基于贝叶斯法的变速箱集中项修过程能力分析  

Analysis on the Process Capability of Gearbox Item Repair Based on Bayesian Theory

作　　者：陈春良[1] 石文华[1] 吕会强[1] 叶红兵[1] 

机构地区：[1]装甲兵工程学院技术保障工程系,北京100072

出　　处：《兵工学报》2012年第7期864-869,共6页Acta Armamentarii

摘　　要：对比了贝叶斯理论方法和传统经典方法应用于变速箱集中项修的不同结果,验证了贝叶斯理论方法应用于部件集中项修过程的实用性和合理性。在对变速箱换件修理流程进行分析的基础之上,确定了集中项修作为整个流程的关键过程。根据集中项修过程中均值和方差随环境而变的实际情况采用了贝叶斯原理,把质量特性的均值和方差作为服从相应分布的随机变量,利用先验分布及历史数据求出后验分布的贝叶斯估计,再进一步应用统计过程控制(SPC)技术分析关键过程能力,给出了过程能力指数的贝叶斯估计及其97.5%置信下限,并与经典理论方法进行了对比分析。对比分析的结果表明,贝叶斯方法分析集中项修过程能力比传统经典方法更加合理、准确。Comparing the Bayesian theory method and the traditional means, the different results validate the rationality and practicability of applying of the bayesian means to the item repair. Based on analysis on the process of replace repair of gearbox, concentrative item repair is confirmed as the key process. According to the fact of replace repair, i. e. , the mean value and standard deviation is changeable with the environment, the Bayesian theory is adopted, in which the mean value and standard deviation is deemed as stochastic. The Bayesian estimate of posterior distribution is obtained from the history data and prior distribution, furthermore, the key process capability is analyzed by using SPC technology with the 97.5% one-sided confidence lower limit of process capacity index given with comparison with classical theory. The comparing result indicates that the applying of Bayesian means to the item repair is more rational and practical than the traditional method.

关 键 词：数理统计学 集中项修 贝叶斯估计 过程能力指数 控制图 

分 类 号：TJ81[兵器科学与技术—武器系统与运用工程]