Increasingly taller and larger timber buildings have become more common. Currently, a 20-storey hotel tower made of cross-laminated timber is under construction in Skellefteå, Sweden. For construction, strength-graded timber with predictable mechanical properties is required. Commonly, strength-grading is done by heuristic methods based on indicating properties which are derived from various measurements, e.g. X-ray, laser scanning and recently also computed tomography (CT) scanning. The available grading methods exploit only a small fraction of the potential strength of timber boards. More accurate strength prediction could increase the efficient use of timber and enable the production of high-strength timber Since the existing approaches used data from only the board surface, the properties inside the board volume had to be reconstructed, e.g. by interpolation or by a mathematical description of the growth surfaces of the tree (Foley 2013). CT scan data provides rich information about the density variations inside the volume of a board, e.g. early- and latewood, knots and pitch pockets. So far, this source of ample data has not been used for mechanical models. The goals of this study are therefore to
– develop an approach to create 3D FE models based on CT images of dried timber boards,
– simulate bending tests on the models and calibrate the results to physical tests, and
– predict the mechanical properties of boards outside the tested sample.
Keywords: cross-laminated timber, strength-grading, wood construction, computer tomography
Authors
J. A. J. Huber
O. Broman
J. Oja
M. Ekevad
Login to download the PDF