In process control, variability in dispersion of the quality characteristic is controlled using a dispersion control chart and it is essential to check variability before applying a location chart. Many dispersion measures including robust measures are exposed to control charting procedure to improve quality. The difference in estimating process dispersion when measures from center points like mean, trimmed mean and Winsor mean and its influence in process control are studied here. This study presents the robust control charts based on trimming and Winsorization and its performance comparison with classical dispersion chart in terms of in-control and out-of-control signals as well as percentage of trimming or Winsorizing. Average Run Length is also studied to ensure the performance of charts. Performance of charts is compared based on real data also.
keywords
Control Limits, Outliers, Robust Control Charts, Modified Trimmed Standard Deviation,
and Average Run Length.