Fatigue damage calculations are traditionally based on the time domain approach. Acceleration time history inputs are used to excite the system and the outputs are in a form of stress time history. This transient dynamic approach, as time history is intuitive to understand, provides straightforward and reasonable result. Nevertheless, a typical automotive proving ground test consists of 20 to 30 road events, it is not only computationally intensive but could be also a grueling process for an engineer to carry out as it requires several iterations for each event in the schedule before fatigue calculation. Alternatively, a frequency domain fatigue calculation is widely used. In this approach, both the dynamic loading and response are expressed in terms of Power Spectral Density (PSD) functions and the dynamic structure is treated as a linear transfer function. The transfer function is then multiplied with the event PSD to get the PSD of the stress.
This PSD approach can simplify the problem considerably as it requires simulation to be run only once. However, the approach itself still has some issues when it comes to large FE models with a wide frequency range of interest. The stress transfer function (TF) can potentially be too big to manage. The objective of this paper is to present an improved technique that utilizes the modal stresses and modal coordinates during the modal superposition process to get around the issue while still maintaining the accuracy of the stress PSD and damage calculation.