Modelado y predicción basados en redes neuronales del rendimiento de ácido acético en la fermentación de Gluconobacter oxydans utilizando aguas residuales lácteas
@limentech, Ciencia y Tecnología Alimentaria · 15.12.2025 · 23 (2) · 5-18
DOI: 10.24054/limentech.v23i2.4359 ↗Kurz gefasst
Fachartikel in der Zeitschrift @limentech, Ciencia y Tecnología Alimentaria, online erschienen am 15.12.2025. Verfasst von Rafael González Cuello und 2 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0).
Abstract (Originalsprache)
Acetic acid (AA) is a valuable bioproduct with broad industrial applications in the food, pharmaceutical, and chemical sectors. In this study, Gluconobacter oxydans was employed to produce acetic acid using a modified medium containing 12% dairy wastewater as a cost-effective substrate. The effects of glucose concentration, incubation time, and temperature on acetic acid production were evaluated, and the process was modeled using an Artificial Neural Network (ANN) based on a multilayer perceptron (MLP) architecture (3–2–1 structure). The experimental acetic acid yield ranged from 1.01 to 4.68 g/100 mL, values consistent with those reported in the literature for biological fermentation systems. The ANN model achieved low prediction errors (SSE = 0.756 and 0.187 for training and testing, respectively) and demonstrated strong generalization capacity without overfitting. Connection weight and relative importance analyses revealed that incubation time and temperature were the most influential variables affecting yield, while glucose concentration had a secondary effect. These findings confirm the suitability of ANN as a reliable computational tool for modeling and optimizing nonlinear bioprocesses. The integration of machine learning approaches with microbial fermentation can enhance process understanding and support the development of sustainable acetic acid production strategies using industrial by-products.
Zitieren
APA
Cuello, R. G., Toro, R. O., & Dunoyer, A. T. (2025). Modelado y predicción basados en redes neuronales del rendimiento de ácido acético en la fermentación de Gluconobacter oxydans utilizando aguas residuales lácteas. @limentech, Ciencia y Tecnología Alimentaria, 23(2), 5-18. https://doi.org/10.24054/limentech.v23i2.4359
Vancouver
Cuello RG, Toro RO, Dunoyer AT. Modelado y predicción basados en redes neuronales del rendimiento de ácido acético en la fermentación de Gluconobacter oxydans utilizando aguas residuales lácteas. @limentech, Ciencia y Tecnología Alimentaria. 2025;23(2):5-18. doi:10.24054/limentech.v23i2.4359
BibTeX
@article{Cuello2025,
author = {Cuello, Rafael González and Toro, Rodrigo Ortega and Dunoyer, Arnulfo Taron},
title = {Modelado y predicción basados en redes neuronales del rendimiento de ácido acético en la fermentación de Gluconobacter oxydans utilizando aguas residuales lácteas},
journal = {@limentech, Ciencia y Tecnología Alimentaria},
year = {2025},
volume = {23},
number = {2},
pages = {5--18},
doi = {10.24054/limentech.v23i2.4359}
}
EUVIMED