Eventos Anais de eventos
ENCIT 2020
18th Brazilian Congress of Thermal Sciences and Engineering
MODELLING AND OPTIMIZATION OF BIOGAS PRODUCTION FROM FOOD WASTES (FW) USING ARTIFICIAL NEURAL NETWORK (ANN)
Submission Author:
Florian Alain Yannick Pradelle , RJ
Co-Authors:
João Gonçalves Neto, Brunno F. Santos, Florian Alain Yannick Pradelle
Presenter: João Gonçalves Neto
doi://10.26678/ABCM.ENCIT2020.CIT20-0260
Abstract
Biogas can be produced by anaerobic digestion from many types of organic residues, in particular from food wastes (FW). The variability of biomasses’ composition and biodigesters’ characteristics makes standardization difficult. Thus, the use of artificial neural networks (ANN) to investigate the effects and their interactions allows exploring several scenarios of biogas production. In the present work, a database was built with available values in 42 references in the literature in order to obtain mathematical models using reactor/feeding type, volatile solid (VS), OLR, temperature and reactor volume as input variables, and the cumulated biogas production as output. Multiple ANN configurations are considered and a statistical analysis of the model robustness is used to define the best topology. A fairly accurate prediction capability is expected. The modeling results can be compared to a more generic model already implemented by the authors for food wastes, fruits and vegetables wastes and codigestion in a former work. As a result of this modelling work, response surfaces of the ANN model are useful tools to assess if the model is predicting coherent behaviors and define optimal combination of conditions in order to maximize biogas production.
Keywords
Biogas, Food wastes (FW), Artificial neural networks (ANN)
DOWNLOAD PDF VIEW PRESENTATION
