Search Results

A procedure for the computation of fatigue life of shafts subjected to variable amplitude independent bending and torsion loading is described. The elastic superposition technique, followed by a Neuber plasticity correction, also allows for the initial and possibly cyclic plasticity dependent residual stress states caused by induction hardening or other surface altering processing effects. The present study documents the formation and alteration of residual stresses caused by initial induction hardening and followed by a straightening process. Sample calculations are presented for two critical finite elements of a prototype shaft, and lives are predicted using a traditional equivalent axial stress and by a new procedure that searches for and computes lives at critical angles.

In design for durability, it is generally believed that a compressive mean stress is beneficial and a tensile mean stress is detrimental. Quantitatively the effect of mean stress on fatigue life however is still inconclusive and may very well depend on both the material used and the loading conditions. Over the years, many models have been proposed to help predict mean stress effects. For example, in the long life region, Goodman's formula is widely used, while the Smith-Watson-Topper damage parameter seems the most popular for use in computerized local strain based fatigue tools. In this paper, several frequently cited mean stress correction methods together with the most recent crack closure based method are compared in various ways. Particular emphasis is given to the effect of yield level mean stresses, which has been traditionally neglected but is of practical importance to the ground vehicle industry.

A large number of published constant amplitude test results, in-phase as well as out-of-phase, on SAE 1045 steels are analysed here in an attempt to evaluate the general applicability of several damage parameters to complex multiaxial cyclic loadings. The critical plane approach is adopted and the calculated results from various constant amplitude tests are directly compared on a damage parameter versus fatigue life plot. It is found that while many damage parameters can predict fatigue life for in-phase constant amplitude loadings within a factor of two of the observed life, only two biaxial type of parameters evaluated here can consistently produce acceptable fatigue life prediction for out-of-phase constant amplitude loadings.

The formulation of an isotropic-kinematic hardening model with a continuous yield surface field is briefly reviewed. Particular emphasis is focused on the redistribution of the yield surface field as a result of plastic loading. It is demonstrated that successive yield surface field redistributions can be concisely recorded by tracking the positions of a very small number of yield surfaces even under a complicated loading history. A FORTRAN program is then developed based on the model with the simple von Mises type of yield criterion. Applications to analysis of some biaxial tests on thin-walled tubes of SAE1045 steels show good agreements between the predicted and measured stress results.

Variable amplitude torsion, bending, and combined torsion and bending fatigue tests were performed on an axle shaft. The moment inputs used were taken from the respective history channels of a cable log skidder vehicle axle. Testing results indicated that combined variable amplitude loading lives were shorter than the lives of specimens subjected to bending or torsion alone. Calculations using strain rosette readings indicated that principle strains were most active around specific angles but also occurred with lesser magnitudes through a wider angular range. Over the course of a biaxial test, cyclic creep narrowly limited the angles and magnitudes of the principal strains. This limitation was not observed in the calculated principal stress behavior. Simple life predictions made on the measured strain gage histories were non-conservative in most cases.