检测过程尺度参数的EWMA非参数控制图  

An EWMA Nonparametric Control Chart for Monitoring Process Variability

作　　者：尚云艳[1] 于蓉蓉[1] 刘鑫 SHANG Yun-yan;YU Rong-rong;LIU Xin(College of Science,Xijing University,Xi’an 710123,China)

机构地区：[1]西京学院理学院,陕西西安710123

出　　处：《控制工程》2021年第5期1033-1040,共8页Control Engineering of China

基　　金：陕西省教育厅专项科研项目(20JK0967)。

摘　　要：当生产过程的分布未知时,非参数控制图是检测过程异常的一种有效手段。为提高对尺度参数的检测效率,提出一种新的非参数EWMA控制图;通过蒙特卡洛随机模拟方法计算控制限,同时得到运行长度的均值、方差和分位数,分析了非参数EWMA控制图受控状态和失控状态时的性能,并与LOG图(基于Logistic分布的Shewhart非参数控制图,简称LOG图)进行比较,模拟结果显示非参数EWMA控制图明显提高了对中小漂移的检测效率;最后引用实例证明非参数EWMA控制图对尺度参数的检测更有效。When the distribution of production process is unknown,nonparametric control chart is an effective method to detect process drift.In order to improve the efficiency of detecting scale parameters.A new exponentially weighted moving average(EWMA)nonparametric control chart is proposed in this paper.The control limits are calculated by Monte Carlo simulation method,and the mean,variance and quantile of the running length are obtained.IC state and OC state of the nonparametric control chart are analysed.Compared with LOG control chart(a Shewhart nonparametric control chart under Logitic distribution,called LOG control chart),the simulation results show that the EWMA nonparametric control chart can significantly improve the detection efficiency of small and medium drift.In the end,an example is cited to prove that the EWMA nonparametric control chart is more effective in the detection of scale parameters.

关 键 词：线性秩统计量 非参数控制图 Logistic分布 

分 类 号：O213.1[理学—概率论与数理统计]