THIS note describes a simple procedure that has been successfully used experimentally to determine the stress intensity factor of a geometry for which no analytically tractable solution is available. The geometry provided by this example is an internally pressurized cylinder with a through-the-wall longitudinal crack, but the technique is entirely general and may be used for practically any specimen geometry.
At the present time, there is no stress intensity factor calibration, analytical or experimental, that satisfactorily predicts K, for cylinders with relatively low ratios of radius to thickness. The solutions that do exist [l, 21 appear to be valid only for higher values of this ratio; i.e., geometries more amenable to treatment by thin shell theory. This note will emphasize the technique rather than the results, as the authors anticipate a later publication devoted to a more thorough treatment of thick-waged cylinders.
The technique employed is extremely simple and, although it is not novel, it has never enjoyed wide publication. (The authors are aware of its presentation to The Boeing Company more than 6 years ago by Paul C. Paris). The idea is based on the fact that the tip of a crack undergoing extension by fatigue "knows nothing" about the gross specimen shape or the loads imposed upon the remote specimen boundaries. It must therefore follow that, for a given material and test conditions (temperature, environment and stress range), the crack growth rate at any instant must be associated with a unique value of K, or more specifically AK. This is borne out by the correlation shown among various configurations[3] and by the several 'power laws' of crack propagation [4,5] which have the general form Compendium same thickness of material should be used in the tests to generate the d&N vs. AK data as will be used in the specimen since there appears to be some effect of thickness on the crack growth rate [3,6].