Predicting the Stiffness of Sawn Products by X-ray Scanning of Norway Spruce Saw Logs

Abstract

The aim of the study was to investigate the possibility of strength grading Norway spruce [Picea abies (L.) Karst.] saw logs on the basis of simulated X-ray LogScanner measurements and to evaluate the potential accuracy of X-ray LogScanner measurements of green heartwood density and percentage of heartwood. The study was based on 272 logs for strength grading and 29 logs for measurements of green heartwood density and percentage of heartwood. The logs were scanned using computed tomography (CT). After sawing, the modulus of elasticity (MOE) of the centre boards was measured using a strength-grading machine. The CT images were used for simulations of an X-ray LogScanner, resulting in simulated measurements of different variables such as diameter, taper, percentage of heartwood, density and density variations. Multivariate models for prediction of MOE were then calibrated using partial least squares (PLS) regression. The MOE of a log was defined as the mean value of the MOE of the two centre boards. The study showed that the simulated X-ray LogScanner measured the percentage of heartwood and green heartwood density with relatively high accuracy (R 2 = 0.94 and R 2 = 0.73, respectively, after removing two outliers) and that these and other variables measured by the simulated X-ray LogScanner could be used to predict the stiffness of the centre boards. These predictions were used to sort the logs according to the predicted MOE. When sorting out 50% of the logs (''high-strength'' logs), the percentage of C30 boards increased from 73% (all logs in the study) to 100% (only ''high-strength'' logs). The rest of the logs could then be divided into two groups, one of them with 100% C24 and C30 boards.