Comportamento hidromecânico da madeira de Eucalyptus durante a secagem por espectroscopia no NIR e velocimetria de imagem por partículas

Abstract

The physical properties of wood such as moisture and shrinkage are among the most important parameters to qualify the material, as they directly affect its performance in structural and industrial applications. However, as it is a heterogeneous material, moisture and wood contractions may vary within the piece. Therefore, the use of non-destructive techniques can be a viable alternative for monitoring the hydromechanical behavior of the material. The objective of this study was to develop a solution to monitor the spatial variation of moisture and wood contractions during drying using near-infrared (NIR) spectroscopy and image analysis. Equidistant points were marked on the radial surface of a freshly felled wood board. NIR spectra were recorded at each marking with a fiber optic probe during the drying of the wood piece. Pre-established calibration for wood moisture estimation was applied to each spectrum to generate moisture information at each point during the drying phases. The spectral analysis was performed by Principal Component Analysis (PCA) while the estimates were obtained by Partial Least Squares Regression (PLS-R) and the moisture classes by Partial Least Squares Discriminant Analysis (PLS-DA). After the spectral acquisition in each drying step, images were obtained by camera and the markings were referenced in the ImageJ software to monitor contractions during drying. Moisture estimates from NIR spectra showed strong correlation with reference values (R²cv=0.92, RMSEcv=9.82%). From the estimates it was possible to generate graphic images to visualize and quantify the spatial variation of moisture and contractions during wood drying. In the initial stages of drying, the ends of the material presented high moisture in relation to the center of the piece. However, the moisture loss was 11% higher at the ends of the part when compared to its interior and the external contraction was 3 times greater than the internal one. Spectral acquisition via NIR and image analysis are efficient tools to estimate wood moisture and shrinkage, respectively.