| Fatigue failure is one of the primary modes of failure for engineering structures. While the frequency domain method can quickly estimate fatigue damage rates, it struggles to effectively account for the influence of the mean stress in each cycle when dealing with non-Gaussian random loads. Therefore, systematically investigating the impact of mean stress on fatigue damage under non-Gaussian random vibrations is crucial. First, Gaussian random stresses are generated using the trigonometric series method, and non-Gaussian characteristics are introduced through a nonlinear transformation model. Then, a comprehensive consideration is given to the power spectral shape, bandwidth, skewness, kurtosis, overall mean, and the manner in which mean stress effects are accounted for. Using the rainflow counting method as the foundation for time-domain fatigue damage calculations, the influence of mean stress on non-Gaussian fatigue damage under various conditions is explored. The results show that for Gaussian and zero-skewness non-Gaussian loads, the frequency domain method considering the global mean stress can effectively predict fatigue damage; however, when subjected to non-Gaussian loads with significant skewness, the mean distribution of stress cycles is influenced by the coupled effects of skewness and kurtosis, necessitating the consideration of the mean stress for each stress cycle at this point. |
