基于二维指数条件分布的加速寿命模型

Accelerated Life Model Based on Bivariate Exponential Conditional Distributions

摘要: 从概率意义上对寿命和应力之间的关系作了探索性研究。为了根据寿命和应力之间的概率相关性建立加速寿命模型，假设在给定应力条件下元件寿命服从指数分布，根据二维指数条件分布的定义和特性推导出一种BEC加速寿命模型。为了检验BEC模型的有效性，利用一组加速寿命试验数据，对BEC模型、Arrhenius模型、逆幂率模型进行了对比分析。结果表明，在加速应力范围，BEC模型能够正确描述寿命与应力之间的关系，与Arrhenius模型、IPL模型具有同等效力；而在正常应力范围，BEC模型的寿命外推精度低于Arrhenius模型和IPL模型。

Abstract: To establish an accelerated life model based on the probabilistic correlation between lifetime and stress, an accelerated life model, named bivariate exponential conditional (BEC) accelerated life model, is derived according to the definition and properties of BEC distribution. Experimental data is employed to valid the BEC model. The results indicate that the BEC model can describe the relationship between the lifetime and stress properly in the accelerated stress domain compared with the Arrhenius and inverse power law (IPL) models, however, the life-extrapolation accuracy of the BEC model is a little lower than that of the Arrhenius and IPL models.