A Fatigue Crack Growth Simulation Algorithm in Consideration of Parametric Variability

To describe fatigue crack growth, Paris law or other equivalent but more complicated laws or models are frequently used. The parametric values of a model are, in general, obtained from fitting the experimental data and presented in an average sense. Sometimes, they may not be sufficient to depict the real fatigue crack growth trend, especially when scattering of the crack growth curves are apparent. To remedy this shortcoming and to study the fatigue reliability in phases of crack growth, a computer simulation algorithm is proposed in the present paper. The algorithm takes not only the average values of model parameters but also their statistical variability into consideration. The validity of the model is studied in detail and examined through comparison made between simulation and experimental results. Both constant-amplitude and random loading conditions are considered. In conclusion, it is found that the simulation algorithm can depict the experimental fatigue crack growth curves very well. Quantitative fatigue reliability with regard to the studied material can be evaluated based on the model.