Specimen Curvature and Size Effects on Crack Growth Resistance From Compact Tension Specimens of CANDU Pressure Tubes

Abstract

CSA Standard N285.8 requires leak-before-break and fracture protection for Zr-2.5Nb pressure tubes in operating CANDU reactors. In-service deuterium uptake causes the formation of hydrides, which can result in additional variability and reduction of fracture toughness. Pressure tube fracture toughness is assessed mainly through rising pressure tube section burst tests. Given the length of the ex-service pressure tubes required for burst testing and the requirement to increase the hydrogen content of irradiated ex-service pressure tubes, only a limited number of burst tests can be performed. Using small-scale compact tension, C(T), specimens are advantageous for obtaining a statistically significant number of fracture toughness measurements while using less ex-service pressure tube material. This work focuses on the study of C(T) geometry designs in order to obtain crack growth resistance and fracture toughness closer to those deduced from burst tests. Because C(T) specimens must be machined from pressure tubes of about 100 mm in diameter and 4 mm in wall thickness, they are out-of-plane curved. As well, they undergo significant tunnelling during crack extension. These two factors can result in a violation of the ASTM standard for fracture toughness testing. The current work examined the influence of specimen curvature and tunnelled crack front on the crack growth resistance curve, or J-R curve. Finite element (FE) models using stationary and growing cracks were used in a detailed numerical investigation. To capture crack tunnelling in the FE models, a damage mechanics approach was adopted, with the critical strain to accumulate damage being a function of crack front stress triaxiality. The J-integral numerically estimated from the domain integral approach was compared to the J-integral calculated from the analytical equations in the ASTM E-1820 standard. In most cases, the difference between the numerical and the standard estimations was less than 10%, which was considered acceptable. It was found that at higher load levels of load-line-displacement, specimen curvature influenced the J-integral results. Crack tunnelling was shown to have a small influence on the estimated J-integrals, in comparison with the straight crack fronts. A modest number of experiments were carried out on unirradiated Zr-2.5Nb pressure tube material using three designs of curved C(T) specimens. It was found that the specimens of both designs that featured a width of 34 mm had more than twice the crack extension of the specimens of the 17-mm width design. The 17-mm width specimens are used mainly to assess the small-scale fracture toughness of pressure tube material. Additionally, the applied J-integral at the maximum load was about 1.4 times higher for the larger-width C(T) specimens. These C(T) specimens also produced J-R curves with greater crack extensions, which were closer to those obtained from the pressure tube section burst tests. Artificially hydrided pressure tube material was not considered in the current work, to avoid any potential source of experimental variability; however, it should be considered in future work.