Reliability of an HTR-module primary circuit pressure boundary : influences, sensitivity, and comparison with a PWR

Abstract The reliability of the HTR-module for electricity and steam generation was analysed for normal operation, as well as accident conditions. The probabilistic fracture mechanics assessment was performed with a modification of the zerberus code on the basis of widely used data. The calculated failure probabilities may thus be compared with similar investigations. The HTR-module primary circuit pressure boundary as a unit showed leak-before-break behaviour in a probabilistic sense, although a break was more probable than a leak for some of its parts. However, the findings may depend greatly on the stochastic data. Therefore a stochastic reference problem is defined and the results are compared with the Japanese round robin on a PWR section. Possible changes of failure probabilities and of the leak-before-break behaviour are discussed for different criteria for the events leading to a leak, and for modifications of the stochastic reference problem such as the inclusion of NDE. The results may be used to identify those stochastic variables which have the greatest influence on the computed failure probabilities, and to perhaps justify further work which would provide more detailed information on these probabilities. Furthermore, there is an obvious need for reduction of the non-statistical reasons for great variations of failure probabilities.