Forecasting the Brazilian demand for biodiesel using artificial neural networks

Detalhes bibliográficos
Autor(a) principal: Caires, Kaique Vitor Louzada
Data de Publicação: 2021
Outros Autores: Simonelli, George
Tipo de documento: Artigo
Idioma: por
Título da fonte: Research, Society and Development
Texto Completo: https://rsdjournal.org/index.php/rsd/article/view/13381
Resumo: Biodiesel is a renewable fuel used as an alternative to totally or partially replace petroleum diesel. The mandatory percentage of this biofuel added to fossil diesel in Brazil has constantly been increasing. Predicting the amount of biodiesel that will be demanded in the future is essential to maintain the national surplus balance and assist in the sector decision-making. Artificial neural networks (ANNs) help forecast different types of demands. Therefore, this study uses artificial neural networks to forecast the Brazilian demand for biodiesel. The ANN proposed in this work encompassed data obtained from a non-parametric demand forecasting model based on time series. The non-parametric model considered the trends and seasonality of the data to forecast the demand for biodiesel. One hundred multilayer perceptron networks were modeled with error propagation for two scenarios of Brazilian biodiesel (use of 15% (B15) or 20% (B20) of biodiesel to diesel). All values ​​of R2 greater than 0.99 for the simulated networks and RMSE <2% prove that the RNA model developed has high precision in predicting the demand for biodiesel. The best network for each scenario was determined by RMSE heuristic analysis. The best-simulated RNA results showed growth in biodiesel demand from 2019 to 2050 of 150.63% for B15. And 229.73% for B20. Both demand growth scenarios are justified by the gradual increase in the mandatory percentage of biodiesel to diesel. Thus, the growing results of biodiesel demand prove the country search for a non-toxic, biodegradable, and renewable fuel in its energy matrix.
id UNIFEI_23fb42e6ca1e3284e84c72bb0b68a39e
oai_identifier_str oai:ojs.pkp.sfu.ca:article/13381
network_acronym_str UNIFEI
network_name_str Research, Society and Development
repository_id_str
spelling Forecasting the Brazilian demand for biodiesel using artificial neural networks Pronóstico de la demanda brasileña de biodiésel mediante redes neuronales artificiales Previsão da demanda brasileira de biodiesel utilizando redes neurais artificiais BiodieselBrasilRedes neurais artificiaisSazonalidadePrevisão de demanda.BiodieselBrazilArtificial neural networksSeasonalityDemand forecast.BiodiéselBrasilRedes neuronales artificialesEstacionalidadPrevisión de demanda.Biodiesel is a renewable fuel used as an alternative to totally or partially replace petroleum diesel. The mandatory percentage of this biofuel added to fossil diesel in Brazil has constantly been increasing. Predicting the amount of biodiesel that will be demanded in the future is essential to maintain the national surplus balance and assist in the sector decision-making. Artificial neural networks (ANNs) help forecast different types of demands. Therefore, this study uses artificial neural networks to forecast the Brazilian demand for biodiesel. The ANN proposed in this work encompassed data obtained from a non-parametric demand forecasting model based on time series. The non-parametric model considered the trends and seasonality of the data to forecast the demand for biodiesel. One hundred multilayer perceptron networks were modeled with error propagation for two scenarios of Brazilian biodiesel (use of 15% (B15) or 20% (B20) of biodiesel to diesel). All values ​​of R2 greater than 0.99 for the simulated networks and RMSE <2% prove that the RNA model developed has high precision in predicting the demand for biodiesel. The best network for each scenario was determined by RMSE heuristic analysis. The best-simulated RNA results showed growth in biodiesel demand from 2019 to 2050 of 150.63% for B15. And 229.73% for B20. Both demand growth scenarios are justified by the gradual increase in the mandatory percentage of biodiesel to diesel. Thus, the growing results of biodiesel demand prove the country search for a non-toxic, biodegradable, and renewable fuel in its energy matrix.El biodiésel es un combustible renovable que se utiliza como alternativa para sustituir total o parcialmente el diésel de petróleo. El porcentaje obligatorio de este biocombustible agregado al diesel fósil en Brasil ha aumentado constantemente. Predecir la cantidad de biodiésel que se demandará en el futuro es fundamental para mantener el saldo excedentario nacional y ayudar en la toma de decisiones del sector. Las redes neuronales artificiales (ANN) son útiles para pronosticar diferentes