摘要
针对航天长寿命产品在研制、生产、试验和使用过程中,存在大量的性能退化数据和少量寿命数据的特点,给出了融合性能退化数据和寿命数据获取其寿命分布的一种Bayes方法,以及利用其性能退化参数来进行可靠性评估的方法.首先,根据Fisher信息量确定维纳过程漂移参数和扩散参数的无信息验前分布;然后,利用性能退化数据结合其性能退化的独立高斯增量特点,获取参数的第一次验后分布,利用寿命数据结合其寿命的逆高斯分布获取参数的第二次验后分布;最后,利用漂移参数和扩散参数的验后分布,对产品进行了可靠性评估.本文模型充分利用了性能数据和寿命数据,在很大程度上提高了评估的精度.
Due to the fact that there are more performance degradation data and less life data during development, producing, test and running, this paper provides a Bayes method to obtain the life distri- bution through fusing the two kinds of data, and the reliability assessment method based on performance parameters. At first, we obtain the noninformative prior distribution of the drift and volatity parameters of the Wiener process by using fisher information. And then, by using the degradation data, this paper gives the first posterior distribution based on the independent Gaussian increment, and the second posterior distribution based on the inverse life distribution by using the life data. The proposed approach is used to assess the reliability of spaceflight longlife product in the end. The results show that the new approach can improve the precision of assessment greatly.
出处
《系统工程理论与实践》
EI
CSSCI
CSCD
北大核心
2012年第11期2517-2522,共6页
Systems Engineering-Theory & Practice
基金
国家自然科学基金(70901024
71071158)
中国博士后科学基金
湖南省优秀研究生创新基金
国防科技大学校优秀研究生创新基金(B0905 01)