Classificação de ácido oleico e ácido palmítico no processo de microfiltração tangencial por meio de redes neurais artificiais e árvores de decisão
Carregando...
Data
2023-01-12
Autores
Rossi, Ivan Sayed
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Estadual Paulista (Unesp)
Resumo
O processo de microfiltração tangencial com membranas cerâmicas tem sido muito utilizado na clarificação do suco de açaí pois tem as vantagens de operar sem conservante químico, separar seletivamente os materiais e possuir baixo custo operacional (usa apenas a energia elétrica para o funcionamento da bomba hidráulica, não é necessário usar trocadores de calor, evitando problemas de poluição térmica e sobrecarga do sistema). Ainda, esta tecnologia consegue estabilizar as características funcionais e microbiológicas dos produtos à base de açaí, pois utiliza baixas temperaturas reduzindo efeitos prejudiciais de alguns tratamentos térmicos e conservando, assim, as características originais da fruta. Sua polpa possui uma composição complexa com elevada porcentagem de lipídios, proteínas, sais minerais, ácidos graxos e fibra alimentar, que contribuem para a formação de incrustações nos poros das membranas. Assim, visando a melhoria do processo, é útil identificar esses compostos e analisar seus impactos individuais sobre os poros da membrana. Através da base de dados disponibilizada por Camimoto (2012), foram desenvolvidos modelos neurais e florestas aleatórias para classificar os ácidos graxos (ácido oleico e palmítico) no processo de microfiltração tangencial. O projeto busca identificar o ácido oleico e ácido palmítico em misturas bifásicas de óleos vegetais e água no processo de microfiltração por meio de Redes Neurais Artificiais (RNAS) e Florestas Aleatórias (Random Forest). Tanto a RNA quanto as Florestas Aleatórias conseguem fornecer resultados altamente precisos com muita rapidez e pouco esforço computacional.
The tangential microfiltration process with ceramic membranes has been widely used in the clarification of the açaí juice because it has the advantages of operating without chemical preservatives, selectively separating the materials and has low operating costs (it uses only electric energy to operate the hydraulic pump, there is no need to use heat exchangers, avoiding problems of thermal pollution and system overload). Furthermore, this technology is able to stabilize the functional and microbiological characteristics of the açaí products, because it uses low temperatures, reducing the harmful effects of some heat treatments and preserving the original characteristics of the fruit. Its pulp has a complex composition with a high percentage of lipids, proteins, mineral salts, fatty acids and dietary fiber, which contribute to the formation of incrustations in the pores of the membranes. Thus, aiming at process improvement, it is useful to identify these compounds and analyze their individual impacts on the membrane pores. Using the database provided by Camimoto (2012), neural models and random forests were developed to classify fatty acids (oleic and palmitic acid) in the tangential microfiltration process. The project seeks to identify oleic acid and palmitic acid in two-phase mixtures of vegetable oils and water in the microfiltration process by means of Artificial Neural Networks (ANN) and Random Forests (Random Forest). Both ANN and Random Forests are able to provide highly accurate results very quickly and with little computational effort.
The tangential microfiltration process with ceramic membranes has been widely used in the clarification of the açaí juice because it has the advantages of operating without chemical preservatives, selectively separating the materials and has low operating costs (it uses only electric energy to operate the hydraulic pump, there is no need to use heat exchangers, avoiding problems of thermal pollution and system overload). Furthermore, this technology is able to stabilize the functional and microbiological characteristics of the açaí products, because it uses low temperatures, reducing the harmful effects of some heat treatments and preserving the original characteristics of the fruit. Its pulp has a complex composition with a high percentage of lipids, proteins, mineral salts, fatty acids and dietary fiber, which contribute to the formation of incrustations in the pores of the membranes. Thus, aiming at process improvement, it is useful to identify these compounds and analyze their individual impacts on the membrane pores. Using the database provided by Camimoto (2012), neural models and random forests were developed to classify fatty acids (oleic and palmitic acid) in the tangential microfiltration process. The project seeks to identify oleic acid and palmitic acid in two-phase mixtures of vegetable oils and water in the microfiltration process by means of Artificial Neural Networks (ANN) and Random Forests (Random Forest). Both ANN and Random Forests are able to provide highly accurate results very quickly and with little computational effort.
Descrição
Palavras-chave
Redes neurais artificiais, Florestas aleatórias, Açaí