Fatigue life prediction of welded joints from nominal system to fracture mechanics

Abstract

This paper presents methods for determining the fatigue life of welded joints with particular emphasis on typical joints. Possible methods for computational determination of the stresses in welded joints are shown, starting with those based on nominal stresses according to different proposals, through structural stresses, including the ‘hot spot’ type, to the most complicated linear fracture mechanics. With this background, the recommendations of the International Institute of Welding (IIW) for determining the fatigue life of welded steel joints of flat components under uniaxial tensile–compression conditions are presented. The paper focuses on the problem of determining and evaluating the fatigue life of selected types of welded joints in accordance with the guidelines of the IIW and the safety analysis of welded structures. The study shows the extent to which the simplest method based on the amplitude of nominal stress can be used for calculating fatigue life based on two selected types of joints with different asymmetry coefficients, R = −1 and 0, and joints without geometrical notch. The analysis was based on steel welded joints of four different materials. The results show that the FAT characteristics are always good except for the FAT 90 category, where, for a lower fatigue life, we approach a dangerous fatigue life estimation. For K-type cruciform joints, the FAT model can be used for all the stress ratios. In other cases, such compliance cannot be accepted. For the category FAT 90 X-type butt joints, a new category FAT '134 and a new slope, m = 5, are proposed. For type X butt joints without notch, the slope m = 12 and a new category FAT' 525 are proposed.