Develop and Evaluate Control Charts
The plot on the Process Control Chart screen shows each control point in
the process control history with optional one-sigma upper and lower
uncertainty limits. Each control point is either a deviation from
the nominal value shown on the
Process Control Data Screen, a deviation from the declared value of
a check standard, a deviation from the mean of sampled process data, or
a deviation between the current measured value and the measured value of
the previous control point.
The control chart also shows the upper control limit (UCL) and lower
control limit (LCL) defined on the Process Control Data screen and
a linear fit to the plotted data. The linear fit, representing the drift
over time of the process of interest, is bounded by upper and lower
standard uncertainty curves. This drift over time provides a tool
for forecasting process trends.
The control data history table is displayed below the control chart.
Records can be added to the table, modified or deleted in the same
manner as on the Process Control Data screen. Upon exiting the
Process Control Chart screen, you will be asked whether or not you
want to implement these changes in the process control data history.
Projecting Trends
One of the central objectives of process control is the detection of drift or
other mechanisms by which process (or subject parameter) bias changes with
time. In SPCView, such drift is shown on the process control chart as a
linear fit of process (or subject parameter) bias vs. time. The fit is
obtained by a weighted or unweighted linear regression analysis of the
control data points. The upper and lower curves bracketing the linear
fit are ± one standard deviation, a standard deviation representing the
uncertainty in the curve fitting process.
If a positive or negative drift mechanism is present, the rate of drift is
estimated as the slope of the linear fit. The "intercept" of the linear fit
comprises an estimated initial value for the process history. Depending
on the drift rate, the time at which the linear fit will intersect either the
UCL or the LCL will be displayed in the Curve Fit Information section
of the screen. The times at which the upper and lower uncertainty curves will
intersect the control limits are also displayed.
Process Evaluation
Evaluating whether a process is in or out of control is essentially a matter
of deciding on whether observed control points are in or out of control. The
occurrence of an out of control point is called an out of control event.
Clicking Evaluate on the Menu bar activates the
Process
Evaluation Worksheet, which can be used to view the confidence for
rejecting a process as being out-of-control at a selected point. This
worksheet also provides visibility of the risk of falsely rejecting an
in-control process.
Interval Analysis
The interval of time suitable for taking the next process control point (e.g.,
for process monitoring or parameter calibration) is determined through
the use of the
SPC Interval Worksheet, which is activated by clicking Compute
Interval on the Menu bar.
Analysis Reports
Hardcopy reports of SPC analysis results and of the control chart can be
generated through the use of the
SPC Report Options Screen
accessed by clicking
Report on the Menu bar and selecting the Print option.
|