Abstract
The problem of detecting a shift of a percentile of a strength distribution in a quality control setting is considered. Shewhart-type charts for lower percentiles are developed based on small samples. Lower and upper control limits for “percentile charts” for Weibull (extreme-value) and lognormal (normal) distributions which were obtained by Monte Carlo simulations are presented. Average run lengths of the Weibull percentile charts and the relationships between them and X̄ and R charts for Weibull data are also investigated. An application to the breaking stress of carbon fibers is presented to illustrate the methods.
Additional information
Notes on contributors
W. J. Padgett
Dr. Padgett is a Professor and Chairman of the Department of Statistics.
John D. Spurrier
Dr. Spurrier is a Professor in the Department of Statistics.