Experimental Study and Numerical Investigation of

Blockhaus Shear Walls Subjected to

In-Plane Seismic Loads
Chiara Bedon 1; Massimo Fragiacomo 2; Claudio Amadio, M.ASCE 3; and Carlotta Sadoch 4
Downloaded from ascelibrary.org by Politehnica University of Timisoara on 05/22/17. Copyright ASCE. For personal use only; all rights reserved.

Abstract: Blockhaus systems represent a traditional construction technology in which structural resistance is obtained by direct contact
between multiple timber surfaces obtained through carvings, notches, and ancient joints. Native of forested areas, this technology is currently
used for the construction of log buildings also in earthquake-prone areas. Their seismic structural behavior hence represents an interesting and
not well known research topic. In the paper, experimental results obtained by cyclic and monotonic tests recently performed on Blockhaus
shear walls with three different types of joints are presented and discussed. At the same time, sophisticated finite-element numerical models
able to take into account the complex structural behavior of log-walls subjected to in-plane seismic forces are described. Performed numerical
simulations highlighted that in general multiple factors affect the global structural response of Blockhaus shear walls, including the presence
of imperfections (small gaps) in the joints, the contribution of friction phenomena, the mechanical properties of timber, and the interaction
between in-plane seismic forces and vertical permanent loads. However, careful calibration of timber mechanical properties and contact
interactions between overlapping logs can provide accurate numerical predictions and good correlation with full-scale experimental results.
DOI: 10.1061/(ASCE)ST.1943-541X.0001065. © 2014 American Society of Civil Engineers.
Author keywords: Wood structures; Timber loghouse shear walls; In-plane seismic loads; Full-scale monotonic and cyclic test; Friction
coefficent; Numerical models; Nonlinear analysis; Wood structures.

Introduction (LVL) (Parida et al. 2013), light-frame walls (Van de Lindt

2004; Van de Lindt et al. 2011; Vogt et al. 2012), and traditional
The paper investigates the seismic response of Blockhaus shear brick masonry filled timber-framed walls (Vasconcelos et al. 2013),
walls. The typical Blockhaus system consists of a series of struc- this is not the case for log-walls, and recent contributions available
tural elements (logs) interacting together by means of traditional in literature confirm this (Hirai et al. 2004; Branco and Araújo
and ancient connections based on carvings, notches, and contacts 2010, 2012). The structural response of log-walls is in fact not well
of multiple surfaces. In this way, the presence of metal fasteners known yet, and further investigations are required to assess the typ-
and steelwork is minimized, with significant advantages in terms ical seismic response and to improve their resistance to seismic
of cost reduction during production (e.g., carving of timber logs) events. For example, no provisions for the design of Blockhaus
and assembling on the building site. buildings are provided by the Eurocode 8 [EN 1998-1:2004,
Native of forested areas (North America, Scandinavia, and European Committee for Standardization (CEN) 2004b], the
Russia), the Blockhaus system is widely used in current practice European Code for Seismic Design.
for the construction of wooden houses in the Italian Alpine region, The paper presents the experimental results obtained from cyclic
Switzerland, and Austria, but also in other earthquake-prone re- and monotonic tests recently performed on three different types
gions. As a result, the seismic characterization of this construc- of shear walls commonly used in practice for the construction of
tion system represents a topic of great interest for researchers and Blockhaus systems. When these shear walls are subjected to in-
designers. plane lateral loads, the adopted timber-timber connections ensure
Whereas numerous researchers have studied the structural significant flexibility of the structural system and enable the attain-
performances of different types of wood shear walls including ment of large displacements. The in-plane flexibility of timber
cross-laminated walls (Ceccotti et al. 2013; Gavric et al. 2014), floors and roofs made of joists and purlins hinged to the walls by
cantilevered walls made of glulam and laminated veneer lumber means of appropriately notched main logs and supporting a floor-
ing made of OSB sheathing nailed on the timber joists wooden
1
Postdoctoral, Dept. of Architecture, Design and Urban Planning, Univ. boards does not ensure stabilization of the shear walls under in-
of Sassari, Palazzo del Pou Salit, Piazza Duomo n.6, 07041 Alghero, Italy plane lateral loads. As a result, the shear strength of Blockhaus
(corresponding author). E-mail: c.bedon@libero.it walls subjected to seismic events depends on the number of inter-
2
Professor, Dept. of Architecture, Design and Urban Planning, Univ. of sections between the main wall, perpendicular walls, and on the
Sassari, Palazzo del Pou Salit, Piazza Duomo n.6, 07041 Alghero, Italy. geometry, thus on the resisting surfaces, of the adopted timber-
3
Full Professor, Dept. of Engineering and Architecture, Univ. of Trieste, timber connection detail [EN 1995-1-1:2009 (CEN 2009a)]. As
Piazzale Europa n.1, 34127 Trieste, Italy.
4 shown in this paper, experimental strengths are in general higher
Rubner Haus AG SpA, Handwerkerzone n.4, 39030 Chienes, Italy.
Note. This manuscript was submitted on November 19, 2013; approved than the analytical shear strengths calculated according to the
on February 25, 2014; published online on July 9, 2014. Discussion period European Code for Timber Structures, the Eurocode 5 [EN
open until December 9, 2014; separate discussions must be submitted for 1995-1-1:2009 (CEN 2009a)]. Nevertheless, the structural response
individual papers. This paper is part of the Journal of Structural Engineer- of a timber log shear wall subjected to in-plane lateral forces is
ing, © ASCE, ISSN 0733-9445/04014118(11)/$25.00. rather complex to predict because it primarily depends on the

