Optimal Simple Ramp-Test Based on Wiener Degradation Processes and Inverse Gaussian Distribution

在Wiener衰變模型與逆高斯分配下之最佳簡單直線加速實驗

李名鏞(Ming-Yung Lee) ； 唐正(Jen Tang)

《智慧科技與應用統計學報》 8卷2期 (2010/12) Pp. 21-34

https://doi.org/10.29807/JTITAS.201012.0002

摘要

面對全球化的挑戰，以及滿足顧客對於產品使用壽命的需求，現在產品的使用年限都非常的長。換句話說，廠商必須評估產品壽命，然而在全球化高度競爭下的電子產業，產品開發週期短。因此對於高可靠度產品，如何快速有效的評估產品壽命，是一項非常重要的課題。傳統的可靠度實驗均採用加速壽命實驗的方式來進行。然而對於高可靠度產品而言，大抵採用隨機衰變模式做為產品壽命評估的方法。本文嘗試結合加速壽命實驗與隨機衰變模式，最小期望成本的考量下，找到一個最佳的直線加速試驗。

關鍵字

直線加速實驗 ； Wiener衰變過程 ； 逆高斯分配

並列摘要

To face intensive global competition and meet higher customer expectations, many today's products are designed to function under normal operating conditions for a long time without a failure. On the other hand, companies need data sooner to estimate the failure time (lifetime, time-to-failure) distributions of their products. Therefore, how to design a degradation test to collect timely data for accessing the reliability of a product are important problems for managers, reliability engineers, and statisticians.Accelerated lifetime test is used to estimate the lifetime distribution in traditional reliability experiment. Therefore, they use stochastic degradation model for high reliability product. In this paper, we will provide an optimal ramp test under minimal expected total cost by combining accelerated lifetime test and stochastic degradation model.

並列關鍵字

ramp-test ； Wiener degradation processes ； inverse Gaussian distribution