DOI
10.1016/0951-8320(95)00117-4
Abstract
A simple, easy-to-use graphical method is presented for use in determining if there is any statistically significant trend or pattern over time in an underlying Poisson event rate of occurrence or binomial failure on demand probability. The method is based on the combined use of both an exponentially weighted moving-average (EWMA) and a Shewhart chart. Two nuclear power plant examples are introduced and used to illustrate the method. The false alarm probability and power when using the combined procedure are also determined for both cases using Monte Carlo simulation. The results indicate that the combined procedure is quite effective in rapidly detecting either a small or large step increase in the Poisson rate or binomial probability over time.
Document Type
Post-print Article
Publication Date
1996
Publisher Statement
Copyright © 1996 Elsevier B.V.
DOI: 10.1016/0951-8320(95)00117-4
The definitive version is available at: https://www.sciencedirect.com/science/article/pii/0951832095001174
Full Citation:
Martz, Harry F., and Paul H. Kvam. "Detecting trends and patterns in reliability data over time using exponentially weighted moving-averages." Reliability Engineering & System Safety 51, no. 2 (1996): 201-207. doi:10.1016/0951-8320(95)00117-4.
Recommended Citation
Martz, Harry F. and Kvam, Paul H., "Detecting trends and patterns in reliability data over time using exponentially weighted moving-averages" (1996). Department of Math & Statistics Faculty Publications. 179.
https://scholarship.richmond.edu/mathcs-faculty-publications/179