Influence of the distribution function on the statistical scatter adjustment of the number of cycles to failure of concrete under fatigue

Abstract

Experimentally determined number of cycles to failure of concrete specimens exhibit significant scatter due to various influencing factors. Previous studies have identified the scatter in compressive strength as a primary factor affecting the scatter of the number of cycles to failure. To address this issue, a statistical approach was applied to correct the stress levels to account for the influence of the compressive strength scatter. This correction significantly reduced the scatter in the number of cycles to failure, improving the reliability of fatigue assessments.

A database of fatigue tests from the literature was compiled, and the stress levels were corrected, resulting in a significant reduction in the scatter of the number of cycles to failure. A normal distribution was assumed for the logarithmic number of cycles to failure. In addition to the normal distribution, the log-normal distribution and the Weibull distribution were also found to be suitable. The choice of distribution function had only a minor influence on the corrected stress levels. This is shown by the comparison of confidence intervals. These findings contribute to a more precise fatigue assessment methodology, facilitating more sustainable and economical structural design.