Estimativa de propriedades de resistência e de rigidez de madeiras tropicais brasileiras pela técnica de colorimetria

Abstract

There is no consensus in the estimation of the number of native species coming from the Amazon Forest, and it is necessary to study the technological properties of tropical wood species for the rational use of this material. Southern and southeastern Brazilian regions are the main consuming centers of the Brazilian tropical wood species from Amazon Forest, for permanent (structures, roofs and floors) or temporary constructions. The aim of this research is to generate regression models to estimate the strength and stiffness properties of wood via colorimetric parameters and density. For this research were used Brazilian tropical wood species, which meet the strength classes established by the ABNT NBR 7190:1997 Standard Code for dicotyledons, according to their characteristic strengths, between them: Caixeta (Simarouba amara Aubl.), Cajueiro (Anacardium giganteum W. Hancock), Cambará (Erisma uncinatum Warm), Tatajuba (Bagassa guianenses Aubl.) and Roxinho (Peltogyne sp.). Density at 12% moisture content (ρap,12) strength and stiffness properties were estimated: compression parallel to the grain strength (fc0), hardness normal to the grain (fH90), modulus of rupture (fM) and elasticity (EM) in static bending. The following colorimetric parameters were estimated according to the CIE L*a*b* system using a Konica Minolta colorimeter: brightness (L*) and shades on the red/green (a*) and yellow/blue (b*) axes. Eighteen specimens were fabricated for each mechanical properties studied for the five Brazilian tropical wood species used. Quadratic multivariable regression models were used to estimate the mechanical properties of wood. For each of the species the regression models were tested using the colorimetric parameters L*, a* and b*. For all species, density of wood was also considered as an independent variable. The coefficient of determination (R²) for each wood specie varied between 45.50% (fH90 for Roxinho) and 85.30% (fM for Tatajuba). Proposed regression models to group involving all species using density as independent variable were excellent, because, R² values were equal to 94.63, 87.62, 94.14 and 96.08% for estimating EM , fM, fc0 and fH90, respectively.