Wood has been a trusted material for construction for millennia, from the revered Japanese cypress to the sturdy ponderosa pine. Today, amid the growing call for sustainable practices, wood is experiencing a revival in contemporary architecture, finding its way into public spaces and multi-storey buildings due to its lower environmental impact compared to conventional materials like steel and concrete.

However, a major challenge remains: when used outdoors, wood is vulnerable to deterioration from sunlight and moisture. To mitigate this, protective coatings are applied. Yet, damage to these coatings often begins before any visible signs appear, making maintenance difficult and leading to premature material failure.

In an innovative effort to address this issue, a research team at Kyoto University has developed a new method to detect early-stage coating deterioration before it becomes visible. This advancement promises to extend the lifespan of wooden structures and enhance the sustainability of wood in the built environment.

Data-Driven Diagnostics for Wooden Structures
The team, led by Yoshikuni Teramoto, combines mid-infrared spectroscopy with machine learning to develop a non-destructive diagnostic technique. Their approach centres on detecting minute chemical changes in wood coatings — changes that are invisible to the naked eye.

Testing included both conventional coatings and those enhanced with cellulose nanofibers, a plant-derived material known to improve durability. The use of sustainable additives like cellulose nanofiber underlines the project’s focus on biobased innovation.

Machine learning algorithms, specifically a partial least square model augmented with a genetic algorithm, were used to identify the most informative infrared signals for predicting the level of deterioration. Impressively, even the subtlest chemical changes were accurately captured, offering early warning signs long before traditional visual inspections could.

Implications for Sustainable Architecture
By enabling early intervention, this technique can significantly reduce maintenance costs and prevent extensive material degradation, offering architects and designers a powerful tool to maintain wooden structures more effectively. Moreover, it supports a longer lifespan for wooden elements, reinforcing the principles of circularity and sustainable building practices.

The researchers are currently expanding their work to real-world buildings, aiming to refine their model for practical application in new paint and coating product development. They also envision potential adaptations for other materials like concrete and metal, which could revolutionise maintenance strategies across a range of construction and design industries.

In bridging traditional craftsmanship with modern data science, this innovation represents a significant step towards smarter, more sustainable material usage in architecture and landscape design.

Source: University of Kyoto via EurekAlert
Image: KyotoU / Whitney Hubbell