Influence of the load level on the development of part-through cracks

Abstract

In assessing the tolerability of flaws in structural components assumptions have to be made of the crack extension during inspection intervals. The estimation of the growth of part-through cracks based on the stress intensity factor distribution K(t) along the crack front and on the growth rule A~/AN = f(~K) for fatigue cracks, however, may be in some cases erroneous. Apart from the difficulties of correctly predicting the absolute value of the crack extension A~(~) at some location on the crack front, when the results of fatigue experiments are frequently subject to considerable scatter, even the prediction of the relative crack growth A~(~I)/A~(~2) (i.e., the development of the crack shape shape) may be unrealistic, as has been shown experimentally [I-3].As the relative stress intensity factor distribution K(¢)/K(~o) does not depend on the absolute value of the load level, it might appear unnecessary to consider load level as a parameter, if examining the crack shape development only.(The dependence of the crack development on geometrical parameters -crack and specimen geometry -and on the loading mode -ratio of the tensile and bending stresses -has recently been investigated [3,4].)Contrary to these expectations fracture surfaces with artificial markings in Fig. 1 [5] show different development of surface cracks in four identical specimens under uniform tensile load. The differences were achieved due to the variation of the load level as schematically depicted in Fig. 2. With increasing values of Smean surface cracks tend to grow to shapes with increasing ratios of length to depth (specimens 1-3). The fourth specimen although monotonically loaded under displacement control seems also to fit with this tendency.The markings on the fracture surfaces were introduced by changing the basic load level for a short time (Fig. 2). The influence of this marking technique on the crack shape development was tested in further experiments and found to be negligible at the low density of markings shown in Fig. i. For higher densities of markings, however, the influence of overloads on the development of the crack shape was significant. This effect of the load level cannot easily be understood. To explain the observed development of surface cracks in plates under uniform fatigue loading Hodulak et al. [6] have suggested a simple model involving layers of higher crack resistance at the specimen surfaces. As the thickness of these layers ("plane stress" region) should depend on material constants and the load, some dependence of the crack shape development on these parameters can be expected. Obviously a similar assumption of the crack resistance distribution would apply in the case of monotonic loading.Along with a change in the crack resistance between the "plane stress" and "plane strain" regions a further mechanism that can inf]uence crack shape development can be considered. The change in the plas-

Fatigue Growth of Surface Cracks in the Elastic‐plastic Region

“…The observed effect of the increased resistance of the material close to the surface against fatigue crack growth is, however, very small compared to the measurements of Cruse et al [4] and Hodulak et al [5]. This is in accord with previous results [5] showing that an effect appears conspicious only if a high mean stress is present. The most probable reason for this effect will be the plane stress state close to the surface compared to plane strain state inside the specimen.…”

“…[2,31) and the Paris growth law can be applied. Deviations from this model are observed experimentally at a certain depth under free surface [4,5]. Dependent on loading conditions the crack runs ahead of a crack front as modelled by an elliptic shape.…”