ICONE23-1277 INVESTIGATION OF THE INFLUENCE OF TURBULENCE MODELS ON THE SIMULATION OF THE GAS DISTRIBUTION
Aniseh A. A. Abdalla,Jiyang Yu,C. H. Zhang,B. X. Hou,M. Saeed
DOI: https://doi.org/10.1299/jsmeicone.2015.23._icone23-1_136
2015-01-01
The Proceedings of the International Conference on Nuclear Engineering (ICONE)
Abstract:During a severe nuclear power plant (NPP) accident, large amounts of hydrogen and steam can be produced in nuclear reactor containment. As the hydrogen concentration in the containment release rates exceed 4%, hydrogen combustion may occur when the hydrogen, air and steam reach certain concentration and a flammable source is present. In the case of hydrogen combustion, there is a possibility of producing short term pressure or detonation force. Therefore, the production of these gas species could threaten containment integrity. Although after the Three Mile Island (1979) accident, study of hydrogen distribution attracted scientists attention, it was after Fukushima (2011) accident, that modeling the gas behavior became an important topic in nuclear safety analyses. In order to mitigate these accidents, we need to understand the hydrogen behavior during NPP accident. Reliable turbulence model must be used in order to have an accurate estimate of the gas concentration distribution and other physical phenomena of the gas mixture. In this study, small scale airfountain test case is selected as a benchmark. Laminar model; standard κ-ε; RNG κ-ε and realizable κ-ε turbulence models are addressed. The computations are performed with HYDRAGON code. The purpose of this study is to test the influence of turbulence models to the gas distribution, and to demonstrate the code thermal-hydraulic simulation capability during NPP accident. HYDRAGON code simulation results show that RNG and κ-ε turbulence models have achieved better agreement with the experimental data on the prediction of dimensionless density distributions. Only at the region near the air-fountain source, RNG turbulence model has better estimation values than κ-ε and realizable models. The results depict that the turbulence model choice has small influence on the result. Especially, at the region near the air-fountain source, which is the region where the turbulence model was strong during air-injection.