Fatigue life prediction under Gaussian type random loading: Nihei, M., Konno, T., Nishijima, S., Shimonaka, T. and Takagi, Y.Trans. Natl. Res. Inst. Met. (Jpn.) 1989 31, (1), 16–21

Abstract

Traditionally fatigue damage is associated with time dependent loading, in the form of local stress or strain histories. But there are load situations both in the lab testing as well as in the real world, where the determination and definition of the loads are more efficiently conducted in the frequency domain. Be it as deterministic sine sweeps or random stationary processes. For uni-axial loads the simulation of these signals on shaker tables has been analyzed in the past.But if we go to multiple loads which are somehow correlated new aspects have to be taken into account as the correlation or phasing have an important impact on the fatigue life – imagine the loads on a truck frame where there is a definite correlation between the loads from the front and rear axles.In this paper we present the methodology to handle multiple loads and also how to correctly handle the local stresses evaluated from multiple correlated signals, if used for a fatigue analysis. We show why traditional equivalent stress approaches bases on von Mises or principal stresses are problematic and the critical plane approach should be used. We discuss on how information on the behavior with respect to in-phase and out-of-phase loading can be analyzed and how it can be used to design better test setups on the shaker table.