Fast and Reliable Characterization and Quality Assurance of Fabricated CLT Panels
Research Team:
Vahid Nasir – OSU, Department of Wood Science and Engineering
Lech Muszynski – OSU, Department of Wood Science and Engineering
Mariapaola Riggio – OSU, Department of Wood Science and Engineering
Autumn Battisti
Ashley McCann
Samson Idoghor
Project Duration:
2024-2027
2024-2027
Research Advisory Group:
A.A. “Red” Emmerson Advanced Wood Products Laboratory
Richardson Hall
Project Goal Statement
This project aims to develop a practical, cost-effective, non-destructive approach for assessing bond integrity in cross-laminated timber (CLT) using ultrasound. The goal is to determine whether ultrasonic measurements can identify variations in bond quality and ultimately support more comprehensive quality assurance during CLT manufacturing than is possible through conventional destructive sampling alone.
Introduction
Reliable adhesive bonding is essential to the performance and durability of CLT, but bond quality can vary within and between panels because of differences in wood properties and manufacturing conditions. Current CLT qualification methods rely on destructive testing of a limited number of specimens and therefore provide only a small sample of the bond conditions within a manufactured product.
This project investigates ultrasound as a non-destructive method for evaluating CLT bond integrity. Unlike destructive testing, ultrasound measurements can be collected without damaging the product and have the potential to be applied at many locations across a panel. The research also examines how natural wood variability and manufacturing conditions influence both bond performance and the ultrasonic response, with the longer-term goal of supporting more comprehensive quality assurance during manufacturing.
Research Methods
The researchers fabricated 3-ply and 5-ply white fir CLT specimens with controlled variations in lamination thickness to introduce realistic variability in bonding conditions. A standardized through-transmission ultrasound procedure was developed in which ultrasonic waves were transmitted perpendicular to the CLT laminations and the received waveforms were recorded before destructive testing. Seventy-two specimens were evaluated.
Rather than relying only on ultrasonic wave velocity, the researchers extracted multiple characteristics from each waveform, including amplitude- and time-of-flight-based measurements. The same specimens were subsequently evaluated using cyclic delamination testing based on ANSI/APA PRG 320 procedures and an additional assessment of wood failure at the bondline. Wood density was also measured to determine how natural material variability affected the ultrasonic response and bond performance.
The resulting data were analyzed statistically and using machine-learning models to determine whether ultrasound measurements collected from intact CLT could predict subsequent pass/fail outcomes from destructive bond testing.
Key Findings
The results demonstrate that ultrasound measurements taken before destructive testing contain useful information about subsequent CLT bond performance. Several characteristics of the transmitted ultrasonic waveform differed between specimens that subsequently passed and failed bond-integrity tests. Machine-learning models using these measurements showed potential for classifying bond performance, with stronger results for 3-ply CLT than for the thicker 5-ply configuration. The study also produced an important finding about the factors controlling bond performance. The deliberately introduced lamination-thickness differences did not have a simple relationship with delamination: larger thickness differences did not necessarily produce poorer bonds. Instead, local wood density was consistently associated with bond outcome, with higher-density specimens generally showing greater susceptibility to bond failure. Density also influenced the ultrasonic response. These findings demonstrate the importance of accounting for natural wood variability when developing non-destructive quality-assurance methods for CLT.
Future Work
The next phase will focus on scaling the approach from controlled specimens toward panel-scale CLT assessment and evaluating its applicability under a broader range of material conditions. The approach will also be evaluated on CLT manufactured with thermally modified wood to determine its performance when material properties differ substantially from conventional CLT.
The project will also investigate and compare supervised and unsupervised machine-learning approaches for delamination prediction, including whether unsupervised methods can identify abnormal or potentially deficient bond conditions when relatively few failure observations are available. The longer-term goal is to develop spatial ultrasonic scanning approaches that can characterize variations in bond integrity across larger CLT panels, providing manufacturers with more comprehensive information about panel quality than sparse destructive sampling, and potentially supporting assessment of CLT after installation or exposure to conditions that may affect bond integrity.
Publications & Presentations
Battisti, A., McCann, A., Idoghor, S., Muszynski, L., Riggio, M., & Nasir, V. (2025). Ultrasonic Characterization for Bondline Integrity in Cross Laminated Timber (CLT). International Conference on Wood Adhesives, Vancouver, Canada, October 22–24, 2025.
Nasir, V. (2025, October 16). Fast and Reliable Characterization and Quality Assurance of Fabricated CLT Panels. 2025 Mass Timber Research & Design Innovation Symposium, Corvallis, OR. Presentation video
McCann, A., Battisti, A., Idoghor, S., Muszynski, L., Riggio, M., & Nasir, V. (2026). Exploring Ultrasound as an Alternative to Destructive Delamination Testing for CLT. International Mass Timber Conference, Portland, Oregon, March 31–April 2, 2026.
McCann, A., Battisti, A., Idoghor, S., Muszynski, L., Riggio, M., & Nasir, V. (2026). Prediction of Cross-Laminated Timber Delamination Using Ultrasound and Machine Learning. Forest Products Society International Conference, Eugene, Oregon, July 14–17, 2026.
Funding & Acknowledgements
This research, titled “Fast and reliable characterization and quality assurance of fabricated CLT panels using an ultrasound method integrated with an AI-based framework”, was supported by the USDA Agricultural Research Service Award No. 58-0204-4-002.
