This document is the accepted manuscript version of the following article:

Ehrhart, T., Steiger, R., Lehmann, M., & Frangi, A. (2020). European beech (Fagus sylvatica L.)
glued laminated timber: lamination strength grading, production and mechanical properties. European
Journal of Wood and Wood Products. https://doi.org/10.1007/s00107-020-01545-6

European beech (Fagus sylvatica L.) glued laminated timber: Lamination strength
grading, production and mechanical properties
Thomas Ehrhart 1) 2), René Steiger 2), Martin Lehmann 3), Andrea Frangi 1)
1) ETH Zurich, Institute of Structural Engineering, Stefano-Franscini-Platz 5, 8093 Zurich, Switzerland
2)
Empa – Swiss Federal Laboratories for Materials Science and Technology, Structural Engineering Research
Laboratory, Ueberlandstrasse 129, 8600 Duebendorf, Switzerland
3)
Bern University of Applied Sciences, Institute for Materials and Wood Technology, Solothurnstrasse 102, 2504 Biel,
Switzerland
*corresponding author: ehrhart@ibk.baug.ethz.ch; +41 44 633 9150

Abstract
This paper presents the results of extensive investigations on the lamination strength grading, the
production and the mechanical properties of European beech (Fagus sylvatica L.) glued laminated
timber (GLT). Based on the analysis of potential influencing parameters on strength and stiffness as
well as subsequent tension tests parallel to the grain on single boards, a combined visual/machine
approach for grading the raw material into tensile strength classes T50, T42, T33 and T22 was
developed. Boards strength graded with the developed procedure were then finger-jointed by a
Swiss GLT producer and the strength of the finger joints was investigated by means of tension and
bending tests. The strength and durability of the bonding was investigated and verified by means of
tensile-shear and delamination tests. It could be shown that the required finger-joint and bondline
strengths for GLT of strength classes GL40 and GL48 can be achieved, but that the process
parameters for finger jointing (in particular the geometrical properties of the finger joint profile) have
to be optimized in order to be able to produce GLT of strength class GL55. Finally, an extensive
experimental testing campaign was performed to investigate the mechanical properties of European
beech GLT produced based on the strength grading rules and production techniques developed
before. Bending, tensile and compressive parallel to the grain, as well as shear tests were carried
out on GLT specimens of strength classes GL40, GL48 and GL55 in different sizes in terms of cross-
section and length. Based on these investigations and complementing numerical simulations,
characteristic strength and stiffness values and formulae for consideration of size effects in bending,
tension and shear were determined.

Keywords
European beech glued laminated timber, Fagus sylvatica L., mechanical properties, strength grading

1 Introduction
To date the share of European beech (Fagus sylvatica L.) amounts to 18.1% of the total timber stock
in the Swiss forests. Along with Norway spruce (Picea abies Karst., 43.7%), European beech is the
second most common tree species and by far the most common hardwood species in Switzerland,
followed by ash (Fraxinus excelsior L., 4.1%), maple (Acer pseudoplatanus L., 3.1%), and oak
(Quercus robur L., 1.9%) (BAFU 2018). In a study conducted between 2004 and 2006 (Brändli 2010),
the species European beech, Norway spruce, ash, maple and oak contributed 18.1%, 44.1%, 3.7%,
2.9% and 2.2% respectively, to the total timber stock in Switzerland.
Hardwoods in general are expected to benefit from the changing climatic conditions in Central
Europe and their share of the forests is predicted to increase further in future (Lindner et al. 2010).
However, dry summers in the years 2018 and 2019 also led to partly severe damage concerning
beech trees dying off because of drought in Central Europe. In the course of the European Research
Project SUSTREE, a decline in the species European beech by 2100 has been forecasted
(Bundesforschungszentrum für Wald bfw Österreich 2019).
The long-term goal of sustainable forest management, a balance between growth and harvest,
cannot be achieved today, especially regarding hardwoods. In 2017, the yearly increase in the
1
hardwood forest stock of 3.45 million m3 in Switzerland was in contrast to a commercial use and
mortality of the hardwood stock of only 2.56 million m3 (BAFU 2018). Additionally, the actually still
predominant use of hardwoods for energy purposes (70% of the hardwood harvested in 2017) is
rated unfavourable by the Swiss economy, forestry agencies and federal offices.
The construction sector has been identified as an economically and in terms of climate policy
interesting option for using beech wood with its excellent mechanical properties. Numerous ongoing
or completed research projects in Austria (e.g. Frühwald and Schickhofer 2004; Hübner 2013;
Linsenmann 2016), Germany (e.g. Glos and Lederer 2000; Frühwald et al. 2003; Blaß et al. 2005;
Westermayr et al. 2018), Slovenia (e.g. Plos et al. 2018) and Switzerland contribute to a steadily
growing knowledge of the characteristics and peculiarities of this wood species with regard to its
application for construction purposes.
In Switzerland, the Swiss Federal Office for the Environment (FOEN) launched the programme
Aktionsplan Holz. The main objective of the programme is to ensure that wood from Swiss forests is
provided, processed and recycled in a sustainable and resource-efficient manner. Resource policy
thus makes a major contribution to forest, climate and energy policy (BAFU 2017). As part of this
research and innovation programme, a project was launched in 2015 under the lead of the Structural
Engineering Research Laboratory of Empa, in collaboration with ETH Zurich and the Bern University
of Applied Sciences BFH/AHB in Biel. The aim of the project was to investigate the entire production
chain of glued laminated timber (GLT) made of European beech wood, as well as to determine the
mechanical properties of the product and to contribute to the implementation of the findings in
relevant standards. A survey among designers, conducted before launching the project, revealed
that highly stressed columns, beams and trusses in residential, office and industrial buildings provide
the biggest potential for the application of beech GLT in load-bearing structures. The research project
focussed on European beech GLT of strength classes GL40, GL48 and GL55. In pilot projects, GL40
and GL48 grade GLT had been successfully produced by the industrial project partner neue Holzbau
AG (Lungern, Switzerland) and applied in practice. In order to be able to evaluate the full potential
of the raw material, the investigations also involved the strength class GL55. The paper gives an
overview on the results of this extensive research project and shows that the GLT strength classes
targeted in the project offer the possibility to expand the currently increasing share of modern timber
construction at the expense of mineral and metallic building materials.

2 Lamination strength grading

2.1 Relevant indicators
Naturally grown wood has to be strength graded when intended to be used for structural purposes,
i.e., the mechanical properties have to be estimated based on the assessment of non-destructively
measurable indicators. Indicators applicable to strength grading (i) shall be highly correlated with the
target parameters (strength and stiffness), (ii) shall be easily and reliably identifiable by appropriate
non-destructive means, and (iii) shall be quantifiable (Ehrhart 2019). All investigations described in
the following were conducted on boards with cross-section dimensions of 160 × 25 mm2. Preliminary
investigations of the industrial project partner producing the GLT had shown that the width of the
boards can be chosen simply based on the capacities in terms of size of the machines involved in
the production process. However, the thickness of the boards, from the point of view of economic
efficiency aimed at being shifted to a value as high as possible, was found to experience limitations
due to the productions process. For thicknesses of 30 and 35 mm distortions formed when drying
the boards led to problems during bonding of the lamellas since the pressure available in the bonding
press was not sufficient to compensate the distortions. In addition, from experience in practice (also
with softwood glulam) it is known that internal stresses in glued-laminated structural members
exceeding a certain limit can provoke premature formation of cracks and hence, compromise the
durability.

