Meaning
Statistical conditions where the average value of a process variable changes over time signal a lack of control in the manufacturing environment. A non stationary process mean occurs when external factors such as ambient temperature or material batch variations cause the central tendency to wander. This behavior complicates the calculation of long term capability indices.
Temporal Shift
Observing the mean move over hours or days indicates that the process is sensitive to time dependent variables. A morning startup might produce parts at the lower end of the tolerance, while the afternoon production shifts toward the upper end. This wandering makes it difficult to set a permanent offset.
Center Instability
Variations in the average value increase the risk of producing scrap even if the process spread remains small. If the mean moves faster than the inspection frequency, the system can produce bad parts before an operator detects the change. Stability requires identifying the root cause of the shift.
Control Charting
Tracking the mean on a shewhart chart reveals patterns that indicate a non stationary state. Points that consistently fall on one side of the center line suggest a bias in the system. This visualization is the first step in stabilizing the output.
