Statistical and Probabilistic Analysis of Fatigue Life Data With Two- and Three-Parameter Weibull Distribution Functions

Product durability and reliability validation requires fatigue testing of limited samples and the corresponding statistical and probabilistic analyses of the test data. The uncertainties introduced into the tests with limited sample size and the assumptions made about the underlying probabilistic distribution function will significantly impact the analysis results and the result interpretation. In this paper, a large amount of fatigue life data collected from several programs are analyzed. The values of the parameters estimated from the two-parameter and three-parameter probabilistic distribution functions, Weibull in particular, are compared. Finally, the observations are summarized and some recommendations are provided.