ASTM D4583-95(2009)

This practice will provide the following: (1) a statistical summary of individual production run data plotted on a control chart; (2) a statistical summary of data from multiple production runs; (3) a procedure to relate the average and variation of these data groups to specification limits, and (4) indexes for comparing different manufacturing units for projecting future capabilities or as historical reference.

1.1 This practice covers (1) a statistical procedure for analyzing the test data generated in the statistical process control of a carbon black manufacturing process; (2) a format for reporting process capability determined from the analysis of control chart data of an individual production run, and (3) a format for reporting process performance determined from the analysis of control chart data of an individual production run.

1.2 This practice specifically applies to the analysis of pelleted carbon black samples taken during the manufacturing process prior to storage. This practice does not apply to shipment samples taken from hopper cars or other containers or packages.

1.3 This practice is specifically designed to be used for those test methods given in Classification D 1765 which specify target values. However, these techniques are applicable to other test methods on carbon black.

1.4 This practice describes the calculation for two methods of determining capability factors from an analysis of process control data.

1.4.1 Process capability (Cp) is a measurement of variation calculated from the process control chart data with the use of an estimated standard deviation (ˆσ) from the mean value of the moving range (R) chart. The calculation of the process capability (Cp and Cpk) indexes can be used as historical information or to predict future performance of the process, but are only valid when the process is in a state of statistical control.

1.4.2 Process performance (Pp) is a measurement of variation calculated from the process control chart data using sample standard deviation(s). The calculation of the process performance (Pp and Ppk) indexes are used for a historical reference of a process' performance and does not require a state of statistical control.