Scholarlink

<p>This paper presents a novel reliability evaluation method that integrates accelerated life testing (ALT) with imprecise probabilistic techniques to address the challenges of uncertainty in life estimation, especially under small sample conditions and censored data. The approach begins by applying the log-rank test to derive interval estimates for the parameters of a power-law acceleration model, capturing the inherent variability in stress-life relationships. These interval estimates are then combined with nonparametric predictive inference (NPI) to generate upper and lower bounds for the product’s survival function, thus forming an imprecise yet informative reliability profile. To demonstrate the practical feasibility and effectiveness of the method, failure data from accelerated life tests conducted on automotive stabilizer bars under varying mechanical loads are analyzed. Compared with other traditional models, this method relies solely on experimental data processing of truncated or small sample data to obtain an imprecise reliability interval through hypothesis testing methods and NPI. This imprecise reliability framework is especially useful in partial information scenarios, offering valuable insights for maintenance planning, warranty decision-making, and product design validation.</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

A hybrid reliability evaluation method combined with accelerated life testing and imprecise models

  • 期刊论文

  • 发表日期：

    2026年1月13日

查看全文
  • Yi-Chao Yin,
  • Yang Wu,
  • Yu Zhou,
  • Yuan-Jian Yang,
  • Shuchen Cao,
  • Mohammad Yazdi

摘要

This paper presents a novel reliability evaluation method that integrates accelerated life testing (ALT) with imprecise probabilistic techniques to address the challenges of uncertainty in life estimation, especially under small sample conditions and censored data. The approach begins by applying the log-rank test to derive interval estimates for the parameters of a power-law acceleration model, capturing the inherent variability in stress-life relationships. These interval estimates are then combined with nonparametric predictive inference (NPI) to generate upper and lower bounds for the product’s survival function, thus forming an imprecise yet informative reliability profile. To demonstrate the practical feasibility and effectiveness of the method, failure data from accelerated life tests conducted on automotive stabilizer bars under varying mechanical loads are analyzed. Compared with other traditional models, this method relies solely on experimental data processing of truncated or small sample data to obtain an imprecise reliability interval through hypothesis testing methods and NPI. This imprecise reliability framework is especially useful in partial information scenarios, offering valuable insights for maintenance planning, warranty decision-making, and product design validation.