Propagation of pressure waves inside the reactor vessel after a large break LOCA is an issue since it affects pressure drop across core internals and, as a result, induces stresses in different components, like core barrel, core structures and even fuel. For reactor safety analysis pressure wave propagation is traditionally performed by systems codes. However, strong dispersion among the calculated results calls for test results to validate the calculations.
The pressure wave propagation following a larger LOCA is being systematically addressed by experiments in the PMK-2 integral-type test facility. In order to capture the high speed propagation of pressure waves special pressure transducers (capable to resolve the pressure variation with a frequency of 4 kHz) have been installed. The first four tests were conducted with rupture disks for opening the break, but a special quick opening valve will be installed for future tests, allowing the adjustment of the opening time between 12 and 50 ms.
The paper presents results of validation of the ATHLET code by the test results. The low-frequency oscillation of the measured system pressure was shown to be caused by flow rate coming from the pressuriser that compensates mass lost via the break: the frequency of the oscillation was slightly under-predicted. The propagation of the first rarefaction wave from the top of the downcomer to the upper plenum is very well calculated by ATHLET: in spite of the first order discretisation no numerical diffusion can be observed. The calculated pressure differences between two different locations in the system are of primary interest, since they define the loads on primary system internals. ATHLET somewhat overestimates the amplitude of the pressure difference pulses, while it fairly well describes the frequency of oscillations. First analyses indicate an effect of the calculated break flow rate. ATHLET calculates a slower attenuation of the pressure oscillations as compared to test results. This can be the consequence of rigid walls assumed in the analysis. The tendency of increasing first pressure peak with increasing system pressure is well predicted by the code.
In summary, it can be stated that ATHLET calculations produce slightly conservative results based on comparison with measured data.