Response Surfaces, Blocking and Split Plots: An Industrial Experiment Case Study

ABSTRACT

This case study is from an experiment done on a step of a manufacturing process. The objective of the experiment is to determine the appropriate levels of three factors to optimize two responses in order to ensure that the manufactured product meets customer requirements. The experiment performed is more complex than standard experimental designs because it has aspects of response surface, blocking, and split-plot elements that pose unique analysis challenges. These challenges include determination of the appropriate error term for significance testing and checking of model assumptions. We illustrate the correct analysis including the residual analysis and multiple response optimization.

KEYWORDS:

• block
• central composite design
• DOE
• equivalent estimation
• multiple response optimization
• residuals
• subplots
• whole plots

ACKNOWLEDGMENTS

We thank Jason Wiersdorf for providing the data used in this case study. We also thank the editor and the reviewers for their helpful suggestions that significantly improved this article.