Analysis of a Large-Scale PWR In-Vessel Retention Device

Abstract

This study assessed the effectiveness of in-vessel retention (IVR) in terminating the progression of an accident sequence initiated by a station blackout and large loss-of-coolant accident in a pressurized water reactor with thermal power of approximately 5000 MW. In the IVR design, external reactor vessel cooling is established by flooding of the reactor cavity. A water channel is introduced into the outer wall of the reactor vessel, and an insulated layered structure is added around the vessel. The amount of heat removed from the corium pool in the vessel lower plenum is limited by the critical heat flux (CHF) at the outer surface of the vessel wall. An integrated assessment was conducted in three steps. First, the responses of the reactor coolant system and containment were simulated using MELCOR. The predicted transient heat load at the vessel wall was then fed into RELAP5-3D, where the flow of natural, buoyancy-driven convection within the IVR water channel was simulated. Finally, the main thermal-hydraulic parameters in the IVR channel were substituted into the ULPU, SULTAN, SBLB, and MELCOR CHF correlations, and the effectiveness of IVR was assessed. The MELCOR simulation demonstrated that the heat load at the vessel wall of the lower plenum is dependent on the configuration of the debris. The heat flux to the vessel wall reached a maximum at 483 min, at an inclination angle of approximately 68 deg. The peak heat flux moved from a small inclination angle to a larger angle as the accident progressed. Both MELCOR and RELAP5-3D calculations predicted a gradual buildup of natural convection flow within the IVR channel following the application of a heat load to the vessel wall. The MELCOR code significantly overpredicts the mass flow of natural convection flow. Both codes predicted that the flow would experience large-amplitude fluctuations as the water in the IVR flow channel reached saturation. These fluctuations were attributed to instability induced by two-phase flow. If the inlet temperature can be kept sufficiently low to obviate boiling in the IVR channel, RELAP5-3D predicts that the channel flow will approach an approximately steady state. The selected CHF correlations predicted significantly different CHFs. The MELCOR correlation, which is a correlation based on pool boiling, produced the most conservative predictions, and the CHFs predicted by SBLB had the highest value. The minimum margin was found between 55 and 75 deg in all correlations. With the exception of the MELCOR correlation, the CHF ratio predicted by the other three correlations is greater than 1.2.