© ASCE 04014118-1 J. Struct. Eng.

J. Struct. Eng., 2015, 141(4): -1--1

 interaction of multiple surfaces in contact and on the possible Again, the main difference between the three examined Block-
interlocking and frictional dissipation of overlapping timber logs. haus specimens is in the type of joint adopted to constrain the main
Therefore, further static shear tests have been performed on shear wall at its ends. Specimen S01 is characterized by the pres-
simple specimens, composed of three timber beams precompressed ence of standard halved joints [Fig. 2(a)]. In specimen S02, double
perpendicular to the grain, to estimate the mean value of the cor- rounded dovetail hardwood shaped joints are used [Fig. 2(b)],
responding static friction coefficient between adjacent beams. At the whereas specimen S03 is characterized by the presence of Tirolers-
same time, further small specimens have been tested under compres- chloss joints [Fig. 2(c)].
sion perpendicular to the grain, to assess the ultimate strength of Each main Hw × W w × Lw shear wall is restrained at a distance
timber-timber joints used in Blockhaus structural systems. Sub- of Dc from its ends by means of two H c × W c × Lc end-constraints
sequently, three-dimensional (3D) finite-element models have been (Table 1 and Fig. 3).
implemented in the ABAQUS software package, with the aim of pre-
dicting the seismic response of the tested full-scale shear walls. Experimental Setup and Loading Protocol
Downloaded from ascelibrary.org by Politehnica University of Timisoara on 05/22/17. Copyright ASCE. For personal use only; all rights reserved.

Nonlinear simulations highlighted, as expected, that several factors Each full-scale specimen was subjected to subsequent cyclic and
affect the accuracy of numerical results, including the presence of monotonic tests for a total number of six tests, in accordance with
possible geometrical imperfections in the timber carpentry joints, suggestions given by standards in use for timber structures [EN
the contribution of friction phenomena, the mechanical properties 12512:2006 (CEN 2006b)]. To assess their shear resistance and
of timber, and the interaction between seismic lateral forces and ver- flexibility under in-plane lateral loads, quasi-static loading condi-
tical permanent loads. Nevertheless, if the main geometrical and tions were applied in all the experiments.
mechanical parameters are appropriately estimated in the adopted Tests were performed, in accordance with the setup depicted in
finite-element models, a good correspondence with test results can Fig. 4, at the temperature and relative humidity conditions of T ¼
be obtained. 20°C (
2°C) and RH ¼ 65 (
5) %, respectively. The lateral force
was applied on the top-end surface of each specimen through a
500 kN hydraulic jack equipped with a load cell connected to a
Experimental Investigation of Blockhaus reaction frame.
Shear Walls To investigate the effect of the gravity loads acting on the wall,
as a part of a real building, the specimens were also vertically pre-
Full-Scale Seismic Tests compressed by means of four Parker hydraulic jacks. Because the
examined construction technology is primarily used for buildings
Description of Specimens up to two levels, a vertical precompression of 10 kN=m on each
Three different specimens of the Blockhaus shear wall concept were full-scale specimen was considered as representative of permanent
experimentally tested under lateral loading. Each specimen had and accidental design loads deriving, in a seismic loading combi-
the same overall geometry but differed in the type of corner joints. nation, from the roof, the second level wall, and second level floor
The typical specimen, produced by Rubner Haus AG SpA, [EN 1998-1:2004 (CEN 2004b)].
consists of a 2.95 × 2.95 m Tirol shear wall made up of 18 basic To ensure a uniform distribution of the applied horizontal and
components, that is 90 mm wide and 160 mm deep timber beams vertical loads on specimens, a steel beam rigidly connected to the
obtained by gluing together using vapor-diffusible adhesive horizontal jack head and joined to the upper side of tested walls was
[EN 301:2006 (CEN 2006a)] two 45 mm thick lamellas of C24
spruce [EN 338:2009 (CEN 2009b)].
The transversal cross section of a Tirol log is displayed in Fig. 1.
Overlapping logs are characterized by the presence of two ≈
0.01 × 0.01 m notches and protrusions, which improve interlock
and friction properties between top and bottom contact surfaces.

