The Uncertainty Analysis of Natural Circulation in Molten Salt Reactor Coupling Dakota With GenFlow

Molten salt reactor is the only one of the liquid fuel reactors in the fourth generation reactors. The flow characteristics of the fluoride makes differences between the liquid fuel reactor and the solid fuel reactor. For licensing of novel design of molten salt reactor, the peak temperature of the molten salt in primary loop has significant influence on the reactor safety. Sensitivity analysis for the peak temperature during pump coast down accident is performed by coupling Dakota and Genflow code. Dakota is a multilevel parallel object-oriented framework for design optimization, parameter estimation, uncertainty quantification, and sensitivity analysis. Dakota generate a plurality of parameters of input files, can greatly simplify the manual optimization process, it will reduce a lot of uncertainty analysis time. GenFlow develop the technical basis for design and licensing of fluoride-salt-cooled, high-temperature reactors, simulate the accident condition and export the instantaneous state. The work coupled Dakota and GenFlow to do the uncertainty analysis, the experience also have important inspiration and reference to coupling Dakota with other code. The work aims at the scenario when the pump coast down and the natural circulation is established. Five key thermal parameters, namely the power load factor, coast down time, the check valve open time, total form losses in the DRACS loop, total form losses in the DHX branch, that affect the natural circulation in the molten salt reactor is sampled within a certain range using the LHS (Latin hypercube sampling) method. The Dakota generate multiple parameters of the input card that are run by GenFlow code automatically. The post-processing results turn back to Dakota to do the uncertainty analysis and sensitivity analysis, which greatly simplify the optimization process and do not need exact mathematical model. It shows the correlativity about the peak temperature with every parameters and the influence between the parameters. Statistical results showed that the effect of the power load factor on peak temperature is very significant, followed by the DHX shell side. After the moment-related statistics, the 95% confidence intervals on the mean and standard deviations are printed. There are 95% confident that the true value of the parameter is in confidence interval between 809.547 and 816.3227. In this sample, 15% of the actual calculation results is higher than the critical point 816°C.