Abraham, Bovas; Whitney, James B. (1990, ASQC) University of Waterloo, Ontario, Canada
This paper discusses the use of the exponentially weighted moving average (EWMA), charting technique when the data are obtained from continuous processes by taking successive measurements at approximately equally spaced time intervals.
The common cause variation is interpreted to have two basic sources one of which may affect the process level, causing it to drift. In such a context plots associated with the EWMA are more appropriate for detecting special causes of variation in the process than either charts for individuals or cumulative sum (CUSUM) charts. Two charts are considered; one to detect special causes of variation in the process and the other to monitor the process level. The discussion also suggests that these charts can be considered as a general purpose tool to signal the occurrence of special causes and trends in data from a variety of continuous processes. The plots associated with the EWMA provide a means for dynamically controlling a process.
Stability of a process is essential before process capability can be defined and calculated. In the present context the process has a drifting mean indicating instability. This leads to a reconsideration of the concept of process capability when a drifting mean is dynamically controlled.