2
2.1.1 Visual indicators
The following visual indicators were investigated, i.e., their presence, location and dimension were
documented and their influence on the tensile strength and stiffness of single laminations was
evaluated:
 Knot area ratio (KAR) & total knot area ratio (tKAR) (e.g. BS 4978, 1996);
 Fibre deviation (Fig. 1a);
 Wavelike annual ring pattern (Fig. 1b);
 Red heart.
Whereas the size of knots and knot groups was quantified by means of the parameters knot area
ratio KAR and the total knot area ratio tKAR respectively, the indicators fibre deviation (Fig. 1a),
wavelike annual ring pattern (Fig. 1b) and red heart were only documented qualitatively. Boards with
visually appearing fibre deviations were permitted but were not allowed to be assigned to the highest
visual grade (Table 1). A more detailed description of the visual indicators evaluated during the
strength grading process can be found in Ehrhart et al. (2016).

(a) (b)

Fig. 1 Example of boards with fibre deviation (a) and wavelike annual ring pattern (b).

2.1.2 Physical indicators

The bulk density (ρ) and the first natural frequency (f0) were measured in order to be able to
determine the dynamic modulus of elasticity MOE (Edyn). The two devices ViScan V2.8 (by
MiCROTEC) and MTG Timber Grader (by Brookhuis) were used. Based on the first natural
frequency f0, the length of the lamination lℓ, and the board’s bulk density ρ, the dynamic MOE Edyn
was calculated with Equation 1. Depending on the actual wood moisture content (u), the resulting
densities were adjusted to the reference moisture content, which had been set to uref = 8% (EN 384,
2019) in this study due to the primarily intended application of beech GLT for indoor construction.
Adjusting the dynamic modulus of elasticity to the reference moisture conditions followed the finding
and recommendations by Unterwieser and Schickhofer (2011).
𝐸dyn = 4 ∙ 𝑓0 2 ∙ 𝑙ℓ 2 ∙ 𝜌 (1)

2.2 Verification by means of tension tests

Tension tests according to EN 408 (2012) were carried out on 294 European beech timber boards
and the influence of the visual and physical indicators determined during the strength grading
process on the tensile strength (ft,0) and the static tensile MOE (Et,0) parallel to the grain was
evaluated. The tensile strength was calculated based on the ultimate force (Fmax) and the
lamination’s width (wℓ) and thickness (tℓ) by means of Equation 2. The tensile MOE was calculated
using Equation 3 based on the displacements (d) measured over a length of lmeas (=
5 × wℓ = 800 mm) at force levels corresponding to 10% (F1) and 40% (F2) of the expected ultimate
force Fmax.
𝐹𝑚𝑎𝑥 𝑙𝑚𝑒𝑎𝑠 ∙ (𝐹2 − 𝐹1 )
𝑓t,0,l = (2) 𝐸t,0,l = (3)
𝑤ℓ ∙ 𝑡ℓ 𝑤ℓ ∙ 𝑡ℓ ∙ (𝑑2 − 𝑑1 )

3
The resulting tensile MOE was adjusted to the reference moisture content (uref = 8%) according to
EN 384 (2019). For the tensile strength, no equation for adjustment is specified in EN 384 (2019)
since the tensile strength parallel to the grain is not subject to marked changes with changing
moisture content (Hoffmeyer 1995). Further information on the test setup and the applied
measurement data recording system is available in Ehrhart et al. (2016).
2.3 Strength grading criteria
Based on correlation analyses, the knots, the fibre deviation and the dynamic MOE were identified
to have the largest influence on the target parameters tensile strength and MOE. Consistent with
findings by Frühwald and Schickhofer (2004) and Aicher and Ohnesorg (2011), no significant
influence of red heart was found.
Based on the analysis of the indicators relevant for strength and stiffness, criteria for the combined
visual/machine strength grading of European beech timber boards into the strength grades T50, T42,
T33 and T22 were developed. The criteria were calibrated based on the 294 European beech timber
boards tested. In Table 1, the visual and physical criteria are summarised. Additional criteria
regarding curvature, cracks and wane, which are mainly of importance for the production of GLT,
may be adopted from DIN 4074-5 (2008).

The resulting 5%-fractile values of tensile strength (ft,0,05) and mean values of MOE (Et,0,mean) parallel
to the grain for the strength grades are summarised in Table 2. The characteristic values were
determined assuming lognormal distribution. The parameters making the lognormal distribution
fitting the experimental results best were estimated with the maximum likelihood method (mlm). It
can be seen that the actual 5%-fractile values of tensile strength are very close to the targeted
values. Further information on the correlation analyses performed when developing the strength
grading rules are available in Ehrhart (2019).

Table 1 Criteria for combined visual and machine strength grading of European beech boards into strength
classes T22, T33, T42 and T50.

Visual grade
Visual criteria Vis. 1 Vis. 2 Vis. 3 Vis. 4
Knot and bark inclusion tKAR ≤ 0.05 tKAR ≤ 0.1 tKAR ≤ 0.2 tKAR ≤ 0.3
Fibre deviation Not perm. Permitted Permitted Permitted
Wavelike annual ring pattern Not perm. Permitted Permitted Permitted
Red heart Permitted Permitted Permitted Permitted
Discolouration
Hardness not reduced Permitted Permitted Permitted Permitted
Hardness reduced Not perm. Not perm. ≤ 0.2 wℓ ≤ 0.2 wℓ
Insect damage Not perm. Not perm. Not perm. Not perm.
Pith Not perm. Not perm. Not perm. Not perm.
Minimum MOE Edyn (GPa) Strength grade (based on visual grade and Edyn)
Visual grade T50 T42 T33 T22
Vis. 1 ≥ 16.5 ≥ 14.0 ≥ 12.0 ≥ 10.0
Vis. 2 - ≥ 16.5 ≥ 14.0 ≥ 10.0
Vis. 3 - ≥ 18.0 ≥ 16.5 ≥ 10.0
Vis. 4 - - - ≥ 15.0

It could be shown that strength grades up to T50 can be achieved with European beech timber
boards when applying the visual and physical indicators presented in Table 1. For a better estimation
of strength and a more efficient strength grading, information on the (local) fibre orientation with
higher precision is necessary. However, as mentioned, for example by Aicher et al. (2001), Frühwald

4
and Schickhofer (2004) and Schlotzhauer et al. (2018), so far no method for the automated detection
and documentation of fibre orientation has been available for European beech wood. Hence, in the
course of the project, a new non-contact method for the determination of fibre direction in European
beech wood, based on the analysis of the wood rays, was developed (Ehrhart et al. 2017, 2018a,
b). This method allows for highly accurate predictions of the fibre orientation, especially for flat-sawn
and semi-rift-sawn boards. Thus, the method can contribute to a more efficient strength grading of
the raw material and the implementation of strength grades even higher than T50.
Table 2 Mean and 5%-fractile values of lamination tensile strength (ft,0,ℓ) and MOE (Et,0,ℓ) and finger joint
bending (fm,j) and tensile strengths (ft,j) grouped by strength grade.