tipos de demandas. Por lo tanto, este estudio utiliza redes neuronales artificiales para pronosticar la demanda brasileña de biodiesel. La ANN propuesta en este trabajo abarcó datos obtenidos de un modelo de pronóstico de demanda no paramétrico basado en series de tiempo. El modelo no paramétrico consideró las tendencias y la estacionalidad de los datos para pronosticar la demanda de biodiesel. Se modelaron 100 redes de perceptrones multicapa con propagación de errores para dos escenarios de biodiesel brasileño (uso de 15% (B15) o 20% (B20) de biodiesel a diesel). Todos los valores de R2 superiores a 0,99 para las redes simuladas y RMSE <2% demuestran que el modelo de ARN desarrollado tiene una alta precisión en la predicción de la demanda de biodiésel. La mejor red para cada escenario se determinó mediante análisis heurístico de RMSE. Los resultados de los mejores ARN simulados mostraron un crecimiento en la demanda de biodiésel de 2019 a 2050 de 150,63% para B15 y 229,73% para B20. Ambos escenarios de crecimiento de la demanda se justifican por el aumento gradual del porcentaje obligatorio de biodiésel a diésel. Así, los crecientes resultados de la demanda de biodiesel evidencian la búsqueda del país de un combustible no tóxico, biodegradable y renovable en su matriz energética.O biodiesel é um combustível renovável utilizado como uma alternativa para substituir de modo total ou parcial o diesel de petróleo. A porcentagem obrigatória desse biocombustível adicionado ao diesel fóssil no Brasil tem sido elevada constantemente. Prever a quantidade de biodiesel que será demandada futuramente é essencial para manter o balanço nacional superavitário e auxiliar nas tomadas de decisões do setor. As redes neurais artificiais (RNAs) são úteis para previsão de diferentes tipos de demandas. Assim sendo, esse estudo utiliza redes neurais artificiais na previsão da demanda brasileira de biodiesel. A RNA proposta neste trabalho englobou dados obtidos de um modelo não-paramétrico de previsão de demanda baseado em séries temporais. O modelo não-paramétrico considerou as tendências e sazonalidade dos dados para previsão da demanda de biodiesel. Foram modeladas 100 redes do tipo perceptron multicamadas com retropropagação do erro para dois cenários do biodiesel brasileiro (uso de 15% (B15) ou 20% (B20) de biodiesel ao diesel). Todos os valores de R2 maiores que 0,99 para as redes simuladas e RMSE<2% comprovam que o modelo de RNA desenvolvido possui alta precisão em prever a demanda de biodiesel. A melhor rede para cada cenário foi determinada por análise heurística do RMSE. Os resultados das melhores RNA’s simuladas mostraram um crescimento da demanda de biodiesel de 2019 a 2050 de 150,63% para o B15, e de 229,73% para o B20. Ambos cenários de aumento de demanda são justificados pela elevação gradual da porcentagem obrigatória do biodiesel ao diesel. Dessa forma, os resultados crescentes da demanda de biodiesel comprovam a busca do país por um combustível não-tóxico, biodegradável e renovável na sua matriz energética.Research, Society and Development2021-05-02info:eu-repo/semantics/articleinfo:eu-repo/semantics/publishedVersionapplication/pdfhttps://rsdjournal.org/index.php/rsd/article/view/1338110.33448/rsd-v10i5.13381Research, Society and Development; Vol. 10 No. 5; e17410513381Research, Society and Development; Vol. 10 Núm. 5; e17410513381Research, Society and Development; v. 10 n. 5; e174105133812525-3409reponame:Research, Society and Developmentinstname:Universidade Federal de Itajubá (UNIFEI)instacron:UNIFEIporhttps://rsdjournal.org/index.php/rsd/article/view/13381/13283Copyright (c) 2021 Kaique Vitor Louzada Caires; George Simonellihttps://creativecommons.org/licenses/by/4.0info:eu-repo/semantics/openAccessCaires, Kaique Vitor Louzada Simonelli, George 2021-05-17T18:20:49Zoai:ojs.pkp.sfu.ca:article/13381Revistahttps://rsdjournal.org/index.php/rsd/indexPUBhttps://rsdjournal.org/index.php/rsd/oairsd.articles@gmail.com2525-34092525-3409opendoar:2024-01-17T09:34:44.275998Research, Society and Development - Universidade Federal de Itajubá (UNIFEI)false
dc.title.none.fl_str_mv Forecasting the Brazilian demand for biodiesel using artificial neural networks
Pronóstico de la demanda brasileña de biodiésel mediante redes neuronales artificiales
Previsão da demanda brasileira de biodiesel utilizando redes neurais artificiais
title Forecasting the Brazilian demand for biodiesel using artificial neural networks
spellingShingle Forecasting the Brazilian demand for biodiesel using artificial neural networks
Caires, Kaique Vitor Louzada
Biodiesel
Brasil
Redes neurais artificiais
Sazonalidade
Previsão de demanda.
Biodiesel
Brazil
Artificial neural networks
Seasonality
Demand forecast.
