Ethanol quantification in pineapple waste by an electrochemical impedance spectroscopy-based system and artificial neural networks

[EN] Electrochemical impedance spectroscopy (EIS) technique has been applied to determine the ethanol concentration in pineapple waste samples. To do this, six different concentrations of ethanol were added to the pineapple samples and were analyzed using the system designed by our research group an...

Descripción completa

Detalles Bibliográficos
Autores: Conesa Domínguez, Claudia, Gil Sánchez, Luís|||0000-0003-3453-5559, Seguí Gil, Lucía|||0000-0002-2711-9445, Fito Maupoey, Pedro, Laguarda-Miro, Nicolas|||0000-0001-6829-7160
Tipo de recurso: artículo
Fecha de publicación:2017
País:España
Institución:Universitat Politècnica de València (UPV)
Repositorio:RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Idioma:inglés
OAI Identifier:oai:riunet.upv.es:10251/99365
Acceso en línea:https://riunet.upv.es/handle/10251/99365
Access Level:acceso abierto
Palabra clave:Electrochemical impedance spectroscopy
Ethanol
Pineapple waste
Artificial neural networks
INGENIERIA QUIMICA
TECNOLOGIA DE ALIMENTOS
TECNOLOGIA ELECTRONICA
Descripción
Sumario:[EN] Electrochemical impedance spectroscopy (EIS) technique has been applied to determine the ethanol concentration in pineapple waste samples. To do this, six different concentrations of ethanol were added to the pineapple samples and were analyzed using the system designed by our research group and consisting of the Advanced Voltammetry, Impedance Spectroscopy & Potentiometry Analyzer (AVISPA) device associated to a stainless steel double needle electrode. Results indicated that phase data in frequencies between 6.0 x 10(5) Hz and 8.0 x 10(5) Hz showed the highest sensitivity to ethanol concentrations. A principal component analysis (PCA) confirmed the potential discrimination and partial least squares (PLS) regression showed mathematical models able to quantify ethanol in samples accurately. In order to implement flexible and precise models in programmable equipment, different types of artificial neural networks (ANNs) have been studied: Fuzzy ARTMAP and multi-layer feed-forward (MLFF) algorithms. As a result, a coefficient of determination (R2) = 0.996 and a root mean square error of prediction (RMSEP) = 0.408 have been obtained. Therefore, it allows us to introduce this technique as an alternative method for ethanol quantification along the fermentation of pineapple waste in an easy, low-cost, rapid and portable way.