失效模式的風險評估一直是企業的一個重要的課題,如何在產品出貨前對失效模
式的風險評估,更是一個需要謹慎處理的重要課題,一旦對問題的評估錯誤對於公司
的經營損失更是難以估計的,尤其在這數位時代,社群媒體傳遞的力量與速度更是不
容小歔,所造成的雪球效應,更讓企業要小心的處理市場上產品品質問題,避免錯誤
評估失效模式,造成後續在市場上的影響力,本論文研究主題將探討台灣電子業界,
企業是否能有機會在產品出貨前對所有的問題,做一正確的風險評估與正確的可靠度
驗證計畫規格。

為何會提到可靠度驗證規格?一般消費性電子產品製造商在產品研發階段,都會
將其研發產品做一系列之可靠度驗證,期許藉由可靠度驗證計劃,盡可能找所有影響
產品壽命之潛藏失效模式問題,並將其問題在設計階段解決,但也可能有一些問題並
未能再量產前解決,或因錯誤的風險評估而決定不處理那些潛藏的失效模式問題,因
此產生出另一課題,即為經由驗證之後找到的問題。企業如何去正確判斷評估問題在
市場上將會造成之影響,這問題是否真正需要正視它,還是只是一個非常微小的問題
對使用者一點影響都沒有,此研究將針對落摔測試之失效模式來探討如何評估,失效
模式在市場影響程度及其風險評估。

本研究有主要三個重點如下:第一個重點為找出落摔測試後面板破片問題與市場
主要面片版破片反映的關係,進而產品設計特徵、可靠度驗證條件及可靠度驗證結果
的相關自變數,與市場主要潛藏面板破片問題應變數之關係,繼而利用相關自變數去
對應變數做相關的迴歸分析,使其應變數可以變成一個風險評估重要因子。也可利用
這樣的迴歸預測分析做更深入的研究,去了解目前產品的可靠度驗證計畫與標準是否
符合市場上使用者對此產品的品質要求與使用者行為模式,若不符合我們可以利用此
模型來調整可靠度驗證規格與標準。

最後,本研究將投入 46 個產品評測資料,其資料包含自變數產品設計特徵、產品
落摔高度、落摔結果、變數面板破片失效模式以及面板破片概率。由於產品可靠度驗
證計畫範圍過廣,本次研究將集中在特定落摔測試計畫、條件、結果及市場失效率,

若研究成功可以將此邏輯概念想法做典範轉移並應用在其他的可靠度驗證計畫及市場
失效模式後續研究。研究主題將著重在討論智慧型手機其市場潛藏面板破片率、產品
設計特徵以及落摔試驗條件及測試結果探討,藉由此研究找到失效模式風險因子的預
測模型可以當作在初期階段對產品的特定失效模式做正確的風險評估,與可靠度驗證
計驗證條件的修正。

The risk assessment of failure mode has always been an important issue for
enterprises. How to assess the failure mode before the product is shipped, which is
an important issue that needs to be handled with caution. Once the evaluation of the
problem goes wrong, it is even more difficult for the company to lose business. It is
estimated that, especially in this Network generation, the power of social media
transmissions should not be underestimated. The snowball effective caused by
enterprises should be more careful to deal with product quality problems in the market
and avoid the mis-evaluation of failure modes. In the market's influence, the research
theme of this paper will explore the Taiwan electronics industry, whether the company
has the opportunity to make a correct risk assessment and correct reliability
verification plan specifications for all problems before the product is shipped.
Why do we mention reliability verification specifications here? In general, consumer
electronics manufacturers will perform a series of reliability verification tests on their
products during the product development phase. It is expected that the reliability
verification plan will find all the potential failure modes that affect the life of the product
as much as possible, and try to solve the problem in the development stage. But there
may be some problems and it is not possible to solve the problem before the mass
production or decide to do not deal with the potential failure mode problems due to the
wrong risk assessment. Therefore, another problem arises, which is the problem found
after verification. How to correctly judge and evaluate the impact of these problems on
the market, whether the issue is a real problem needs to face it and solve it, or is it just
a very small problem that has no any impact on the user, so the research will be based
on drop test verification. The failure mode explores how to assess the impact of failure
modes on the market and its risk assessment.

Therefore, this paper has three main focuses. The first one focuses on finding the
relationship between the failure mode problem and the main quality problems of the
market after drop test verification and finding out its related to product design features,
reliability verification conditions and reliability verification results. The relationship
between the relevant independent variables and the number of strains related to the

market's main return problem, and then using the relevant independent variables to
correspond to the variables to make relevant regression analysis, so that the strain
number can become an important factor of risk assessment, and can also use such
regression prediction analysis to do more In-depth research to understand whether
the current product reliability verification plan and standards could meet the quality
requirements and user behavior patterns of users on the market. If so, we can use this
model to adjust the reliability verification specifications and standards.
Finally, this study will invest 46 sets of marketing products test data from related
forums and websites, including the independent variable product design features,
product drop test verification conditions, product verification results and strain number
product failure mode potential display broken rate from user data. Due to the wide
scope of product reliability verification programs, this study will focus on the drop test
plan, conditions, results and market failure rate. If the research is successful, this
logical, concept and idea can be transferred as a model and applied to other reliability.
Follow-up study on verification plans and market failure modes. The research topic will
focus on discussing the potential display broken ratio on the smartphones, the product
design features, and the drop test conditions and test results in the research and
development stage. The prediction model of the failure mode which risk factors is also
studied. It can be used as a correct risk assessment for the specific failure mode of
the product at the initial stage, and adjust with the drop test method, conditions and
judgement criteria.

