摘要
为探究在实际生产中,历史数据权重对控制过程的影响,以及减小权重,仅影响控制图一部分偏移灵敏的情形,提出双平滑系数控制图(DRES控制图).新的控制图用两个单边统计量重置的指数加权移动平均变异系数控制图来监测向上或向下的偏移,即选取两个大小不同的平滑系数,对原有控制图进行改进,重新构造了控制限,使得对大小偏移都有很好的监测能力.研究结果表明:利用平均链长的对比,新的控制图在变异系数发生无论大偏移还是小偏移都相对灵敏.
In order to explore the effect of different historical data weights on the control process in practical production,as well as to alleviate the situation that the control chart is only sensitive to a part of the shifts due to different weight selection,a double smoothing coefficient control chart(DRES control chart)is proposed in this paper.The two one-sided resetting exponentially weighted moving average coefficient of variation(CV)control charts are used to monitor the increase(or decrease)shifts of CV,that is,two different smoothing coefficients are applied to improve the performance of original control chart.The new control chart reconstructs the control limit so as to gain good monitoring ability for both shift cases.The findings show that the new chart is relatively sensitive in both large shifts and small shifts of CV.
作者
李翠
宋向东
LI Cui;SONG Xiangdong(School of Science,Yanshan University,Qinhuangdao 066000,China)
出处
《辽宁工程技术大学学报(自然科学版)》
CAS
北大核心
2019年第5期486-490,共5页
Journal of Liaoning Technical University (Natural Science)
关键词
控制图
变异系数
平均链长
偏移量
指数加权移动平均值
control chart
the coefficient of variation
average run length
the shifts
exponentially weighted moving average