| Abstract: | This paper describes a probabilistic fracture mechanics (PFM) computer program using the parallel Monte Carlo (MC) algorithm. In the stratified MC algorithm, a sampling space of probabilistic variables such as fracture toughness value, the depth and aspect ratio of an initial semi-elliptical surface crack is divided into a number of small cells. Fatigue crack growth simulations and failure judgements of those samples are performed cell by cell in parallel. The developed PFM program is implemented on a massively parallel computer composed of 512 processors. As an example, some life extension simulations of aged reactor pressure vessel material are performed, taking analysis conditions of normal and upset operations of PWRs. The results show that cumulative breakage probabilities of the analyzed model are of an order of 10−7 (1/crack), and that parallel performance always exceeds 90% owing to an employed function of dynamic workload balancing. It is also demonstrated that the degradation of fracture toughness values due to neutron irradiation and the probabilistic variation of fracture toughness values significantly influence failure probabilities. |
