|
Contents
| Quality Encyclopedia
| Discussion Blogs
Interpreting
an EWMA Chart
Standard Case (Non-wandering
Mean)
Always look at Range chart
first. The control limits on the EWMA chart are derived from the average
Range (or Moving Range, if n=1), so if the Range chart is out of control,
then the control limits on the EWMA chart are meaningless
On the Range chart, look
for out of control points. If there are any, then the special causes must
be eliminated. Remember that the Range is the estimate of the variation
within a subgroup, so look for process elements that would increase variation
between the data in a subgroup. Brainstorm and conduct Designed Experiments.
Note that Auto Drop is not invoked for EWMA charts.
After reviewing the Range
chart, interpret the points on the EWMA chart relative to the control limits.
Run Tests are never applied to a EWMA chart, since the plotted points are
inherently dependent, containing common points. Never consider the points
on the EWMA chart relative to specifications, since the observations from
the process vary much more than the Exponentially Weighted Moving Averages.
If the process shows control
relative to the statistical limits for a sufficient period of time (long
enough to see all potential special causes), then we can analyze its capability
relative to requirements. Capability is only meaningful when the process
is stable, since we cannot predict the outcome of an unstable process.
Wandering Mean Chart
Look for out of control
points. These represent a shift in the expected course of the process, relative
to its past behavior. The chart is not very sensitive to subtle changes
in a drifting process, since it accepts some level of drift as being the
nature of the process. Remember that the control limits are based on an
exponentially smoothed prediction error for past observations, so the larger
the prior drifts, the more insensitive the chart will be to detecting changes
in the amount of drift.
See also:
When
to Use an EWMA Chart
EWMA
Control Limits
EWMA
Wandering Mean Control Limits
|
