Orthogonal arrays are a specific type of fractional factorial experiment carefully selected to maintain orthogonality (independence) among the various factors and certain interactions. It is this orthogonality that allows for independent estimation of factor and interaction effects from the entire set of experimental results. Using orthogonal arrays for fractional factorial design reduces the analysis result resolution (i.e., factor effects are aliased with interaction effects as more factors are added to a given array); however, the significant reduction in the required number of experiments (cost) can often justify this loss in resolution as long as some of the interaction effects are assumed negligible. The designer must make an assumption as to which quantities are insignificant (typically the highest-order interactions) so that a single contributing quantity can be identified. In essence, for an orthogonal array of a given size, the more factors and interactions you want to study, the greater the confounding. Confounding occurs when columns are “shared” by quantities. The result of confounding is that you cannot determine which quantity in a given column produced the effect on the outputs attributed to that column (from postprocessing analysis).This results in lower confidence in the analysis results (because more assumptions of insignificant factors must be made). Orthogonal arrays have been used in design since as early as the 1940’s by Plackett and Burman, who used saturated designs (only studying factor effects), and were really popularized by Taguchi, who developed a family of 2- and 3-level orthogonal arrays to study interaction effects (Ross, 1988). The 3-level arrays also allow for an estimation of second-order effects (i.e., design space curvature). The use of orthogonal arrays provides a systematic and efficient method to study the design space and provide suggestions for improving the design. However, the actual tasks of selecting the appropriate orthogonal arrays to use and assigning the factors and interactions to columns can be tedious and overwhelming. The automation of this procedure in Isight allows a designer with little or no knowledge of orthogonal arrays to efficiently and effectively study the design space using this formal DOE methodology. | |||||||||