Estimativa da resistência e da elasticidade à compressão paralela às fibras da madeira de Eucalyptus grandis e E. urophylla usando a espectroscopia no infravermelho próximo

Abstract

The use of wood for structural purposes, safely and economically, requires knowledge of its mechanical properties. The mechanical test is relatively quick, however, sample preparation is a laborious and timeconsuming processes which causes waste of wood due to the difficulty to obtain free of defects samples with standard dimensions. Furthermore, to obtain data about stiffness (modulus of elasticity) the use of machine testing with a large capacity and for measuring deformation is required. One alternative to circumvent these problems is the near infrared spectroscopy (NIRS), which is an indirect, fast and accurate method for the prediction of many wood properties. The use of this technique for the evaluation of some mechanical properties can reduce the time, cost and material waste, especially in parts with a high incidence of defects, but requires a calibration step that is carried out from reference data. The objective of this work was to evaluate the efficiency of the near infrared spectroscopy to estimate the elasticity and strength in compression parallel to fibers of Eucalyptus grandis and E. urophylla wood. The trials of parallel to fibers compression were performed in a universal test machine following the procedures of the ASTM D-143-94 (1997). The absorbance spectra were acquired from a scanning spectrometer with 4,000 to 12,500 cm-1 in diffuse reflection mode. Thus, spectral information measured on the transverse face of the samples was correlated with the values of mechanical properties by partial least square regression. The strength and elasticity in compression parallel to fibers of Eucalyptus grandis and E. urophylla wood can be estimated by models that present coefficients of determination in the validation of 0.78 and 0.75 respectively. These values are satisfactory considering the simplicity and the rapidity required for NIRS analysis.