Fig. 1. Typical cross section of a Tirol timber log with notches (dimen-
sions in mm); dashed area: simplified cross section assumed in Fig. 2. Carpentry joints of specimens S01, S02, and S03: (a) standard;
FE-models (b) rounded dovetail; (c) tirolerschloss

© ASCE 04014118-2 J. Struct. Eng.

J. Struct. Eng., 2015, 141(4): -1--1

 Table 1. Geometrical Properties of Full-Scale Blockhaus Specimens
Main wall End-constraints
Height Width Length Height Width Length Distance from
Specimen H w (m) W w (m) Lw (m) Type H c (m) W c (m) Lc (m) wall-end Dc (m)
S01 2.95 0.09 2.95 Orthogonal log-wall 2.95 0.09 0.6 0.1
S02 Pillar 2.95 0.12 0.12 0.1
S03 Orthogonal log-wall 2.95 0.09 0.6 0
Downloaded from ascelibrary.org by Politehnica University of Timisoara on 05/22/17. Copyright ASCE. For personal use only; all rights reserved.

Fig. 3. Specimen S01: (a) elevation; (b) top view; (c) lateral view (dimensions in mm)

Fig. 4. Full-scale tests performed on Blockhaus shear wall: Experimental setup and detail of cable connection to the top timber log

© ASCE 04014118-3 J. Struct. Eng.

J. Struct. Eng., 2015, 141(4): -1--1

 used. Moreover, to ensure the shear collapse of the walls and to residual lateral resistance. Because of this reason, the in-plane lat-
avoid their overturning because of the applied in-plane lateral eral loads were applied to each wall by taking into account the dis-
loads, two 10-mm diameter steel cables were attached to the ends placement history depicted in Fig. 5(b). A first loading-unloading
of the specimen upper log (see the detail in Fig. 4). cycle up to a load level slightly above the yielding point of the
The investigated shear walls were then attached to the steel specimen, as assessed from the previous cyclic test, was performed.
foundation by means of two steel angle bracket connectors, as Then the lateral load was almost linearly increased up to either the
usually done in practice for Blockhaus systems installed by Rubner collapse of each specimen, or to a maximum lateral displacement
Haus. The typical shear steel bracket is 3-mm thick and is joined to of the top log δ max ¼ 200 mm (≈ Hw =15, maximum available
the bottom timber log of the main shear wall and to foundation by displacement of the hydraulic jack used).
means of twelve Anker nails and two SKR RothoBlaas M10 metal
Discussion of Full-Scale Cyclic and Monotonic Test Results
bolts, respectively. In these hypotheses, using the expressions given
by the Eurocode 5 [EN 1995-1-1:2009 (CEN 2009a)], a value of Specimen S01. Fig. 6 displays the cyclic response for the shear
Downloaded from ascelibrary.org by Politehnica University of Timisoara on 05/22/17. Copyright ASCE. For personal use only; all rights reserved.

