolant Accident in PWRs.

So far the so-called system codes such as RELAP and ATHLET are routinely applied to deal with thermal-hydraulic problems in the nuclear energy field, since it always indicates large duration and geometrical scale. These codes are based on one-dimensional approach and rely on a large number of component-specific empirical correlations (Lucas et al., 2011). All these features allow them to do numerical analyses with quite low computation costs. However, they also restrict the capacity and transferability of these codes for complicated 3D local phenomena like the above mentioned flashing. On the other hand, Computational Fluid Dynamics (CFD) simulations are able to provide geometry-independent results. For this reason CFD is getting more and more attention in the safety analyses of two-phase scenarios in nuclear reactors. Unfortunately, CFD technology is still not mature for two-phase flows especially with phase change, although it has a relatively long history concerning single phase flows. In order to get reliable results from CFD simulations sub models for the description of interphase mass, momentum and energy exchanging processes are still needed to be qualified. To meet this need experimental data with high resolution in space and time, so-called CFD-grade data, are required.

Experimental database

In the frame of this project, the capacity and limitation of the current CFD codes such as ANSYS CFX for the prediction of flashing boiling in modules with large vertical dimensions will be tested. The background of the project is the containment cooling condensers (CCC) of the KERENA™ reactor, which is a generation III+ BWR concept originally proposed by AREVA NP. ( The CCC cooling circuit is one of the passive safety systems driven by natural circulation. To avoid overpressure inside reactor vessel (RV) during disturbances or accidents, the surplus steam generated in the core will be vented passively to the flooding pool vessel (FPV) through an emergency condenser (EC), which is immersed in water at the lower part of the FPV. Steam can also be released directly to the upper free space of the FPV, where the containment cooling condensers (CCC) are located. Heat possessed by the steam from the evaporation of water outside the EC or directly from the RV has to be transported by the cooling water inside the CCC tubes to the SSPV by natural circulation.

KERENA(TM)reactor Schnittbild

Fig. 1 Sketch of passive safety systems in KERENA™ reactor (Leyer and Wich, 2012)
Click on the picture for full view.

The thermal dynamic behaviour and heat removal capacity of the CCC system is investigated experimentally on the INKA (the Integral Test Facility Karlstein) test facility (Leyer and Wich, 2012), see Fig. 2(a).

Fig. 2a: Components of the INKA test facility (Leyer and Wich, 2012)
Click on the picture for full view.
Fig. 2b: Schema of TOPFLOW pressure release experiment
Click on the picture for full view.

Before the starting of simulations for CCC, the data of pressure release experiments, which were carried out on the TOPFLOW test facility, are used for the verification of closure models and simulation setups. This is because the phenomena observed in this experiment are inherently similar to that occurring inside the CCC passive system. On the other hand, TOPFLOW experiments are designed especially for the qualification of CFD codes and has high time and space resolution. In addition, the geometry of axisymmetric pipe is much simpler than that of the CCC circuit.

During the TOPFLOW pressure release experiment, nearly saturated water is circulated with a velocity of about 1 m/s through upwards the test section, see Fig. 2(b). The depressurization is realized not only by the elevation of the vertical pipe but also by the blow-off valve mounted above the steam drum. As the water superheat exceeds a certain degree caused by depressurization, water begins to evaporate at the top of the test section and the evaporation wave spreads immediately downwards to the bottom. The generation of steam will lead to a temperature drop in the water to overcome the latent heat. As it drops below the saturation point, the evaporated steam will disappear again (Lucas et al., 2011).

Results of the simulations

For one example of the TOPFLOW experiments, the predicted time-dependent steam volume fraction, water temperature, absolute pressure and mass flow rate are compared with measured ones. In general, good agreement is observed. However, for the triggering of flashing boiling in the simulation requires a higher water superheat than that in the experiment. This might be caused by models and assumptions used in the simulation, which is one topic of future work.

(a) Steam volume fraction
Click on the picture for full view.
(b) Water temperature
Click on the picture for full view.

(c) Absolute pressure
Click on the picture for full view.
(d) Mass flow rate
Click on the picture for full view.

Fig. 3 Examples of CFX results for the TOPFLOW pressure release experiment (20 bar)

The generation and disappearance of steam during the flashing is shown in the following movies for TOPFLOW vertical test section and INKA rising pipe, respectively.

TOPFLOW vertical test section
Click on the picture for video.
INKA rising pipe
Click on the picture for video.


  • Laurien, E., 2005.
    Kavitation, Volumensieden, Ausgasung – Vergleich neuer Modellierungsansätze.
    Technologietag des ERCOFTAC Pilot Center Germany South, Stuttgart, 30. Sept. 2005

  • Leyer, S. and Wich, M., 2012.
    The Integral test facility Karlstein.
    Science and Technology of Nuclear Installations, Hindawi Publishing Corporation, Volume 2012, Article ID 439374, doi: 10.1155/2012/439374

  • Lucas, D., et al., 2011.
    Experiments on evaporating pipe flow.
    The 14th International Topical Meeting on Nuclear Reactor Thermalhydraulics, NURETH-14, Toronto, Ontario, Canada, September 25-30, 2011

  • Manera, A., et al., 2005.
    Out-of-phase flashing-induced instabilities in natural circulation two-phase system with parallel channels.
    4th International Conference on Transport Phenomena in Multiphase Systems. Gdańsk, Poland, June 26-30, 2005


This work is carried out in the frame of a running research project funded by funded by EON Kernkraft GmbH.


Dr. Dirk Lucas

Head Computational Fluid Dynamics
Phone: +49 351 260 2047