Assessment of the Gasflow Spray Model Based on the Calculations of the Tosqan

Within the framework of the EU (European Union) project SARNET (Severe Accident Research Network), spray experiments were performed in the TOSQAN facility in France. The experiments, Test 101 (thermal hydraulic part, without He release) and Test 113 (dynamic part, with He release), were selected for benchmark calculation with different CFD (Computational Fluid Dynamics) and LP (Lumped Parameter) codes. The CFD participants performed their calculations with a 2D geometry of the TOSQAN vessel without considering the pre-steam injection phase, but using the average values from the experiment for the simulation of the initial condition at the start of the spray. AREVA performed two GASFLOW post-test analyses of the Test 101 with the full 3D geometry simulation of the vessel: • Part simulation (PS): without considering the pre-steam injection phase, but using the average values from the experiment for the simulation of the initial flow conditions at the start of the spray. • Full simulation (FS): considering the pre-steam injection phase resulting in determination of the initial flow conditions at the start of the spray. FZK performed 2D calculations simulating the spray injection phase with and without He release in Tests 101 and 113, respectively. Furthermore FZK simulated the pre pressurization of the vessel with He in Test 113. The calculations of the TOSQAN spray experiments were necessary for validating the GASFLOW spray model because the effect of sprays on hydrogen risk had been investigated with this code for different LOCA (Loss of Coolant Accident) events, with and without delayed depressurization resulting in fast homogenization of the EPR containment atmosphere. The comparison of the calculations with the measured experimental results showed good agreement between calculation and measurements for Test 101 in all four phases; namely, vaporization phase, fast condensation phase, and slow condensation phase, as well as the equilibrium phase. Furthermore, it showed the same tendency for fast homogenization of the vessel atmosphere as mentioned above in the spray calculation for the EPR. The comparison showed maximum deviations of only about 3% for all three global values; namely, pressure, gas mean temperature, and gas mole number. Simulation of Test 113 showed less then 6% deviation for the He volume concentration after 250 s spray time while the He stratification is fully destroyed as was seen in the experiment. The selected mesh or control volumes for TOSQAN geometry provides an extrapolation from the compared results to prototypic containment scale. The linear scale of the TOSQAN vessel relative to the EPR nuclear power plant is about 1:22. The GASFLOW code is a finite volume computer code, which has been developed at the Los Alamos National Laboratory in USA and the research centre Karlsruhe in Germany. The code is designed to be a best-estimate tool for predicting the transport of steam/hydrogen/air mixtures, with/without spray, as well as the recombination and combustion of hydrogen with the possibility of adding other gases for simulating design basis or severe accidents in nuclear reactor containments. GASFLOW models are validated with numerous experiments describing different thermohydraulic phenomena in the CFD4NRS-3: Experimental Validation and Application of CFD and CMFD Codes to Nuclear Reactor Safety Issues OECD/NEA & IAEA Workshop Hosted by United States NRC Washington D.C., USA 14–16 September 2010 2/14 containment. These simulation analyses have shown that the GASFLOW spray model is applicable and reliable for the containment analyses due to good agreement with experimental results and can be applied for simulating all relevant phases during spraying, particularly during the thermodynamic equilibrium phase with and without light gas release.