11.28 kN can be obtained for the design shear resistance of each wall S01 with standard halved joints, as horizontal in-plane load
adopted steel angle bracket. V versus lateral displacement δ curve.
During the experiments, seven linear variable differential trans- Specimen S01 showed an overall appreciable dissipative capac-
formers (LVDTs) with a resolution of 0.01 mm were used (Fig. 4). ity {equivalent damping ratio ν eq ¼ 0.102 [EN 12512:2006 (CEN
The horizontal displacement at the top (1 LVDT) and the bottom 2006b)]}. In accordance with [EN 12512:2006 (CEN 2006b)] and
(1 LVDT) of each shear wall and the vertical uplift (2 LVDTs), the Fig. 5(a), three fully reversed cycles were performed for each value
horizontal slip between the top timber log and the steel profile rig- of imposed lateral displacement δ. As highlighted in Fig. 6, no
idly connected to the hydraulic jack (1 LVDT), and the deforma- strength degradation can be noticed in each series of displacement
tions of the walls along the diagonals (2 LVDTs) were measured cycles. Specimen S01 proved to be very flexible and the cyclic test
during the horizontal loading phase. was stopped at the attainment of a maximum displacement at the
Cyclic Tests. Cyclic tests were performed first on vertically f top log δ max ¼ 80 mm. The corresponding horizontal load resulted
loaded S01, S02, and S03 specimens, to estimate the seismic resis- V max ¼ 30 kN, and no sign of failure could be observed.
tance of Blockhaus shear walls and its possible decrease after a Specimen S01 was then subjected to a subsequent monotonic
certain number of cycles. According to EN 12512:2006, the in- lateral test. Yield slip resulted equal to δ y ≈ 3.5 mm. In this case,
plane horizontal cyclic load was applied following the displace- the collapse of the shear wall occurred at the attainment of a maxi-
ment history displayed in Fig. 5(a). A reference slip δ y ¼ 5 mm mum lateral load V max ¼ 43.28 kN, corresponding to a maximum
(≈ Hw =590), estimated based on engineering judgment, was taken displacement of the top log δ max ¼ 173.5 mm (Fig. 6).
into account to evaluate the amplitude of cycles. All the tests were In Blockhaus shear walls, the shear resistance depends on the
stopped at the attainment of a maximum lateral displacement at the material properties and on the number/geometry of intersections
top log equal to δ max ¼ 80 mm (≈ H w =36). with orthogonal walls, thus on the shear and compressive strengths
Monotonic Tests. Because cyclic experiments did not lead of the adopted carpentry timber joints. Based on formulas given
specimens to collapse, as discussed in detail in the section “Dis- by the Eurocode 5, the characteristic shear strength V k can be
cussion of Full-Scale Cyclic and Monotonic Test Results,” in a estimated as
subsequent experimental phase the same specimens were subjected V k ¼ minðV k;comp ; V k;shear Þ ð1Þ
to quasi-static monotonic in-plane lateral loads, to estimate their
where the values V k;comp and V k;shear are, respectively, given by
V k;comp
≤ kc;90 × fc;90;k ð2Þ
n × Aef

2 V k;shear
≤ fv;k ð3Þ
3 n × Ashear

Fig. 5. Displacement history: (a) cyclic tests, according to [EN Fig. 6. Horizontal in-plane load-lateral displacement curve for speci-
12512:2006 (CEN 2006b)]; (b) monotonic tests men S01 (cyclic and monotonic tests)

© ASCE 04014118-4 J. Struct. Eng.

J. Struct. Eng., 2015, 141(4): -1--1

Downloaded from ascelibrary.org by Politehnica University of Timisoara on 05/22/17. Copyright ASCE. For personal use only; all rights reserved.

