Methodologies allowing for reliable and informative experimental error or uncertainty calculations are well documented. These methodologies allow researchers to compute the uncertainty associated with experimental results and to determine how specific experimental details contribute to the total uncertainty. While previous statistical studies within the fatigue community have focused on multiple specimen variability, uncertainty analyses will allow single specimen errors to be addressed.
Uncertainty analyses and conventional statistical methodologies were applied to high load ratio fatigue crack growth rate data generated using AA 7075-T651 compact tension [C(T)] specimens. Levels of single specimen uncertainty for the crack growth rates, stress intensity factor ranges, and crack lengths were subsequently determined. Measures of multiple specimen scatter were also statistically quantified.
Multiple specimen uncertainties were found to be greater than single specimen uncertainties. This demonstrates that quantities other than measurement error are dominating the scatter in crack propagation data. Crack mouth opening displacement measurements were observed to exert a dominating influence on the single specimen uncertainty analyses, forcing significant variations in the total uncertainty for crack length, crack growth rate, and the stress intensity factor range.
Uncertainty analyses are not commonly used in the fatigue laboratory. Uncertainty methodologies may gain more acceptance in the testing community because of the valuable measure of error that can be obtained and the numerous improvements that can be made to current procedures as a result.