This paper evaluates the empirical Dirlik method for predicting fatigue life in structures subjected to random vibrations across various standardized loading spectra. Using a surrogate model, the study compares the Dirlik approach against the Narrow Band (NB0), Benasciutti-Tovo (BT), Zhao-Baker, and Lalanne methods. Results demonstrate that spectral bandwidth and frequency response placement dominate fatigue life outcomes. While all evaluated models converge closely for near-narrow-band spectra , performance diverges significantly in wide-band conditions. Specifically, the BT method risks non-conservative under-prediction for wide-band spectra , whereas NB0 and Lalanne tend to yield conservative over-predictions. Ultimately, the findings establish that the Dirlik and Zhao-Baker as well as the Benasciutti-Tovo (BT) models are the most robust, accurate, and reliable general-purpose choices across diverse operational conditions.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

The Use of Dirlik Method for Fatigue Life Calculation in Terms of Different Standardized Loading Spectra

  • Michał Böhm

摘要

This paper evaluates the empirical Dirlik method for predicting fatigue life in structures subjected to random vibrations across various standardized loading spectra. Using a surrogate model, the study compares the Dirlik approach against the Narrow Band (NB0), Benasciutti-Tovo (BT), Zhao-Baker, and Lalanne methods. Results demonstrate that spectral bandwidth and frequency response placement dominate fatigue life outcomes. While all evaluated models converge closely for near-narrow-band spectra , performance diverges significantly in wide-band conditions. Specifically, the BT method risks non-conservative under-prediction for wide-band spectra , whereas NB0 and Lalanne tend to yield conservative over-predictions. Ultimately, the findings establish that the Dirlik and Zhao-Baker as well as the Benasciutti-Tovo (BT) models are the most robust, accurate, and reliable general-purpose choices across diverse operational conditions.