Estimation of moisture content of trembling aspen (Populus tremuloides Michx.) strands by near infrared spectroscopy (NIRS)

In a preliminary study on improving manufacturing control of composite wood products, near infrared spectroscopy (NIRS) was tested to determine moisture content (MC) of trembling aspen (Populus tremuloides Michx.) flakes that are used in oriented strand board (OSB) manufacturing. Three drying cycles (of 1 kg each) of aspen flakes were scanned at different drying levels by NIRS in the 1,300-2,200 nm spectral region. The study showed an influence of MC on NIR spectra. A partial least squares regression model was developed between NIR spectra and gravimetric-based moisture contents. The statistics achieved R-2 and root mean square error (RMSE) for the calibration model ranging from 0.97 to 0.99 and from 2.5 to 5.9 %, respectively. The validation statistic models achieved R-2 and RMSE ranging from 0.96 to 0.99 and from 2.7 to 6.01 %, respectively. These preliminary results show that NIRS can be a useful tool for monitoring MC of OSB flakes.