Estimativa da resistência e da rigidez à compressão paralela às fibras pela dureza e pela densidade aparente de espécies de madeira provenientes de florestas nativas

• Ano de defesa: 2022
• Autor(a) principal: Pietrobon, Isabela Matias
• Orientador(a): Christoforo, André Luis
• 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 Civil - PPGECiv
• Palavras-chave em Português: Madeira; Folhosa; Densidade; Compressão; Modelos de regressão; Dureza
• Palavras-chave em Inglês: Wood; Hardwoods; Density; Compression; Regression models; Hardness
• Área do conhecimento CNPq: ENGENHARIAS::ENGENHARIA CIVIL::ESTRUTURAS; ENGENHARIAS::ENGENHARIA CIVIL::GEOTECNICA
• Link de acesso: https://repositorio.ufscar.br/handle/20.500.14289/15690

When using wood in structures, it is necessary to know its properties to make a safe and economical structural dimensioning, using the document of ABNT NBR 7190 (1997). However, carrying out the complete characterization tests requires large machines with high costs. Thus, the standard itself estimates the properties due to the various properties necessary for the development of a structural project. Therefore, this research aimed, considering 15 wood species from tropical forests and equally divided into the strength classes of the standard, to evaluate, via regression models, the feasibility or accuracy of the use of apparent density (ρap) and hardness (parallel ( fh0) and perpendicular (fh90)) as estimators of strength (fc0) and stiffness (Ec0) to compression in the direction parallel to the fibers. The regression model considering the ρap as an estimator obtained the best fit through the cubic model (R² adj between 73 and 91%), however, the linear ones were also considered significant by the ANOVA. Considering fh0 as an estimator, the best fits were the linear models with R² adj of 93,98% (fc0) and 64,15% (Ec0). When fh90 was used as an estimator, greater precision was noted in the estimate of fc0 (R² adj = 89,53% - quadratic model) than in Ec0 (R² adj = 61,59% - linear model), both significant by the analysis of variance. In an attempt to use the multiple regression models, the best linear fits were obtained, however, when compared with the regression models of a single variable, the R² adj was not significantly higher. To verify the effectiveness of the regression models, the absolute error in estimating the properties of other wood species was between 0,09 and 60%.