Statistical Process Control

Statistical process control is the application of statistical methods to monitor and control the quality of a production process. Statistical process control is appropriate to support any repetitive process, including financial auditing and accounting, IT operations, health care processes, and clerical processes such as loan arrangement and administration, customer billing etc.

ISO 9000

Walter A. Shewhart pioneered SPC at Bell Laboratories in the early 1920s, developed the control chart in 1924.

W. Edwards Deming invited Shewhart to speak at the Graduate School of the U.S. Department of Agriculture and served as the editor of Shewhart’s book Statistical Method from the Viewpoint of Quality Control (1939), which was the result of that lecture.

• Articles by Donald J. Wheeler
• The only reason to collect data is to take action

Process Behavior Charts (Shewhart individuals control chart)

Use

• Find out when a change has occurred in a process
• Evaluate if your process changes made any significant difference to the process
• Determine whether a process is predictable or not
• In predictable processes, predict the range within further measurements will fall if the conditions do not change
• Assess a process performance against a requirement e.g. customer demand and expectations.

References

• Deming Alliance: Process Behaviour Charts - An Introduction
• 2013 Donald J. Wheeler: When the XmR Chart Doesn’t Seem to Work
• 2023 Commoncog: How to become data driven
• 2023 Commoncog: Process Behavior Charts
• Shewhart individuals control chart

Calculations

• Some value with ideally over 23 data points
• calculate average(X)
• calculate mR
• calculate average(mR)
• calculate UNPL = average(x) + (2.66 * average(mR))
• calculate LNPL = average(x) - (2.66 * average(mR))
• calculate URL = 3.269 * average(mR)

Terms

• mR – moving range, the difference between two successive values
• UNPL – Upper Natural Process Limit, gives us the best case scenario forecast
• LNPL – Lower Natural Process Limit, gives us the worst case scenario forecast
• URL – Upper Range Limit

Interpreting

1. Any point outside the limits indicates the presence of an exceptional cause that has a dominant effect.
   • If not accounted by cyclical or seasonal patterns
2. Run of eight successive values all on the same side of the central line may indicate the presence of an exceptional cause that has a weak but sustained effect.
3. Runs near the limits three successive values all within 25% of the upper or lower limits may indicate the presence of an exceptional cause that has a moderate but sustained effect.
4. Alternating at short intervals likely means two systems of causes that need to be separated.
5. Wide limits are a noise warning that it would be a mistake to react to changes in these values as it is virtually impossible to assign an explanation to why a point goes up or why it goes down.

In a stable process, most of values fall within the 50% area between the limits.