Detalhes bibliográficos
Ano de defesa: |
2019 |
Autor(a) principal: |
Cardoso, Wilson Junior |
Orientador(a): |
Não Informado pela instituição |
Banca de defesa: |
Não Informado pela instituição |
Tipo de documento: |
Dissertação
|
Tipo de acesso: |
Acesso aberto |
Idioma: |
eng |
Instituição de defesa: |
Universidade Federal de Viçosa
|
Programa de Pós-Graduação: |
Não Informado pela instituição
|
Departamento: |
Não Informado pela instituição
|
País: |
Não Informado pela instituição
|
Palavras-chave em Português: |
|
Link de acesso: |
https://locus.ufv.br//handle/123456789/29233
|
Resumo: |
The main goal of this work was to propose a new method to collect NIR spectrum from dehydrated sugarcane juice to predict soluble solids (BRIX) and dextro-rotatory substances (POL), fiber, sucrose, glucose, and fructose, using multivariate regression models. A total of 133 different genotypes harvested m different periods were analyzed. The BRIX, POL and fiber values ranged from 5.4 to 23.6; 8.3 to 89.8; and 8.4 to 22.5 %, respectively. Sucrose (SUC), glucose (GLU), and fructose (FRU) were quantified using high performance liquid chromatography (HPLC). The SUC, GLU and FRU values ranged from 25.7 to 215.7; 2.8 to 25.5, and 3.2 to 21.7 mg/mL. As the main constituent of the sugarcane juice, water absorbs strongly the NIR radiation and hinders the detection of other chemical species. For that reason, two new methods were proposed to remove water: one using an oven (OVD) and another using a heating plate (HPD). Among the 9 support materials studied, thick couche paper (GRO) was the best support for the OVD method and glass cover slips (GCS) was the support chosen for the HPD method. For the OVD method, the temperature, time, and sample volume were optimized using a central composite design (CCD). The optimum parameters were, 113.6 ºC, 19.55 minutes, and 51.6 uL. For the HPD method, once there is more control of the drying process, the temperature was fixed at 120 ºC and the time at 2.5 minutes. NIR spectra were acquired mn the range from 1000 to 2500 nm. PLS models were built for OVD and HPD methods using two ranges of NIR spectra, i.e., the full, from 1000 to 2500 nm and the short, from 1000 to 1700 nm. The short one 1s normally used 1n portable NIR instruments. The root mean square error of prediction (RMSEP) values for SUC, GLU, FRU and fiber using the OVD method were respectively 17.33; 3.98; 1.85; and 1.95 for full spectra and 18.52; 4.37; 3.04; and 2.20 for short spectra. The RMSEP for SUC, GLU, FRU and fiber using the HPD method were respectively 19.91; 3.77; 1.97; and 2.58 for full spectra and 26.86; 3.89, 2.60; and 3.56 for short spectra. BRIX and POL models were not presented. For OVD and HPD methods , the analytical frequency were 2 samples per min and 1.2 samples per min, respectively. For NIR, the analytical frequency was 1 sample per min. The best sucrose model presented to be superior to the traditional methods (BRIX and POL) when selecting genotypes. |