Property Symbol Unit T50 T42 T33 T22

Number of tension tests n - 64 82 45 60
Lamination

Tensile strength, mean value ft,0,ℓ,mean MPa 90.3 71.1 61.0 47.4
Tensile strength, 5%-fractile value ft,0,ℓ,05 MPa 55.1 43.8 32.6 21.8
Number of MOE measurements n - 43 43 19 37
Tensile MOE, mean value Et,0,ℓ,mean GPa 16.8 15.4 14.4 13.7
Tensile MOE, 5%-fractile value Et,0,ℓ,05 GPa 15.0 12.6 12.0 11.3
Number of bending tests n - 29 30 31 -
Bending strength, mean value fm,j,mean MPa 85.9 83.3 78.9 -
Finger joints

Bending strength, 5%-fractile value fm,j,05 MPa 66.8 66.1 65.2 -

Number of tension tests n - 36 46 48 -
Tensile strength, mean value ft,j,mean MPa 73.4 62.0 59.1 -
Tensile strength, 5%-fractile value ft,j,05 MPa 52.2 43.7 43.5 -
Ratio between fm,j,mean and ft,j,mean - - 1.17 1.34 1.34 -
Ratio between fm,j,05 and ft,j,05 - - 1.28 1.51 1.50 -

3 Production of European beech GLT

In general, the process steps in the production of beech GLT do not differ from those for softwood
GLT: Logs are sawn, boards are dried, strength graded, finger-jointed, planed and bonded on the
faces. However, in order to minimize cracking due to internal stresses occurring during the
production process, a smaller lamination thickness (in this project: final thickness tℓ = 25 mm) as well
as a wood moisture content, which already during production approximately corresponds to the later
equilibrium moisture content (in this project: 8 ± 2%), were chosen.
When launching the project, it has been agreed on with the industrial partner producing the glulam
that beside information on the type of adhesive products applied to finger jointing and bonding of the
laminations, the process parameters are kept confidential.
3.1 Finger jointing
Since being of limited length when produced in the sawmills, European beech boards are joined
longitudinally by finger joints. The strength of these joints plays a central role regarding the load-
bearing capacity of GLT elements, especially in components subjected to tensile stress and
laminations located in the tensile zone of beams. The strength of the finger joints was investigated
by researchers at the Bern University of Applied Sciences BFH/AHB in Biel. In the following, the
conducted tests are briefly summarised and the main results are presented. Further information is
available in Lehmann et al. (2018).
The finger-jointed specimens were produced on a production line for softwood GLT (i. e. without
optimising the geometrical properties of the finger joints for hardwood GLT) with a one-component
polyurethane adhesive. The parameters of the adhesive application and the pressing process were
adjusted for beech wood according to the experience of the project partner neue Holzbau AG
(Lungern, Switzerland) in order to achieve satisfying performance of the finger joints. The cross-
section dimensions of the laminations after planing were 150 × 20 mm2 in the preliminary
5
investigations and 150 x 25 mm2 when producing the GLT beams to be investigated. To get a full
representation of the material quality with regard to density and dynamic MOE, the tested specimens
were equally chosen from the three strength classes T50, T42 and T33. However, it was made sure
that zones with knots or with visually appearing fibre deviations were located at least 15 cm away
from the finger joints.
The finger joint strength was assessed by means of 90 bending tests and 130 tension tests
performed according to EN 408 (2012). The bending tests were conducted displacement-controlled;
the tensile tests were carried out force-controlled. In Table 2, the resulting bending strengths (fm,j)
and tensile strengths (ft,j) of the tested finger joints are shown grouped by T-class. For all grades,
the finger joint tensile strengths exceed the boards’ tensile strengths. Further information on the test
setups and the results are provided by Clerc et al. (2017) and Lehmann et al. (2018).
3.2 Bonding
During the GLT production, the laminations were bonded using a one-component polyurethane
adhesive and a primer, following the findings of investigations regarding appropriate bonding
procedures conducted by the project partners Henkel Engineered Wood Adhesives and BFH/AHB
Biel (Clerc et al. 2017). The chosen adhesive system exhibited sufficient strength in the tensile shear
tests in the dry (A1) and wet state (A4) according to EN 302-1 (2013). The delamination behaviour
proved to meet the requirements specified in EN 15425 (2017) when tested according to EN 302-2
(2017). Further information on the investigations of the bond line strength of European beech GLT
specimens bonded with different adhesive systems and varying the production parameters can be
found in Clerc et al. (2017) and Lehmann et al. (2018).

4 Experimental investigations on European beech GLT

The experiments on GLT specimens were carried out at Empa Dübendorf and ETH Zürich. The test
programme consisted of full-scale testing of European beech GLT specimens subjected to bending,
tension parallel to the grain, compression parallel to the grain and shear. In order to be able to
quantify the influence of the size of the stressed volume on the mechanical properties (size effect,
e.g. Weibull, 1939), the tested specimens differed not only with regard to strength class (GL40, GL48
and GL55) but also in terms of size, i.e., cross-section area and length. Thus, equations for the
quantification of the size effect could be deducted from the experiments. To keep the costs for the
specimens as low as possible, the reference height for the test campaign was chosen 400 mm. Data
analysis was conducted assuming normal distribution of the density and lognormal distribution of
strength and stiffness, i.e., MOE and shear modulus. The parameters making the respective
distribution fitting the experimental results best were estimated with the maximum likelihood method
(mlm).
4.1 Bending strength and MOE
4.1.1 Material and methods
Four-point bending tests according to EN 408 (2012) were performed on European beech GLT
beams with heights of 200 mm (width w = 120 mm), 400 mm (w = 160 mm), 600 (w = 160 mm) and
800 mm (w = 180 mm). Thus, the span varied between (18 × 0.20 =) 3.60 m and 14.40 m.
Specimens of the strength classes GL40c (with outer 50% of the cross-sections consisting of T33
grade laminations and inner 50% of the cross-section consisting of T22 grade laminations), GL48c
(T42/T22) and GL55c (T50/T33) were investigated (Fig. 2).

6
 T-class Layup 180 mm
T22 0.25 h
T33
h 0.50 h T42
T42 160 mm
T50 0.25 h

T42

800 mm
160 mm
T22

600 mm
T33 T42 T50
T22
120 mm

400 mm
T22 T22 T33
200 mm

T42
T22 T42
T42
T42 T33 T42 T50

GL48c GL40c GL48c GL55c GL48c GL48c

n=7 n=7 n=7 n=7 n=7 n=4

Fig. 2 Sizes and cross-section layups (used T-classes) of the specimens investigated in four-point bending
tests.

Seven specimens were tested per series. In each test, the local (Em,loc) and global MOE (Em,glob), and
the bending strength (fm) were determined according to EN 408 (2012). In the tests on beams with
heights of 600 and 800 mm, additionally the shear modulus (Gg) was determined. Em,loc was
assessed based on deflection measurements in the neutral axis on both lateral sides of the beams.
To prevent the sensors from damage, deflection measurements relative to the strong floor on top of
the beams were used to determine Em,glob. Hence, the measured deflections include deformations
perpendicular to the grain due to respective compression stresses at the supports (Bogensberger et
al. 2006). This is why the global MOE, since being affected by the test and the measurement set up,
is known to be a stiffness indicator rather than a material property.
4.1.2 Results and discussion
The resulting bending strength (fm) and local MOE (Em,loc) determined in the four-point bending tests
are summarised in Fig. 3. In the boxplots, the upper and the lower margin of the boxes are defined
by the first and third quartiles. The horizontal line within the box represents the median and horizontal
lines outside the box represent minima and maxima. Additionally, dots represent outliers that are
located more than 1.5 times the interquartile distance above or below the box in Fig. 3. The 5%-
fractile values were calculated (i) assuming infinite sample size and (ii) according to
EN 14358 (2016) accounting for the sample size.
Considering only the beams of strength class GL48c (boxes with grey fill), a significant decrease in
strength with increasing beam height is observed (Fig. 3a). For beams with a height of 200 mm,
bending strengths of up to 96 MPa were achieved. The pronounced size effect can be explained by
the increasing amount of potentially weak points, especially of finger joints, in the highly stressed
outer zone of the beams.
When analysing the test results of the series with a beam height of 400 mm, the differences in
bending strength are very small between the strength classes GL40c, GL48c and GL55c. This finding
is attributed to the influence of the finger joints, which were identified as predominant cause of
fracture and limiting the strength of the beams independent of the strength class. Assuming infinite
sample size, the test results exceeded the targeted characteristic values (5%-fractile values) of
bending strength of 40 MPa, 48 MPa and 55 MPa in all strength classes. When introducing a penalty
factor according to EN 14358 (2016) to account for the actual number of specimens tested per
sample, for beams with heights higher than the reference height of 400 mm due to the size effect
(Equation 4) and the small sample size the target values of the strength class GL48 are not met. The

7
GL55 sample even for beams with heights of 400 mm revealed 5%-fractile values fm,k,EN14358 lower
than the target value. In order to be able to produce GLT of strength class GL55, the process
parameters for finger jointing (in particular the geometrical properties of the finger joint profile) would
have to be optimized.

