The computation of mass flow rates through crack-like defects in piping systems of light water reactors requires typically the description of two-phase flow conditions. The computed discharge rate depends on the crack opening area, the thermal-hydraulic modeling of the flow, and the flow resistance of the crack. Several models have been proposed to characterize the critical flow through crack-like defects. An evaluation of advantages and shortcomings of the different models with regard to the interaction of the three different parts (crack opening area, thermal-hydraulic modeling, flow resistance) has been performed.
In this paper, the flow resistance modeling from several approaches is discussed, and compared with a database from eight different testing programs. Five different flow models are applied to analyze a database of more than 800 leak rate measurements for subcooled water from twelve different experimental programs. It is shown that the correct modeling of the flow resistance is crucial for a best estimate reproduction of the measured data. It turns out that generally, equilibrium models are about as good as non-equilibrium models. The data were processed with the GRS software WinLeck which includes different analytical approaches for the calculation of crack sizes and leak rates in piping components. The most reliable results within the model selection are produced by the CDR model (Critical Discharge Rate) of the ATHLET code (Analysis of Thermal-hydraulics of Leaks and Transients) developed by GRS.
As a conclusion, the accurate modeling of form losses and frictional pressure losses for critical discharge flow rates through crack-like leaks are essential for a reliable prediction of flow rates. Uncertainties in leak rate computations results arise due to the lack of information about the flow geometry and its associated drag. The assumed flow resistance of a through-wall crack influences the computed leak rate as significant as the phase-change- and flow-models. The manifest difference between equilibrium models (Pana, Estorf) and non-equilibrium models (Henry, ATHLET-CDR) seems to be less significant than the pressure loss issue. One can conjecture that, for crack-like through-wall defects, friction effects play a more important role than non-equilibrium effects.