Uso de propriedades colorimétricas na estimativa das propriedades físicas e mecânicas de madeiras tropicais

• Autor(a) principal: Ruthes, Heloiza Candeia
• Data de Publicação: 2023
• Tipo de documento: Tese
• Idioma: por
• Título da fonte: Repositório Institucional da UFSCAR
• Texto Completo: https://repositorio.ufscar.br/handle/ufscar/18992
• topic: Regression models; Colorimetry; Tropical Species; Physical properties; Mechanical properties; Modelos de regressão; Colorimetria; Madeiras tropicais; Propriedades físicas; Propriedades mecânicas; ENGENHARIAS::ENGENHARIA CIVIL::CONSTRUCAO CIVIL::MATERIAIS E COMPONENTES DE CONSTRUCAO; ENGENHARIAS::ENGENHARIA CIVIL::ESTRUTURAS; ENGENHARIAS::ENGENHARIA CIVIL; CIENCIAS EXATAS E DA TERRA::PROBABILIDADE E ESTATISTICA::ESTATISTICA::REGRESSAO E CORRELACAO; ENGENHARIAS::ENGENHARIA CIVIL::ESTRUTURAS::ESTRUTURAS DE MADEIRAS

Research into alternative methods for estimating wood properties has gained prominence due to the difficulties inherent in this process. Such difficulties include the lack of laboratories and specialized equipment necessary for the complete characterization of wood. Therefore, this study aims to develop equations using regression models, to estimate fifteen physical and mechanical properties by Brazilian standards for wooden structures. To accomplish this purpose, experimental data relating to the physical, and mechanical properties and colorimetric parameters of ten new species of tropical wood were used. Initially, the physical and mechanical properties and colorimetric parameters of these species were obtained and subjected to statistical analysis. Additionally, data from ten species resulting from related and previous research were also incorporated into the analysis. Linear and quadratic equations were obtained using regression models, considering colorimetric parameters and apparent density as independent variables. The independence of these variables was evaluated using the Pearson correlation test. Initially, the models were built with experimental data from just fifteen species. Subsequently, after validating these results, the same equations were generated considering experimental data from all species. In all models, the significance of the equations and their terms was assessed using analysis of variance (ANOVA). The adequacy of the models obtained was evaluated by the adjusted coefficient of determination, and in the validation phase, the percentage error obtained was investigated. The results indicated that the developed equations are statistically significant and present satisfactory adjustments for most estimated properties. Furthermore, apparent density proved to be the most influential parameter in these models, although colorimetric parameters have considerable significance.