Abstract
Wood coatings play a vital role in prolonging the lifespan of timber structures by protecting them from environmental degradation. However, conventional evaluation methods rely on visual inspections, which cannot detect latent deterioration before visible damage occurs. This study integrates attenuated total reflectance-Fourier transform infrared (ATR-FTIR) spectroscopy with partial least squares (PLS) regression to quantitatively predict the deterioration of wood coatings subjected to accelerated weathering using a xenon lamp method. The study further explores the impact of cellulose nanofibre (CNF) as an additive to enhance coating durability. A genetic algorithm-based wavenumber selection (GAWNSPLS) was employed to identify key spectral regions contributing to model accuracy. The developed models demonstrated strong predictive performance, achieving coefficients of determination (R2) of 0.95 and 0.92 for coatings with 3.8% and 24.9% CNF, respectively, in leave-one-out cross-validation (LOOCV). These findings highlight the potential of mid-infrared spectroscopy combined with machine learning as a non-destructive technique for assessing the durability of wood coatings and optimising formulations for enhanced performance.