Accurate prediction of crack initiation life is of critical importance in designing turbo-machinery. To improve this accuracy, more sophisticated prediction techniques are required. A large number of materials have shown a correlation between low cycle fatigue initiation life and strain range, as represented by the well-known Manson-Coffin equation. Testing has shown that tensile mean stress has a negative impact on life. This effect has been noted for many years when applied to high-cycle fatigue, resulting in the use of Goodman or Haigh diagrams to account for the impact of both stress range (and therefore strain range) and mean stress. Morrow proposed a methodology for accounting for mean stress in low cycle fatigue. Noting that as plasticity increases, the effect of mean stress decreases, the correction was applied only to the elastic strain versus life line. Use of the Morrow mean stress corrections improves the accuracy of life predictions, but there are limitations. The most significant of these limitations are situations in which the correction may be non conservative for high compressive mean stresses or very high tensile mean stresses. While a benefit from compressive mean stress is to be expected, at some point further increasing the compressive mean stress should have a negative impact on life. At very high tensile mean stresses near the material yield, the calculated impact of mean stress on life is non conservative. To overcome these limitations, the analyst may place limits on the acceptable range of R-ratios used based on actual test data, but this would do little more than highlight when the user is outside of his database limits. Alternatively a life system may be made in which mean stresses are conservatively expected to be tensile. Neither of these methods are useful for calculating different lives in compressive or tensile regimes. This difficulty is especially seen in bolted joints, which are subjected to high mean stresses and small stress amplitudes. A technique is proposed here in-which limits are placed on the mean stress correction, directly analogous to those used in the creation of so-called modified Goodman diagrams. This technique has been successfully applied at PSM to improve the accuracy of life prediction without increasing the risk of non conservatism. A review of some literature is made to show examples where this effect may be taking place. A small number of tests provide additional validation.

This content is only available via PDF.
You do not currently have access to this content.