Abstract
This study aims to reveal the possibility of a crack due to the temperature gradient on RPV head in Maanshan Nuclear Power Plant in Taiwan. The temperature distribution inside RPV and on the RPV wall is calculated by Computational Fluid Dynamics (CFD) software STAR-CCM+. The computational domain starts from the bottom part of the upper plenum and continues throughout the entire dome region, so that the high temperature coolant leaking to the dome region through gaps between solids can be included to calculate the leak flow and temperature distribution in the dome region. In the present study, Computational cells are generated for half of the upper plenum region and dome region using symmetry assumption to save computer resources. The total number of cells is ~65 millions including 3 prism layers. The calculations not only provide the detailed information of flow and temperature distribution inside the dome region but also show 2% of high-temperature coolant entering to the dome region by the adiabatic assumption of all solid structures. Total effective degradation years (EDY) and reinspection years (RIY) for investigating the crack probability in Maanshan Nuclear Power Plant are applied to this study. The EDY and RIY factor (<1.0) based on the calculated maximum dome temperature indicate that the previous inspection is conservative. Some scenarios with different magnitude of the flow velocity ejected from head cooling nozzles are calculated to determine the bounding case according to EDY and RIY value. The probability of cracking is small if there is enough coolant (>0.4%) ejected from head cooling nozzles. RKE, SKE, and k-ωSST model are compared in this paper. Results obtained from RKE model are similar for those from SKE model, but slightly different for those form KWSST model.