Improvements to the RELAP5/MOD3 reflood model and uncertainty quantification of reflood peak clad temperature

Assessment of the original REAP/N4OD3.1 code against the FLECHT SEASET series of experiments has identified some weaknesses of the reflood model, such as the lack of a quenching temperature model, the shortcoming of the Chen transition boiling model, and the incorrect prediction of droplet size and interfacial heat transfer. Also, high temperature spikes during the reflood calculation resulted in high steam flow oscillation and liquid carryover. An effort had been made to improve the code with respect to the above weakness, and the necessary model for the wall heat transfer package and the numerical scheme had been modified. Some important FLECHT-SEASET experiments were assessed using the improved version and standard version. The result from the improved REAP/MOD3.1 shows the weaknesses of REAP/N4OD3.1 were much improved when compared to the standard MOD3.1 code. The prediction of void profile and cladding temperature agreed better with test data, especially for the gravity feed test. The scatter diagram of peak cladding temperatures (PCTs) is made from the comparison of all the calculated PCTs and the corresponding experimental values. The deviation between experimental and calculated PCTs were calculated for 2793 data points. The deviations are shown to be normally distributed, and used to quantify statistically the PCT uncertainty of the code. The upper limit of PCT uncertainty at 95% confidence level is evaluated to be about 99K