Journal article
Robustness properties of Cuscore statistics for monitoring a nonstationary system
Quality and reliability engineering international, Vol.24(7), pp.817-841
11/2008
DOI: 10.1002/qre.931
Abstract
We investigate the performance of cumulative score (Cuscore) charts in monitoring a nonstationary process subject to minimum mean-squared error (MMSE) feedback control. Specifically, we have previously designed Cuscore statistics to discover spike or step signals hidden in a nonstationary noise disturbance for feedback-controlled processes. Here, we study the properties of the proposed policies in terms of their robustness to process model misspecification, to deviations from assumed values for their parameters, to model misspecification for the disturbance as well as to model misspecification for the signal. We found that, in general, the Cuscore charts are fairly robust to process and signal model misspecification. However, they are less robust to departures from its true parameter values. We present a numerical example to demonstrate the importance of understanding the Cuscore's robustness as well as its construction and application. Copyright © 2008 John Wiley & Sons, Ltd.
Details
- Title: Subtitle
- Robustness properties of Cuscore statistics for monitoring a nonstationary system
- Creators
- René Valverde-Ventura - Eli Lilly and Company, Tippecanoe Laboratories, Lafayette, IN 47909, U.S.AHarriet Black Nembhard - The Pennsylvania State University, University Park, PA 16802, U.S.A
- Resource Type
- Journal article
- Publication Details
- Quality and reliability engineering international, Vol.24(7), pp.817-841
- Publisher
- John Wiley & Sons, Ltd
- DOI
- 10.1002/qre.931
- ISSN
- 0748-8017
- eISSN
- 1099-1638
- Number of pages
- 25
- Language
- English
- Date published
- 11/2008
- Academic Unit
- Engineering Administration; Industrial and Systems Engineering
- Record Identifier
- 9984121961002771
Metrics
8 Record Views