ASTEC FPT1 and FPT2 sensitivity studies on bundle nodalization

The PHEBUS integral in-pile experiment series, a unique source of representative integral source term data for severe accidents, are performed at CEA/IRSN (Cadarache, France), being financially supported by the European Commission on behalf of the European Reactor Safety Community. The PHEBEN2 Project ('Validating Severe Accident Codes Against Phebus FP for Plant Applications') was a part of the EC 5th Framework Programme dedicated to the validation of integral codes for LWR severe accident and is coordinated by the European Commission JRC/ISIS. The main contractors are from France, Germany, Great Britain, Greece, Italy, the Netherlands, Spain and Switzerland. Some partners were also accepted from Bulgaria, Hungary and Romania. Codes such as ASTEC, ECART, MELCOR and MAAP were applied to Phebus tests in 'best-shot' analyses, supplemented by sensitivity studies, as is usual in plant assessments. This paper presents the first task assigned to the Romanian PHEBEN2 team, namely, to perform ASTEC sensitivity studies related to bundle nodalization, taking into account the PHEBUS FPT1 and FPT2 tests. The main variables whose sensitivity to meshing changes was checked are the following: hydrogen production, UO2 mass balance at the end of fuel degradation phase, control rod material release and fission product released fraction. Five cases will be compared for FPT1: 0: original discretization (fpt1.dat file from IRSN ): 18 mesh points; 1: 27 mesh points; 2: 33 mesh points; 3: 41 mesh points; 4: 10 'regular' mesh points for the fuel region (0.0 - 1.0 m) and the rest as in the reference case, resulting in 17 'coarser' mesh points. Also, five cases will be considered for FPT2: 0: original discretization (fpt2.dat file from IRSN ): 27 mesh points; 1: 27 'regular' mesh points; 2: 36 mesh points; 3: 42 'regular' mesh points; 4: 20 'regular' mesh points for the fuel region (0.0 - 1.0 m) and 13 'regular' mesh points for the rest, resulting in 33 'coarser' mesh points. For FPT1, the changes in axial mesh definition have a significant effect on the hydrogen instantaneous generation rate at the end of the heat-up period. However, the effect on the hydrogen cumulated mass is only important for the time range when most of the hydrogen is produced. Refining the meshing leads to greater values than in the 'reference' case, but using a coarser meshing (10 'regular' points) in the fuel region gives the closest cumulated mass to the reference one. For FPT2, the hydrogen instantaneous generation rate is more influenced by mesh changing than the cumulated mass. It should be noted that mesh refining generally leads to the 'time-shifting' of the moment when the maximum hydrogen instantaneous generation rate occurs. The mesh changing effect on the total hydrogen cumulated mass is less important (± 2%). When simulating FPT1, using a 'regular' finer axial meshing in the fuel bundle region underestimates the UO2 volatilized mass with more than 20%, while using a 'proper' meshing underestimates the target parameter with about 50%. For FPT2, refining the axial meshing overestimates the UO2 volatilized mass with less than 4%, while using a 'regular' axial meshing lowers it with about 30%. For FPT1, axial mesh refining underestimates the Ag volatilized mass fraction with more than 60%. Using a 'coarser' meshing leads to a strange behavior of the target parameter after 16500 s. Axial mesh changing underestimates the In and Cd volatilized mass fraction. Using 'proper' meshing leads to close values for all the control rod materials. The Sn volatilized mass fraction is less influenced by the mesh changing. For FPT2, the axial mesh refining overestimates the Ag, In and Cd volatilized mass fractions with at least 10%. Using a 36 mesh 'proper' scheme gives the same values as in the reference case, because the axial mesh points in the fuel bundle region are the same for the two cases. Mesh refining underestimates the Sn volatilized mass fraction with less than 10%. The 36 mesh 'proper' scheme gives again the same results as in the reference case. Changing the axial mesh points in the bundle does not lead to significant discrepancies in estimating the FPs released mass fractions for both experiments. The simulated hydrogen production and uranium release during FPT1 were in good agreement with the measured data. Therefore, we may affirm the module's robustness when dealing with physically-sensed problems. (authors)