Multiaxial variable amplitude fatigue life analysis using the critical plane approach, Part II: Notched specimen experiments and life estimations

- Multiaxial variable amplitude fatigue behavior studied for notched specimens.
- Modified Fatemi-Socie parameter and critical distance approach used for analysis.
- Life estimation trends were consistent between notched and un-notched specimens.
- Linear damage sum of 0.3 improved life estimations compared to a value of 1.
- Most fatigue lives estimated within a factor of 3 of experimental results.

While part I of this paper was focused on evaluating multiaxial variable amplitude fatigue life estimations for un-notched specimens, part II extends the same critical plane-based analysis procedures to situations involving notched specimens. In addition to the factors considered in the un-notched analyses, local stress concentrations, stress gradient effects, and changes in local stress state must also be accounted for in the presence of a notch. This was accomplished in the current study by coupling the Theory of Critical Distances point method with a pseudo stress-based plasticity modeling technique. Then, a modified version of the Fatemi-Socie parameter was used to calculate fatigue damage, and changes in life estimation accuracy were studied with respect to the consideration of transient material deformation behavior, crack initiation definition, and damage summation rule. Results from the notched specimen analyses were also compared to those for un-notched specimens, and some discussion is provided. While the effect of transient deformation behavior and crack initiation definition were found to be relatively small for the loading histories used in this study, changing the critical damage sum at failure had a much greater impact on life estimations. Although some of the analysis procedures investigated were able to estimate nearly all fatigue lives within a factor of 3 of experimental results, several areas were identified where there is potential for even further improvements to be made. These include issues related to the accuracy of life estimation curves, damage calculation models, and/or the modeling of material deformation behavior.