Avaliação da produção de açúcar utilizando reconciliação de dados, estatística multivariada e modelagem fenomenológica da cristalização

• Ano de defesa: 2019
• Autor(a) principal: Castro, Bruno José Chiaramonte de
• Orientador(a): Bernardo, André
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de São Carlos; Câmpus São Carlos
• Programa de Pós-Graduação: Programa de Pós-Graduação em Engenharia Química - PPGEQ
• Palavras-chave em Português: Perdas indeterminadas; Reconciliação de dados; Indústria de açúcar; Análise multivariada; Cristalização da sacarose; Modelagem fenomenológica; Simulação
• Palavras-chave em Inglês: Undetermined losses; Data reconciliation; Sugar industry; Multivariate analysis; Sucrose crystallization; Phenomenological modelling; Simulation
• Área do conhecimento CNPq: ENGENHARIAS::ENGENHARIA QUIMICA::OPERACOES INDUSTRIAIS E EQUIPAMENTOS PARA ENGENHARIA QUIMICA; ENGENHARIAS::ENGENHARIA QUIMICA::PROCESSOS INDUSTRIAIS DE ENGENHARIA QUIMICA; ENGENHARIAS::ENGENHARIA QUIMICA::TECNOLOGIA QUIMICA
• Link de acesso: https://repositorio.ufscar.br/handle/20.500.14289/11254

Sugarcane, grown in more than a hundred countries, is one of the major crops of the world. Brazil is the world’s largest producer. Sugarcane industries are constantly seeking to reduce the industrial process losses, to reduce the process variability, and to greater understand and control the phenomena involved in the sucrose crystallization. Thus, the first approach of this work aimed at quantifying the contribution of the sugar manufacturing sector to the total undetermined losses of a sugar and ethanol plant. To this end, a method based on data reconciliation for the process flow rates and concentrations was used. This method could be applied to any equipment or sector of the industry in order to rapidly identify sugar losses that are not currently quantified. The second approach of the work aimed at identifying the major sources of variation in the quantity and quality of the produced sugar, by applying two multivariate statistical techniques (PCA – Principal Component Analysis, and PLS – Partial Least Squares) to the data from the same industry. Lastly, the third objective of the work was to implement a model to the two-massecuite system for sucrose crystallization, based on the first principles of mass and energy conservations, and on the crystal population balance, including the phenomena of nucleation and growth rate dispersion. Microsoft Excel was used for the process data reconciliation, software Minitab for the multivariate statistical analyses, and EMSO simulator for the implementation of the phenomenological model of the crystallization process. From data reconciliation, 37.3% of undetermined losses were found to occur in the sugar manufacturing sector, between juice concentration and sugar bagging. The PCA highlighted the high correlation between the presence of alcoholic flocs in sugar and the concentrations of starch and dextran in it. Both PCA and PLS showed that the color of the sugar was highly correlated to its moisture content, indicating that the phenomena of inclusion and occlusion of molasses in the crystals contributed significantly to increased crystal color. In the simulation of the two-massecuite crystallization process, the crystal mean size reached 0.64 mm, with coefficient of variation equal to 30.58%. These results approximated the real data of the studied plant, in which the crystal mean size was 0.65 mm, with coefficient of variation equal to 25.36%.