Application of the Monte Carlo thermal design analysis to evaluate uncertainties of the PWR core using the THALES subchannel code

In order to maintain the safety of the reactor core, the minimum DNBR (Departure from Nucleate Boiling Ratio) in the PWR (Pressurized-Water Reactor) core remains higher than the DNBR limit during Condition I and II events. Therefore, it is important to adequately evaluate the thermal performance of the PWR core. To realistically evaluate the relationship among the uncertainties and reduce the conservatism resulting from the unknown phenomena, the Monte Carlo method is being used in many areas requiring the statistical approach. Especially, the Monte Carlo method is drawing attention as the method for the evaluation of the thermal performance of the PWR core. For the best estimate evaluation of the uncertainties in the PWR core, KEPCO Nuclear Fuel (hereinafter KEPCO NF) has been developing the thermal design analysis based on the Monte Carlo method. For the Monte Carlo thermal design analysis, various studies are conducted as follows. To generate the Gaussian random numbers, Gaussian random number generators are investigated. In this paper, Box-Muller, Polar, GRAND, and Ziggurat method are briefly reviewed. The random numbers are generated on the basis of the nominal value and uncertainty of the parameter. If the normal distribution is acceptable at 5% significance level through the normality tests, the random numbers are used for the Monte Carlo thermal design analysis. Using the subchannel code THALES (Thermal Hydraulic AnaLyzer for Enhanced Simulation of core) developed by KEPCO NF, the subchannel analyses are carried out considering the core operating parameters randomized, and then DNBR distribution is derived. Finally, if the DNBR distribution is statistically combined with the uncertainties of the other parameters, the DNBR_T distribution can be obtained. From the DNBR_T distribution, the DNBR limit is determined to avoid DNB (Departure from Nucleate Boiling) at a 95% probability at a 95% confidence level. Through the example calculation, it is verified that this analysis produces reasonable results. If more studies of this analysis are done, the Monte Carlo thermal design analysis is useful to determine the DNBR limit. (author)