Validation data set for testing the criteria for multiaxial fatigue strength estimation

Abstract

AbstractThis paper describes the way in which a specialized validation set intended for quick validation of new multiaxial fatigue strength estimation models in the high‐cycle fatigue regime was prepared. The rules established for accepting various data sets into the complete FatLim database are briefly described. This final FatLim database contained 282 data items at the moment this paper was written. The paper describes how FatLim has been reduced to a smaller validation set, marked as AMSD25. The criteria used for accepting the data items for this small‐scale test set are based on evaluating the mean and the standard deviation of the fatigue index errors obtained for each data item on the basis of 18 different prediction criteria. The items that got into the AMSD25 validation set are described, and potential issues are noted. The comparison between the original FatLim validation set and the shortened AMSD25 data set is presented. It shows that the in‐phase load cases without any mean stress effect are generally well mastered by most criteria, and they can be excluded as insignificant from any such analysis. However, the mean stress effect is generally poorly incorporated. Future multiaxial fatigue strength criteria should focus on load cases including mean stresses.