David A. Freedman, Statistical Models: Theory and Practice, Cambridge University
Press (2005)
Draper, N.R.; Smith, H. (1998). Applied Regression Analysis (3rd ed.). John Wiley.
R. Dennis Cook; Sanford Weisberg Criticism and Influence Analysis in Regression,
Sociological Methodology, Vol. 13. (1982), pp. 313–361
Francis Galton. Presidential address, Section H, Anthropology. (1885) (Galton uses
the term "regression" in this paper, which discusses the height of humans.)
Lindley, D.V. (1987). "Regression and correlation analysis," New Palgrave: A
Dictionary of Economics, v. 4, pp. 120–23.
Hazewinkel, Michiel, ed. (2001) [1994], "Regression analysis", Encyclopedia of
Mathematics, Springer Science+Business Media B.V. / Kluwer Academic Publishers.
Fox, J. (1997). Applied Regression Analysis, Linear Models and Related Methods.
Sage
Meade, N. and T. Islam (1995) "Prediction Intervals for Growth Curve Forecasts"
Journal of Forecasting, 14, pp. 413–430.
Kvalseth, Tarald O. (1985). "Cautionary Note about R2". The American Statistician.
39 (4): 279–285.
Devore, Jay L. (2011). Probability and Statistics for Engineering and the Sciences
(8th ed.). Boston, MA: Cengage Learning. pp. 508–510.
Richard Anderson-Sprecher,“ Model Comparisons and R 2”, The American

Statistician,Volume 48,Issue 2,1994,pp.113-117。
Imdadullah, Muhammad. "Coefficient of Determination".
Ronald Fisher, Statistical Methods for Research Workers, 1925
Fisher 1925, p. 47, Chapter III. Distributions.
標準差維基百科連
結:https://zh.wikipedia.org/wiki/%E6%A8%99%E6%BA%96%E5%B7%AE

52

回歸分析 維基百科連結:https://zh.wikipedia.org/wiki/%E8%BF%B4%E6%AD%B8%E5%88%86%E6%9E%90
P-Value 維基百科連結:
https://en.wikipedia.org/wiki/P-value
R
2維基百科連結:https://en.wikipedia.org/wiki/Coefficient_of_determination
TCO (Total Cost Ownership)維基百科連
結:https://en.wikipedia.org/wiki/Total_cost_of_ownership
MIL-STANDARD 810F, Test Method 516.6 Shock, Procedure II – Material to be
packaged
IEC 68-2-32 Test Ed: Free Fall
EBC NEWS 手機維修資訊
Web Link: https://fnc.ebc.net.tw/FncNews/Content/60444
周文賢(2005)。統計學。台北市:智勝文化事業公司。
林震岩(2006)。多變量分析 SPSS 的操作與應用。台北市:智勝文化事業公司。
吳宗正(1993)。迴歸分析。台北市:三民書局。
陳耀茂(2004)。時間數列分析的 SPSS 使用手冊。台北市:鼎茂圖書。
陳志安(2005)。應用線性迴歸模型。台北市:華泰文化事業公司。
黃俊英(2000)。多變量分析。台北市:華泰文化事業公司。
鄧聚龍、郭洪(1996)。灰預測原理與應用。台匇市:全華科技圖書股份有限公司。
張貴琳,薄型化液晶顯示器模組之強化結構,國立中興大學機械工程研究所碩士論
文,2007。
南寧漢、黃智傑,折疊式行動電話 LCD 模組於落下測試之有限元素分析,明基電通股
份有限公司網通事業群,2005。
林茂興,TFT-LCD 模組 LGP 衝擊強度之有限元素分析,奇美電子股份有限公司
陳伯群,TFT-LCD 掉落模擬分析及驗證研究,國立中央大學機械工程研究所碩士論
文,2003。
陳君明,TFT-LCD 衝擊模擬分析及驗證研究,國立中央大學機械工程研究所碩士論
文,2002。

53

王志榮,行動電話之掉落有限元素分析模型研究,國立中興大學應用數學研究所碩士
論文,2007。
劉彥麟,應用六標準差於中小型尺寸 TFT LCD 面板強度之改善,國立清華大學工業工
程與工程管理學系碩士論文,2009。
Huet, R. and Ming Wu,LCD Display Strength: Why Edge Preparation Matters,
Portable Information Devices,2007。