New processing to improve timber strength grading with a vibration method

摘要

Strength grading of timber structures is required in a new European directive (EN 14081-2). Common grades are the following: C18, C24 and C30. It is difficult to predict strength of timber with non-destructive techniques because wood is a heterogeneous material; its strength depends on the rigidity of the wood fibre but also on the presence of singularities (nodes, fissure...). A well-known technique (so called "bing") consists in shocking the timber at its extremity, measuring the longitudinal resonance frequencies and estimating the modulus of elasticity. The correlation between strength and modulus of elasticity is poor so that the efficiency is between 50 and 60% for the grade C30 and Douglas fir. INNODURA has improved the processing to increase the efficiency. Two kinds of parameters are distinguished: (1) Grading parameters that have to be correlated with strength for all samples; modulus of elasticity is the first one of these parameters, (2) Qualitative parameters to detect timbers with singularities; these parameters indicate that the modulus of elasticity is too high by comparison with strength. Data processing acts as a neural network; it includes the two kinds of parameters with non-linear functions. The results for Spruce and Douglas fir are given for green and dried conditions. The efficiency for Douglas fir in C30 was improved by more than 20%. A prototype, based on this method, is presented in the last paragraph of this paper.