March 23, 2011
It's Out of Control
In my last blog, we discussed the importance of using control limits that are calculated from a true measure of the process variation. Now that we have the correct limits, we can use them to monitor the process behavior. They tell us if a subgroup is within the expected process behavior (within the control limits) or if the subgroup is outside of the expected process behavior.
Statistical Process Control is simply using data or subgroups to make a process behave the way that we would like for it to behave and the control limits are the tool that we use to do that. Our control limits represent the expected behavior of the process so we can say that any subgroup that is outside of the control limits (or outside of the expected behavior) is “out of control” or unexpected. If a subgroup is “out of control” we should investigate to determine the reason or assignable cause. If needed, an action should be taken. But be careful when considering what actions to take! The best action may not be to return the system to a state of “in control”!
If you know the reason for the process behavior change, and that change in behavior is an improvement which you can expect the change to continue in the future, then the proper action would be recalculation of the control limits to reflect the process change. Just don’t get too trigger happy adjusting the control limits! Be sure that these conditions are met first.
On the other hand, if the reason for the behavior change is not known and/or the behavior change is not an improvement and/or if you don’t know whether the change to continue in the future, then the goal becomes finding the reason and removing it so that the process can return to its desired behavior!