| Summary: | When an assembly (such as a shrink-fitted shaft/hub configuration) is subjected to oscillating loads, two components of the assembly may experience oscillating slip relative to each other, giving rise to fretting on the interface between the two components. If, subsequently, a fatigue failure originates in that fretting region, the nominal failure stress is often considerably lower than the fatigue strength of the appropriate material. The object of the research reported in this thesis was to investigate a hypothesis which explains the (relatively) low magnitude of fretting fatigue strengths as a consequence of macroscopic stress concentrations in the fretted region. Such stress concentrations are a function of the overall geometry of the assembly and it was suggested that, as a result, stress distributions in assemblies should be similar to stress distributions in geometrically similar one-piece (solid) components. In an assembly, high stress concentrations near an interface are partially relieved by slip (which gives rise to fretting) and it was thought that an assembly should be stronger, therefore, than the equivalent solid. Even so, it was expected that a detailed analysis of the stress distribution in an assembly would predict stress concentrations of sufficiently high magnitude to explain failure under the low nominal stresses found in practice, thus demonstrating that fretting, itself, does not greatly affect fatigue strengths. Fatigue tests were carried out using solid and assembled specimens of various shapes and a method was developed, using finite element techniques, to analyse the stresses in the assemblies, taking full account of slipping. The theoretical stress analysis showed that, for certain geometries and systems of loading, the stress distributions in an assembly may differ considerably from those in a geometrically similar solid. Furthermore, when the fatigue results were examined in the light of the stress analysis, it was apparent that fretting can reduce the fatigue strength of a material by more than 60%. To explain this result, a new hypothesis is proposed. It is suggested that under conditions of oscillating slip, the discontinuous contact between two surfaces of an assembly results in large stress concentrations in the regions of actual contact. It is shown that such stress concentrations can be of sufficient magnitude and affect a sufficiently large volume of material to cause failure at a nominal stress considerably lower than the fatigue strength of the material concerned. Such a hypothesis is supported by the results of this investigation and also by the results of several other workers.
|
|---|
