Abstract:
Demand for natural fibers as an alternative to synthetic fibers is on the rise. Constraints
of current fiber extraction units are limited to single raw material and has low
energy efficiency. Hence, this study was taken up for the design and development of
a multiple fiber extractor, modelling and optimization of the machine operating
parameters, and economic analysis of the developed prototype. The designed
machine operates based on the raspador principle. Box Behnken design (BBD) and
artificial neural network (ANN) were used to evaluate the performance and optimize
the variables of the designed multiple fiber extractor. The input variables included
roller speed (550, 735, and 920 rpm), gap/clearance of the roller (2, 3, and 4 mm) and
the pitch/distance between the spline slots (0.5, 1.5, and 2.5 mm), while the output
variables were yield (kg), capacity (kg/h), energy consumption (kJ), and efficiency (%).
The prediction accuracy of quadratic models (R2 = 0.962–0.985) developed in BBD
were better than the ANN models (R2 = 0.831–0.946). The optimized combination
was found to be 714.23 rpm roller speed, 3.12 mm gap, and 1.49 mm pitch of the
blade, which yielded (1.48 kg) a desirability value of 0.95. The XRD results indicated
the highest crystallinity index was recorded for Pineapple leaf fiber compared to
banana fiber. SEM results revealed that extracted BF having typical network structure
(cellulose, hemicellulose, lignin, and waxes) compared to smooth PLF (lignin and
hemicellulose). The fibers extracted using the developed machine was utilized for
making biodegradable cutlery.
