Large-scale PANDA facility – radiation experiments and CFD calculations

Abstract CFD modelling of the thermo-hydraulic phenomena in the containment during the various phases of a severe accident necessarily requires consideration of radiative heat transfer – in the presence of steam. These radiative phenomena include (i) energy transfer within the gas mixture and (ii) between the gas and surrounding structures. Preliminary calculations carried out for these types of experiments within the OECD/NEA HYMERES-2 project with the CFD code containmentFOAM using a Monte Carlo solver for thermal radiation, demonstrated that the radiative heat transfer is significant even for very small amounts of vapour in the range of ≈0.1 % to ≈2 %. For this reason, the test matrix was tailored to the two opposite extremes: either gas compositions with a low vapour/steam content, where radiative heat transfer can be neglected, or gas mixtures with higher vapour contents, so that radiative heat transfer plays a dominant role. For the selected experiments of the H2P2 series and the corresponding CFD calculations, a vessel with a diameter of 4 m and a height of 8 m was preconditioned with different air-vapour mixtures (a) at room temperature and (b) elevated temperatures. A stable helium layer was then built-up in the upper part of the vessel. The gas was then compressed by injecting helium from above which resembles with best efforts a compression with a piston in a cylinder. This results in a height-dependent and transient increase of the gas temperature. These experiments and the associated CFD calculations were developed to isolate the phenomena of thermal radiation as good as possible from convective and diffusive effects – within the always present experimental limitations. For the reference experiment with ‘dry conditions’ corresponding to the lowest experimentally possible humidity of ≈0.1 %, we show that the use of a model without radiation provides the best agreement between the experimental and numerical results. For the much higher steam content of ≈60 %, the statistical narrow band correlated-k model (SNBCK), non-gray gas model, is the best candidate for future calculations – with computationally forgivable additional effort. We also provide with the Filtered Rayleigh Scattering technique (FRS) an outlook for a possible future instrumentation approach to better meet the requirements of the CFD community.