Ethyl Alcohol Production Optimization by Coupling Genetic Algorithm and Multilayer Perceptron Neural Network

roomycankerblossomΤεχνίτη Νοημοσύνη και Ρομποτική

23 Οκτ 2013 (πριν από 3 χρόνια και 7 μήνες)

59 εμφανίσεις

Poster Presentation 6
-
34


Ethyl Alcohol Production Optimization by Coupling Genetic Algorithm and
Multilayer Perceptron Neural Network


Elmer Ccopa Rivera*
, Aline C. da Costa and Rubens Maciel Filho


Faculty of Chemical Engineering, State University of Cam
pinas (UNICAMP),

CP 6066, Campinas, CEP 13081
-
970, SP, Brazil

Phone: 19
-
37883970 email: elmer@feq.unicamp.br



In this work, genetic algorithms (GA) and multilayer perceptron neural network (MLPNN) were used in
combination with modeling and simulation

to optimize an extractive alcoholic fermentation process. The
objective is to maximize productivity and conversion in the fermentor. A comparison is made between the
performances of two models when the process is optimized using a real
-
coded and binary
-
co
ded genetic
algorithms. This first one is a deterministic model, whose kinetics parameters were experimentally
determined as functions of the temperature, the second is a non
-
linear model represented for a MLPNN.
The optimization results are compared and t
he efficiency of the genetic algorithm is demonstrated.
Besides, the result demonstrated the prediction accuracy of MLPNN. Thus, MLPNN showed to be a very
powerful and flexible tool well suited for modeling the fermentation process due to an implicit corre
ctive
action arising from the training methodology and the associated estimation procedure. In addition to
establish optimal conditions for operation, the present methodology also makes possible to predict both
conversion and productivity when the system i
s disturbed in some way. This is useful no only for the
additional knowledge supplied about the process, but also for fermentation monitoring and control.