Fig. 3 Bending strength fm (a) and local bending MOE Em,loc (b) of the tested GLT specimens of strength classes
GL40c, GL48c and GL55c. Beam heights of 200, 400, 600 and 800 mm were investigated (u = 8 ± 2%).

The coefficient of variation (cov) is around 0.10 in all series, except in series GL48c / h =400 mm
(cov = 0.05). For softwood GLT, Brandner and Schickhofer (2008) report a cov of 0.10 - 0.20. Fink
(2014) found cov of 0.14 and 0.13 for softwood GLT beams of strength classes GL24h and GL36h,
respectively. In the JCSS Probabilistic Model Code (2006), a cov of 0.15 is listed for fm,g of softwood
GLT. The smaller cov of beech GLT found in this project may be due to (i) differences in the strength
grading procedures and (ii) the predominant role of the finger joint strength with respect to the
bending strength of beech GLT.
Based on the experimental and additional numerical investigations (Ehrhart 2019), the size effect in
bending can be described with Equation 4 with an upper strength limit at a reference height of
400 mm. As hardly any finger joints were present in the outermost laminations of the beams with a
height of 200 mm, the results of the respective test series were not accounted for when investigating
the size effect.
400 0.14
𝑓m,g,k = min {𝑓m,g,k,ref ∙ ( ) ; 𝑓m,g,k,ref } (4)
ℎ

Significant differences between the strength classes can be observed regarding the bending MOE.
The calculated mean values of the local bending MOE are Em,loc,mean = 14.2 GPa (GL40c), 15.3 GPa
(GL48c average over all heights) and 16.2 GPa (GL55c) and can be very well estimated based on
the dynamic MOE determined during strength grading of the laminations. According to EN 384
(2019), the (local) bending MOE of softwood GLT may be calculated based on the global bending
MOE by means of Equation 5. For tropical hardwoods and chestnut, Ravenshorst and van de Kuilen
(2010) reported ratios between the local and global MOE of 1.16 and 1.14, respectively. Based on
the results found in the present project, Equation 6 was found to describe the relationship between
the local and the global MOE best for European beech GLT (coefficient of determination, r2 = 0.88).

𝐸m,loc,EN384 = 1.30 ∙ 𝐸m,glob − 2.60 (5) 𝐸m,loc,Beech = 1.17 ∙ 𝐸m,glob − 1.89 (6)

8
In Fig. 4, the global and local MOE are plotted. For comparison purposes, the relation specified in
EN 384 (2019) (Equation 5), the one proposed by Ravenshorst and van de Kuilen (2010) and the
relation found in the current experiments (Equation 6) are plotted.
18
40c / 400
48c / 200
17 48c / 400
48c / 600
Local bending MOE [GPa]

48c / 800
16
55c / 400

15

14

Em,loc = Em,glob
Em,loc = 1.17 Em,glob – 1.89 (r 2 = 0.88)
13
Em,loc = 1.30 Em,glob – 2.69 (EN384, 2019)
Em,loc = 1.16 Em,glob (Ravenshorst, 2010)
12
12 13 14 15 16 17 18
Global bending MOE [GPa]

Fig. 4 Local vs. global bending MOE: Experimental data and relationship determined in the project and
comparison with the relationship specified in EN 384 (2019) and the one published in Ravenshorst and van de
Kuilen (2010).

4.2 Tensile strength and MOE parallel to the grain

4.2.1 Material and methods
In order to evaluate the tensile strength parallel to the grain of European beech GLT, tests were
conducted on lamination bundles. Each specimen consisted of three laminations. The cross-section
dimensions were 160 × 75 mm2 and the free testing length was 3.54 m (i. e. 22.125 b). The number
of finger joints per specimen was between four and eight and on average 5.9. All specimens, in each
of the three laminations, contained at least one finger joint within the free testing length. For each of
the target strength classes GL55h (consisting of T50 laminations), GL48h (T42) and GL40h (T33),
seven specimens were tested. The tensile strength (ft,0) and the tensile MOE (Et,0) parallel to the
grain were determined in the tests. Further information on this test series is available in Ehrhart
(2019).
In addition, it was investigated whether tensile tests on lamination bundles are suitable for
supplementing or even replacing the time-consuming and much more expensive four-point bending
tests in future. In bending tests, failure predominantly occurs in the outermost, tensile-stressed
laminations. By “cutting out” the highly stressed edge area of a beam and conducting a tensile test,
material savings of about 95% per test would be possible. Assuming that the height of the critical
zone is 1/8 of the beam’s height (h), i.e., the part of the cross-section in which the bending stress
m ≥ 0.75 max, and its length is 6 h, i.e., the inner distance between the force application points in
bending tests according to EN 408 (2012), the volume of the tension test specimen is 1/24 (4.2%) of
the specimen’s volume in a bending test. The tests were conducted only on specimens consisting of
3 laminations, i.e. representing a beam with the reference height h = 600 mm. Hence, no conclusions
regarding the evaluation of size effects in bending by means of tensile tests can be drawn.

9
4.2.2 Results and discussion
The resulting tensile strength (ft,0) and tensile MOE (Et,0) parallel to the grain determined in the
tension tests on beech GLT lamination bundles are summarised in Table 3 grouped by strength
class. The mean and 5%-fractile values of strength and MOE are listed. The characteristic values of
the tensile strength of the specimens of strength classes T50, T42 and T33 exceed the target fractile
values of the respective strength classes with 52.9 MPa (> 50 MPa), 44.3 MPa (> 42 MPa) and
36.6 MPa (> 33 MPa). In EN 14080 (2013), the ratio of the characteristic tensile strength parallel to
the grain and the characteristic bending strength is defined as ft,0,g,k / fm,g,k = 0.8 for GLT made of
softwood and poplar. Compared to the nominal bending strengths of 55 MPa, 48 MPa and 40 MPa,
ratios ft,0,g,k / fm,g,k of 0.96 (GL55), 0.92 (GL48) and 0.92 (GL40) were found in this project, exceeding
the ratio specified in EN 14080 (2013).
With a value of 15.5 GPa, the mean MOE in strength class GL48h (specimens of all beam heights
considered) is only slightly higher than that of strength class GL40h (15.2 GPa). Compared to the
results of the bending tests (GL48c, all beam heights: 15.3 GPa / GL40c: 14.2 GPa, Fig. 3b), the
MOE determined for GL40h is above the average and the MOE determined for GL48h is
representative. Amounting to 17.0 GPa, the mean MOE for strength class GL55h is significantly
higher compared to the mean value found in the bending tests (GL55c: 16.2 GPa, Fig. 3b).
The analysis of the failure patterns of the tensile tests showed that at least one finger joint was
involved in the failure of all test specimens. In some specimens, failure of one or more finger joints
was observed in combination with a wood fracture close to a knot. Fracture patterns, in which finger
joints in all three layers combined led to failure, were observed particularly in specimens of strength
class GL55h.Hence, the tensile tests on the lamination bundles confirmed the results of the bending
tests letting conclude the finger joints to limit the bending strength of the GLT.
Table 3 Mean and 5%-fractile values of tensile strength (ft,0, in MPa) and tensile MOE (Et,0, in GPa) parallel to
the grain determined in tensile tests on beech GLT specimens of strength classes GL55h, GL48h and GL40h
with three laminations (u = 8 ± 2%).

