多阶段可校正系统退化建模与可靠性评估

退化数据在小子样、高可靠、长寿命产品可靠性评估中的作用日益突显。针对工程实践中多阶段可校正电子装备可靠性评估难题,综合考虑任务的多阶段性和校正行为影响,建立了基于Wiener过程的性能退化模型,采用贝叶斯方法得到退化模型中漂移参数的后验分布估计,构建了光滑函数解决系统校正行为导致的函数不连续性问题,在此基础上根据首次到达时的概率分布提出了多阶段可校正系统的可靠性评估方法,给出相关数值算例,验证了模型、方法和分析结果的正确性和实用性。

Degradation modeling and reliability assessment of multi-stage system with calibrations

Degradation data are becoming more and more important in reliability assessment of products with small samples, high reliability and longevity. Aiming at reliability assessment of multi-stage electronic products with calibrations in the practice of engineering, a degradation model is developed in terms of degradation signals of multi-stage systems based on Wiener diffusion processes under pre-specified periodical calibrations, and the drift coefficient in the degradation model is inferred by the Bayesian method. Then, a smooth function is built to solve the problem of discontinuity resulted from pre-specified calibrations, based on which a method for Bayesian reliability assessment is presented for multi-stage system with calibrations in terms of the distribution of first passage time. Finally, a numerical example is given to validate the validity and practicability of the model, method and analysis results.