Application-Oriented NIR Calibration Models as Digital Enablers in Circular Wood-Based Resource Chains

The analysis of biogenic materials represent a central challenge for quality assurance and further processing. Particularly when using compost material for the production of extruded wood fibers, an efficient and reliable method is essential. Near-infrared spectroscopy (NIR) offers a promising solution, as it operates contact-free, quickly, and non-destructively. Through specific absorption patterns, both organic and inorganic material components can be identified.
In this study, applied research was conducted to develop NIR calibration models for key parameters such as nitrogen content, moisture content, and contamination identification in extruded fibers derived from composting. Using laboratory reference analyses and partial least squares (PLS) regression, various NIR prediction strategies were examined under varying moisture conditions and material heterogeneity.
Nitrogen values could be predicted from the NIR spectra, with moisture-dependent calibration proving necessary. A comparison of three model scenarios showed that targeted selection of spectral regions is more advantageous in terms of model quality and generalizability than using the entire spectrum. A two-step modeling approach, separating nitrogen and moisture, did not provide benefits. The scenario with selected spectral regions thus proved to be the most robust and practical strategy, particularly regarding transferability to new samples.
The results highlight the potential of NIR spectroscopy as a reliable, fast, and non-destructive method for quality assurance in the processing of biogenic residues. The additional washing step offers a simple way to further improve the reliability of the analysis, especially for critical material batches. This approach directly supports the integration of digital process chains in production and provides a scalable methodology for other stakeholders working with organic or lignocellulosic residues.
This research project demonstrates how applied NIR calibration, embedded in a transnational innovation initiative (DRWO4.0) can drive the introduction of intelligent, resource-efficient systems in the circular bioeconomy of the wood sector.