Objective Introduction Variable Charts Attribute Charts Examples
Statistical process control methods are analytical and graphical methods for for monitoring process variation sequentially in time. Shewhart control charts are the principal procedures for assessing process stability. They are based on the premise that control limits provide bounds for natural process variability (i.e., variation due to common causes) and that variation beyond these bounds is due to process changes (i.e., variation due to special causes). The goal of quality control methodologies is to identify special causes and to eliminate (or minimize) their effects on process variation.
Control charts are of two types: variable control charts for numeric variables and attribute control charts for binary (or binomial) variables. These charts are constructed by plotting summary statistics computed on each of a series of samples obtained sequentially over time.
Variable control charts are plots of the sample means (a measure of process location) and standard deviations or ranges (measures of process variability) over time. Other measures of location and variability could be used but the mean (x-chart), standard deviation (s-chart), and range (R-chart) are nearly universally used.
The following control chart applet plots the sample means together with control limits.
Process variability is gauged relative to the control limits, which provide probability bounds for the natural variability of the process. Various methods are used to compute the control limits, but a simple statistical approach will be used here.
Concrete compressive strengths (Hogg and Ledolter, p. 192)
Coating weights