00303 S6f5ec79c7f009f5b2911cd990f7402

306

Peterson & Parvey

of four large U.S. corporations were examined to determine how they were organized to apply designed experiments. The benchmarked companies were chosen because they are considered "best in class" or "world class" implementors. The ultimate goal of this investigation was to identify a preferred approach for applying DoE methods in a multi-divisional company.

Introduction

Designed experiments play an important role in the development and improvement of products and processes. A designed experiment is a series of tests in which changes are madę to input variables in order to observe changes in an output response. Classical DoE and Analysis of Variance (ANOVA) techniąues are used to determine significant factors in problem solving and product/process optimization. Factorial experiments and response surface techniąues are used to study relationships between factors and identify optimum combinations.

Many ąuality engineering and robust design methodologies feature the use of designed experiments. Techniąues championed by Genichi Taguchi were introduced in the United States in the 1980's at AT&T Bell Labs, Ford Motor Company, ITT Defense, and the Xerox Corporation. Taguchi methods are used to enhance design robustness and minimize variation in products and processes. Because there has been a well documented, ongoing debate regarding the superiority of various DoE methods, specific DoE/Taguchi methods will not be described or endorsed in this chapter. The different DoE techniąues are all based in the same underlying philosophies and scientiflc analyses -- basie textbooks may be referenced for additional information on the technical subject matter.

Developing A DoE Program

Many companies have realized dramatic ąuality improvements through the use of DoE techniąues. Success stories have been published in ąuality publications and presented at conferences. Many approaches have yielded good results, but there are several essential elements to implementing a successful DoE program. Because of the technical naturę of the tools, there is a need for technical experts with detailed knowledge of the topics. There is also a need for training; employees at all levels of an organization need to receive education regarding the use of these methods. Finally, there is a need for applications. The benefits of hiring technical experts and conducting specific training courses can never be realized if the tools are not applied.