Pre-Test Analysis of Natural Circulation Test of PHENIX End-of-Life With the MARS-LMR Code

PHENIX, a prototype sodium-cooled fast reactor (SFR), has demonstrated a fast breeder reactor technology and also achieved its important role as an irradiation facility for innovative fuels and materials. In 2009 PHENIX reached its final shutdown and the CEA launched a PHENIX end-of-life (EOL) test program, which provided a unique opportunity to validate an SFR system analysis code. The Korea Atomic Energy Research Institute (KAERI) joined this program to evaluate the capability and limitation of the MARS-LMR code, which will be used as a basic tool for the design and analysis of future SFRs in Korea. For this purpose, pre-test analyses of PHENIX EOL natural circulation tests have been performed and one-dimensional thermal-hydraulic behaviors for these tests have been analyzed. The natural circulation test was initiated by the decrease of heat removal through steam generators (SGs). This resulted in the increase of intermediate heat exchanger (IHX) secondary inlet temperature, followed by a manual reactor scram and the decrease of secondary pump speed. After that, the primary flow rate was also controlled by the manual trip of three primary pumps. For the pre-test analysis the Phenix primary system and IHXs were nodalized into several volumes. Total 981 subassemblies in the core were modeled and they were divided into 7 flow channels. The active 4 IHXs were modeled independently to investigate the change of flow into each IHX. The cold pool was modeled by two axial nodes having 5 and 6 sub-volumes, respectively. The reactor vessel cooling system was modeled to match the flow balance in the primary system. The flow path of vessel cooling system was quite complicated. However, it is simplified in the modeling. For a MARS-LMR simulation, the dryout of SGs have been described by the use of the boundary conditions for IHTS as a form of time-to-temperature table. This boundary condition reflects the increase in IHTS temperature by SG dryout during the initial stage of the transient and the increase in heat removal by the opening of the two SG containments at 3 hours after the initiation of the transient. Through the comparison of the pre-analysis results with the prediction by other computer codes, it is found that the MARS-LMR code predicts natural circulation phenomena in a sodium system in a reasonable manner. The final analysis for validation of the code against the test data will be followed with an improved modeling in near future.