Classificação de resíduos madeireiros da Amazônia e carvões derivados por espectroscopia no infravermelho próximo

• Autor(a) principal: Lima, Michael Douglas Roque
• Data de Publicação: 2022
• Tipo de documento: Tese
• Idioma: por
• Título da fonte: Repositório Institucional da UFLA
• Texto Completo: http://repositorio.ufla.br/jspui/handle/1/50279
• topic: Biomassa residual; Distinção de espécies; Qualidade do carvão; Temperatura de carbonização; Madeira - Classificação; Waste biomass; Species distinction; Charcoal quality; Carbonization temperature; Wood - Classification; Tecnologia e Utilização de Produtos Florestais

Solutions to discriminate and classify logging wastes and amazon charcoals quickly and reliably are necessary to optimize the use of natural resources. Thus, the main goal of this study was to develop a methodology to identify and classify woods and charcoals of native species produced in brick kilns and on a laboratory scale, based on the use of near-infrared spectroscopy (NIR). For this, logging wastes were used, specifically branches, from twelve species logged in a forest management plan certified in the Paragominas town, Pará. Discs with a thickness of 20 cm were used to determine the wood properties (moisture and basic density), carbonization in laboratory-scale at four different final temperatures (400, 500, 600, and 700°C), and identification by the xyloteque. 30 cm thick discs were carbonized in the charcoal production unit where the wastes were sampled. The charcoals produced on a laboratory scale and in brick kilns were submitted to NIR recordings. Principal component analysis (PCA) and partial least squares-discriminant analysis (PLS-DA) were used to identify native woods based on spectral data and classify them into density classes. This thesis has been divided into four chapters to facilitate understanding. The first describes the basic density and moisture content of wood wastes from 12 tropical species, as well as the accuracy of multivariate models in classifying these wastes in terms of basic density through their spectral signatures. The second presents the potential of NIR to quickly identify wood wastes of tropical species from Amazonia. The third investigated the effects of the final carbonization temperature and the quality of the species on the carbonization efficiency and on the apparent relative density of charcoals derived from Amazonian wastes. The last chapter contains data on the quality of charcoal from wood wastes produced in brick kilns and the discrimination of charcoals in terms of origin by means of the NIR spectral signature. The results show important variations in the basic density (0.509 to 0.916 g cm-3) and moisture (9.5 to 10.6%, wet basis) of the woods at the time of the spectral readings. The PLS-DA model adjusted with the first derivative of the spectra measured on the radial surface of the woods showed 97.9% accuracy in the discrimination of species (Chapter 1). The PLS-DA model calibrated with radial surface spectra treated with the first derivative and validated by the independent validation method presented 97.9% correct answers in the classification based on wood density (Chapter 2). The yields of charcoal (GYC) produced in the laboratory decreased with the increase of the final temperature (400 – 700°C), in which the species T. guianensis (25.6%), Manilkara sp. (25.4%), and D. excelsa (24.7%) showed the most significant reductions in GYC (Chapter 3). The number of correct classifications regarding the origin of charcoal samples produced in brick kilns using the PLS-DA model reached about 70% (Chapter 4). Therefore, the NIR proved to be promising in the discrimination and classification of wood wastes from forest management for charcoal production. In addition, the NIR has the potential to discriminate charcoal from Amazonian waste in terms of origin.