Controlling Change: Process Monitoring and Adjustment During Transition Periods

Statistical control implies stable random variation about a target value that is produced by some common cause. In many manufacturing systems, the majority of the processing time is spent in the steady-state phase. Hence, it makes sense that a system is in statistical control most of the time and the operator or decision-maker should avoid introducing unnecessary variation by making adjustments to it. However, a transition period is common in many operations. Examples of transitions are process start -up (or ramp-up) periods, process changeovers, and product or raw material grade changes. In situations such as these, the process is not operating in its steady-state phase; hence, it is also unlikely that a system is in statistical control as a transition begins. Out-of-specification product made during a transition must be disposed of or scrapped, causing economic loss due to raw material and labor costs. During operation, whether in transition or steady state, any real system will need some method of regulation or adjustment. In the introduction to Statistical Control by Monitoring and Feedback Adjustment, Box and Lucefio (I) pointed out why this is true: