| Article ID | Journal | Published Year | Pages | File Type |
|---|---|---|---|---|
| 776533 | International Journal of Fatigue | 2015 | 8 Pages |
•Fatigue life prediction of a dumbbell cylindrical rubber component was performed.•Fatigue life and normal direction angle of crack growth plane were obtained.•We establish the relation between exponent F and R ratio in Mars–Fatemi model.•Fluctuation of fatigue life is due to the crack precursor size inhomogeneity.•We confirm the agreement of precursor sizes from relaxing and non-relaxing tests.
Fatigue life prediction for a dumbbell cylindrical natural rubber component under uniaxial tensile loading conditions was performed based on the Thomas fatigue crack growth model for relaxing (R = 0) load cycles and the Mars–Fatemi model for non-relaxing (R > 0) load cycles. By using a self-written program, we proposed a new approach to establish the relation between the power law exponent F and the R ratio in the Mars–Fatemi model. The approach is based on rubber fatigue life (S–N) data rather than crack growth rate and tearing energy (da/dN–T) data, avoiding certain difficulties often encountered using the crack growth method. The results indicate that the relation between F and R is a quadratic or cubic function over the range 0 < R < 0.3. Finally, the quantitative effect of initial crack size on fatigue life was studied. We found that the inferred mean size of crack precursors in the rubber component is around 30–40 μm under both relaxing and non-relaxing loading conditions, and the fluctuation of fatigue life is due to the inhomogeneity of crack precursor size except the factors such as unavoidable variations in testing conditions and specimen variations. The good agreement of inferred crack precursor sizes from different R ratio loading conditions is a strong indication that the Mars–Fatemi model provides a proper accounting for the effects of strain crystallization, and it confirms yet again the understanding that nucleated cracks originate from similarly sized precursors in both relaxing and non-relaxing fatigue experiments.
