Methodology for Correlating Fatigue Data Obtained under Different Test Conditions

Abstract

Abstract A number of relatively simple methods are presented which can be used to correlate fatigue data obtained under different test conditions and so collapse the data into a single fatigue-life curve. The test variables considered are stress concentration, mean stress, temperature, and residual stress. The cyclic stress-strain curve in conjunction with Neuber's rule is used to estimate actual stresses and strains in notched specimens. Mean stress effects are accounted for by using empirically determined relationships that have been shown to work for a variety of structural materials. A simple ratio of the ultimate tensile strength at room temperature to that at the test temperature is used to correct for temperature effects. Finally, residual stress is handled by assuming an equivalence with an applied mean stress of the same sign and magnitude. The combined use of these corrections can result in more efficient use of existing fatigue data and allow prediction of the fatigue life of a material under a wide variety of operating conditions. The methodology is illustrated using high-cycle fatigue data on a 403 stainless steel, and accuracy of the methods is ascertained by comparison with a baseline fatigue curve and via statistical criteria.