Uncertainty evaluation of reliability of safety grade decay heat removal system of Indian prototype fast breeder reactor

Highlights: • Uncertainty analysis of failure frequency of SGDHRS of a medium sized fast reactor is studied. • Lognormal distribution of failure rate of components is taken with error factor of 3. • The error factor in the distribution of failure frequency in most cases is 3. • The relative importance of the safety components is brought out. - Abstract: Deterministic and probabilistic safety assessment of nuclear power reactor technology is very important in assuring that the design is robust and safety systems perform as per requirement. The parameters required as input data for such analysis have uncertainties associated with them. Their impact is to be assessed on the results obtained for such analyses and it affects the overall decision making process. Safety Grade Decay Heat Removal System (SGDHRS) is one of the safety systems in fast breeder reactors and itremoves decay heat after reactor shutdown. It is a critical safety system; hence failure frequency for SGDHR is targeted to be less than 1.0 × 10⁻⁷ per reactor year. By bringing diversity in some of the components of SGDHRS, such as sodium-to-sodium decay heat exchanger (DHX), sodium to air heat exchanger (AHX) and valves, one can achieve the targeted low failure frequency of SGDHRS. We perform uncertainty analysis of the reliability of such SGDHRS here. Uncertainty in failure rate (of components of SGDHRS) is assumed to follow the log-normal distribution with error factor of three. Monte Carlo method of sampling is used in MATLAB environment. Results are obtained in terms of mean, median and standard deviation values of failure frequency. Percentile and confidence interval analysis of mean values are also obtained. These provide 95 and 98 percentile and confidence interval values of 98%, 99% and 99.8%. It is found that error factor of failure frequency of SGDHRS is found to be less than 3 in all the cases except the one in which DHX, AHX and Valves are designed with diversity in design. It is to be noted here that error factor of all input parameters distribution is taken as 3