Inverse Gaussian process based reliability analysis for constant-stress accelerated degradation data

The constant stress accelerated degradation test has been widely used to evaluate product reliability when degradation measurements of product can be observed. This paper considers interval estimation procedures of the constant stress accelerated degradation model based on the inverse Gaussian process. The exact confidence interval is established for the shape parameter of the inverse Gaussian accelerated degradation model. Using the generalized pivotal quantity procedure, the generalized confidence intervals of other model parameters and some quantities of interest such as the pth percentile and reliability function of lifetime at the normal using stress level are derived. In addition, the generalized prediction interval is developed for the future degradation levels at the normal stress level used. The bootstrap confidence intervals of the proposed inverse Gaussian degradation model are also discussed, and the two interval estimation methods are compared. The performances of the proposed interval estimation procedures are assessed by Monte Carlo simulation in terms of coverage percentage and average interval length. Finally, an example is provided to illustrate the proposed procedures.(c) 2021 Elsevier Inc. All rights reserved.

Automated summary

This paper focuses on interval estimation procedures for the constant stress accelerated degradation model based on the inverse Gaussian process. Exact confidence intervals are established for the shape parameter of the inverse Gaussian accelerated degradation model. Generalized confidence intervals for other model parameters and quantities of interest are derived using the generalized pivotal quantity procedure. Additionally, generalized prediction intervals are developed for future degradation levels at the normal stress level. The performances of the proposed interval estimation procedures are evaluated using Monte Carlo simulation. A comparison between bootstrap confidence intervals and the proposed inverse Gaussian degradation model is also conducted. The importance of this study is to provide reliable interval estimates for the constant stress accelerated degradation model.