Scientific Liaison Officer at the London Branch of the Office of Naval Research, London,
Mehl, R. F.
Director of Metals Research Laboratory and Head of the Department of Metallurgical Engineering, Carnegie Institute of Technology, Pittsburgh, Pa.
Pages: 33 Published: Jan 1953
It is now well established that the fatigue properties of metals are statistical in nature. Earlier work showed that the fatigue life is variable, and very recently it was shown for the first time that the endurance limit is statistical. Since both the fatigue life and endurance limit are subject to marked variability, these quantities cannot be stated as single values, but must be represented statistically. This fact is of scientific interest as well as practical importance. In the design of parts for service in applications which involve cyclic stresses, the engineer must, in choosing a proper factor of safety, recognize the fact that some failures can occur prematurely or even at stresses below the normally determined endurance limit. In the field of research, many of the long accepted conclusions drawn from the usual investigations can be questioned since the statistical approach, now known to be necessary, was not used.
While the statistical nature of fatigue properties has been established, the influence of metallurgical factors on this behavior has not been known. The dependence of fatigue variability on such factors as composition, microstructure, hardness, and inclusion rating has not been evaluated, and it was toward this objective that considerable experimental work was directed.
The statistics of the fracture curve and endurance limit were obtained for a variety of materials, and analysis and comparison of the results by statistical methods indicated the influence of the metallurgical variables on the statistical behavior.
Paper ID: STP45670S