From the Japanese cypress to the ponderosa pine, wood has been used in construction for millennia. Though materials like steel and concrete have largely taken over large building construction, wood is making a comeback, increasingly being used in public and multi-story buildings for its environmental benefits.
Of course, wood has often been passed over in favor of other materials because it is easily damaged by sunlight and moisture when used outdoors. Wood coatings have been designed to protect wood surfaces for this reason, but coating damage often starts before it becomes visible. Once the deterioration can be seen with the naked eye, it is already too late.
To solve this problem, a team of researchers at Kyoto University is working to create a simple but effective method of diagnosing this nearly invisible deterioration before the damage becomes irreparable.
“If we can ‘see’ what the eye cannot, we can extend the life of wooden structures and improve sustainability in the building industry,” says corresponding author Yoshikuni Teramoto.
The team is endeavoring to bring data-driven tools into traditional wood maintenance by combining mid-infrared spectroscopy with machine learning. They’ve started by testing artificially weathered wood coatings along with coatings containing cellulose nanofiber, a plant-derived additive that can improve the durability of these coatings.
Their machine learning component uses a technique called partial least square, which they employed to build a model to predict the extent of deterioration. They also used a genetic algorithm to identify the most informative infrared signals, improving both accuracy and interpretability.
“We were surprised to find that very subtle chemical changes — far too small to detect visually — could be captured by infrared spectroscopy and predicted by the model,” says Teramoto.
This approach allows the researchers to detect subtle chemical changes and estimate the level of deterioration with high accuracy. By making it possible to diagnose early coating deterioration quickly and without damaging the wood, their method could also reduce the need for costly visual inspections by detecting early warning signs of deterioration and preventing further decay.
With their study, the researchers have also demonstrated how chemistry and data-driven modeling techniques can work together to support smarter maintenance of sustainable buildings. “We hope this technology will help bridge the gap between traditional craftsmanship and modern data science,” continues Teramoto.
The research team is now conducting tests on real wooden buildings, with plans to improve their model for application in new paint and coating product development.
Beyond wood, the team’s method may also be applied to materials like concrete or metal to unlock new possibilities for diagnosing other kinds of early material failure, improving the sustainability of other applications and industries in the process.
