A method for obtaining conservative S−N data for welded structures

A method of fatigue testing is proposed to simulate the behavior of large-sized welded structures having high tensile residual stresses by means of ordinary small width specimens containing a low level of residual stresses. The method involves the varying of the stress range from test to test while always maintaining the maximum stress at the yield strength of base metal. The results obtained by the proposed method agreed with those for slit welded joints containing high tensile residual stresses fatigued at constant amplitude at a stress ratio of zero. However, the fatigue strength of small width welded specimens as determined by the proposed method was lower than that obtained by the conventional method at a stress ratio of zero. It is emphasized that the proposed method is effective in obtaining conservative S-N data to be used for design of welded structures, where local fluctuating stresses were considered to pulsate downwards from tensile yield strength regardless of the applied stress ratio. It was also found that in the presence of a high tensile residual stress the grinding of the toe of welds which contain no undercut was not effective in improving the fatigue strength of welded joints.