ABSTRACT
This chapter discusses the techniques available for minimising such scatter or process variability when the manufacturing process is linear in nature. It describes the role of noise-design factor interactions in determining process variability within the simple framework of a 22 design. These ideas are generalised to any factorial or fractional factorial design using a first order response surface model. The technique is applied to the 25 experiments of the ausforming process. The chapter discusses the role of the prediction error of the response surface model in minimising process variability. It introduces a simple generalised linear model which can be used to minimise process variability when replicate tests from a linear design are available. A comparison is made between the response surface and generalised linear model approaches to minimising process variability.
