This paper presents a methodology for brittle fracture probability assessment of WWER-1000 RPVs subjected to PTS. The main attention of the study is focused on the definition of the stochastic input data: fracture toughness, CTB, size and shape of the defects.
Fracture toughness of RPV metal in the initial state is determined, separately for the base metal and the welds, as the mean curve lines of relevant experimental data. Metal embrittlement is taken into account by increasing CTB.
Based on the results of the Ukrainian surveillance specimens program, a CTB database of WWER-1000 RPVs is created. As a result of processing these data, the CTB temperature dependence on the neutron fluence and chemical composition are obtained. The probability of the fracture toughness curve position on the temperature axis is defined by the CTB standard deviation in accordance with normal distribution law.
The distribution laws of depth and shapes of defects are taken according to the NRC data and verified on the base of the statistics of the defects, which were found in the WWER-1000 RPVs of the Zaporozhye NPP.
In fact, probabilistic calculation of the RPV brittle fracture is performed similarly to the deterministic one. Fields of temperature and stresses in different RPV zones in each timepoint of considered emergency scenario are calculated by the high-effective variant of transfer matrix method in axisymmetric elastic formulation. Semielliptical axial cracks of different sizes and proportions are conservatively considered on the inner surface of RPV. SIF along the crack front is defined using the original variant of Weight Function Method. The probability of failure of a particular defect is determined as the probability that CTB of the appropriate metal reaches the maximum allowable CTB value for the scenario.
The presented method is used for the renewal of operating licenses process of Unit 2 RPV of the South Ukrainian, and Units 1 and 2 RPVs of the Zaporozhye NPPs.