Abstract
ABSTRACT The corrosion of medium carbon steel (MCS) is examined with respect to the three factors cooling temperature, HCl concentration, and holding temperature. The factor levels consist of a 23 factorial arrangement along with six center runs. The response is the weight loss of MCS for each of the fourteen specimens. Specimens are grouped together in a furnace for cooling temperature and in an oven for holding temperature according to a split-block design. Specimens are also assigned to HCl concentration according to a split-plot design with respect to the other two factors. This case study shows how to incorporate this design structure into the analysis of variance in order to determine important effects and optimal settings of the factor levels. Recommendations are also given for obtaining suitable replication and for fitting second-order models with respect to this design.