摘要
The accuracy of predicting the reactor core water level through best estimate codes during low-flow and low-pressure conditions is essential for assessing the effectiveness of cooling strategies in the event of a loss of RHR (Residual Heat Removal system) incident during mid-loop operations. Some best estimate codes, which are based on the two-fluid model, leverage drift-flux parameters to mitigate the difficulties associated with predicting the interfacial area concentration in the interfacial drag formulation. Previous investigations utilized extensive two-phase flow data collected in an 8 × 8 rod bundle at atmospheric pressure to formulate the INSS (Institute of Nuclear Safety System) drift-flux correlation, a drift-flux correlation designed for application in low-flow and low-pressure conditions. In the current study, the uncertainty related to the distribution parameter in the INSS drift-flux correlation was evaluated, employing the 8 × 8 rod bundle data. The identified uncertainty of the distribution parameter was then applied to perform uncertainty quantification analyses on void fraction data collected in 6 × 6, 5 × 5, and 3 × 3 rod bundles. The results indicated that the void fraction values calculated using the INSS drift-flux correlation were consistent with experimental data, supporting the correlation's applicability across a wide range of flow conditions and rod bundle sizes. Moreover, a comparison between void fraction values obtained from the INSS correlation and those calculated using the EPRI drift-flux correlation, as utilized in the RELAP5 code, revealed that the INSS drift-flux correlation provided improved predictive accuracy over the EPRI drift-flux correlation. The integration of the INSS drift-flux correlation into the RELAP5 code allowed for validation of its predictive capabilities concerning reactor core water levels under low-flow and low-pressure conditions. Uncertainty analyses were conducted using the ROSA/LSTF experimental data, which simulated a loss of RHR event during mid-loop operations. The outcomes of these analyses demonstrated that simulated results effectively predicted water levels in both the upper plenum and reactor core, with the exception observed during the initial phase of the experiment. It was concluded that the INSS drift-flux correlation was suitable for application in accident analyses under low-flow and low-pressure conditions.