GL55h GL48h GL40h

Symbol
= T50 = T42 = T33
Number of specimens n 7 7 7
Tensile strength, mean value ft,0,mean 57.7 53.3 43.7
Tensile strength, 5%-fractile value ft,0,05 52.9 44.3 36.6
Coefficient of variation [-] cov 0.05 0.09 0.10
Tensile MOE, mean value Et,0,mean 17.0 15.5 15.2
Tensile MOE, 5%-fractile value Et,0,05 16.3 14.7 13.9
Coefficient of variation [-] cov 0.03 0.03 0.06

4.3 Compressive strength and MOE parallel to the grain

4.3.1 Material and methods
Compression tests parallel to the grain according to EN 408 (2012) were performed on European
beech GLT specimens with quadratic cross-sections of the dimensions 150 × 150 mm2,
200 × 200 mm2 and 280 × 280 mm2. The lengths of the specimens were six times the cross-section
width and, thus, 900, 1200 and 1680 mm, respectively. The tested specimens belonged to the target
strength classes GL55h (assembled with T50 laminations), GL48h (T42) and GL40h (T33). Each
sample consisted of seven specimens.
4.3.2 Results and discussion
The mean values of compressive strengths (fc,0,mean) are very similar in all series and lie between
58.2 MPa (GL48h/280 mm) and 65.8 MPa (GL55h/200 mm) (Fig. 5a). When comparing the strength
10
classes GL40h (fc,0,mean = 60.4 MPa), GL48h (200 mm: fc,0,mean = 63.8 MPa), and GL55h (fc,0,mean =
65.8 MPa), a small (+3 to +6%) increase in compressive strength parallel to the grain for specimens
of higher strength classes was determined. The analysis of the test results of the specimens of
strength class GL48h with different cross-section size revealed no clear trend regarding size effect.
These results correspond well with those reported by Westermayr et al. (2018). Although the majority
of the beech laminations tested by Westermayr et al. were of very low quality, the authors report
compressive strengths between 49.7 and 70.8 MPa (fc,0,mean = 57.0 MPa, u ≈ 12% ) for the 60
specimens tested.
In addition to the high level of compressive strength parallel to the grain and the low variation of the
test results, the failure behaviour is particularly noticeable. Local compression of fibres, often in
zones with finger joint accumulations, local fibre deviations and knots, lead to a markedly ductile
failure behaviour.
The mean value of the compressive MOE (Ec,0,mean) parallel to the grain determined on the specimens
of strength class GL40h is 15.1 GPa, for specimens of grade GL48h 15.7 GPa (all geometries), and
for those of grade GL55h 17.0 GPa (Fig. 5b). The marked differences in MOE between the different
strength grades most likely results from the strength grading process, in the course of which the
dynamic MOE – a very good indicator for the static MOE – is considered.
75 18
(a) (b)

70
17
fc,0 [MPa]

Ec,0 [GPa]

65
16
60

15
55

50 14
GL
40h 48h 48h 48h 55h 40h GL 48h 48h 48h 55h
h
200 150 200 280 200 200 h 150 200 280 200
7n 7 7 7 7 7 n 7 7 7 7
fc,0,mean 60.4 59.9 63.8 58.2 65.8 Ec,0,mean 15.1 15.6 16.0 15.5 17.0
fc,0,05 59.7 56.4 62.5 55.9 63.7 Ec,0,0.05 14.4 14.6 15.4 14.8 16.7
cov 0.01 0.04 0.01 0.03 0.02 cov 0.03 0.04 0.02 0.03 0.01

Fig. 5 Compressive strength fc,0 (a) and compressive MOE Ec,0 (b) parallel to the grain of the tested GLT
specimens of strength classes GL40h, GL48h and GL55h with cross-section dimensions of 150 × 150,
200 × 200 and 280 × 280 mm2 (u = 8 ± 2%).

40h / 200 48h / 150 48h / 200 48h / 280 55h / 200

70 70
(a) (b)
Compressive strength [GPa]
Compressive strength [GPa]

65 65

60 60

r2 = 0.51 r 2 = 0.57
55 55
650 675 700 725 14 15 16 17 18
Density [kg/m3] Compressive MOE [GPa]

Fig. 6 Relationship between density and compressive strength parallel to the grain (a) and between
compressive MOE and compressive strength parallel to the grain (b).

11
A coefficient of determination of r2 = 0.51 was found for the correlation between density and
compressive strength parallel to the grain (Fig. 6a). The correlation between the compressive MOE
and the compressive strength parallel to the grain is even higher (r2 = 0.57, Fig. 6b). If both the
density (ρ) and the MOE (Ec,0) have been determined by any non-destructive measurement in
advance, the compressive strength parallel to the grain of European beech GLT can be estimated
using Equation 7 with a coefficient of determination r2 = 0.65.
𝑓c,0 = −30.2 + 8.61 ∙ 𝜌 ∙ 10−2 + 2.00 ∙ 𝐸c,0 ∙ 10−3 (7)

4.5 Shear strength and modulus

4.5.1 Material and methods
For the determination of shear strength of full-size GLT beams, to date, no harmonized test
procedure and set up is available in European standards. Furthermore, according to the current
version of the European standard for the design of timber structures Eurocode 5 (2010), size effects
in shear are not considered, disregarding numerous studies that confirmed the influence of the size
of the stressed wood volume on the shear strength (e.g. Keenan 1974; Longworth 1977; Foschi and
Barrett 1980; Colling 1986; Rammer et al. 1996; Gehri 2010 and Brandner et al. 2012).

(a) (b) (c)

Fig. 7 Short span three-point bending test (type 3P) (a), EN 408-alike compression shear test (type EN408)
(b) and asymmetric four-point bending test (type 4P) (c) in the style as presented by Basler et al. (1960).

