摘要
液体火箭发动机试验费用昂贵、可靠性要求高,有必要采用分析和决策技术对可靠性增长进行综合管理。传统的可靠性增长分析与决策是采用可靠性增长模型,包括经典可靠性增长模型和Bayes可靠性增长模型。可靠性增长模型以可靠度的点估计或置信下降作为决策标准,缺陷是没有考虑可靠度本身的不确定性以及由决策损失导致的严重后果。本文根据液体火箭发动机的特点,采用信息融合技术,建立了基于增长数据折合的Bayes指数分布可靠性增长模型,评估可靠性水平。针对评估结果存在的不确定性,采用Bayes风险决策方法确定停止可靠性增长试验时间的标准。工程应用表明:该方法可操作性强,得出的结果科学、可靠。
Liquid rocket engine demands high reliability but test expenditure is costly. It is necessary to manage reliability growth using analysis and decision technology. Reliability growth model includes classical and Bayesian reliability model. Reliability growth model is used by traditional reliability growth analysis and decision. Reliability growth model make decision based on reliability point estimate and has shortcomings without taking uncertain reliability and decision loss into consideration. According to the characteristics of liquid rocket engine, information fusion has been used. Bayesian exponential reliability growth model based on converting growth has been built. To the uncertainty of reliability, Bayesian risk method has been used to decide the stopping time for reliability growth test. Example shows the model has high operability, accuracy and feasibility.
出处
《宇航学报》
EI
CAS
CSCD
北大核心
2004年第6期655-658,662,共5页
Journal of Astronautics
