Split and Conquer! Using SAS/QC to Design Quality Into Complex Manufacturing Processes

From sasCommunity

[edit] Author

José Ramírez & Randy Tobias

[edit] Abstract

A complex manufacturing process has many stages, with different factors active at different stages. How do you discover which factors are important at a particular stage, and what factor, or factors, from an early stage interacts with one, or more, at a later stage (e.g. the source of raw material versus final packaging)? Split-plot experiments come to the rescue! They are a key tool for improving final product quality by studying how all stages affect final product quality. Until recently the design of fractional factorial experiments with more than one stage was not trivial.

In this paper we show how the ADX Interface for design of experiments in SAS/QC makes it very easy to design and analyze the most common experiments of this type. For more complex situations, recent advances in PROC FACTEX enable you to custom design experiments for processes with several stages. Two examples are presented showcasing these new features in ADX and PROC FACTEX.

[edit] Paper

[Paper in PDF form