Random Outcome and Stochastic Analysis of Some Fatigue Crack Growth Data

Fatigue crack propagation data of a batch of AISI 4340 steel specimens are released in the present paper. The statistical nature of the data is specially emphasized, and a probabilistic fracture mechanics model is introduced to analyze the data. The stochastic differential equation associated with the model is then solved. The solution gives us the crack exceedance probability as well as the probability distribution of the random time to reach a specified crack size. These quantities are useful in the reliability assessment of structures made of the tested material. Comparing the analytical result with the experimental result, it is found that the proposed probabilistic fracture mechanics model can reasonably explain the experimental data. For those data that cannot be fitted well by the proposed model, methods of improvement are proposed in the present paper as well.