Predicting the total fatigue life in metals

Abstract When a metal specimen is subjected to cyclic loading, a large number of initiated cracks will initiate in its volume. The specimen forms a sample of initial cracks: the larger specimen, the larger is the sample. In previous work of the author it was shown that the fatigue limit can be predicted by estimating the largest expectable crack depth with the help of statistics of extremes. In this paper it is shown that the fatigue crack initiation time above the fatigue limit can be predicted in an analogous manner. Instead of estimating the largest crack size with the distribution of maxima, prediction of the shortest expectable initiation time is obtained using the distribution of minima. Good agreement with extensive set of experimental data was obtained. The presented method offers a new way for estimating the total fatigue life of a component. When estimates of the crack initiation life and the critical crack size are obtained, the stable crack growth can be computed using Paris law. The estimate of the total fatigue life is obtained as the sum of initiation and crack growth lives. A method for constructing design curves for finding the crack initiation life for any material is presented.