基于性能退化可靠性分析的混合Brown运动模型

Mixed Brownian Motion Model Based on Reliability Analysis of Performance Degradation

对于高可靠、长寿命的产品,很难通过寿命试验或加速寿命试验方法来获取产品的失效时间,这给以产品失效数据为基础的传统可靠性分析方法带来了极大的困难。基于产品性能退化数据的可靠性分析不需要将产品做到失效,而是寻找产品的性能退化规律对其可靠性进行分析,这为高可靠、长寿命产品的可靠性评定提供了一条可行的途径。由于不同的试验产品其性能参数的退化可能服从不同的随机过程模型,在一般Brown运动退化模型的基础上,提出了带漂移的混合Brown运动模型,依据Brown运动的概率性质,研究了混合Brown运动模型的一些概率性质,并给出了模型参数的估计式及其可靠度函数表达式。以GaAs退化数据为例,对混合模型进行了验证。

For some high reliability and long life electronic products, it is hard to get their failure times through life tests or accelerated life tests, which brings difficulties for the conventional failure-data-based reliability analysis. The reliability analysis based on performance degration data provides a feasible approach for the reliability evaluation of high-reliability and long-life products because they don＇t need the failure data of a product but the degradation pattern. For the degradation of each product performance may follow a different random process, a mixed Brownian motion model with drifts was proposed based on the general Brown motion degradation model. The parameter estimation of the proposed model and its reliability function were provided based on the analysis of its probabilistic nature. Finally, the model was verified with the GaAs degradation data as an example.