Towards a Proper Understanding of Fatigue Crack Growth and Crack Closure

Abstract

Fatigue is the weakening of a material or structure due to cyclic loading and unloading. As such fatigue is very important in engineering fields like aerospace, where structures are exposed to cyclic loads. Therefore theories on the initiation and propagation of fatigue have been extensively researched In the past century. The main Fatigue Crack Growth (FCG) prediction models are based on a Stress Intensity Factor (SIF), ΔK. Representing FCG data as a function of this SIF causes a stress ratio (R-)effect, which is accounted for by means of plasticity induced crack closure. Crack closure causes the crack to close before a zero tensile load is applied, as such influencing the effective SIF. If FCG data is presented as a function of the effective SIF ΔKeff instead of ΔK, the data correlates very well and the R-effect disappears.
However, the theories based on a SIF and crack closure have recently been subject of discussion. The prediction models using ΔK for similitude are empirically derived and thus do not have any physical explanation. Furthermore ΔK has been derived for quasi-static loading conditions and as such it cannot be blindly adopted for fatigue loading conditions. Besides there is a lot of confusion about the phenomenon (plasticity induced) crack closure, load-displacement observations attributed to crack closure could for instance also be attributed to residual compressive stresses. Especially early test results in vacuum environment do not agree with theories based on SIF and crack closure.
An alternative proposed in literature is to approach fatigue from a Strain Energy Release (SER) perspective. This theory approaches fatigue conform the laws of thermodynamics, and claims that not the amount of energy released under quasi-static load conditions should be considered, but the energy released during a complete fatigue load cycle. It also claims that the R-effect is only an artefact of choosing ΔK for similitude. Treating fatigue as a SER dominated phenomenon instead of a SIF dominated phenomenon, might result in a proper description, prediction and understanding of fatigue.
As such the first goal of this research was to investigate if ΔK is a correct similitude parameter for FCG. This was investigated by comparing experimental results in air and vacuum. According to conventional fatigue theories based on SIF, there should be no difference between air and vacuum. Experiments were designed and similar test conditions were applied with the only difference the environment tested in. Experimental results in air showed -as expected- a clear R-effect that could be accounted for by the plasticity induced crack closure corrections proposed in literature. However, the results in vacuum did not show this R-effect, while the results of crack opening experiments and plasticity did not differ compared to results of experiments in air. It was therefore concluded that ΔK is improper to use as similitude parameter for FCG prediction...