Many quality improvement projects require some form of experimentation on a process whereby an investigator is able to examine the effect of purposeful changes on the settings of the input variables on the response. Today, statistical methods of experimental design are increasingly being adopted in optimization of food technological processes and in the development of new, health-supporting food of high quality. Statistical techniques, mostly standard central composite designs (CCD), are often used because of the belief that the experimental factors can be completely randomized. However, most industrial experiments often involve factors with hard to change (HTC) levels and those with easy to change (ETC) levels, whereby the HTC factor cannot be completely randomized and this leads to a split-plot design. This work designs and conducts a split-plot central composite experiment with oven temperature as the HTC factor and, amount of flour, baking powder, and amount of milk as ETC factors. Restricted Maximum Likelihood (REML) and generalized least squares (GLS) estimation techniques were used to analyze the generated data. Each main factor effect and the quadratic effects of temperature, baking powder, and milk, were highly significant. The stationary point was determined to be y ̂=11.047 at the points: A=〖250〗^o C,B=1.5 cups,C=1.5 teaspoonfull,and D=0.75cup.
Key words: Keywords: Experiment, split-plot CCD, Cake height, design, stationary point
|
