Control Charts

Custom Search

Control Charts (Page 2 of 2)

To implement control charting for a given process, the following steps are usually taken:

1) identify the process and/or equipment to be subjected to control charting;

2) identify the process output parameter to be charted;

3) check the chosen parameter's data if they constitute a normal distribution;

4) determine the sampling method and plan;

5) construct the preliminary control chart indicating just the upper and lower specification limits of the process;

6) conduct the preliminary data collection to gather baseline data upon which the characteristics of the control chart will be based;

7) calculate the appropriate statistics and the control limits of the chart from the initial data collected;

8) complete the control chart by including the process mean and control limit lines; and

9) initiate the actual control charting.

The premise of control charting is that the output data charted constitute a normal distribution, which is a symmetrical bell-shaped distribution described by two numbers: its center (the mean of the data) and its spread. The spread of a normal distribution on the left of its center is equal to that on the right. For the purpose of control charting, this spread on either side is equal to 3 standard deviations, such that data falling outside 3 standard deviations on either side are considered outliers.

Thus, referring to Figure 1 again, the process mean line of a control chart is a horizontal line whose y-coordinate is equal to the mean of the charted data. The upper/lower control limit lines (UCL and LCL) are the horizontal lines whose y-coordinates are equal to the mean of the charted data plus/minus 3 times the standard deviation of the charted data, respectively, or: UCL = Mean + 3 Stdev and LCL = Mean - 3 Stdev. Note that the process mean line is exactly between the control limit lines.

In further reference to Figure 1, the lines for the upper and lower specification limits (USL and LSL) of the control charts are simply the horizontal lines whose y-coordinates are equal to the maximum and minimum output data allowed by the process (oftentimes the customer specifications), while the target line is the horizontal line exactly between these two specification limit lines.

Interpretation of control charts is not difficult. However, the engineer has to be aware of the common guidelines used in control chart interpretation. Some 'symptoms' that indicate that a process is out of control are: 1) one or more points are outside the control limits; 2) nine (9) consecutive points are on one side of the average; 3) six (6) consecutive points are increasing or decreasing; and 4) fourteen (14) consecutive points are alternating up and down. If any of these out-of-control symptoms are observed, the engineer has to initiate an out-of-control process investigation.

See Also: SPC

HOME