Chaughule, Thorat - Statistical Analysis/Design of Experimentsruns may have to be added if an estimate of experimental error is to be obtained. These
limitations of one-factor-at-a-time optimization process can be eliminated by optimizing
all the affecting parameters collectively by statistical experimental design (Mason et al.,
2003 ; Dean & Voss, 1999).
8.8. STEPWISE PROCEDURE FOR EFFECTIVE DESIGN OF EXPERI-
MENTS
Optimization of a process is an ultimate goal in statistical design of experiment. In
typical drying experiment, researchers are usually interested in determining which
process variables affect the response (drying parameters). A logical next step is to de-
termine the region of the important factors that leads to the best possible response. For
example, if the response of a drying process is the retention of micronutrients then one
would look for region of retention of micronutrients. Several authors reported optimiza-
tion of experiments by various statistical drying process using full factorial, fractional
factorial, response surface methodology and one factor at a time.
Figure 8. 5 shows the steps involved in the process of optimization. Before starting
any process, it is prerequisite to know the current operating conditions in the process.
Commonly goal for the food drying process is based on maximum retention of micronu-
trients, color, texture, sweetness, principal constituents and minimum solid gain in the
case of pretreatments like osmotic dewatering process. Once the response objectives are
clear, various statistical tools can be employed to change the current factors settings or
sometimes change of drying systems. By studying various techniques finally optimum
settings can be obtained.
Figure 8. 5. Steps in involved in optimization of a process
Figure 8. 6 illustrates optimization scheme for any food, vegetable or fruit dehydra-
tion process. The first step is to study current operational settings. It is then important
to select appropriate factors with most possible influential levels for the process. If the
drying process is less complex and comprise of only two factors then factorial design or
fractional factorial design can be employed. The first order model can be generated for
the process to study the optimum. If the desired response is obtained then the process
can be said to be approaching optimum. These designs are handy tool for optimizing less
complicated process in short time. However, they are very cumbersome for the process
affected by many factors at the same time and hence not practical. If the process is more
Current experimental settingsImprovements & Modifications through Experimentation and ModelingOptimum experimental settings