Testing for the validity of the assumptions in the exponential step-stress accelerated life-testing model

Zhejiang Gongshang University, Hangzhou, PR China
Computational Statistics & Data Analysis (Impact Factor: 1.4). 05/2009; 53(7):2702-2709. DOI: 10.1016/j.csda.2009.01.008
Source: RePEc


In the application of the exponential step-stress accelerated life-testing model, there are usually three assumptions required: (1) for any stress level, the lifetime distribution of a test unit is exponential; (2) for any stress level, the mean life of a test unit is a log-linear function of stress; (3) a cumulative exposure model holds. This paper explores the validity of assumptions 1 and 3. It is proved that assumption 3 is unnecessary to the exponential step-stress accelerated life-testing model. A test statistic is proposed to test the validity of the assumptions 1. The null distribution of the test statistic is derived. A Monte Carlo simulation is given to study the power of the proposed test procedure. Finally, an example is given to illustrate the proposed test procedure.

Download full-text


Available from: Bingxing Wang, Jan 22, 2014
  • Source
    • "Bhattacharyya and Soejoeti [19] proposed the tampered failure rate model. It is worth mentioning that Wang [20] gave a necessary condition to decide whether or not a given model such as the cumulative exposure model is rational. Miller and Nelson [21], as well as Bai et al. [22], discussed optimum plans for simple SSALT. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Constant-stress procedures based on parametric lifetime distributions and models are often used for accelerated life testing in product reliability experiments. Maximum likelihood estimation (MLE) is the typical statistical inference method. This paper presents a new inference method, named the random variable transformation (RVT) method, for Weibull constant-stress accelerated life tests with progressively Type-II right censoring (including ordinary Type-II right censoring). A two-parameter Weibull life distribution with a scale parameter that is a log-linear function of stress is used. RVT inference life distribution parameters and the log-linear function coefficients are provided. Exact confidence intervals for these parameters are also explored. Numerical comparisons of RVT-based estimates to MLE show that the proposed RVT inference is promising, in particular for small sample sizes.
    IEEE Transactions on Reliability 09/2014; 63(3):807-815. DOI:10.1109/TR.2014.2313804 · 1.93 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: In this article the parameter estimation method of Step-Stress Accelerated Life Testing (SSALT) model is discussed by utilizing techniques of Generalized Linear Model (GLM). A multiple progressive SSALT with exponential failure data and right censoring is analyzed. The likelihood function of the SSALT is treated as being a censoring variate with Poisson distribution and the life-stress relationship is defined by a log link function of a GLM. Both the maximum likelihood estimation and the Bayesian estimation of GLM parameters are discussed. The iteratively weighted least squares method is implemented to obtain the maximum likelihood estimation solution. The Bayesian estimation is derived by applying Jeffreys' non-informative prior and the Markov chain Monte Carlo method. Finally, a real industrial example is presented to demonstrate these estimation methods.
    IIE Transactions 08/2010; 42(8-8):589-598. DOI:10.1080/07408170903459976 · 1.37 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: In simple step-stress experiments under Type-II censoring with the cumulative exposure model and exponentially distributed lifetimes, maximum likelihood estimates (MLE) of the expected lifetimes may not exist due to the absence of failure times either before or after the stress change point. For this reason, when planning a step-stress experiment, the change point could be chosen so as to minimize the probability of non-existence of the MLE. These non-existence probabilities are examined and compared in the one- as well as the two-sample situations. Moreover, the optimal allocations of the change points are discussed and the effects of the use of non-optimal choices for the change points are assessed.
    Computational Statistics & Data Analysis 01/2011; 55(1):236-247. DOI:10.1016/j.csda.2010.04.013 · 1.40 Impact Factor
Show more