Scholarlink

<p>Inherent uncertainties in mechanisms pose significant challenges to motion reliability assessment, especially in strongly nonlinear systems. To address the limitations of traditional approximate analytical methods in terms of accuracy and efficiency, this study proposes a dynamic reliability analysis framework based on a generalized F-discrepancy (GF-discrepancy)–based representative point selection method under hybrid random–interval uncertainties. By introducing interval variables to characterize epistemic uncertainty, the proposed approach extends reliability evaluation from point estimates to interval-based reliability bounds and better reflects practical engineering conditions. Numerical examples involving both purely random and hybrid random–interval uncertainties demonstrate that the proposed method achieves accuracy comparable to Monte Carlo simulation while requiring significantly fewer samples, thereby substantially improving computational efficiency.</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

Mechanism reliability under hybrid uncertainties via generalized F-discrepancy

  • 期刊论文

  • 发表日期：

    2026年7月11日

查看全文
  • Chongqing Qian,
  • Yongjuan Wang,
  • Yuzhao Yang

摘要

Inherent uncertainties in mechanisms pose significant challenges to motion reliability assessment, especially in strongly nonlinear systems. To address the limitations of traditional approximate analytical methods in terms of accuracy and efficiency, this study proposes a dynamic reliability analysis framework based on a generalized F-discrepancy (GF-discrepancy)–based representative point selection method under hybrid random–interval uncertainties. By introducing interval variables to characterize epistemic uncertainty, the proposed approach extends reliability evaluation from point estimates to interval-based reliability bounds and better reflects practical engineering conditions. Numerical examples involving both purely random and hybrid random–interval uncertainties demonstrate that the proposed method achieves accuracy comparable to Monte Carlo simulation while requiring significantly fewer samples, thereby substantially improving computational efficiency.