Scholarlink

<p>A modified exponentially weighted moving average (EWMA) control chart is presented to monitor carbon monoxide (CO) levels in the air. There were two modifications made to this chart to differentiate it from the conventional EWMA method. First, in the conventional EWMA method, it is assumed that the data are uncorrelated. However, air quality data of CO concentrations are autocorrelated. Therefore, the correlation structure in the data were first modeled using the Holt–Winters method, and then the EWMA control chart was applied to the residuals. Second, in the conventional EWMA chart, alarms are used to indicate a distributional shift, such as an increase in the mean, when the charting statistic exceeds a predetermined control limit. In this method, only a binary decision regarding whether the monitored process is in control is provided. With the modified EWMA method, <i>p</i> values are computed at each time point, assuming that the process is in control. Unlike conventional methods, which use control limits, the modified method uses <i>p</i> values, which indicate the strength of the signal when the process is out of control. In addition, even when the process is in control, <i>p</i> values indicate how stable the process is. The modified EWMA control chart was applied to monitor CO concentrations during a series of forest fires from 2018–2020 in Butte County, California.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Using P Values to Design an EWMA Control Chart to Monitor Carbon Monoxide Levels in the Air

  • 期刊论文

  • 发表日期：

    2025年10月1日

查看全文
  • Sesha Dassanayake

摘要

A modified exponentially weighted moving average (EWMA) control chart is presented to monitor carbon monoxide (CO) levels in the air. There were two modifications made to this chart to differentiate it from the conventional EWMA method. First, in the conventional EWMA method, it is assumed that the data are uncorrelated. However, air quality data of CO concentrations are autocorrelated. Therefore, the correlation structure in the data were first modeled using the Holt–Winters method, and then the EWMA control chart was applied to the residuals. Second, in the conventional EWMA chart, alarms are used to indicate a distributional shift, such as an increase in the mean, when the charting statistic exceeds a predetermined control limit. In this method, only a binary decision regarding whether the monitored process is in control is provided. With the modified EWMA method, p values are computed at each time point, assuming that the process is in control. Unlike conventional methods, which use control limits, the modified method uses p values, which indicate the strength of the signal when the process is out of control. In addition, even when the process is in control, p values indicate how stable the process is. The modified EWMA control chart was applied to monitor CO concentrations during a series of forest fires from 2018–2020 in Butte County, California.