In order to (i) investigate the shear strength of European beech GLT, (ii) to assess the influence of
the size of the stressed wood volume on the shear strength and (iii) to evaluate the influence of the
test configuration on the shear strength, three different test configurations were adopted in this
project (Fig. 7). Furthermore, specimens of different sizes in terms of cross-section dimensions and
lengths were tested. Specimens belonging to the strength classes GL48c and GL55c were
investigated. However, aiming at reducing the probability of premature bending failures, the
outermost laminations were always of strength grade T50.
Some 29 of the 42 shear tests were conducted using a short-span three-point bending test
configuration with a span of 2.5 times the beam height (type 3P, Fig. 7a). This test configuration is
similar to the one used by Büeler (2011) and Steiger and Gehri (2011). Glued-in steel rods were
used to transfer the forces at the load application points and at the supports. As described by Steiger
and Gehri (2011), accompanying compressive stresses perpendicular to the grain occur in this test
configuration, contributing to an apparently higher shear strength. Beam heights of 200, 400 and
600 mm were investigated. In 24 of the above mentioned sample of 29 tests, specimens with an I-
shaped cross-section were investigated in order to increase the probability of shear failures and
decrease the number of bending failures (web width = 0.75 × girder width, girder height =
0.20 × beam height). Five specimens were tested with rectangular cross-sections (400 × 160 mm2).
For comparative purposes, six specimens were tested with an EN 408-alike compression-shear test
configuration (type EN408, Fig. 7b). The cross-section dimensions of these specimens were
200 × 120 mm2 and the length was 520 mm. Another seven specimens were tested with a newly
developed asymmetric four-point bending test configuration (type 4P, Fig. 7c), which is based on a
12
shear test configuration presented by Basler et al. (1960). Further information on all shear test setups
and the measurements is available in Ehrhart et al. (2018c) and Ehrhart (2019).
4.5.2 Results and discussion
Not in all tests, shear failures could be achieved. Consequently, censoring of data was considered
in the data analysis, following the recommendations by Steiger and Köhler (2005). In 29 of 31 tests
on specimens with I-shaped cross-sections, a shear failure could be achieved (percentage of shear
failures: ηshear = 94%). Compared to tests by Lam et al. (1995) on softwood GLT (ηshear ≈ 40%) and
own tests with rectangular cross-sections (ηshear ≈ 20%), the proportion of shear failures when using
I-shaped cross-sections is much higher and thus this type of cross-section is advisable when
investigating specimens with particularly high shear strength, as already recommended by, for
example, Larsen (1987) and Schickhofer (2001). In the EN 408-alike shear tests, mostly
compression failures were observed. The resulting shear strengths (fv) for the different test
configurations, cross-section shapes, beam heights and strength classes are shown in Fig. 8.
Depending on the test configuration and the beam size, the shear strengths were found to be in a
range of between 8 to 17 MPa. While van de Kuilen et al. (2017) reported shear strengths on a
similar level (fv,mean = 13.4 MPa, cov = 0.12), Aicher and Ohnesorg (2011) found much lower shear
strengths (fv,mean = 6.1 MPa, cov = 0.19). However, in about 50% of the tests carried out by Aicher
and Ohnesorg, a bondline failure occurred, indicating that the shear strength of the used adhesive
system was the limiting factor regarding the shear strength of the GLT specimen.
In agreement with the test results published by Schickhofer (2001), no significant difference in shear
strength was found for the different strength classes investigated. However, digital image correlation
measurements showed that failure had often been initiated in zones with local stress concentrations
near knots and bark inclusions.
18

Failure type
Shear
16

Bending
Short span 3-point bending tests Compression
14
fv [MPa]

12
10
8
6

Type 3P 3P 3P 3P 3P 4P EN408
GL 48c 48c 55c 48c 48c 48c 48c
CS I I I I R I R
h 200 400 400 600 400 400 200
n 7 3 7 7 5 7 6
fv,mean 16.0 13.2 12.4 9.4 - 10.8 -
fv,05 14.5 - 10.2 7.7 - 9.0 -
cov 0.06 - 0.15 0.12 - 0.11 -

Fig. 8 Shear strength grouped by test configuration type (type), strength class (GL), cross-section type (CS)
and beam height (h). Censoring of the data was taken into account when analysing the test results. The
markers indicate the type of failure (u = 8 ± 2%).

13
A pronounced size effect was observed, i.e., for increasing beam heights, the mean and
characteristic values of shear strength decrease significantly. This finding is consistent with previous
studies, in which an influence of the beam height (Brandner et al. 2012), the shear area (Keenan
1974; Gehri 2010; Rammer et al. 1996; Soltis and Rammer 1994) or the beam volume (Longworth
1977; Colling 1986; Foschi and Barrett 1980) on the shear strength had been reported. Ehrhart
(2019) presented a volume-based approach for the consideration of the size effect. Assuming a
conservative ratio between bending and shear strength and identifying critical loading conditions,
the volume based approach can be simplified to Equation 8 based on the beam height.
400 0.4
𝑓v,g,k = min {6.0 ∙ ( ) ; 6.0} (8)
ℎ

In 71 shear field measurements, a mean value of shear modulus Gg,mean = 1.17 GPa and a 5%-
fractile value of Gg,05 = 0.97 GPa were determined (cov = 0.11). These values are 25% higher
compared to the values reported by Büeler (2011). However, Büeler used beech laminations of very
low quality with a moisture content of u ≈ 12%. In the technical approval Z-9.1-679 (2009) for
European beech GLT, a mean value of shear modulus of 1.00 GPa and a 5%-fractile value of
0.80 GPa are indicated.
The asymmetric four-point bending test presented in this paper is based on the configuration used
by Basler et al. (1960) for the investigation of web buckling of welded steel plate girders. It has
proven to be very suitable for determining the shear strength of GLT. Shear failures occurred in all
tests and the shear field is almost free of simultaneously acting perpendicular to grain stresses. The
test configuration is a realisation of a loading situation in practice (multi-span beam) where shear
stresses are most likely governing the design. Given a certain shear force, the simultaneously acting
maximum bending moment and the forces to be introduced are considerably lower compared to a
three-point bending test. Consequently, shear failures are more likely to occur, even without the
need of changing from the rectangular to an I-shaped cross-section. Thus, this test configuration
should be further investigated, applied and, ultimately, evaluated to be included in the test standard
EN 408 (2012).

5 Mechanical properties of European beech GLT

Based on the experimental investigations presented in this paper and additional numerical
simulations targeting the bending and tensile strength parallel to the grain (Ehrhart 2019), the
mechanical properties and densities summarised in Table 4 have been determined for European
beech (Fagus sylvatica L.) GLT of the strength classes GL40, GL48 and GL55 with combined and
homogeneous layups (Fig. 2). The presented values are valid for a wood moisture content of
u = 8 ± 2% and refer to a reference height of 400 mm. Size effects on the bending, tension and shear
strength shall be considered using Equation 4 (fm,g,k and ft,0,g,k) and Equation 8 (fv,g,k).
Whilst the MOE in tension and compression parallel to the grain were almost identical for all strength
classes GL40h (mean values of 15.2 vs. 15.1 GPa), GL48h (15.5 vs. 15.7 GPa) and GL55h (17.0 vs
17.0 GPa), slightly lower values were found for the local bending MOE (14.2, 15.3 and 16.2 GPa).
However, as the range of bending MOE within strength class GL48c was between 14.2 and
16.2 GPa, these differences can rather be attributed to the natural variation of the material properties
than to the type of loading. Whereas this finding is of interest in research, the differences are small
and hence, the design of beech GLT structural members in practice may be based on only one MOE
value, which is valid independent of the loading situation. The MOE values listed in Table 4
additionally take into account the results of numeric simulations (Ehrhart 2019).
Table 4 Mechanical properties (strength in MPa; MOE and shear moduli in GPa) and density (in kg/m3) of
European beech (Fagus sylvatica L.) GLT of strength classes GL40, GL48 and GL55 with homogeneous and

14
combined layups according to Fig. 2. The specified values refer to a beam height of 400 mm and a moisture
content of u = 8 ± 2%.

