COMPARISON OF EXPERIMENTAL DESIGNS USED TO STUDY VARIABLES DURING HAMMER MILLING OF CORN BRAN

Corn bran at three moisture contents (6%, 13%, and 24% d.b.) and feed rates (11.89, 23.86, and 35.86 g s(-1)) was ground using a hammer mill operated at two different tip speeds (19.0 and 47.0 m s(-1)). Geometric mean diameter (GMD) and energy consumption were observed within a split-plot experimental design. Additionally, the statistical precision of the split-plot experimental design was compared to the statistical precision of the experiment if it had been conducted as a randomized complete block design (RCBD) with a factorial treatment structure. Moisture content (MC) had a significant effect on GMD (final particle size) but not on energy consumption. Tip speed (TPS) had significant effects on both GMD and energy consumption. Feed rate (FR) had a significant effect on energy consumption but no significant effect on GMD of corn bran. Energy consumption was observed to increase with increasing TPS and FR. Conducting the experiment as a split-plot design substantially increased the precision for energy consumption with regard to FR and TPS, whereas the precision for MC was decreased when compared to an RCBD. Particle size GMD precision was only slightly affected by using a split-plot design as compared to conducting the experiment as an RCBD. Finally, if hard-to-change (HTC) and easy-to-change (ETC) factors exist, it is important to consider different-sized experimental units. Neglecting to consider this issue can result in inefficient use of resources and possibly lead to inaccurate and biased results.