Premium
Using Profile Monitoring Techniques for a Data‐rich Environment with Huge Sample Size
Author(s) -
Wang Kaibo,
Tsung Fugee
Publication year - 2005
Publication title -
quality and reliability engineering international
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.913
H-Index - 62
eISSN - 1099-1638
pISSN - 0748-8017
DOI - 10.1002/qre.711
Subject(s) - quantile , homogeneity (statistics) , sample size determination , control chart , computer science , sample (material) , data mining , chart , statistical process control , standard deviation , statistics , plot (graphics) , process (computing) , mathematics , machine learning , chemistry , chromatography , operating system
In‐process sensors with huge sample size are becoming popular in the modern manufacturing industry, due to the increasing complexity of processes and products and the availability of advanced sensing technology. Under such a data‐rich environment, a sample with huge size usually violates the assumption of homogeneity and degrades the detection performance of a conventional control chart. Instead of charting summary statistics such as the mean and standard deviation of observations that assume homogeneity within a sample, this paper proposes charting schemes based on the quantile–quantile (Q–Q) plot and profile monitoring techniques to improve the performance. Different monitoring schemes are studied based on various shift patterns in a huge sample and compared via simulation. Guidelines are provided for applying the proposed schemes to similar industrial applications in a data‐rich environment. Copyright © 2005 John Wiley & Sons, Ltd.