Statistical properties of control charts for the mean

摘要

Control charts are statistical methods widely applied to improve the quality and productivity of products and processes. In this research work we study the statistical properties of these procedures, with the aim of determining their optimal application to control the mean of a quality characteristic. The sampling frequency and size are analyzed, evaluating the performance of the control chart for the sample mean as a function of the sampling scheme under several out of control situations and process capabilities. The statistical properties of this chart are studied when supplementary runs rules are considered, and compared with the CUSUM chart and the sample means chart with only one out-of-control signal (a point outside the control limits). The statistical performance of the moving averages chart are determined as a function of the grouping size. The effectiveness of this chart is compared with that of the individuals chart, and CUSUM and EWMA procedures. Finally, the effect of serial correlation on the statistical properties of all the control charts mentioned above, is analyzed.