The development of crack closure during the plane strain extension of large and small fatigue cracks has been investigated in a 2124 aluminum alloy using both experimental and numerical procedures. Specifically, the growth rate and crack closure behavior of long (-17--38 mm) cracks, through-thickness physically-short (50-400 pm) cracks, and naturally-occurring microstructurally-small (2-400 p m) surface cracks have been examined experimentally from threshold levels to instability (over the range 10-i2-10-6 m/cycle). Results are compared with those predicted numerically using an elastic-plastic finite element analysis of fatigue crack advance and closure under both plane stress and plane strain conditions. It is shown that both the short through-thickness and small surface cracks propagate below the long crack threshold at rates considerably in excess of long cracks, consistent with the reduced levels of closure developed in their limited wake. Numerical analysis, however, is found consistently to underpredict the magnitude of crack closure for both large and small cracks, particularly at near-threshold levels; an observation attributed to the fact that the numerical procedures can only model contributions from cyclic plasticity, whereas in reality significant additional closure arises from the wedging action of fracture surface asperities and corrosion debris. Although such shielding mechanisms are considered to provide a prominent mechanism for differences in the growth rate behavior of large and small cracks, other factors such as the nature of the stress and strain singularity and the extent of local plasticity are shown to play an important role.

No abstract

No abstract

The influence of the very first compression cycle on subsequent Mode I crack growth in notched plates of ductile solids subject to fully compressive cyclic loads has been examined in this work. Whereas stable crack growth occurs from the root of the notch over hundreds of thousands of compression cycles, the amplitude of the very first compression cycle is found to have a decisive effect on the total distance of crack advance. Specifically, a three-fold increase in the stress amplitude during the first compression cycle alone leads to more than a ten-fold increase in the total distance of crack growth during subsequent cycles. Finite element calculations of the crack closure stress levels during cyclic compression are in concurrence with the experimental results of closure stresses (measured from changes in compliance) for a lower strength steel. A brief discussion is also presented of the effects of periodic compressive overloads on crack growth behavior under far-field cyclic compression.

The development of crack closure during the extension of long and short fatigue cracks has been investigated in a 2124 aluminum alloy using both experimental and numerical procedures. Specifically, the growth rate and closure behavior of long (∼17 to 38 mm) cracks in compact C(T) specimens and through-thickness physically-short (∼50 to 40 μm) cracks in single-edge-notched SEN(B) bend specimens have been examined experimentally from threshold levels to instability (over the range ∼10−12 to 10−7 m/cycle), and results compared with those predicted numerically using an elastic-plastic finite-element analysis of fatigue crack advance in plane strain. It is shown that the numerical analysis consistently underpredicts the magnitude of crack closure for both long and short cracks, particularly at near-threshold levels; an observation attributed to the fact that the numerical procedures can only model contributions to closure from cyclic plasticity, whereas, in reality, significant additional closure arises from the wedging action of fracture surface asperities and corrosion debris. Such closure is shown to provide the predominant mechanism for rationalizing differences in the growth rate behavior between long and physically short cracks, although other factors, such as the nature of the singularity and the extent of local plasticity, are deemed potentially to be important.