Biodiésel
Brasil
Redes neuronales artificiales
Estacionalidad
Previsión de demanda.
title_short Forecasting the Brazilian demand for biodiesel using artificial neural networks
title_full Forecasting the Brazilian demand for biodiesel using artificial neural networks
title_fullStr Forecasting the Brazilian demand for biodiesel using artificial neural networks
title_full_unstemmed Forecasting the Brazilian demand for biodiesel using artificial neural networks
title_sort Forecasting the Brazilian demand for biodiesel using artificial neural networks
author Caires, Kaique Vitor Louzada
author_facet Caires, Kaique Vitor Louzada
Simonelli, George
author_role author
author2 Simonelli, George
author2_role author
dc.contributor.author.fl_str_mv Caires, Kaique Vitor Louzada
Simonelli, George
dc.subject.por.fl_str_mv Biodiesel
Brasil
Redes neurais artificiais
Sazonalidade
Previsão de demanda.
Biodiesel
Brazil
Artificial neural networks
Seasonality
Demand forecast.
Biodiésel
Brasil
Redes neuronales artificiales
Estacionalidad
Previsión de demanda.
topic Biodiesel
Brasil
Redes neurais artificiais
Sazonalidade
Previsão de demanda.
Biodiesel
Brazil
Artificial neural networks
Seasonality
Demand forecast.
Biodiésel
Brasil
Redes neuronales artificiales
Estacionalidad
Previsión de demanda.
description Biodiesel is a renewable fuel used as an alternative to totally or partially replace petroleum diesel. The mandatory percentage of this biofuel added to fossil diesel in Brazil has constantly been increasing. Predicting the amount of biodiesel that will be demanded in the future is essential to maintain the national surplus balance and assist in the sector decision-making. Artificial neural networks (ANNs) help forecast different types of demands. Therefore, this study uses artificial neural networks to forecast the Brazilian demand for biodiesel. The ANN proposed in this work encompassed data obtained from a non-parametric demand forecasting model based on time series. The non-parametric model considered the trends and seasonality of the data to forecast the demand for biodiesel. One hundred multilayer perceptron networks were modeled with error propagation for two scenarios of Brazilian biodiesel (use of 15% (B15) or 20% (B20) of biodiesel to diesel). All values ​​of R2 greater than 0.99 for the simulated networks and RMSE <2% prove that the RNA model developed has high precision in predicting the demand for biodiesel. The best network for each scenario was determined by RMSE heuristic analysis. The best-simulated RNA results showed growth in biodiesel demand from 2019 to 2050 of 150.63% for B15. And 229.73% for B20. Both demand growth scenarios are justified by the gradual increase in the mandatory percentage of biodiesel to diesel. Thus, the growing results of biodiesel demand prove the country search for a non-toxic, biodegradable, and renewable fuel in its energy matrix.
publishDate 2021
dc.date.none.fl_str_mv 2021-05-02
dc.type.driver.fl_str_mv info:eu-repo/semantics/article
info:eu-repo/semantics/publishedVersion
format article
status_str publishedVersion
dc.identifier.uri.fl_str_mv https://rsdjournal.org/index.php/rsd/article/view/13381
10.33448/rsd-v10i5.13381
url https://rsdjournal.org/index.php/rsd/article/view/13381
identifier_str_mv 10.33448/rsd-v10i5.13381
dc.language.iso.fl_str_mv por
language por
dc.relation.none.fl_str_mv https://rsdjournal.org/index.php/rsd/article/view/13381/13283
dc.rights.driver.fl_str_mv Copyright (c) 2021 Kaique Vitor Louzada Caires; George Simonelli
https://creativecommons.org/licenses/by/4.0
info:eu-repo/semantics/openAccess
rights_invalid_str_mv Copyright (c) 2021 Kaique Vitor Louzada Caires; George Simonelli
https://creativecommons.org/licenses/by/4.0
eu_rights_str_mv openAccess
dc.format.none.fl_str_mv application/pdf
dc.publisher.none.fl_str_mv Research, Society and Development
publisher.none.fl_str_mv Research, Society and Development
dc.source.none.fl_str_mv Research, Society and Development; Vol. 10 No. 5; e17410513381
Research, Society and Development; Vol. 10 Núm. 5; e17410513381
Research, Society and Development; v. 10 n. 5; e17410513381
2525-3409
reponame:Research, Society and Development
instname:Universidade Federal de Itajubá (UNIFEI)
instacron:UNIFEI
instname_str Universidade Federal de Itajubá (UNIFEI)
instacron_str UNIFEI
institution UNIFEI
reponame_str Research, Society and Development
collection Research, Society and Development
repository.name.fl_str_mv Research, Society and Development - Universidade Federal de Itajubá (UNIFEI)
repository.mail.fl_str_mv rsd.articles@gmail.com
_version_ 1797052672390463488