Scholarlink

<p>The crack propagation behavior of aerospace metals under high-temperature conditions is critical for designing structurally damage-tolerant components. To evaluate the combined effects of temperature and stress ratio on the titanium alloy Ti-6Al-4V, fatigue crack growth tests were conducted at three temperatures (25 °C, 100 °C, 200 °C) and four stress ratios (R = 0.1, 0.3, 0.5, 0.7). Based on the experimental data, four predictive models —back propagation neural network (BPNN), K-nearest neighbors (KNN), random forest (RF), and extreme gradient boosting (XGBoost)—were developed to estimate fatigue crack growth rates. The results indicate that the crack growth rate (da/dN) increases with the stress ratio, leading to an earlier onset of the Paris law region and final failure region, while it decreases significantly with rising temperature. All models effectively capture the nonlinear characteristics of crack growth, outperforming the traditional Walker equation. Among them, BPNN exhibited superior extrapolation capabilities, establishing it as a reliable tool for predicting crack growth beyond the tested parameter ranges.</p>

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

Fatigue crack growth rate models of Ti-6Al-4V alloy considering temperature and stress ratio

  • 期刊论文

  • 发表日期：

    2025年5月6日

查看全文
  • Fan Cao,
  • Shuanglong Tao,
  • Rong Chen,
  • Shuchun Huang,
  • Guangwei He,
  • Wei Li,
  • Zhexu Liu

摘要

The crack propagation behavior of aerospace metals under high-temperature conditions is critical for designing structurally damage-tolerant components. To evaluate the combined effects of temperature and stress ratio on the titanium alloy Ti-6Al-4V, fatigue crack growth tests were conducted at three temperatures (25 °C, 100 °C, 200 °C) and four stress ratios (R = 0.1, 0.3, 0.5, 0.7). Based on the experimental data, four predictive models —back propagation neural network (BPNN), K-nearest neighbors (KNN), random forest (RF), and extreme gradient boosting (XGBoost)—were developed to estimate fatigue crack growth rates. The results indicate that the crack growth rate (da/dN) increases with the stress ratio, leading to an earlier onset of the Paris law region and final failure region, while it decreases significantly with rising temperature. All models effectively capture the nonlinear characteristics of crack growth, outperforming the traditional Walker equation. Among them, BPNN exhibited superior extrapolation capabilities, establishing it as a reliable tool for predicting crack growth beyond the tested parameter ranges.