FEED PROFILE OPTIMIZATION OF A BATCH FED ALCOHOLIC FERMENTATION BY USING GENETIC ALGORITHM TO MAXIMIZE ETHANOL PRODUCTION
Lucas Costa Barreto1; Samuel Vitor Saraiva1; Frede Oliveira Carvalho1; João Nunes Vasconcelos1; Jessika da Rocha Silva1
1 Federal University of Alagoas
doi:10.20906/CPS/CILAMCE2017-0911
Resumo
This paper aims to develop a mathematical modeling and computer simulations of an alcoholic fermentation with inconstant batch feed. Fermentation is a process with large application in the industry, which objective is to obtain valuable products such as ethyl alcohol, also known as ethanol. Ethyl alcohol production in Brazil is growing fast due to politic investments to reduce pollution levels caused by combustion gases. Produced mainly by the black treacle (molasses), the Brazilian ethanol fermentation may seem simple. However, it is a microbiological process, which means there are many variables to be analyzed as well as the metabolic biochemical reactions and various types of transport phenomena. Economically speaking, evaluate the variables that compose this process becomes an important step in optimizing the production and reducing costs. One of the most important aspects in a batch feed fermentation is the substrate feed profile on the bioreactor, which determines how much ethanol is produced. In this work, a strategy to optimize the parameters of each substrate feeding profile is developed by using objective functions. The goal is to maximize the ethyl alcohol concentration in the product. The Genetic Algorithm (GA) approach is a heuristic method based on the natural selection principle postulated by Darwin as well as the genetic succession principle postulated by Mendel. The GA method has been highlighted in the literature due to its good performance. Thus, in most optimization cases, a GA follows the steps: ability, selection, crossing, mutation, acceptation, exchange, and test. The algorithm was implemented in MATLAB interacting with the objective function to adjust the feed profile parameters, aiming to the maximum ethanol production.
Palavras-chave: fermentation; genetic; algorithm; bath; fed; modeling; simulation; alcohol