Facts And Fallacies Of Stress Determination

Author:  Almen, J. O.
Source:  SAE Journal, Vol. 5, No. 2, February, 1942
Doc ID:  1941003
Year of Publication:  1941
Abstract:  
It is in machines where weight is all-important, such as in airplane engines, or where weight and cost must both be considered, as in automobiles, that the inadequacy of our means of stress determination and our ignorance of fatigue strength of materials are most keenly felt, the author of this paper believes. Calculated stresses, he contends, are in themselves meaningless and are of value only when they are interpreted in terms of experience. Although conceding that the accuracy of stress data from photo-elastic and extensometer readings is usually greater than that obtained from the most involved mathematical analysis he shows by fatigue tests that these methods are far from reliable. In the major part of this paper, S-N diagrams with linear and logarithmic coordinates; combinations of these coordinates; and three-dimensional coordinates, are discussed. This discussion brings out the effect on fatigue strength of varying degress of stress concentration; of surface treatment of the test specimens; of stress range; and presents considerable fatigue data on ball bearings. It is shown that, when fatigue-test results are run on a large number of commercially identical parts over a sufficiently large load range, the scatter of the test points, when plotted on logarithmic coordinates, falls within a well-defined pattern which tends to radiate from a point at high stress and low number of stress cycles and to diverge to a broad band at low stress and high number of stress cycles.