Comparison of uniaxial spectral methods for weld fatigue

Spectral fatigue methods have gained attention due to lower computational cost compared to time-domain fatigue methods for variable amplitude stress-time signals. Multiple spectral methods have been proposed in the literature, and several studies have compared their performance. The results of such comparisons depend on the S-N curve slope and the considered power spectral density spectrum. Thus, the comparative studies have resulted in different recommendations regarding which spectral methods should be used for fatigue assessment. In this paper, an overview is given of the recommendations from various comparative studies. Moreover, a comparison of the recommended spectral methods is made. The comparison focuses on weld fatigue, where only a few S-N curve slopes (3 and 5) are normally considered when following relevant guidelines, and low shear stress ranges are negligible. In addition, the comparison is based on stress response spectra defined from simple structural dynamics systems to make them relevant for weld fatigue. Furthermore, only the stress ranges are considered for weld fatigue, and the formulas in this paper are given in terms of stress ranges rather than stress amplitudes to keep the format used in weld fatigue. The damage results from spectral methods are compared against the damage from rainflow counting and Palmgren-Miner damage accumulation, considering simulated time-series. The results show a dependency between the damage error and the spectral width parameter. The Jun-Park and Dirlik spectral methods are shown to give the best correlation with standard damage assessment using rainflow counting and Palmgren-Miner. • An overview of selected spectral fatigue comparative studies is provided. • Spectral methods are compared for structural dynamics stress response spectra. • The Jun-Park and Dirlik methods perform best for uniaxial weld fatigue.