摘要
针对传统的 Shewhart 控制图使用当前观测数据进行测量系统分析(MSA)的稳定性分析,不易检测出过程均值中存在的小偏移,而指数加权滑动平均(EWMA)和累计和(CUSUM)控制图考虑了历史数据,可以克服Shewhart 控制图在小波动情况下检测能力不足的缺陷:首先通过 Montle-Carlo 模拟计算控制图性能评价指标ARL(average run length),分析了 EWMA 和 CUSUM 控制图的优越性;然后给出了 EWMA 控制图在 MSA 稳定性分析中的实施步骤.示例分析表明 EWMA 在小波动时比 CUSUM 和 Shewhart 控制图更为有效.
Traditional Shewhart control chart is often used to analyze the stability of measurement system analysis （MSA） based on current observation data. However, it is difficult to detect small shift using the control chart. By using history data instead, exponentially weighted moving average （EWMA） and cumulative sum control （CUSUM） control chart improve the detection ability to small fluctuation. This paper firstly analyzes the advantages of EWMA and CUSUM control chart by means of ARL（average run length） obtained by Montle-Carlo simulation. Then, an implementing procedure of EWMA control chart is presented for stability analysis of MSA. A case study shows that EWMA control chart is more efficient than CUSUM and Shewhart chart in the case of small fluctuation.
出处
《系统工程学报》
CSCD
北大核心
2008年第3期381-384,共4页
Journal of Systems Engineering
基金
国家高技术研究发展计划(863计划)资助项目(2006AA020204)
新世纪优秀人才支持计划资助项目(NCET-05-0485)
关键词
测量系统分析
指数加权滑动平均
稳定性分析
measurement system analysis （MSA）
exponentially weighted moving average （EWMA）
stability analysis
