Verification of analytical models of the S-N curve within limited fatigue life

This paper presents analytical methods for determination of the curve for a high-cycle fatigue. It has been found based on qualitative and quantitative verification that the error introduced by these methods can be as large as 3-fold length of the experimental life. In addition, the wrong result can lay on either the safe or the unsafe side with equal probability. Therefore, an analytical-and-experimental (hybrid) method has been proposed. Verification
of computed characteristics vs. experimental data demonstrated that the “proprietary” proposal reduced the error. In addition, an approximate error depending on the number of experiments has been determined based on the computations.