K-calibration and advantages for fatigue testing of single edge-notch tensile specimens with restrained ends

Abstract

In studies of fatigue crack propagation, it is desirable that the portion of the width over which crack growth measurements are feasible be as large as possible and that K not vary strongly with crack length for a constant load. A long specimen can also be desirable since the space available on both sides of the crack plane facilitates the utilization of a travelling microscope to follow the crack tip, the positioning of displacement gauges on the sides of the specimen in order to measure crack opening, or the fixing of an environmental chamber to the specimen. From these points of view, the single edge-notch (SEN) tensile specimen and the central notch (CN) specimen are more interesting than the more often employed compact tension (CT) specimen. Of course, both the SEN and CN specimens require larger loads for given values of a/W and K than the CT specimen. In comparison to the SEN specimen, the CN specimen presents the practical difficulty of necessitating the initiation and propagation of two symmetrical cracks, which can lead to an important percentage of rejects. The major difficulty encountered with SEN tensile specimens is that the non-symmetrical line of action of the applied load with respect to the non-cracked ligament can, depending on the loading or gripping conditions employed, lead to modifications in the stress distribution. This results in uncertainty in the true value of K obtained. The loading or gripping conditions employed with SEN specimens can be assumed to lie between two extremes. At one limit, the specimen is pin-loaded, employing pins which are considered to be frictionless, so that complete freedom of rotation of the ends of the specimen can be assumed. At the other limit, rigidly fixed grips are employed and are assumed to completely restrain the ends of the specimen and prevent any rotation.