TFT-LCD面板壽命研究

本論文係針對TFT-LCD (thin film transistor-liquid crystal display)面板的壽命或稱之爲平均故障間隔時間(mean time between failures, MTBF)進行研究。研究內容分爲兩個部分，第一部份針對某特定14.1吋筆記型電腦使用之TFT-LCD面板於產品開發階段完成，於產品量產前，隨機抽取生產線上產品100片，於實驗室以高溫應力加速模式，進行加速壽命鑑定試驗，採用固定時間法蒐集失效數據(type I censored data)，並以Arrhenius加速壽命模式所得到的加速係數，將加速壽命所得到的失效時間轉換爲正常溫度環境使用下之失效時間後，進行壽命分析，分析方法係利用最大概似法找出最大的對數概似(log-likelihood)值之失效分佈，進而估計該失效分佈函數之參數，計算出產品的鑑定試驗壽命值。第二部份，針對同型之TFT-LCD面板於量產銷售階段，所售出之182萬片面板，產品於壽命週期期間因功能失效而退回維修中心的3600筆失效數據，進行壽命分析，先利用Kaplan-Meier方法之無母數分佈存活分析，得到TFT-LCD面板的存活函數(survival function)及變異數估計(variance estimation)，另同樣使用最大概似法找出最佳失效機率分佈之參數，計算產品壽命值。最後本論文將探討上述兩者壽命值之間的關係並分析其差異的原因。

關鍵字

TFT-LCD ；壽命；失效分佈；最大概似法

並列摘要

The calculation of MTBF (mean time between failures) is an important task in reliability life data analysis. Using different distributions, the values of mean time between failures are always different. This paper investigates two different approaches to obtaining the TFT-LCD panel life. One is the accelerated demonstration life testing which is performed at the reliability lab during the panel development stage, and the other approach is based on field return data from the repair center. The maximum likelihood estimation method is used to get the log-likelihood value of the different distributions, including the Weibull distribution, normal distribution, lognormal distribution and exponential distribution. The distribution of the maximum log-likelihood is chosen for the optimum TFT-LCD panel life distribution. We also use the maximum likelihood estimates to get the parameters of the distribution and calculate the panel life value. The deviation of the life got from the accelerated demonstration life testing data and the field return data analysis is also discussed in this paper.

並列關鍵字

TFT-LCD ； life ； failure distribution ； maximum likelihood

參考文獻

王宗華(1998)。可靠度工程技術手冊。台北:中華民國品質學會。

Google Scholar

柯輝耀(1997)。可靠度保證─工程與管理技術之應用。台北:中華民國品質學會。

Google Scholar

Chen, Y. C.,Chen, C. M.,Chen, Y. C.(2007).Accelerated life estimation for LCD modules using artificial neural network based on Taguchi method.International Journal of Industrial Engineering Theory-Applications and Practice.14(3),289-297.