| Abstract | This paper presents a probabilistic short crack growth model to estimate the fatigue life distribution for bare 2024-T351 coupons. Fractographic analysis has clearly shown that the majority of cracks nucleate from the constituent particles present in bare 2024-T351 coupons. The existing fracture mechanics models have difficulty in 'growing' a small-particle-induced crack (i.e. infinite life), and cannot correlate the size of a crack-nucleating particle to the length of the fatigue life in cases where a small (large) particle results in a short (long) life. In this probabilistic model, particle width and height distributions were first used to account for the randomness of the particle size. Then another random variable, the stress intensity factor limit, ΔKIDS, for particle-induced cracks, was introduced to account for the combined effect of microstructure features (e.g. grain size, grain orientation, and grain boundary) on short crack growth. A Microsoft Excel VBA (Visual Basic for Applications) program was developed and linked with the AFGROW COM (Component Object Model) server to estimate the life distribution using the Monte Carlo technique. The results showed that by including the ΔKIDS for particle-induced cracks, the model was able to 1) correlate a crack-nucleating particle to the individual fatigue life, including "smaller (larger) particle, shorter (longer) life" cases; and 2) better estimate the fatigue life distribution. |
|---|
