An advanced virtual grading method for wood based on surface information of knots

Abstract

Strength grading of timber boards is an important step before boards can be used as lamellas in glued laminated timber. Grading is generally done visually or by machine, whereby machine grading is the faster and more accurate process. Machine grading gives the best strength prediction when dynamic modulus of elasticity (MoE_dyn) is measured and some kind of knot assessment algorithm is included as well. As access to the actual density for the measurement of MoE_dyn may be impossible in some conditions, this grading method may face some problems in strength prediction. As the strength of a board is related to its natural defects and the ability of the stress waves to propagate around these defects, a virtual method for more accurate strength predictions is developed based on the knot information on the surface. Full 3D reconstruction of the boards, based on knot information on the surfaces of these boards, is an important step in this study. Simulations were run for 450 boards of spruce, Douglas fir, beech, ash and maple, covering a large quality range. Abaqus and Python were used for the numerical simulations. From the numerical simulations, three different stress concentration factors were calculated in the vicinity of the defects. Additionally, a virtual longitudinal stress wave propagation was modeled for the determination of the dynamic modulus of elasticity. The FEM results were used to predict the tensile strength. By means of a regression analysis, the correlation with actual visual and machine readings was validated. An improvement in the strength prediction was observed based on the virtual method, when compared to currently available grading machines. It shows the potential of numerical methods for strength prediction of wood based on visual knot information.