Abstract
NBR 7190 (ABNT, 1997) defines the simplified characterization, where there is the use of mathematical relationships between the properties of strength and stiffness. Due to the great influence of porosity and apparent density on wood properties and the search for increasingly accurate mathematical models, this work aimed to generate regression models for the estimation of static bending properties (strength " f M " and modulus " E M "), hardness parallel to the grain " f H,0 " and normal to the grain " f H,90 " and toughness "W" of Brazilian tropical woods from the porosity values, obtained through the analysis mercury instruction porosimetry, apparent density and based on both properties in multivariate regression models. All regression models presented coefficients of determination "R 2 " greater than 50%, where f M , f H,90 and W were best estimated by density apparent, E M and f H,0 were better estimated considering apparent density and porosity.
Keywords:
Estimation of wood properties; Regression models; Porosity; pparent density
Thumbnail
Thumbnail
Thumbnail
