Keywords: fractional design, Taguchi design, two-level design, dispersion effects.
An industrial Taguchi experiment was performed to study the influence of several controllable factors on the mean value and the variation in the percentage of shrinkage of products made by injection moulding. For studying the variation, three noise factors were also included in the design. All factors were set at two levels.
The problem is a `nominal-is-best' problem where the aim is to reach a certain tartet for the percentage shrinkage, at the same time having as small as variation as possible about the target value. The design that was applied is a so-called Taguchi L8(27)-design with seven controllable factors. At each setting of the controllable factors, the noise factors were varied according to a Taguchi L4(23)-design.
Data File (tab-delimited text)
|Engel, J. (1992). Modelling variation in industrial experiments. Applied Statistics 41, 579-593.|
|Steinberg, D. M., and Bursztyn, D. (1994). Confounded dispersion effects in robust design experiments with noise factors. Journal of Quality Technology 26, 12-20.|
|Engel, J., and Huele, A. F. (1996). A generalized linear modeling approach to robust design. Technometrics 38, 365-373.|
|Lee, Y., and Nelder, J. A. (1998). Generalized linear models for the analysis of quality-improvement experiments. Canadian Journal of Statistics 26, 95-105.|
|Huele, A. F., and Engel, J. (1998). Response to Nelder and Lee. Technometrics 40, 172-175.|