"Non-scientific" experimentation in design of product and processes

Abstract

College students in engineering and physical sciences are often educated in ©scientific" methods of experimentation. For many years there have been advocates of statistical tools for research and development in design as well as prototyping and testing, yet their acceptance into engineering and technical curricula has been limited. The discussions in this paper explain the rationale for the use of statistical methods such as statistical design of experiments for superior product and process performance. Both educators and students in technical fields would realize that data-driven tools are indeed useful and important for designed-in quality and reliability. In fact, savings and opportunities for enhanced performance via virtually cost free techniques could often be discovered as well. The approach taken, largely based on statistical design of experiments, could be perceived to go against ingrained concepts of "scientific methods" but the efficacy of new product and process design can be readily explained and demonstrated. Such an understanding would be extremely useful in formulating effective design curricula at both graduate and undergraduate levels. Copyright © 2007 by ASME.