Date: July 23, 1999
Creator: Balazs, B
Description: A sixteen-run, Resolution IV, fractional factorial screening design has been used to evaluate the relative significance of seven independent material and process variables for an electrically deposited commercial acrylic paint. A Resolution IV design was chosen so that variable interactions could be detected without any interference from the effects of the variables themselves. Because resource limitations did not permit use of a Resolution V design, the two-way interactions were confounded with one another in groups of three, which unfortunately made interpretation of the results more complicated. Each design point was replicated once and a pair of centerpoints were run at the beginning, in the middle and at the end of the design points. A total of 38 test panels were prepared. Independent variables used in this experiment were total charge, current density, mixer speed, silica content, crossover point, temperature and solids fraction. The magnitude of each independent variable's effect on the dependent response variables was estimated by fitting a first-order model with two-way interaction terms. Three-way interactions were assumed to be insignificant. Interaction confounding was eliminated by using the significance of the independent variables to determine which of the three possible interactions was significant. Data analysis was performed using the ...
Contributing Partner: UNT Libraries Government Documents Department