Large sampling uncertainty is generally introduced in the calculation of a low percentile of fatigue crack growth life due to a small number of coupon tests. It is often desirable to estimate a low percentile (for example, 10th percentile) with a certain coverage probability (for example, 95%) using the confidence bound approach. An equally competing objective is not to have overly conservative bounds.The performance of two bootstrap-based methods are investigated for calculating a one-sided 95% confidence bound on a low percentile of lognormal and gamma fatigue crack growth life distributions. A comparison is also made with the classical tolerance interval method and nonparametric (Hanson–Koopmans) method. These confidence bounds methods are tested to estimate -basis fatigue crack growth design life using material properties estimated from coupon test samples ranging from 8 to 64.
Comparison of Methods for Calculating -Basis Crack Growth Life Using Limited Tests
AIAA Journal ; 54 , 4 ; 1287-1298
2016-03-10
12 pages
Aufsatz (Zeitschrift)
Elektronische Ressource
Englisch
Comparison of Methods for Calculating
Online Contents | 2016
|Calculating 3-D Crack-Opening Behavior of a Fatigue Crack
Online Contents | 1995