透過您的圖書館登入
IP:3.145.46.18

並列摘要


In reliability analysis, the models used to aggregate the effect of time-varying stresses are generally invariant to the order of stress application, or stress sequence. A model is referred to as being insensitive to the stress sequence if the application of stress z_1 for t_1 units of time followed by stress z_2 for t_2 units of time is equivalent to applying stress z_2 for t_2 units of time followed by stress z_1 for t_1 units of time in terms of the estimated reliability at the end of the test. The use of such insensitive models may lead to inaccurate reliability estimation when units are subject to stress sequencing. This paper proposes new classes of reliability models that have the ability to account for the stress sequence effect based on Exponential and Weibull models. This effect has been extensively investigated in the analysis of fatigue testing, and models from this area are first reviewed. The optimal design of a test plan for the estimation of the parameters of reliability models subject to stress sequence is examined under type I censoring. General guidelines for conducting the test are also discussed.

延伸閱讀