Fatigue behaviour analysis of mechanical components subject to random bimodal stress process: frequency domain approach

The greatest advantages of this method are the computational efficiency and the capability to give a synthetic representation of a random process. The importance of this method is clear, especially when it is used in the postprocessing phase of a numerical analysis, for example Finite Element (FEA) or Multibody (MBS) simulation. The weak step in this procedure is the direct evaluation of the probability density function (pdf) of the cycles amplitude from stress power spectral density (PSD). In fact, the translation to frequency domain of the Rainflow cycle counting is a serious problem but there is a body of important literature concerning this. After reviewing the most important literature from this research field, a new original index to compare the reliability of the frequency domain methods is proposed. This index efficiently achieves this purpose, and a future objective could be to use such index to define a correction factor in order to exactly evaluate the fatigue damage. A very important class of random processes was considered: the bimodal power spectral density (PSD) random processes. The aim of this paper is to identify independent parameters in these processes and through them evaluate the cycles probability density function (pdf) amplitude.