The introduction of an overload or underload within a constant amplitude loading fatigue test leads to a retardation or acceleration of the Fatigue Crack Growth Rate (FCGR). The understanding of the causes of these effects is essential in the context of variable amplitude fatigue loading, since in principle any loading history can be represented as a sequence of overloads and underloads. In the case of overload, along with some other minor causes, the residual stress changes at the crack tip and crack closure behind the tip can be thought of as the main factors that affect the fatigue crack growth rate. Whilst this has been recognised and accepted for many decades, controversy persists regarding the relative significance and presence of these two effects, and consensus is yet to emerge. The effect of crack closure, when the baseline loading ratio is high enough, can be inhibited so that the main cause of retardation becomes doubtless the residual stress present ahead the crack tip.In the present paper we report our attempt to deconvolve the contributions of crack closure and residual stress on crack retardation following an overload. To accomplish this task we analyse the results of fatigue tests run at two baseline load ratios, namely R=0.1 and R=0.7. At the load ratio of R=0.7 the crack closure effect is not operative, as confirmed by Digital Image Correlation analysis of the crack flanks close to the tip, and post mortem fractographic analysis of crack surfaces. Therefore, for R=0.7 the compressive residual stress region created by the overload ahead of the crack tip is the sole mechanism causing crack retardation. Therefore, for R=0.7 the focus must be placed entirely on the strain field around the crack tip. To this end, line profiles along the crack bisector of elastic strain in the crack opening direction were collected at several stages of crack propagation past the overload using in situ Synchrotron X-ray Powder Diffraction (SXRPD) technique.By performing comparison between the two loading conditions (R=0.7 and R=0.1), information was extracted regarding the role of residual stress alone, and then, by subtracting this effect for the R=0.1 sample, for crack closure alone. To enable this analysis, we propose a introducing the concept of equivalent effective stress intensity factor range, ∆Keq,eff proposed by Walker. Afterwards, the SIF range reduction ratio, β, which represents the “knock down” factor with respect to the steady state growth was assessed. It is in terms of these newly introduced parameters that the magnitude and extent of the overload-induced crack growth rate retardation can be plotted, fitted and decomposed into closure and residual stress effects, respectively. It is concluded that although the residual stress effect is present at all values of the load ratio R, its effect is relatively short-lived, whilst the closure effect that is dominant at low values of R causes longer range retardation.
Read full abstract