From fractographic observations of specimens that have failed due to rolling contact fatigue, it has been concluded that the first stage of damage is the formation of mode 11 fatigue cracks parallel to the contact surface due to the cyclic shear stress component of the contact stress. Although these initial subsurface cracks, in both metals and ceramics, are produced in a direction parallel to the cyclic shear stress, cracks eventually grow in a direction close to the plane of the maximum tensile stress if we apply a simple mode I1 loading to them. The difference between crack growth in simple mode I1 loading and crack growth due to rolling contact fatigue is, we suppose, whether or not there is a superimposed compressive stress. Based on this hypothesis, we developed an apparatus to obtain the intrinsic characteristics of mode I1 fatigue crack growth, and developed a simplified model of subsurface crack growth due to rolling contact fatigue.Some results in terms of da/dN versus AKII relations have been obtained using this apparatus on specimens of steel and aluminum alloys. Fractographs of the mode I1 fatigue fracture surfaces of the various materials are also provided.

No abstract

Static and cyclic fatigue tests were made on glass, glass ceramics and silicon nitride under static tensile loading and tension-compression cyclic loading at various stress ratios between R=-1 and 1, by using an apparatus specially designed with a device to minimize the load eccentricity.Crack growth tests were also made under static loading and cyclic loading in order to obtain the crack growth rate da/dt, as a function of stress intensity factor K. Test results on smooth round bar specimens show that almost no effect of stress cycling was seen on silicon nitride A and borosilicate glass, while appreciable degradation in strength due to cyclic loading was observed on glass ceramics and silicon nitride B. The relation, da/dt versus K, of small cracks obtained from the tensile test on smooth round bar specimens showed higher values than those of large cracks obtained from the crack growth tests. The increase in crack growth rate due to cyclic stress was observed in both silicon nitride B and A, in the latter of which no effect of stress cycling was observed in the tests on smooth round bar specimens.

No abstract