Statistical process control (SPC) is the use of statistical methods, mainly control charts, to monitor a production process and detect when its variation goes beyond what its normal, stable behavior would produce.
SPC was developed by Walter Shewhart at Bell Laboratories in the 1920s and later spread by W. Edwards Deming. Its core idea is two kinds of variation. Common-cause variation is the normal noise of a stable process. Special-cause variation comes from something specific, such as a worn tool, a new material lot or a drifted sensor. SPC aims to react to special causes and leave common-cause noise alone, since adjusting for noise makes things worse.
The main tool is the control chart. Samples are measured over time and plotted against a center line and upper and lower control limits, conventionally three standard deviations from the mean. A point outside the limits, or a non-random pattern such as a run of points on one side, signals a likely special cause. Common charts include X-bar and R for subgroup means and ranges, individuals charts, p and c charts for defect counts, and CUSUM or EWMA charts for small shifts.
A frequent confusion: control limits describe what the process does; specification limits describe what the customer requires. A process can be in control and still make out-of-spec parts. That gap is measured by capability indices such as Cpk.
Questions about Statistical process control
What is the difference between control limits and specification limits?
Control limits describe what the process does. Specification limits describe what the customer requires. A process can be in control and still make out-of-spec parts; capability indices such as Cpk measure that gap.
What is the difference between common-cause and special-cause variation?
Common-cause variation is the normal noise of a stable process. Special-cause variation comes from something specific, such as a worn tool, a new material lot or a drifted sensor. SPC reacts to special causes and leaves common-cause noise alone.
How does SPC relate to anomaly detection?
Control charts are among the methods used for anomaly detection, which range from statistical limits and control charts to clustering, isolation forests and autoencoders. Many machine-learning anomaly detectors are, in effect, multivariate versions of the control chart idea.
Related terms
Anomaly detection
Anomaly detection is the identification of observations or patterns that deviate significantly from normal behavior in data.
Root cause analysis
Root cause analysis (RCA) is a structured method for finding the underlying cause of a failure or defect, so the corrective action prevents recurrence instead of only treating the symptom.
Computer vision inspection
Computer vision inspection is the automated checking of parts or products for defects using cameras and image analysis, increasingly with deep learning models.
OEE (overall equipment effectiveness)
OEE (overall equipment effectiveness) is a measure of how effectively a machine or line is used during planned production time.