Assessment of fracture resistance data using p-SPT specimens

Abstract

Determination of fracture properties of aged materials is important to assess degradation of structural materials subjected to in-service loads and also to assess residual life of the component. The pre-cracked small punch test is an alternative method for the determination of fracture properties in case of availability of limited quantity of materials insufficient for conducting standard ASTM tests. In this work, a simplified methodology is developed to assess J-R data of structural steels 20MnMoNi55 and T91 using pre-cracked small punch test (p-SPT) specimens. The size of the p-SPT specimen is taken as 10 × 10 × 0.5 mm. Wire EDM technique is used to fabricate specimens having through thickness crack with a/W equals to 0.4, 0.45 and 0.5. The tests are conducted to get load v/s displacement data. FE elastic-plastic analyses of the p-SPT specimen are also carried out in parallel to obtain load v/s displacement data computationally. A comparison of computed and experimental data showed that the peak loads calculated numerically are significantly higher than the experimental values. To resolve this issue, it is presumed that there is some extent of crack growth in the p-SPT specimens during experiment before the onset of peak loads. To make an assessment of this crack growth at peak load, a series of FE elastic-plastic analyses of p-SPT specimens are carried out with different crack lengths. Such FE results are then used to obtain two empirical correlations representing (i) variation of peak loads and crack length (ii) variation of CTOD with peak loads. The first empirical correlation is then used to assess crack lengths (and hence the crack growth) corresponding to experimental peak loads for all the tested specimens. Similarly, the second empirical correlation is used to calculate CTOD corresponding to the experimental peak loads. These two sets of data finally generated CTOD as a function of crack growth. Using the conventional relations between CTOD and J-integral, the data is then converted to J-R curve. This procedure is used to calculate J-R curves for 20MnMoNi55 and T91 materials. The J-R curves generated for both the materials are found to be in reasonable agreement with the experimental data reported in the literature. The methodology described in this paper has the potential to determine J-R data of aged structural materials using pre-cracked small punch test specimens.