Feedback - Advantages of Statistical Process Control
SPC is a clear example of negative-feedback system control, providing management with timely information on the state of the process, allowing corrective action to be taken if required. Equally, SPC indicates when a process should be left to run, thereby saving on unnecessary adjustments to the process and, perhaps, avoiding non- conformance which could result from such adjustments. The following simple example may help to make the latter point clear. Consider a process with a target output value of 10.0 which, when running in a stable manner, has an inherent variability (see common cause variability, later) such that 50 per cent of the output is at the mean output value, 25 per cent measures 0.1 below, and the balance of 25 per cent measures 0.1 above the mean output value. Assume that the initial process setting is such that the mean output value matches the target value of 10.0. Twelve consecutive measurements might produce the following results (see table). Note that none of the measured values are outwith the expected range: with SPC, no adjustment of the process would be recommended.
Now consider the same situation in the absence of SPC. The operator can adjust the process setting but cannot directly determine its mean value: he can only measure individual instances of the output, which, as we have seen, may differ from the mean as a result of the inherent variability of the process. Suppose, now, that the operator measures each output, calculates the deviation from the target value and adjusts the output by that amount. In the next table, the first output value (10.1) is above the target value (10.0) by 0.1. He adjusts the process setting to reduce the output value by this amount but, unknown to him, the process setting is now 9.9. As a result, the second output value is 9.8, not 9.9 (as it was in the previous table). The deviation is now 0.2 below target. The operator adjusts the process to increase the output by 0.2, taking the mean setting from 9.9 to 10.1, and so on.
Notice the effect: under SPC, as shown in the first table, six of the values would be at the target value and the others would range between 0.1 above to 0.1 below target. In the absence of SPC, as shown in the second table, only two of the values are on target and the others range from 0.1 above to 0.2 below target. The process adjustments have increased the variability of the output.