A Bayesian approach to degradation-based burn-in optimization for display products exhibiting two-phase degradation patterns. (November 2016)

• Record Type: Journal Article
• Authors: Yuan, Tao; Bae, Suk Joo; Zhu, Xiaoyan
• Is Part Of: Reliability engineering & system safety. Volume 155(2016:Nov.)
• Page Start: 55
• Page End: 63
• Publication Date: 2016-11
• Subjects: Burn-in -- Degradation model -- Gibbs sampling -- Hierarchical Bayesian model -- Reliability; Reliability (Engineering) -- Periodicals; System safety -- Periodicals; Industrial safety -- Periodicals; Fiabilité -- Périodiques; Sécurité des systèmes -- Périodiques; Sécurité du travail -- Périodiques; 620.00452
• DOI: 10.1016/j.ress.2016.04.019
• Languages: English
• ISSNs: 0951-8320

Abstract

Abstract: Motivated by the two-phase degradation phenomena observed in light displays (e.g., plasma display panels (PDPs), organic light emitting diodes (OLEDs)), this study proposes a new degradation-based burn-in testing plan for display products exhibiting two-phase degradation patterns. The primary focus of the burn-in test in this study is to eliminate the initial rapid degradation phase, while the major purpose of traditional burn-in tests is to detect and eliminate early failures from weak units. A hierarchical Bayesian bi-exponential model is used to capture two-phase degradation patterns of the burn-in population. Mission reliability and total cost are introduced as planning criteria. The proposed burn-in approach accounts for unit-to-unit variability within the burn-in population, and uncertainty concerning the model parameters, mainly in the hierarchical Bayesian framework. Available pre-burn-in data is conveniently incorporated into the burn-in decision-making procedure. A practical example of PDP degradation data is used to illustrate the proposed methodology. The proposed method is compared to other approaches such as the maximum likelihood method or the change-point regression. Abstract : Highlights: We propose a degradation-based burn-in test for products with two-phase degradation. Mission reliability and total cost are used as planning criteria. The proposed burn-in approach is built within the hierarchical Bayesian framework. A practical example was used to illustrate the proposed methodology.