Two-phase pressurized thermal shock analysis with CFD including the effects of free-surface condensation

Abstract Pressurized thermal shock (PTS) is the stress experienced by a reactor pressure vessel (RPV) due to a rapid temperature and pressure change, or large spatial variations in temperature because of plumes and stripes of cold water surrounded by hotter water. PTS resulting from a loss-of-coolant accident was investigated for the nuclear Power Plant Gosgen-Daniken, which is a German-type three-loop pressurized-water reactor. The most critical regions in the RPV during such PTS transients are the most irradiated regions in the downcomer wall under the cold leg in the vicinity of the first weld of the first ring of the RPV. Cooling plumes or stripes develop at the RPV wall or core barrel wall depending on parameters like the water level in the downcomer, the mass flow rate of the injected cooling water and condensation effects. For fracture-mechanics analysis of the RPV, temperature profiles and heat-transfer coefficients corresponding to either cooling plumes or stripes at the wall (here mainly at the inner RPV wall) are needed. Various methodologies can be followed for generating the temperature profiles and heat-transfer coefficients resulting from the mixing of cold injection water with the hot RPV water. One is the calculation of these data by analytical fluid-mixing codes validated with experiments, such as KWU-MIX. Alternatively, computational fluid dynamics (CFD) tools can be used after suitable validation. One advantage of CFD is the possibility to show attachment of plumes and stripes to the inner wall of the RPV or detachment towards the core barrel. A free-surface condensation model was implemented in ANSYS CFX for simulating two-phase transients when the water level is below the cold-leg nozzle. For the purpose of validating free-surface condensation and the maximum mass flow rate for the stripe of liquid water before detachment from the RPV wall, experimental data from the full-scale Upper Plenum Test Facility (UPTF) were used. The comparison of CFD and KWU-MIX analyses of the limiting transient for the core weld obtained from the selection process with KWU-MIX shows that the inherent conservatism in the standard KWU-MIX approach leads to higher thermal loading than the best-estimate CFD approach and demonstrates the benefits of a CFD tool to quantify potential margins.