Effect of Specimen Size and Preparation on the Fatigue Strength of a Plain Carbon Steel Tested in Rotating Bending and in Torsion

Abstract

A survey of the literature indicates that, in the presence of a stress gradient, the fatigue strength of large components is generally less than the fatigue limit exhibited by small specimens of similar material. In some cases the ‘size effect’ reported is of alarming proportions. However, a closer examination of the literature and the results of some exploratory tests in rotating bending suggest that the effects observed may not be attributable simply to size but may also be dependent on the final stages of specimen preparation.An extensive test programme carried out on notch-free 0·32 C steel specimens, ranging in minimum diameter from 0·05 to 05 in in torsion and from 0·05 to 1·6 in in bending, provides results showing four main features. As other workers have found, very small specimens have generally high fatigue limits and as-polished specimens have higher fatigue limits than as-heat-treated specimens. For specimens that were normalized, carefully machined and polished, then stress-relieved and finally lightly repolished, the fatigue strength shows a significant increase as the diameter decreases, the maximum increases being about 15 per cent in torsion and 20 per cent in bending. For specimens finally stress-relieved the size effect, over the ranges from 0·25 in diameter upwards, is negligible.