Sensitivity analysis of in-vessel accident progression behavior in Fukushima Daiichi Nuclear Power Plant Unit 3

The Great East Japan earthquake and the subsequent tsunami which occurred on March 11th 2011 put the operating Units 1–3 at Fukushima Daiichi Nuclear Power Plant (NPP) in severe accident conditions resulting from loss of offsite power and AC power. It is believed that the Station Blackout (SBO) and loss of heat sink led to core meltdown in all Units 1–3. Despite past research efforts on the severe accident progression in Fukushima NPP Units 1–3, there are still knowledge gaps and uncertainties existing in understanding of the severe accident scenarios and consequences. Hence, this study aims at identifying the modeling uncertainties and addressing the sensitivity parameters in Fukushima NPP Unit 3. A more detailed Control Volume (CV) division model of the reactor core region has been developed to better simulate the thermal-hydraulic behavior of liquid water and steam, which is considered to be crucial in simulating the core uncovery and degradation process. The boundary conditions such as the water injection rates by the Reactor Core Isolation Cooling (RCIC) system, the High Pressure Core Injection (HPCI) system and Alternative Water Injection (AWI) to the reactor core were determined based on the available reactor water level and pressure measurement data. The current study suggested that the local vapor heatup behavior could influence the core melting and relocation behavior, which can lead to different core degradation scenarios. With the current modeling assumptions in MELCOR, the best estimate conditions for RPV pressure history of Unit 3 suggested that 6 SRVs could have remained open when the major core slumping took place at ca. 45:20 h (ca. 12:00, March 13) with 50 to 80 tons of water inventory in the lower plenum. The current analysis also suggested that the efficiency of the AWI to the reactor core could have been only 15% as of reported by TEPCO with the current modeling conditions if debris dryout was assumed to have occurred at around ca. 54.0 h (20:40 h, March 13th). As for lower head failure, there is still large uncertainty in predicting lower head failure time with Larson-Miller creep rupture model in the current MELCOR modeling.