In this study, we present the development of DecAI, an artificial intelligence (AI) model designed to optimise and accelerate performance evaluations of wood protection products in the European EN 330:2014 "Field test: L-joint method." The Danish Technological Institute (DTI) compiled a dataset of over 100,000 data points from approx.10,000 L-joint samples collected over +15 years at field sites in Denmark and Malaysia. By integrating biannual moisture content measurements with annual EN330 degradation ratings, DecAI enhances the predictive capabilities of traditional testing methods. DecAI combines advanced neural network architectures, including bidirectional Long Short-Term Memory (LSTM) networks with attention mechanisms, to analyse both static features, such as wood species and treatment protocols, and temporal features, such as degradation progression and environmental conditions. The model was validated through retrospective analysis using completed test series, where predictions based on early-stage data (within 1.5 years) were compared to observed long-term degradation patterns. DecAI consistently demonstrated high predictive accuracy, as measured by Mean Squared Error (MSE), across diverse wood protection systems. A key strength of DecAI is its practical application as an add-on to biannual EN330 test reports. By providing updated future projections on the effectiveness of wood protection systems, DecAI enables stakeholders to make data-driven decisions, refine performance evaluations, and accelerate product development timelines. This innovative integration of predictive analytics into standardised testing offers a scalable and actionable approach to modernising wood protection strategies.

Download document (611 kb)
free for the members of IRG. Available if purchased.

Purchase this document