针对复杂机械系统可靠性评估未考虑可靠性实验故障数据具有多故障模式的不足,建立两重混合威布尔分布模型。为解决可靠性实验小子样故障数据问题,利用模糊聚类的方法实现对故障数据的分类,求得混合威布尔分布的形状参数,以此为基础将混合威布尔分布转换为具有共轭先验分布的混合指数分布。用贝叶斯估计求解混合指数分布各参数值,然后将求得的混合指数分布各参数值转换为混合威布尔分布的各参数值。最后以某型号卧式加工中心为例,求得加工中心的可靠性评估结果。
To solve the reliability assessment' s problem of complex mechanical system that not considering for the characteris- tics of multiple failure modes, this paper built a mixed double Weibull distribution model. To solve the small sample failure data' s problem, it used fuzzy clustering method to achieve fault data classification and obtained the sub-population' s shape parameter of the mixed Weibull distribution. Then the mixed Weibull distribution converted into mixed exponential distribution which had conjugate prior distribution. By using the Bayes estimation, it obtained the mixed exponential distribution' s parameters, then the mixed Weibull distribution' s parameters converted into the mixed exponential distribution' s parameters. At last, it took an example of a certain type machining center, and obtained the result of reliability evaluation.