162 Chapter 7
sensory characteristics at the end of the chop-
ping step. These characteristics were evalu-
ated as a global index called the chopping
degree (CD). The processing conditions
established at the end of the Simplex algo-
rithm (six trials only) were 3 minutes and
2000 rpm. They achieved a high value for the
CD (4.8/5; 5 being the maximum value)
(Curt et al. 2004a ). This result was confi rmed
by another study using response surface
methodology, where the effects of four
process parameters — chopping duration,
speed, temperature, and pressure — on the
chopping degree were studied (Curt et al.
2004b ).
Chopping is often performed as a batch
operation. Another strategy to determine the
optimal processing conditions is to use the
repetitive nature of batch processes in batch -
to - batch methodologies. Curt et al. (2007)
showed that a batch - to - batch algorithm using
human knowledge was able to control the
process to obtain the desired sensory proper-
ties at the end of the chopping process. Ten
runs were carried out independently from
each other to validate the algorithm in various
processing situations. For each of the ten run
tests, only one batch was necessary to achieve
the targeted chopping degree.Conclusion
Emulsifi cation control is based on smart
combinations between ingredients ’ choice
and processing defi nition. Although commi-
nuted meat products are traditional products
and their manufacturing follows ancient rules
of thumb, new combinations have to be
invented to face changing requirements, such
as lowering cost, improving nutritional
balance, and decreasing energy consumption.
To achieve this, a better knowledge of ingre-
dients ’ properties and their behavior in com-
minuted meat products is required; the use of
new ingredients and technologies can be
useful; and the development of on - line
sensors and control strategies is necessary.ogy was used to determine the optimum salt
level (1.3% – 2.1%) and pectin level (0.25% –
1.0%) when olive oil replaced pork backfat
(0% – 100%) for the production of highly
acceptable low - fat frankfurters (9% fat, 13%
protein) (Pappa et al. 2000 ). Gunvor et al.
(2005) used a cross - mixture design to con-
struct the sensory attributes model for sau-
sages ’ fi rmness and color. The color and
fi rmness were instrumentally measured and
modeled as mathematical functions of bio-
chemical composition (protein, connective
tissue, and fat) and muscle content. These
models were constrained by acceptability
limits found through a consumer test.
Constraints were then applied in a nonlinear
least - cost optimization model. The objective
function to be minimized was the cost func-
tion of the meat ingredients, which were
varied. Constraints for protein, fat, and con-
nective tissue contents were also made
according to legal restrictions. Three optimal
solutions were compared. A least - cost solu-
tion was found fulfi lling consumer accept-
ability, without fulfi lling the legal restrictions.
In the second optimal solution, a bit more
expensive solution fulfi lling the legal restric-
tion without fulfi lling the consumer accept-
ability was found. In the third optimal
solution, the biochemical composition (legal
restrictions) and linear sensory attributes
were restricted but the total cost became sig-
nifi cantly higher compared to the previous
solutions. These results illustrate the diffi -
culty in fulfi lling several quality require-
ments (legal, sensory, and cost) using only
formulation parameters (quantities of the bio-
chemical components and protein sources).
Few studies deal with process optimiza-
tion. One diffi culty that has been encountered
in the optimization of processing conditions
is the measurement of certain food product
properties and the lack of suitable on - line
sensors. Curt et al. (2004a) used the Simplex
method to determine the value of two process
parameters, mixing duration and mixer rota-
tion speed, to obtain a product with desired