Modelling of the extraction of phenolic compounds from beer malt bagasse using artificial neural network / Modelagem de extração de compostos fenólicos de bagaço de malte de cervejaria usando redes neurais artificiais

• Autor(a) principal: Costa Júnior, Luís Fernando
• Data de Publicação: 2020
• Outros Autores: Silva Valente, Gerson de Freitas, Costa Silva, Marisa da Mota
• Tipo de documento: Artigo
• Idioma: por
• Título da fonte: Revista Veras
• Texto Completo: https://ojs.brazilianjournals.com.br/ojs/index.php/BRJD/article/view/17598
• topic: Utilization of by-products. RSM. ANN.

One of the major problems in the beer industry is the high generation of solid waste, which is rich in antioxidant phenolic compounds. To extract the phenolic compounds from beer malt bagasse, we tested it for the extraction time, pH, and ratio solvent: sample. It characterized brewery malt bagasse in terms of: moisture, lipids, protein, crude fiber, mineral salts and phenolic compounds. We used assays in tests according to a Box-Behnken design. The surface methodology response (RSM) and artificial neural network (ANN) were used to modelling the extraction factors of phenolic compounds of beer malt bagasse. With ANN, the 15 assays, of which 70% were used for network training, 30% for neural network testing. The Garson method was used to determine the relative importance of the variables. The neural network technique was better for modeling to the extraction of phenolic compounds from beer malt bagasse. Among the variables of highest relevance were the ratio solvent: sample. The best extraction conditions were: pH of acidified water of 2.5; 45 min extraction time and solvent: sample ratio was 10 mL 5g?1. Modeling with ANN was better that RSM for extraction of phenolic compounds from beer bagasse malt.