GL strength class
Property Symbol
40c 40h 48c 48h 55c 55h
Bending strength fm,g,k 40.0 40.0 48.0 48.0 55.0 55.0
Tension strength ft,0,g,k 26.0 32.0 30.0 38.4 36.5 44.0
Compression strength fc,0,g,k 40.0 45.0 45.0 50.0 50.0 55.0
Shear strength fv,g,k 6.0
E0,g,mean 14.0 14.2 15.2 15.4 16.4 16.6
Modulus of elasticity
E0,g,05 13.0 13.2 14.2 14.4 15.4 15.6
Gg,mean 1.10
Shear modulus
Gg,05 0.90
ρg,k 660
Density
ρg,mean 690

6 Conclusion
In the presented study, based on extensive experimental and numerical investigations, it was shown
that it is possible to produce GLT of strength classes GL40, GL48 and GL55 with homogeneous and
combined layups from European beech wood. Rules for a combined visual/machine strength grading
approach were developed and presented. The knot size, described by the total knot area ratio, fibre
deviations and the dynamic modulus of elasticity were determined to have the largest influence on
the tensile strength and stiffness parallel to the grain of single laminations. For further improving the
quality of the strength estimation and in order to get a more efficient strength grading process, a non-
contact method for the determination of the fibre orientation was developed within this project.
The finger-joint strength and the bondline strength and durability were tested and evaluated
according to current European standards. Although these standards were developed for softwood
GLT, all requirements were fulfilled using optimised adhesive systems and process parameters.
Extensive experimental investigations on European beech GLT have shown that the mechanical
properties are considerably higher compared to softwood GLT. Using European beech wood for
structural applications allows to extend the range of GLT from currently GL32 (according to
EN 14080, 2013) up to GL55 and, thus, to increase the bending, tensile, compressive and shear
strengths by more than 70%. However, for the production of European beech GLT of strength class
GL55 finger jointing of the boards has to be optimized, particularly regarding the profile of the finger
joints.
Similar to prescriptions in Eurocode 5 (2010), Equation 6 was developed to consider size effects on
the bending strength of European beech GLT. However, compared to Eurocode 5 (2010), a slightly
higher exponent of 0.14 (instead of 0.10) was found. Furthermore, it is suggested not to increase the
bending strength for beams smaller than the reference height, but to decrease the strength of beams
larger than the reference height. Additionally, Equation 8 was presented for consideration of the size
effect in shear, taking into account results of this project and previous studies confirming the
influence of the size of the stressed wood volume on the shear strength.
The asymmetric four-point bending test presented in this paper has proven to be very suitable for
determining the shear strength of GLT. Shear failures occurred in all tests and the shear field is
almost free of simultaneously acting perpendicular to grain stresses.

15
Acknowledgments
The authors gratefully acknowledge the funding received from the Swiss Federal Office for the
Environment FOEN within the framework of the Aktionsplan Holz. Gratitude for the excellent
cooperation is expressed to the Swiss sawmillers Corbat SA, Konrad Keller AG, Richard Lötscher
AG and Koller AG, to the suppliers of the strength grading devices MiCROTEC and Brookhuis and
to the adhesive experts of Henkel Engineered Wood Adhesives. Finally, the assistance of the
technicians at Empa and ETH Zürich in preparing and carrying out the experiments as well as the
input and support by Holzindustrie Schweiz, the hardwood GLT specialists of neue Holzbau AG, who
took care of producing the GLT specimens, and the experts Prof. em. E. Gehri and M. Zimmermann
are thankfully acknowledged.

Conflict of interest statement

On behalf of all authors, the corresponding author states that there is no conflict of interest.

References
Aicher S, Höfflin L, Behrens W (2001) A study on tension strength of finger joints in beech wood laminations. Otto-Graf-
Journal, 12, pp. 169–186.
Aicher S, Ohnesorg D (2011) Shear strength of glued laminated timber made from European beech timber. European
Journal of Wood and Wood Products, 69, pp. 143–154. doi: DOI 10.1007/s00107-009-0399-9.
BAFU, BFE, SECO (Publisher) (2017) Ressourcenpolitik Holz - Strategie, Ziele und Aktionsplan Holz (in German; Wood
Resource Policy - Strategy, Objectives and Wood Action Plan). Federal Office for the Environment FOEN, Bern,
Switzerland.
BAFU (Publisher) (2018) Jahrbuch Wald und Holz 2018 - Umwelt-Zustand Nr. 1830 (in German; Swiss Statistical Yearbook
of Forestry). Federal Office for the Environment FOEN, Bern, Switzerland.
Blaß HJ, Denzler J, Frese M, Glos P, Linsenmann P (2005) Biegefestigkeit von Brettschichtholz aus Buche - Karlsruher
Berichte zum Ingenieurholzbau - Band 1 (in German; Bending strength of beech glulam). Karlsruhe, Germany:
Universitätsverlag Karlsruhe.
Bogensberger T, Schickhofer G, Unterwieser H (2006) The mechanical inconsistency in the evaluation of the modulus of
elasticity according to EN 384. CIB-W18 Meeting Thirty-Nine. Florence, Italy. Paper 39-21-2.
Brandner R, Gatternig W, Schickhofer G (2012) Determination of shear strength of structural and glued laminated timber.
CIB - Meeting Forty-Five. Växjö, Sweden. pp. 237–254. Paper 45-12-2.
Brandner R, Schickhofer G (2008) Glued laminated timber in bending: New aspects concerning modelling. Wood Sci
Technol 42, pp. 401–425. doi: 10.1007/s00226-008-0189-2.
Brändli, UB (2010) Schweizerisches Landesforstinventar - Ergebnisse der dritten Erhebung 2004–2006 (in German; Swiss
National Forest Inventory - Results of the third survey 2004–2006). Eidgenössische Forschungsanstalt für Wald, Schnee
und Landschaft WSL, Birmensdorf, Switzerland and Bundesamt für Umwelt, BAFU, Bern, Switzerland.
Basler K, Yen BT, Mueller JA, Thürlimann B (1960) Web buckling tests on welded plate girders. Welding Research Council
Bulletin Series.
BS 4978 (1996) - Specification for Visual strength grading of softwood. BSI British Standard Institute.
Büeler M (2011) Schubverhalten von Brettschichtholzträgern aus Buche und Esche (in German; Shear behaviour of beech
and ash glulam beams). Master’s Thesis, ETH Zürich.
Bundesforschungszentrum für Wald bfw, Österreich (2019) Zukünftige Baumartenverteilung bis 2100 (in German; Future
tree species distribution until 2100). Zuschnitt, 75, p. 14.
Clerc G, Lehmann M, Volkmer T (2017) Brettschichtholz aus Laubholz - Technische Grundlagen zur Marktimplementierung
als Bauprodukt in der Schweiz - Abschlussbericht Modul 3 (in German; Hardwood glued laminated timber - Technical
requirements for the market implementation as a building product in Switzerland). Biel, Switzerland.
Colling F (1986) Influence of volume and stress distribution on the shear strength and tensile strength perpendicular to
grain. CIB - Meeting Nineteen, Volume 2. Florence, Italy. pp. 254–276. Paper 19-12-3.
DIN 4074-5 (2008) Strength grading of wood - Part 5 - Sawn hard wood. DIN German Institute for Standardization.
Ehrhart T, Fink G, Steiger R, Frangi A (2016) Experimental investigation of tensile strength and stiffness indicators
regarding European beech timber. World Conference on Timber Engineering. Vienna, Austria.

