Meaning
Statistical procedure for partitioning total variation into constituent parts. Using an anova capability analysis identifies whether observed variance stems from the manufacturing process itself or from the measurement system. This analysis determines the statistical weight of factors such as operator shift and batch number.
It measures the signal to noise ratio to ensure that a process shift is detectable against background fluctuation.
Variance Component
Quantification of individual source contributions defines the stability of the production environment. A variance component represents the percentage of total variation attributable to a single factor. High variance in the machine factor suggests equipment maintenance is required.
Low variance in the operator factor indicates training is uniform.
Measurement System
Evaluation of the gauge itself prevents the reporting of false negatives. The measurement system must contribute less than ten percent of the total process variation to be considered acceptable. If the system consumes thirty percent or more of the tolerance, the underlying process capability remains unknown.
Reliability depends on separating measurement error from actual part variation. Accuracy in these findings dictates whether the production line enters full rate operation or returns to the design phase. Precision is not the same as accuracy and this distinction is maintained throughout the study.
Tolerance Interval
Calculation of the range containing a specific proportion of the population provides a production safety margin. The tolerance interval accounts for both the mean and the standard deviation of the sample. Wide intervals relative to specifications indicate a high probability of producing nonconforming units.
Narrowing this interval requires reducing common cause variation.