SAFE-CASK first-year activity: Methodology to relate failure probabilities of cast iron for deep disposal canisters to manufacturing defects

This report summarizes the contribution of JRC - IE to the analysis and testing of the cast iron insert I26, which has been manufactured as part of the Swedish KBS -3 programme. The main objective has been to relate scatter in elongation at fracture for tensile specimens to presence of casting defects. This has done by performing tensile tests, non - destructive evaluation of specimens by radiography before tests and fractographic post - test analysis. A clear relation between elongation at fracture and defects was established by comparing the measured elongation at fracture and presence of defects for individual specimens. A simple probabilistic model based on fracture mechanics using distributions of elongation at fracture and fracture toughness is presented. This model is then used to predict defects distributions as well as distribution of elongation at fracture in specimens. An important conclusion is that the variation in fracture toughness has a very small effect compared to the scatter in defects. Because the model assumes that failure is controlled solely by a crack - like defect, the model over - predicts the size of the defects. The report also describes an investigation to assess the transferability between specimen and full-sized canister a pressure a mock-up was designed and loaded in a cold - isostatic press. The mock - up was loaded to 1350 bars, which corresponds to three times the pressure exerted during glaciation, and the mock -up retained its overall integrity. (authors)