摘要
针对及时修正试验策略,根据Fries可靠性增长模型和Duane学习曲线特性,基于不同的失败类型对模型进行了扩展。用进化算法对似然函数求极值,获得了参数极大似然估计,以及最终研制阶段可靠性点估计。对某固体火箭发动机可靠性增长实例的分析结果表明,与经典评估方法相比,本文模型更具优越性。
According to the test strategy of correction in time, the reliability growth model for various failure categories was extended based on the Fries＇s reliability growth model and Duane＇s learning-curve property in this paper. The extreme value of likelihood was calculated through evolutionary algorithm. The maximum likelihood estimation of the parameters and reliability point estimator of the last development stage were obtained. The analysis result of reliability growth for some solid rocket engine showed that the method in this paper had more advantages than the classical estimation,
出处
《上海航天》
北大核心
2005年第6期48-50,共3页
Aerospace Shanghai
基金
国防预研基金资助项目(41328010504)
关键词
固体火箭发动机
可靠性增长
进化算法
极大似然估计
可靠性点估计
Solid rocket engine
Reliability growth
Evolutionary algorithm
Maximum likelihood estimation
Reliability point estimation
