Scholarlink

<p>Step-Stress Accelerated Degradation Test (SSADT) has been extensively employed in reliability assessment of degradation-induced failure products, with growing attention focused on its optimal design. In this paper, the M-optimality criterion is applied to the optimal design of SSADT for Inverse Gaussian (IG) process. Different from the traditional optimization criteria that consider prediction accuracy or evaluation accuracy, this criterion takes minimizing the asymptotic variance of the Mechanism equivalence factor (<i>Mef</i>) as the objective function in order to improve the equivalence of failure mechanism. Using a widely used stress relaxation failure data set of electrical connectors, it is proved that compared with the traditional optimization criteria, the M-optimality criterion does not significantly reduce the model parameter estimation accuracy or product life prediction accuracy, but improves the failure mechanism equivalence.</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

Step-stress accelerated degradation test for Inverse Gaussian process based on M-optimality criterion

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年8月25日

查看全文
  • Jiayin Tang,
  • Hao Zhou

摘要

Step-Stress Accelerated Degradation Test (SSADT) has been extensively employed in reliability assessment of degradation-induced failure products, with growing attention focused on its optimal design. In this paper, the M-optimality criterion is applied to the optimal design of SSADT for Inverse Gaussian (IG) process. Different from the traditional optimization criteria that consider prediction accuracy or evaluation accuracy, this criterion takes minimizing the asymptotic variance of the Mechanism equivalence factor (Mef) as the objective function in order to improve the equivalence of failure mechanism. Using a widely used stress relaxation failure data set of electrical connectors, it is proved that compared with the traditional optimization criteria, the M-optimality criterion does not significantly reduce the model parameter estimation accuracy or product life prediction accuracy, but improves the failure mechanism equivalence.