Multiaxial fatigue crack growth in rail steel.

In the introduction to a recent symposium on rolling contact fatigue,
R.A.Smith stated that it was difficult to apply our greatly increased
understanding of metal fatigue, to rolling contact fatigue, because of "the
apparent lack of alternating tensile stresses to drive the cracks." He went
on to say "alternating shear stresses are easily found, but the
reproduction of continuous crack growth controlled by shear <Mode II in
fracture mechanics terms), has proved to be near impossible." This project
has demonstrated that under specific conditions this mode of growth does
occur.
The project began by studying rolling contact fatigue defects, in
particular the 'squat' defect in railway lines, and the stress analyses that
have been performed on them. It was concluded that the largest stress
cycle experienced by the cracks must be a shear stress. It. series of tests
were then performed that loaded a crack in pure shear, or a mixture of
tension and shear, looking at the effects of using fully reversed shear
loading, and the effects of applying tensile mean stresses to reduce the
friction on the crack flanks. However these tests all produced less than
one millimetre of mode II growth, before the cracks arrested or branched.
The final series of tests however applied a tensile load cycle before
each shear load cycle. This time coplanar growth was produced, that is the
crack grew in the direction of the maximum shear stress. This type of load
cycle is a simplification of the load cycle calculated by Bower and Johnson
of Cambridge University, where the tensile load is produced by fluid
trapped in the crack.
Two crack growth rate formulae were produced that fitted the data,
indicating that the growth rate was dependent on both the tensile and the
shear parts of the cycle.