摘要
为精确描述多故障模式下设备寿命数据的变化规律,采用混合威布尔分布对其进行研究。由于混合威布尔分布包含的参数较多,其参数估计较为困难。为方便快捷地估计各参数,基于极大似然估计理论,建立混合威布尔分布参数优化模型,并采用布谷鸟搜索算法进行参数寻优。案例研究以柴油机喷油器寿命数据为对象,以赤池信息量准则为拟合优度评价指标,对比分析三参数和两重三参数威布尔分布的拟合效果。研究结果表明:所提参数寻优方法可行、有效,且两重威布尔分布的拟合效果良好。
For the aim of modeling the life data that have failure modes accurately,fits the data by mixture Weibull distributions. Due to too many parameters included in the mixture Weibull distributions, their parameter estimation is rather difficult. To estimate every parameters conveniently, a parameter optimization model for mixture Weibull distributions which is based on maximum likelihood estimation is established. And, Cuckoo Search （CS） algorithmis used to solve the model. The case study takes the life- time data of diesel injectors as the research object,and akaike information criterion as the evalution index of goodness of fit. The fitting effects of three parameters Weibull and two-component three-parameters Weibull are compared. The result proves that the proposed method in the paper is feasible and effective, and the two-component three parameters Weibull is better than three parameters Weibull.
作者
池阔
康建设
王广彦
吴坤
Chi Kuo Kang Jianshe Wang Guangyan Wu Kun(Ordnance Engineering College,Shijiazhuang 050003, Chin)
出处
《现代制造工程》
CSCD
北大核心
2016年第12期149-153,共5页
Modern Manufacturing Engineering
关键词
可靠性分析
两重威布尔分布
极大似然估计
布谷鸟搜索算法
赤池信息量准则
reliability analysis
two-component Weibull distribution
maximum likelihood estimation
cuckoo search algorithm
akaike information criterion
