Chan, L.K.; Gan, E.H.; Mak, T.K. (1992, ASQC) The University of Manitoba, Winnipeg, Manitoba, Canada R3T 2N2
The successful design of experiments (DOE) requires an appropriate management procedure and an effective technique. This paper describes the 7-Step Mdox management procedure and four useful DOE techniques.
The 7-Step MDox management procedure, which can be fitted into the Deming cycle, includes: (1) Identify the project and the DOE objectives; (2) Define the response quality characteristic to be improved and identify the controllable and hard-to-control factors which might affect the response; (3) Select the design technique; (4) Conduct the experiment; (5) Validate the experimental results through confirmation experiments; (6) Implement necessary process changes; and (7) Document the entire management procedure.
The four DOE techniques presented are: (1) Parameter Design (Taguchi, 1986), which uses inner array and outer array to give a clear representation between controllable and noise factors; (2) Transformation of the response quality characteristic (Box, 1988), which offers a more reliable analysis; (3) Dual response surface methodology (Vining and Myers, 1990); and (4) Regression approach (Chan and Mak, 1991). To illustrate these techniques, the authors provide an example of an electrical circuit design.
Design of experiments (DOE),Design,Management,Total Quality Management (TQM),Taguchi method,Statistics