Scholarlink

<p>The three-parameter Weibull distribution is pivotal in reliability engineering for characterizing the lifespan of safety-critical components, such as aero-engine turbine disk, where parameter estimation is severely challenged by the scarcity of failure data arising from prohibitive testing costs and operational limitations. To address this problem, this study proposes a data-driven artificial intelligence algorithm named multi-physics coupled multi-source fatigue crack network synergistic optimization (MPC-MCNSO), which reformulates the complex three-parameter Weibull estimation as an optimization problem inspired by the synergistic propagation of multi-source fatigue crack network. Subsequently, comprehensive benchmarking against classic and reinforcement learning-based optimizers is conducted through 160000 Monte-Carlo simulations across various distributions and sample sizes with several metrics, revealing the MPC-MCNSO’s superior accuracy, stability, fitness degree, and reliability assessment capability. Finally, validation on experimental fatigue life data obtained from turbine disk test specimens combined with discussions of experimental details and fusion mechanisms further confirmed its effectiveness in reliability assessment.</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

MPC-MCNSO: Artificial intelligence algorithm for Weibull parameter estimation with limited data-application to aero-engine turbine disk reliability assessment

  • 期刊论文

  • 发表日期：

    2026年5月30日

查看全文
  • Jianyi Gu,
  • Xiangwei Kong,
  • Jin Guo,
  • You Guo,
  • Zinan Wang,
  • Heran Zhang

摘要

The three-parameter Weibull distribution is pivotal in reliability engineering for characterizing the lifespan of safety-critical components, such as aero-engine turbine disk, where parameter estimation is severely challenged by the scarcity of failure data arising from prohibitive testing costs and operational limitations. To address this problem, this study proposes a data-driven artificial intelligence algorithm named multi-physics coupled multi-source fatigue crack network synergistic optimization (MPC-MCNSO), which reformulates the complex three-parameter Weibull estimation as an optimization problem inspired by the synergistic propagation of multi-source fatigue crack network. Subsequently, comprehensive benchmarking against classic and reinforcement learning-based optimizers is conducted through 160000 Monte-Carlo simulations across various distributions and sample sizes with several metrics, revealing the MPC-MCNSO’s superior accuracy, stability, fitness degree, and reliability assessment capability. Finally, validation on experimental fatigue life data obtained from turbine disk test specimens combined with discussions of experimental details and fusion mechanisms further confirmed its effectiveness in reliability assessment.