Strength Grading of Timber

Timber exhibits large variations in mechanical properties, which makes grading of timber into appropriate strength classes important for the competitiveness towards other construction materials. Modelling and strength grading of timber has thus for more than ten years been an intense research area within the field of timber engineering and wood mechanics at Linnaeus University.

Our research

Timber is a naturally grown material that exhibits much larger variations in mechanical properties than what other common structural materials do. In order to utilize structural timber efficiently, it is therefore important to grade each piece into an appropriate strength class. The more accurate the grading, the more competitive structural timber becomes compared to other, less eco-friendly materials.

For more than ten years, modelling and strength grading of timber has been an intense research area within Wood Building Technology at Linnaeus University.

Models from detailed data

Accurate grading of structural timber is based on indicating properties (IPs) of strength, stiffness and density. IPs, in turn, are often based on comparatively inelaborate data from measurements, like axial resonance frequency, weight and dimensions of boards.

However, this is not sufficient for very accurate prediction of strength. Instead, detailed data from optical surface scanning, and/or data from X-ray scanning, should be collected and translated into models of individual timber boards. Using these, grade determining properties like strength can be predicted with higher accuracy.

Part of our research has already led to practical and accurate methods for machine strength grading. Other parts are more of basic research, expected to contribute to improved utilization of wood in a longer perspective. PhD theses presented by the group are Oscarsson (2014) and Hu (2018a).

Ongoing research comprise sub areas as follows.

Assessment of methods for machine strength grading

In collaboration with industry, the research group has developed a now patented machine strength grading method (Olsson et al. 2013; Olsson and Oscarsson 2017; Olsson et al. 2018). The method is based on data of local fiber direction on timber surfaces in combination with axial resonance frequency. By utilizing data from a wood scanners in a relatively simple mechanical model, it is possible to calculate local bending stiffness along boards. This, in turn, gave basis for more prediction of strength than what is obtained using other methods available on the market.

For a method to fulfil all requirements laid down in the standard EN 14081-2 and be approved for the European market, at least 450 timber boards must be assessed and subjected to destructive testing. For the abovementioned method, this has been done in two different campaigns, to obtain settings to grade structural timber into C-classes (destructive bending tests) and to grade glulam laminations into T-classes (destructive tensile tests), respectively. The campaigns gave comprehensive non-destructive and destructive data of a large number of boards, which in turn gave basis for a critical assessment of the standard EN 14081-2. Results of such an assessment are currently under review.

Modelling of sawn timber

Further improvement of grading accuracy would require models of timber boards, by which the location and geometry of knots and fiber orientation in the surroundings of knots are accurately represented in 3D. Ongoing work aims at development and successive improvements of such models.

For results achieved so far, see e.g. Briggert et al. (2016), Hu et al. (2018b) and Lukacevic et al. (2019). Improved knowledge of location of pith, annual ring width and geometry of knots, to be obtained using combinations of data from computer tomography X-ray scanning, surface scanning and mathematical models, is expected to enable more accurate models of timber boards.

Detection of annual rings and location of pith

Knowledge of the location of pith in relation the cross section of a timber board is important for prediction of stiffness, strength and shape stability, and necessary to establish an accurate board model. For example, modelling of knot geometry requires knowledge of location of pith.

A project carried out in collaboration with industry aims at determining location of pith and annual ring distance based on grey scale images of longitudinal board surfaces, obtained from optical scanning. By wavelet transform, performed on images of the four surfaces, the annual ring wavelength on the surfaces can be determined. On basis of this information, the most likely location of pith, locally along the boards, is calculated. Preliminary results were presented in Habite et al. (2019).

Applications on engineered wood products

Data of timber boards, collected by industry scanners, give basis for value adding processing of wood other than machine strength grading of structural timber. Finger-joints in structural timber and glulam lamella enable production of long members and elimination of weak sections. Detailed data of local fiber orientation and location of knots enable reduced waste of wood cut off prior to finger-jointing.

Results presented in Olsson et al. (2019) show that a criterion for margins between knots and finger joints, utilizing knowledge of local fiber orientation, give considerably less waste compared to a criterion based on knot diameter alone.

Another area of application, new to the research group, concerns grading of laminations for cross-laminated timber (CLT). Currently, C-classes developed for structural timber are used also for grading of CLT laminations.

As part of a project titled Improving the competitive advantage of CLT-based building systems through engineering design and reduced carbon footprint, the research group will investigate if other requirements than those applied for C-classes would give better material utilization in CLT panels than what is achieved today.

Branch to trunk integration – a study in 3D on micrometer scale

A branch can carry large loads without being broken from the tree trunk. The explanation to this lies in the complicated cell structure in the area just above the branch where it connects to the trunk. However, no experimental studies have so far been presented showing the actual 3D structure, on fiber level, of the branch-trunk integration.

However, within an ongoing research project titled CT scanning and the influence of knots in structural timber, the research group utilize high-resolution X-ray scanning, performed successively on smaller and smaller specimen, cut in the interesting area where a branch integrate with the trunk in Norway spruce.

An earlier study showing fiber orientation in 3D on a millimeter scale is presented in Hu et al. (2018c). However, the new study utilizing CT X-ray scanning, will give resolution in the most interesting areas of only a few micrometers.