Two materials, an Inconel 718 and a 1045 steel, are used to verify the extension of a shear strain-based parameter developed to account for out-of-phase cyclic strain hardening to multiaxial mean-stresses. Shear strain amplitude on the maximum shear strain amplitude plane and the maximum stress normal to this plane are the nominal stress-strain parameters considered in this approach. Tension-torsion and axial-internal pressure loadings using tubular specimens are employed to investigate stress-strain states that exhibit mean-strains and/or mean-stresses. Deformation response relevant to the proposed fatigue damage algorithm such as mean-stress relaxation is discussed. Adequate fatigue life correlations are obtained by implementing the proposed analysis. It is also demonstrated that methodologies successful for correlating uniaxial mean-stress data often lead to erroneous multiaxial life predictions.

No abstract

Bending and shearing stresses in many structures are out-of-phase. A typical example is the evaluation of a fixed position in a bridge beam as a moving load crosses it. Traditional fatigue design approaches for welded structural details are generally based on inphase laboratory data. Thus the effect of nonproportional stress histories on the fatigue life of welded structural details is usually not considered. Stress-relieved tube-to-plate weldments were subjected to multiaxial in-phase and out-of-phase stress states using combinations of bending and torsion. The fatigue test results were evaluated using three local stress damage criteria modified for multiaxial loading. Finite element techniques were used to calculate the local stresses and strains. The local shear-stress based approach of Findley provided the best correlation of all test data for the loading paths considered. The unique aspects of nonproportional loading are reflected by the Findley model fatigue damage calculations, and the test data were correlated to within a factor of two. Since the scatter in the experimental data was also a factor of two for identical tests, further improvements in predictive ability are not possible. Also, the predicted direction of fatigue damage for the Findley model coincided with surface cracking characteristics.

Partial coronal surface acid etch before use of flowable composite maintained a strong bond, yet resulted in easy material removal with a low complication rate.