Coolant mixing studies of natural circulation flows at the ROCOM test facility using ANSYS CFX


Coolant mixing studies of natural circulation flows at the ROCOM test facility using ANSYS CFX

Höhne, T.; Kliem, S.

Coolant mixing in the cold leg, downcomer and the lower plenum of pressurized water reactors is an important phenomenon mitigating the reactivity insertion into the core. Therefore, mixing of the de-borated slugs with the ambient coolant in the reactor pressure vessel was investigated at the ROCOM mixing test facility. The Rossendorf Coolant Mixing Model (ROCOM) was constructed for the investigation of coolant mixing in PWR. The ROCOM facility has four loops each with an individually con¬trol¬led pump. This allows to per¬form tests in a wide range of PWR flow conditions from natural con¬vec¬tion flow up to forced con¬vec¬tion flow at nominal flow rates including flow ramps. ROCOM is operated with wa¬ter at ambient tem¬pe¬ra¬tures because the re¬ac¬tor pressure ves¬sel (RPV) mock-up and its internals are made of perspex. The model has a linear scale 1:5 to the prototype, the water inventory in the loops is kept in scale 1:125 and the traveling time of the coolant is identical to the original reactor.

Thermal hydraulics analyses showed, that weakly borated condensate can accumulate in particular in the pump loop seal of those loops, which do not receive safety injection. After refilling of the primary circuit, natural circulation in the stagnant loops can re-establish simultaneously and the de-borated slugs are shifted towards the reactor pressure vessel (RPV). In the ROCOM experiments, the length of the flow ramp and the initial density difference between the slugs and the ambient coolant was varied. During inherent dilution the slug could have a higher temperature and a lower density. Additionally the boron content influences the fluid density. In ROCOM this density difference is adjusted by the addition of ethyl alcohol. The acquisition of the concentration fields is performed with high spatial and temporal resolution mea¬sure¬ments of the tracer con¬¬¬centration.

Experiments with 0 up to 2% density differences between the de-borated slugs and the ambient coolant were used to validate the CFD software ANSYS CFX. A Reynolds stress turbulence model was employed and a hybrid mesh consisting of 3.6 million nodes and 6.4 million elements was used. The Best Practice Guidelines were applied to ensure the quality of the calculations. In the calculations, the High-Resolution discretisation scheme of ANSYS CFX was used to discretize the convective terms in the model equations. A second-order implicit scheme was utilized to approximate the transient terms. The time step size was set to 0.1 s. The ethyl alcohol water, which had a lower density, was applied as a tracer. It was modelled with the multi-component model of ANSYS CFX. The ethyl alcohol water was modelled as a component with different density and viscosity compared to water.

Depending on the degree of density differences the experiments and CFD calculations show a more or less intense stratification in the cold leg and downcomer. The ANSYS CFX calculations show a good qualitative agreement with the data. At some local positions differences in the predicted and measured concentration fields occur. The obtained experimental and numerical results can be used for further studies of the core behaviour using coupled thermo-hydraulic and neutron-kinetic code systems.

Keywords: CFD; ROCOM; boron dilution; PWR

  • Contribution to proceedings
    OECD/NEA International Workshop on The Benchmarking of CFD Codes for Application to Nuclear Reactor Safety (CFD4NRS), 05.-09.09.2006, Garching, Germany
  • Lecture (Conference)
    OECD/NEA International Workshop on The Benchmarking of CFD Codes for Application to Nuclear Reactor Safety (CFD4NRS), 05.-09.09.2006, Garching, Germany

Permalink: https://www.hzdr.de/publications/Publ-8083
Publ.-Id: 8083