Fatigue of structural connections: A recent research on multiaxial fatigue

The effect of stress concentration in connections between primary members of structures was found to be one of the main reasons for fatigue damage. Most of these connections are subjected to multiaxial fatigue in both high cycle and low cycle ranges. In case of high cycle fatigue, the equivalent effective stress with Miner's rule is the most popular multiaxial fatigue criterion in service life estimation of structures. However under many variable amplitudeloading conditions, Miner's predictions have been found to be unreliable since it does not properly take into account the loading sequence effect. Recently, a number of damage models have been proposed to capture the loading sequence effect of variable amplitude loads more precisely. The scale of metal grains of a metallic aggregate is generally defined to be within the mesoscopic scale (grain scale). Many of the more precise high cycle fatigue theories involve grain scale than macroscopic scale. In the high cycle fatigue region, some grains undergo local plasticity while the rest of the matrix behaves elastically. Recently, one mesoscopic scale fatigue model was developed to obtain more precise estimation to multiaxial fatigue and it has been found that this model is applicable to common structures since the associated material parameter determination procedures are dependent on commonly available fatigue test data. However, experimental comparisons of this theory (model) have exhibited a certain amount of deviation for multiaxial fatigue life due to variable amplitude loading conditions since the considered damage law is Miner's rule. Therefore the main objective of this chapter is to propose a new mesoscopic scale (grain scale) fatigue model to obtain a more precise estimation to multiaxial high cycle fatigue life for variable amplitude proportional loading. The accumulated inelastic meso-strain is considered as the damage variable in this model, and a new damage indicator is defined as an alternative to the damage factor of the previous model. Initially chapter describes the existing multiaxial high cycle fatigue models and their shortcomings. Then, the chapter describes the new fatigue model. The verification of the proposed fatigue model is performed by comparing the experimental fatigue lives with theoretical predictions. The proposed model is applied to estimate the fatigue life of a riveted bridge connection and obtained results are compared with previous estimations based on other known models. Finally, it is shown that the proposed multiaxial fatigue model gives a much more realistic fatigue life to variable amplitude proportional loading situation where detailed stress histories are known.