A process is defined as being in control when the only source of variation is common cause. Through the use of statistical process control, a manufacturer can review production data and determine when common and special cause variations are at work in the process.
W. Edwards Deming estimated that 94% of all process variation is a result of common causes while 6% could be attributed to special causes. It is the identification and elimination of special cause variation that leads to a predictable, stable process output resulting in reduced scrap, improved product flow through the plant and increased profitability. Some examples of special cause variation include injecting the wrong parts for a specific order, damaging wax dies or heats with the wrong chemistry. However, some special cause variations are subtler and require a methodology in order to identify, monitor and control key input variable(s).
This paper will present a methodology for identifying sources of special cause variation along with systems which can be implemented to reduce or eliminate these issues resulting in improved process stability. A case study conducted with Wisconsin Precision Castings will be presented demonstrating the use of this methodology to identify and reduce variation within the investment casting process.