Chaughule, Thorat - Statistical Analysis/Design of Experiments
ous applications of statistical designs used in food dehydration so far available in the
literature.
8.9.1. Case Studies
8.9.1.1. Optimization of spray drying of ginger extract
Jangam and Thorat (2010) have optimized conditions for spray drying of water ex-
tract of ginger. RSM with CCD was used to design the experiments. The design consisted
of 30 experiments (interested readers please refer to Jangam and Thorat, 2010)with
each factor varied over five levels. Air inlet temperature (120-160°C), air flow rate (40-
60 Nm^3 /hr), atomization pressure (1.5-2.25 kg/cm^2 ) and liquid feed flow rate (2.5-4
ml/min) were selected as independent variables. Moisture content, water activity, per-
centage retention of 6-gingerol, flowability and porosity of spray dried powder were the
responses. The software used was Design expert 6.0.10.
Jangam and Thorat (2010) fitted the experimental data to a second order polynomial
model and regression coefficients were obtained. The generalized second-order poly-
nomial model used in the response surface analysis was as follows:
= +∑= +∑= +∑ ∑= =+
4
i 1 i j4
ji 1 i,j4
i 1(^4) i (^2) i
(^0) i 1 i i
Y B BX BX B XX
Where, Bo, Bi and Bi,j are the coefficients of regression and Xi and Xj are the inde-
pendent variables. The mathematical models for each response have been evaluated us-
ing multiple regression method. The significance of the equation parameters for each
response variable has analyzed using the analysis of variance. Design expert software
used for regression analysis, generated regression coefficients, analysis of variance and
correlation coefficients for the model. The model adequacy can be checked on the basis
of correlation coefficient (R^2 ). In the article, the optimization was carried out using nu-
merical optimization tool, by setting the desired goal for each response variable. Opti-
mized conditions were obtained for the criteria of minimum moisture content and water
activity and maximum porosity, flowability and retention of 6-Gingerol by assigning
same importance to all responses. Figure 8.7 shows the effect of air temperature and air
flow on flowability. It can be seen that there exists optima at the center for flowability.
The inner circle shows the set of factor levels responsible for optimum response. Like-
wise, many response surface plots have been obtained for the different responses. Table
Jangam and Thorat (2005) also have reported the R^2 values of the response studied
which signified the accuracy of the models adapted.