Obtaining a neural model for ethanol´s concentration estimation at Héctor Molina distillery

Authors

  • Luis Eduardo López de la Maza Grupo de Análisis de Procesos, Facultad de Ingeniería Química, Universidad Tecnológica de La Habana “José Antonio Echeverría” (CUJAE), La Habana, Cuba
  • Lourdes Zumalacárregui de Cárdenas IGrupo de Análisis de Procesos, Facultad de Ingeniería Química, Universidad Tecnológica de La Habana “José Antonio Echeverría” (CUJAE), La Habana, Cuba
  • Osney Pérez-Ones IGrupo de Análisis de Procesos, Facultad de Ingeniería Química, Universidad Tecnológica de La Habana “José Antonio Echeverría” (CUJAE), La Habana, Cuba
  • Orestes Llanes-Santiago Facultad de Ingeniería Automática, Universidad Tecnológica de La Habana “José Antonio Echeverría” (CUJAE), La Habana, Cuba

DOI:

https://doi.org/10.1590/2224-6185.2018.2.%25x

Keywords:

neural networks, models, fermentation.

Abstract

The ethanol production for alcoholic beverages manufacturing based on fermentation at “Héctor Molina” distillery is performed in a bioreactor tank operating at discontinuous mode. Currently the biological processes’ modeling using artificial intelligence techniques has become a trend. In this paper, models with artificial neural networks (RNA), multilayer perceptron specifically, were created. These models were able to estimate the concentration of ethanol in the fermenters of the Héctor Molina distillery using MATLAB 2013. The best neural topology presents four input variables and six neurons in the hidden layer with a mean of the mean square error of 4,34 10-4 and a correlation factor with the experimental data of 0,916.

Downloads

Published

2018-05-09

How to Cite

López de la Maza, L. E., Zumalacárregui de Cárdenas, L., Pérez-Ones, O., & Llanes-Santiago, O. (2018). Obtaining a neural model for ethanol´s concentration estimation at Héctor Molina distillery. Chemical Technology, 38(2), 315–325. https://doi.org/10.1590/2224-6185.2018.2.%x

Issue

Vol. 38 No. 2 (2018)

Section

Artículos

License

This work is licensed under the Creative Commons CC-BY-NC-ND

Most read articles by the same author(s)