Improvement of reflood model in RELAP5 code based on sensitivity analysis

• Sensitivity analysis is performed on the reflood model of RELAP5.
• The selected influential models are discussed and modified.
• The modifications are assessed by FEBA experiment and better predictions are obtained.

Reflooding is an important and complex process to the safety of nuclear reactor during loss of coolant accident (LOCA). Accurate prediction of the reflooding behavior is one of the challenge tasks for the current system code development. RELAP5 as a widely used system code has the capability to simulate this process but with limited accuracy, especially for low inlet flow rate reflooding conditions. Through the preliminary assessment with six FEBA (Flooding Experiments with Blocked Arrays) tests, it is observed that the peak cladding temperature (PCT) is generally underestimated and bundle quench is predicted too early compared to the experiment data. In this paper, the improvement of constitutive models related to reflooding is carried out based on single parametric sensitivity analysis. Film boiling heat transfer model and interfacial friction model of dispersed flow are selected as the most influential models to the results of interests. Then studies and discussions are specifically focused on these sensitive models and proper modifications are recommended. These proposed improvements are implemented in RELAP5 code and assessed against FEBA experiment. Better agreement between calculations and measured data for both cladding temperature and quench time is obtained.