The effect of crack growth retardation when comparing constant amplitude to variable amplitude loading in an aluminium alloy

It has often been observed that the growth of short fatigue cracks under variable amplitude (VA) cyclic loading is not well predicted when utilising standard constant amplitude (CA) crack growth rate/stress intensity data (da/dN v δK). This paper outlines a coupon fatigue test program and analyses, investigating a possible cause of crack growth retardation from CA-only testing. Various test loading spectra were developed with sub-blocks of VA and CA cycles, then using quantitative fractography (QF) the sub-block crack growth increments were measured. Comparison of these results found that, after establishing a consistent uniform crack front using a VA load sequence, the average crack growth rate then progressively slowed down with the number of subsequent CA load cycles applied. Further fractographic investigation of the fracture surface at the end of each CA and VA sub-block crack growth, identified significant crack front morphology differences. Thus it is postulated that a variation or deviation from an efficient crack path is a driver of local retardation in short crack growth during CA loading. This may be a source of error in analytical predictions of crack growth under VA spectra loading that may need to be considered in addition to other potential effects such as less closure whilst cracks are small. For aircraft designers, using solely CA data for fatigue life predictions this may result in non-conservative estimates of total crack fatigue life, producing unexpected failures or an increased maintenance burden.