Quest for Fatigue Limit Prediction Under Multiaxial Loading

The paper discusses origins of fatigue data used for evaluating the prediction quality of various multiaxial methods intended for estimating the fatigue limit. The paper defines some requirements useful for processing similar data sets. It focuses on four most often used data sets, analyzes the way they were prepared and the way they are transmitted over various research papers. These sets are commented and objections to some their items stated. Weight coefficients enabling their application in validation studies are proposed. The final reduced total set has approximately one half of its original size. The reduced set is re-analyzed by several fatigue limit calculation methods and the results are compared to the origin. Though the relative probability distribution of the output is quite well kept, the importance of evaluating the second hand data sets carefully is clearly highlighted.