Premium
An analysis of failure‐time distributions for product design optimization
Author(s) -
Field Dennis,
Meeker William Q.
Publication year - 1996
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/(sici)1099-1638(199611)12:6<429::aid-qre60>3.0.co;2-6
Subject(s) - replication (statistics) , computer science , design of experiments , product (mathematics) , process (computing) , path (computing) , reliability engineering , industrial engineering , data mining , engineering , statistics , mathematics , geometry , programming language , operating system
Analysis of experimental data is often a problem facing design and manufacturing engineers. Many experiments are run for the express purpose of making a decision between manufacturing process or material alternatives. Statisticians recommend replication in experimental design; however, methods of analysing experimental data, as presented in a majority of engineering curricula, generally review only the most basic situations (checking for a statistically significant difference between the means or variances of two samples, for example). If means and variances change with time, group comparisons may require more sophisticated analyses. This paper presents one method that takes into account shifts in group means and variances over time. Resistance temperature sensor drift data generated from six different design configurations are used as an illustration. This procedure takes into account all drift path information from multiple sensors in multiple groups.