Font Size: 

A series of in-phase axial-torsional smooth tube fatigue experiments wereperformed using regular intermittent overloads in otherwise constant amplitude load historiesto achieve crack-face interference-free (crack closure free) fatigue crack growth.Observations were made of not only the fatigue life but if and when the initial crackgrowth on a plane of maximum shear stress range changed to crack growth on a planeof maximum tensile stress range. Although there was a considerable scatter in the strainand length at which changes in crack growth mode occurred, the general trend in thedata was that the crack length at which change occurred increased with increasing strainand strain ratio. Two separate criteria for crack growth mode change from shear planeto tensile plane crack growth were inserted into a strain based short crack fatigue crackgrowth model. When the crack length at the change in mode was predicted using the firstcriterion (based on choosing the plane that exhibited the maximum crack growth rate)the boundary, which is formed on a strain versus crack length plot, fell at lower strainsthan the data. Use of the other criterion (based on choosing the plane with the higherstrain energy release rate) yielded a boundary that indicated shorter than observed cracklengths and an upper bound to the strain at which changes in mode were observed. Fatiguelives predicted using the two forms of the model fell very close to each other and tothe experimental fatigue life data. The closeness in the life predictions produced by thetwo forms of the model and the scatter in observed strain and crack length at the pointof mode change are assumed to be a consequence of nearly equal crack growth rates fortensile and shear mode crack growth.