Box-BehnkenDesign, Central Composite Design, Experimental Design, Model, Response Surface Method, Steepest Ascent/Descent

Year 2014, , 35 - 45, 01.01.2014

Mustafa Agah Tekindal Hülya Bayrak Berrin Özkaya Yasemin Yavuz

https://doi.org/10.17557/tjfc.78787

Cited By: 7

Abstract

SECOND-ORDER RESPONSE SURFACE METHOD: FACTORIAL EXPERIMENTS AN ALTERNATIVE METHOD IN THE FIELD OF AGRONOMY

Abstract

The main purpose in all experimental designs is to take into account the factors that are considered likely to have an effect on the response variable emphasized, and to minimize the error of experiment in this way. Bread, which is the staple human food, cannot have any negative effect on human beings as long as it is produced by using suitable materials under appropriate conditions. However, when inappropriate amounts of raw materials are used (e.g. non-optimal amounts of bran, yeast or other additives), bread threatens health. In this study, Box-BehnkenDesign (BBD) and Central Composite Design (CCD), the two different designs of the response surface method, were applied to a single dataset. Two designs were evaluated in terms of the results obtained. The purpose in the second-order factorial experiments is to identify the optimum levels of independent variables for the dependent variable. In this study, the implementation of second-order response surface model and interpretation of the results were based on 2k CCD (Central Composite Design) and BBD (Box-Behnken Design) with one replicate. In the CCD, the amount of bran added, flour type, the ratio of yeast added, furnace temperature, the duration of remaining in the furnace, and fermentation time were accepted to be significant factors that affected volume yield. In addition, R2 = 80.7% shows that the regression equation explains variables by 80.7%. In the BBD, the ratio of bran added, the type of flour, the ratio of yeast added, furnace temperature (only in quadratic form), the duration of remaining in the furnace (only in quadratic form), and fermentation time (only in quadratic form) were accepted to be significant factors that affected volume yield. Furthermore, R2 = 89.64% shows that the regression equation explains variables by %89.64. This method provides savings in terms of time and the amount of material by limiting the area at particular levels. Researcher may use the results of either CCD or BBD (whichever s/he deems suitable) according to the volume s/he wants to obtain

Keywords

Box-BehnkenDesign, Central Composite Design, Experimental Design, Model, Response Surface Method, Steepest Ascent/Descent