16
Ehrhart T, Steiger R, Palma P, Frangi A (2018a) Estimation of tensile strength of European beech timber boards based on
the analysis of density, dynamic MOE and local fibre direction. World Conference on Timber Engineering. Seoul, Republic
of Korea.
Ehrhart T, Steiger R, Frangi A (2018b) A non-contact method for the determination of fibre direction of European beech
wood (Fagus sylvatica L.). Eur J Wood Prod 76(3), pp. 925–935. doi: 10.1007/s00107-017-1279-3.
Ehrhart T, Steiger R, Palma P, Frangi A (2018c) Mechanical properties of European beech (Fagus sylvatica L.) glued
laminated timber. International Network on Timber Engineering Research - Meeting Fifty-One. Tallinn, Estonia. pp. 343–
360. Paper 51-12-4.Ehrhart T (2019) European beech glued laminated timber. Doctoral Thesis No. 26173, ETH Zürich.
Ehrhart T, Steiger R, Frangi A (2017) Impact and detection of grain direction in European beech wood. International
Network on Timber Engineering Research - Meeting Fifty. Kyoto, Japan. pp. 479–482. Technical Note.
EN 302-1 (2013) Adhesives for load-bearing timber structures - Test methods - Part 1 - Determination of longitudinal tensile
shear strength. CEN European Committee for Standardization.
EN 302-2 (2017) Adhesives for load-bearing timber structures - Test methods - Part 2 - Determination of resistance to
delamination. CEN European Committee for Standardization.
EN 384 (2019) Structural Timber - Determination of characteristic values of mechanical properties and density. CEN
European Committee for Standardization.
EN 408 (2012) Timber structures - Structural timber and glued laminated timber - Determination of some physical and
mechanical properties. CEN European Committee for Standardization.
EN 1990 (2010) Basis of structural design. CEN European Committee for Standardization.
EN 14080 (2013) Timber structures - Glued laminated timber and glued solid timber - Requirements. CEN European
Committee for Standardization.
EN 14358 (2016) Timber structures - Calculation and verification of characteristic values. CEN European Committee for
Standardization.
EN 15425 (2017) - One component polyurethane (PUR) for load-bearing timber structures - Classification and performance
requirements. CEN European Committee for Standardization.
Eurocode 5 (2010) Design of timber structures - Part 1-1: General - Common rules and rules for buildings. CEN European
Committee for Standardization.
Fink G (2014) Influence of varying material properties on the load-bearing capacity of glued laminated timber. Doctoral
Thesis, ETH Zurich.
Foschi RO, Barrett JD (1980) Consideration of size effects and longitudinal shear strength for uncracked beams. CIB -
Meeting Thirteen. Otaniemi, Finland. Paper 13-6-2.
Frühwald A, Ressel JB, Bernasconi A (2003) Hochwertiges Brettschichtholz aus Buchenholz (in German; High quality
glued laminated timber made from beech wood). Hamburg, Germany.
Frühwald K, Schickhofer G (2004) Strength grading of hardwoods. World Conference on Timber Engineering - Vol. IIIb.
Lahti, Finland, pp. 675–679.
Gehri E (2010) Shear problems in timber engineering - analysis and solutions. World Conference on Timber Engineering.
Riva del Garda, Italy.
Glos P, Lederer B. (2000) Sortierung von Buchen- und Eichenschnittholz nach der Tragfähigkeit und Bestimmung der
zugehörigen Festigkeits- und Steifigkeitskennwerte (in German; Strength grading of beech and oak sawn timber and
determination of associated strength and stiffness characteristics). Munich, Germany.
Hoffmeyer P (1995) Wood as a building material, Lecture A4. In: Timber Engineering Step 1: Basis of design, material
properties, structural components and joints. Centrum Hout, Almere, The Netherlands, pp A4/1 – A4/21
Hübner U (2013) Mechanische Kenngrössen von Buchen-, Eschen-und Robinienholz für lastabtragende Bauteile (in
German; Material parameters of beech, ash and robinia wood for load-bearing structures). Doctoral Thesis, Graz University
of Technology.
JCSS Probabilistic Model Code Part 3: Resistance Models (2006). doi: ISBN 978-3-909386-79-6.
Keenan FJ (1974) Shear strength of wood beams. Forest Products Journal, 24(9), pp. 63–70.
van de Kuilen JWG, Gard W, Ravenshorst G, Antonelli V, Kovryga A (2017) Shear strength values for soft- and hardwoods.
International Network on Timber Engineering Research - Meeting Fifty. Kyoto, Japan. pp. 49–65. Paper 50-6-1.
Lam F, Yee H, Barrett JD (1995) Shear strength of Canadian softwood structural lumber. CIB - Meeting Twenty-Eight.
Copenhagen, Denmark. pp. 118–134. Paper 28-6-1.
Larsen HJ (1987) Determination of shear strength and strength perpendicular to grain. CIB - Meeting Twenty. Dublin,
Ireland. pp. 155–160. Paper 20-6-3

17
Lehmann M, Clerc G, Lehringer C, Strahm T, Volkmer T (2018) Investigation of the bond quality and the finger joint strength
of beech glulam. World Conference on Timber Engineering. Seoul, Republic of Korea.
Lindner M, Maroschek M, Netherer S, Kremer A, Barbati A, Carcia-Gonzalo J, Seidl R, Delzon S, Corono P, Kolström M,
Lexer MJ, Marchetti M (2010) Climate change impacts, adaptive capacity, and vulnerability of European forest ecosystems.
Forest Ecology and Management, 259(4), pp. 698–709. doi: 10.1016/j.foreco.2009.09.023.
Linsenmann P (2016) European hardwoods for the building sector (EU Hardwoods) – Final report. Available at: http://eu-
hardwoods.eu.
Longworth J (1977) Longitudinal shear strength of timber beams. Forest Products Journal, 27(8), pp. 19–23.
Plos M, Fortuna B, Straze A, Turk G (2018) Visual Grading of Beech Wood - a Decision Tree Approach. World Conference
on Timber Engineering. Seoul, Republic of Korea.
Rammer D, Soltis L, Lebow P (1996) Experimental shear strength of unchecked solid-sawn Douglas-fir. Forest Service -
Research Paper FPL-RP-553.
Ravenshorst G, van de Kuilen JWG (2010) Relationships between local, global and dynamic modulus of elasticity for soft-
and hardwoods. CIB - Meeting Forty-Two. Duebendorf, Switzerland. Paper 42-10-1.
Schickhofer G (2001) Determination of shear strength values for GLT using visual and machine graded spruce laminations.
CIB - Meeting Thirty-Four. Venice, Italy. pp. 383–409. Paper 34-12-6.
Schlotzhauer P, Wilhelms F, Lux C, Bollmus S (2018) Comparison of three systems for automatic grain angle determination
on European hardwood for construction use. Eur J Wood Prod, 76(3), pp. 911–923. doi: 10.1007/s00107-018-1286-z.
Soltis L, Rammer D (1994) Shear strength of unchecked glued-laminated timber beams. Forest Products Journal, 44(1),
pp. 51–77.
Steiger R, Gehri E (2011) Interaction of Shear Stresses and Stresses Perpendicular to the Grain. CIB - Meeting Fourty-
Four. Alghero, Italy, pp. 135–149. Paper 44-6-2.
Steiger R, Köhler J (2005) Analysis of censored data - examples in timber engineering research. CIB - Meeting Thirty-
Eight. Karlsruhe, Germany. Paper 38-17-1.
Unterwieser H, Schickhofer G (2011) Influence of moisture content of wood on sound velocity and dynamic MOE of natural
frequency- and ultrasonic runtime measurement. Eur J Wood Prod 69(2), pp. 171–181. doi: 10.1007/s00107-010-0417-y.
Weibull W (1939) A statistical theory of the strength of materials. Royal Technical University, Stockholm, Sweden.
Westermayr M, Stapel P, Van de Kuilen JWG (2018) Tensile and Compression Strength of Small Cross Section Beech
Glulam Members. International Network on Timber Engineering Research - Meeting Fifty-One. Tallinn, Estonia. pp. 307–
322. Paper 51-12-2.
Z-9.1-679 (2009) Allgemeine bauaufsichtliche Zulassung für BS-Holz aus Buche und BS-Holz Buche-Hybridträger (in
German; National technical approval for beech glulam and hybrid glulam). Studiengemeinschaft Holzleimbau eV.; DIBt
Deutsches Institut für Bautechnik.

18