Should Observations Be Grouped for Effective Process Monitoring?
- Journal of Quality Technology
- October 2004
- Volume 36 Issue 4
- pp. 343-366
- Reynolds, Marion R., Stoumbos, Zachary G.
- Virginia Polytechnic Institute and State University, Blacksburg, VA, Rutgers, The State University of New Jersey, Piscataway, NJ
[This abstract is based on the authors' abstract.] During process monitoring, it is assumed that a special cause will result in a sustained shift in a process parameter that will continue until the shift is detected and the cause is removed. In some cases special, however, causes may produce a transient shift that lasts only a short time. Control charts used to detect these shifts are usually based on samples taken at the end of the sampling interval d, but another option is to disperse the sample over the interval. For this purpose, combinations of two Shewhart or two cumulative sum (CUSUM) charts are considered. Results demonstrate that the statistical performance of the Shewhart chart combination is inferior compared with the CUSUM chart combination.