Abstract:
The application of stepwise multivariate-linear regression models for determination of
particleboard properties based on structural factors was studied. Poplar (Populus alba),
Beech (Fagus orientaleis) and Hornbeam wood (Carpinus betulus) with dry density of 460,
630 and 790 kg/m3
, respectively, were used as raw materials. Three levels of boards target
density (520, 620 and 720 kg m-3) and urea formaldehyde (UF) resin (6, 7, and 8%) were
compared. The variables were included in the regression equations of modulus of rupture
(MOR), modulus of elasticity (MOE), shear strength, and thickness swell (TS) after 24
hours immersion based on the degree of importance. In order to obtain the optimum
board density and resin content for each species, contour plots were drawn by Minitab 13
software. Regarding the results from contour plots, particleboards with density ranging
from 520 to 620 kg m-3 and 6% resin had most of their mechanical properties within those
required by the corresponding standards. Thickness swell values were higher than
requirements. We suggest additional treatments such as using adequate amount of water
resistant materials to improve TS after 24 hours immersion.
Keywords: Board density, Particleboard, Regression models, Resin content, Wood density.
