Skip to Main Content
 

Global Search Box

 
 
 

ETD Abstract Container

Abstract Header

Semi-parametric Bayesian Inference of Accelerated Life Test Using Dirichlet Process Mixture Model

Abstract Details

2015, Doctor of Philosophy (PhD), Ohio University, Mechanical and Systems Engineering (Engineering and Technology).
Accelerated life testing (ALT) is commonly used to estimate the reliability of highly reliable products. This dissertation develops statistical models to predict useful life of nano devices with data collected under constant-stress ALT and step-stress ALT. As an example of nano devices, nc-MoOx embedded ZrHfO high-k dielectric thin film is studied with respect to its physical properties, failure mechanisms, and long-term stability. The devices used for ALT and reliability prediction demonstration have identical structure with this nc-MoOx embedded device. This research develops a semi-parametric Bayesian method to analyze ALT. The model assumes a log-linear lifetime-stress relationship, without assuming any parametric form of the failure-time distribution. The Dirichlet Weibull mixture model is employed to model the failure-time distribution under a given stress level. The model is fitted with a simulation-based algorithm, which implements Gibbs sampling to analyze ALT data and predicts the failure-time distribution at a normal stress level. Two practical examples related to the reliability of nanoelectronic devices are presented for constant-stress ALT, including one right-censored data and one complete data set. One right-censored practical example is demonstrated for simple step-stress ALT. All three examples illustrate the capability of the proposed methodology to provide accurate prediction of the failure-time distribution at a normal stress level.
Tao Yuan (Advisor)
100 p.

Recommended Citations

Citations

  • Liu, X. (2015). Semi-parametric Bayesian Inference of Accelerated Life Test Using Dirichlet Process Mixture Model [Doctoral dissertation, Ohio University]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=ohiou1447193154

    APA Style (7th edition)

  • Liu, Xi. Semi-parametric Bayesian Inference of Accelerated Life Test Using Dirichlet Process Mixture Model. 2015. Ohio University, Doctoral dissertation. OhioLINK Electronic Theses and Dissertations Center, http://rave.ohiolink.edu/etdc/view?acc_num=ohiou1447193154.

    MLA Style (8th edition)

  • Liu, Xi. "Semi-parametric Bayesian Inference of Accelerated Life Test Using Dirichlet Process Mixture Model." Doctoral dissertation, Ohio University, 2015. http://rave.ohiolink.edu/etdc/view?acc_num=ohiou1447193154

    Chicago Manual of Style (17th edition)