PSI - Issue 13

Hans-Jakob Schindler / Procedia Structural Integrity 13 (2018) 398–403 Author name / Structural Integrity Procedia 00 (2018) 000 – 000

402

5

as shown above, it serves well as a basis to estimate FT in terms of K Ic or J Ic by empirical or semi-analytical correlations. So, it is likely that better estimates of K Ic or J Ic would be obtained by using modified CV-test. The modification should be such that they are closer to FT-tests, in terms of notch acuity, notch-tip constraints and loading rate, while keeping the simplicity of the loading and measurement system, and the straightforward evaluation of test data as far as possible. Some corresponding possibilities are roughly outlined below. 4.1. Standardized test procedure and specimens A key issue is the standardization of the test specimen geometry, loading system and test evaluation. There is no point in leaving these up to the user. With a standardized specimen, there will no longer be “invalid” tests, because any result can be used at least for correlations, lower bounds, etc. A promising test specimen appears be a three point-bend specimen, with, for example, length and thickness the same as the CV-specimen (i.e. 55 mm and 10 mm, respectively), but possibly somewhat larger in height (W = 15 mm instead of 10 mm) and an EDM cut of about half the cross section instead of a machined V-notch of only 2 mm depth. This is associated with larger in-plane constraints and a sharper notch (radius  = 0.1 mm instead of 0.25 mm). There is no need to pre-fatigue the specimens, since the corresponding effect can be accounted for by appropriate analytical or empirical correction terms, as shown by Schindler (2014a). Such a specimen serves well in the upper shelf as well as in the DBT-range, as discussed below. Its use on the lower shelf would surely be possible, too, but not further discussed here, because this regime is of minor importance. 4.2. Testing and evaluation on the upper shelf As mentioned above, estimates J Ic or K JIc from the upper shelf CV impact energy is usually sufficient for the purpose. But there can be a need for more accurate near-initiation, e.g. in material development. In order to obtain quasi-static K Ic and J I , a specimen as suggested in section 4.1 should be loaded quasi-statically and monotonically in displacement-control. As shown by Schindler (2000), at the instant of maximum force or somewhat beyond, the stable crack extension by ductile tearing,  a s , has reached an amount  a se that is suitable for estimating the J-R curve, so the test should be stopped there. From the J-integral J e at this stop of the test and  a se measured subsequently on the fracture surface after having broken it fully apart, the J-R-curve in the form of J = C ∙  a p is readily determined as p The exponent p is given in eq. (2). From (4), near-initiation J Ic values can be determined by the intersection with the 0.2 mm-offset blunting line according to ASTM E1820 or ISO 12135. To account for the finite notch root radius of the EDM-cut, a correction term as suggested by Schindler et al. (2014) has to be subtracted from the value of J at the intersection. 4.3. Testing and evaluation in the transition regime It is a fundamental misconception of ASTM E 1921 that a valid T 0 can only be obtained from a series of “ valid ” K Jc tests. Actually, the finding that a single reference temperature is sufficient to characterize FT in the transition regime implies that T 0 can be determined from any test that exhibits cleavage, and that any suitably defined reference temperature from this arbitrary test, denoted in the following as T x , is related to T 0 by a relation of the form T 0 = T x –  T 0//x (5) For a SENB specimen as suggested in Sec. 4.1, there are many possibilities to define T x , for example as the temperature where the energy consumed up to unstable cleavage reaches 50% of the energy at maximum force in an upper-shelf test, or by a certain level of J. In contrast to the procedure according to ASTM E1921, no validity criteria are required, except that a cleavage fracture has to be initiated without a significant amount of prior ductile tearing. As shown by Schindler (2014) the scatter band in the K Jc -vs.-T-plane is crescent-shaped. Its horizontal width decreases with increasing K Jc -level and becomes minimum at the transition to upper shelf. Thus, if the reference temperature T x is defined at a K Jq -level at the transition to upper shelf, then the measurement uncertainty of T x se e a J a J   ( ) a          (4)