Page 379 - Biaxial Multiaxial Fatigue and Fracture

The Environment Effect on Fatigue Crack Growth Rates in 7049 Aluminium Alloy at ___ 363

the stress intensity amplitude AK due to premature contact of crack surfaces. Vasudkvan and
Sadananda claimed, however, that plasticity in the wake of a crack has a minor influence on
crack closure [33-351 and argued that fatigue propagation in vacuum does show load-ratio
effects. It has experimentally been observed that da/dN near the threshold up to higher growth
rates is nearly independent of the R-ratios in a high vacuum [36,37], which leads to conclude
that plasticity does not induce closure. They found that crack closure effects were less
important when compared with microstructure and environment effects and do not influence
the overall crack growth behaviour. Thus, crack growth rate data generated in laboratory
should not be used to predict crack growth rates in a structure unless one has measured crack
closure of the structure in service, which depends on the component geometry, load,
environment, and crack length.
Since crack closure is usually assumed as one of the major causes for retardation effects,
the examination of the magnitude of crack closure and its relative role in crack growth
processes is very important. The existence of plasticity-induced crack closure has been
severely questioned by Vasudevan and Sadananda [33-351, while asperity-induced closure,
which includes roughness due to crack tortuosity, oxide or chemical reaction debris, etc. is
more accepted. Paris [38] have provided a theoretical justification for modified criteria for
crack closure. According to Sadananda et al. all methods used to date for measuring crack
closure, tend to overestimate the crack closure levels. Therefore they proposed a crack closure
measurement based on the shape of the load-displacement-curve [39].
Keeping in mind that the role of environment in the near-threshold regime is strongly more
significant than any mechanical contributions such as plasticity, roughness, oxide, etc. effects,
one may conclude according to Vasudevan and Sadananda that AKth decreasing with R is an
intrinsic fatigue property of the material for that environment.
Overload effects have predominantly been attributed to either plasticity induced crack
closure behind the crack tip, residual stresses ahead of a crack tip, or a combination of both.
The mechanisms such as crack tip blunting, crack deflection, branching and secondary
cracking, as well as crack tip strain hardening or residual stresses ahead of the crack tip,
involve mainly transient conditions at or ahead of the crack tip [40]. Mechanisms such as
plasticity-induced closure and roughness-induced closure operate behind the crack tip
indirectly affecting the crack driving force.
Vasudevan and Sadananda presented an 'Unified Approach to Fatigue" which considers
closure as a minor factor for crack advance [40]. They modelled fatigue crack propagation for
a wide variety of materials by assuming two stress intensity parameters, AK and Kmax, as the
relevant crack tip driving forces. They showed that for most situations, the description in
terms of AK and Kmx is necessary and sufficient for fatigue crack growth without the need of
crack closure. According to this Unified Approach to Fatigue, K,, and AK are two intrinsic
parameters simultaneously required for quantifying fatigue crack growth data. These two
driving forces are intrinsic parameters of each material and are valid for short or long cracks,
having no anomalies between these two regimes [41]. The apparent different behaviour of
short and long cracks is related to residual stresses. They claim that the residual stress cannot
be crack closure. Crack closure exists only behind the crack tip and is induced by roughness,
oxides, plastic deformation, etc. Crack tip plasticity can only produce compressive residual
stresses at the crack tip and not crack closure effects according to Vasudkvan and Sadananda
as consequence of overloads for instance. The two parameters AK and Kmx lead to two
intrinsic thresholds that must be simultaneously exceeded for a fatigue crack to grow [41].
Crack retardation owing to residual stresses ahead of the crack tip is reported in [42,43]. Ling
and Schijve [44] confirmed that residual stresses play a major role in the retardation by
demonstrating that the effects can be eliminated by annealing after the overloads.