00234 �35b630d34288277a734379e56dd94a

236

Basley

(actual) sigma of .369. There is seen to be opportunity to reduce variability through elimination of special causes. The Y₂ sigma for the old modified Shewhart adjustments was .45, which is somewhat better than the open-loop sigma (.531) but much worse than the current EWMA controller performance.

Conclusions and Acknowledgements

It has been shown that algorithmic SPC based on the EWMA is practical to implement for control of laboratory data either in a manuał or automatic form. The adjustment strategy reduces process variability by absorbing drifts and is robust to errors in estimation of X and process gain, while the Shewhart chart on forecast errors provides an opportunity for operator involvement aimed at continuous improvement via eliminating special causes. The success of these new control systems is the result of team efforts. Credit for implementation of the Greenwood system goes to Bobby Burdette, Joe Crum, and Fred Sanders. The manuał Pensacola system was installed by John Pace, Jerry Smith, and Greg Ward.

References

Alexander, M. T., and Macklin, C. (1989). “Process Prediction with Geometrie Moving Averages,” 1989 ASQC Quality Congress Transactions, Milwaukee, WI: ASQC, 140-149.

Alwan, L. C., and Roberts, H. V. (1988). “Time Senes Modeling for Statistical Process Control,” Journal of Business and Economic Statistics, 6, 87-95.

Baxley, R. V. (1990). “A Simulation Study of Statistical Process Control Algorithms for Drifting Processes,” Statistical Process Control in Manufacturing, Keats, J. B., and Montgomery, D. C., Editors, New York: Marcel Dekker, Inc., and Milwaukee: ASQC Quality Press, 247-298.

Baxley, R. V. (1990). “Discussion”, Technometrics, 32, 13-16.

Box, G. E. P., and Jenkins, G. M. (1963). “Further Contributions to Adaptive Quality Control: Simultaneous Estimation of Dynamics, Non-Zero Costs”, Buli. Int. Statist. Inst., 34th Session, 943-974.