Design an Optimal Accelerated-Stress Reliability Acceptance Test Plan Based on Acceleration Factor
Design an Optimal Accelerated-Stress Reliability Acceptance Test Plan Based on Acceleration Factor
复制标题
基于加速因子设计最优加速应力可靠性验收试验方案
DOI：
10.1109/tr.2018.2828605
复制
发表时间：
2018-09-01
影响因子：
5.9
通讯作者：
Zhou, Yuan
中科院分区：
文献类型：
--
作者：
Wang, Hao-Wei;Teng, Ke-Nan;Zhou, Yuan
可靠性验收测试用于在买方接受生产商交付的产品之前调查其可靠性。然而，传统的基于寿命数据的可靠性验收测试是在相对较低的应力水平下进行的，非常耗时。研究如何设计最佳的加速应力可靠性验收测试（ASRAT）以提高测试效率是一个有趣且有用的问题。应用逆高斯过程对产品的降解数据进行建模。应用加速因子常数原理推导了逆高斯退化模型的加速因子表达式，从而通过加速因子得到了ASRAT的终止时间。选择终止时间的渐近方差作为目标函数，建立ASRAT计划的优化模型。通过在总成本约束下最小化终止时间的渐近方差来优化决策变量，包括加速应力水平、样本大小和测量次数。然后，开发了一种基于 MATLAB 程序的自动算法，以帮助高效地设计最佳 ASRAT 计划。最后，利用所提出的方法设计了一种电连接器的最优ASRAT方案。敏感性分析表明最优ASRAT方案对于参数估计的小偏差具有鲁棒性。此外，还进行了模拟测试，以证明 ASRAT 的实用性，并研究退化模型错误指定对设计最佳 ASRAT 计划的影响。
A reliability acceptance test is used to investigate the reliability of products delivered by producer before buyer accepts them. However, the traditional reliability acceptance test based on lifetime data, which is conducted under a comparative low stress level, is time consuming. It is an interesting and useful issue to study how to design an optimal accelerated-stress reliability acceptance test (ASRAT) to improve test efficiency. An inverse Gaussian process was applied to model degradation data of products. Acceleration factor constant principle was applied to derive the expression of acceleration factor for the inverse Gaussian degradation model, so the termination time of ASRAT was obtained through acceleration factor. The asymptotic variance of termination time was selected as an objective function to establish the optimization model of the ASRAT plan. The decision variables, including accelerated stress level, sample size, and the number of measurement were optimized by minimizing the asymptotic variance of termination time under a total cost constraint. Then, an automatic algorithm based on MATLAB procedure was developed to help efficiently design optimal ASRAT plans. Finally, the proposed method was used to design an optimal ASRAT plan for a type of electric connector. Sensitivity analysis indicates that the optimal ASRAT plan is robust to small bias of parameter estimation. Moreover, simulation tests were conducted to demonstrate the usefulness of ASRAT and investigate the impact of degradation model misspecification on designing an optimal ASRAT plan.