Fig. 7. Deformed shape of Blockhaus shear walls at the end of monotonic test

In Eqs. (2) and (3), f c;90;k and fv;k denote the characteristic
compressive strength perpendicular to grain and the shear strength,
respectively; n = number of intersections between main and
orthogonal walls; Aef and Ashear = compressive and shearing
resisting surfaces in each joint; kc;90 ¼ 1.25 for spruce and for
the geometrical properties of the joints. In the specific case of speci-
men S01, the in-plane resistance is governed by the shear strength
of standard halved joints, with V k ¼ 33.6 kN. Because this value is
78% of the experimental mean strength, therefore in the typical
range of 70–80% for the characteristic-mean strength values of tim-
ber structures, it can be concluded that the use of the Eurocode 5
leads to acceptable accuracy. Further investigations highlighted that
the collapse of specimen S01, unlike the other specimens, occurred
because of the sequential shear cracking of the five bottom logs in
the main wall, as can be noticed from Fig. 7(a).
Specimen S02. The same test procedure discussed for specimen Fig. 8. Horizontal in-plane load-lateral displacement curve for speci-
S01 was applied also to specimen S02. Also in this case, the shear men S02 (cyclic and monotonic tests)
wall demonstrated high capacity to dissipate energy (ν eq ¼ 0.190)
and large flexibility because of relative slidings between overlap-
ping logs [Fig. 7(b)]. Results obtained from cyclic and monotonic
in-plane tests are shown in Fig. 8. at the end of experiments highlighted a series of local crushing
The cyclic test of specimen S02 was stopped at the attainment of failures in the hardwood dovetails, attributable to compression
a maximum lateral displacement at the top log of δ max ¼ 80 mm. perpendicular to grains. In this case, the use of the Eurocode 5
No damage was noticed in the shear wall. Nevertheless, because of formulas is slightly less accurate but always conservative.
the geometry of the adopted rounded dovetail shaped joints, in Specimen S03. As expected, specimen S03 was also very
comparison with specimen S01, the shear wall S02 showed a lower flexible and dissipative (ν eq ¼ 0.173), as displayed in Fig. 9. At
strength (V max ¼ 18.48 kN). It is interesting to notice that for the the end of the cyclic test (δ max ¼ 80 mm), no damage was noticed
adopted connection, based on Eurocode 5 formulas [Eqs. (1)–(3)], in the shear wall. Also in this circumstance, however, the load
the lateral resistance of the shear wall is governed by the compres- V-displacement δ curve obtained from the monotonic test does
sion strength of the rounded dovetail joint (V k ¼ 19.2 kN). Results not exactly coincide with the backbone curve of the cyclic test
obtained by the monotonic test confirmed the high flexibility of in the first loading phase (Fig. 9).
specimen S02. At the attainment of a maximum displacement at This effect could be justified based on the outcomes from cyclic
the top log δ max ¼ 200 mm (V max ¼ 27.12 kN), no sign of collapse and monotonic tests performed subsequently on the same speci-
was visible in the shear wall [Fig. 7(b)]. Disassembly of specimen men, namely a possible initiation of failure, although not visible,

© ASCE 04014118-5 J. Struct. Eng.

J. Struct. Eng., 2015, 141(4): -1--1

 experimental tests were performed on small Blockhaus components
to mechanically characterize the behavior of each component
under the effects of a seismic event. As is known, the seismic re-
sponse of a Blockhaus shear wall under in-plane loads strongly
depends on the interaction between multiple logs in contact, thus
it depends on friction phenomena and on local behaviors of the
adopted joints.
Prediction of the Static Friction Coefficient
Friction is not considered a resistant mechanism in timber log
shear walls by Eurocode 5. Nevertheless, friction mechanisms
have an important role in the structural behavior of these structural
Downloaded from ascelibrary.org by Politehnica University of Timisoara on 05/22/17. Copyright ASCE. For personal use only; all rights reserved.

systems. For this reason, Branco and Araùjo (2010) performed an

experimental program on simple timber log specimens, having
cross section geometry similar to Fig. 1, to assess the magnitude
of friction developed along two notched contact surfaces subjected
Fig. 9. Horizontal in-plane load-lateral displacement curve for speci-
to in-plane lateral loads. Their experiments resulted in a static fric-
men S03 (cyclic and monotonic tests)
tion coefficient equal to 0.66 [coefficient of variation ðCoVÞ ¼
8.4%] and 0.64 (CoV ¼ 12.8%) for log walls subjected, respec-
tively, to a vertical precompression of 3 and 3.9 kN. Further
friction experiments, resulting in a static friction coefficient of
in a few joints (e.g., crushing mechanisms), resulting in a corre- 0.67, were also performed by Hirai et al. (2004) on timber log
sponding lower stiffness of the shear wall under monotonic test. walls tied with vertical pretightened through bolts. Both contribu-
For specimen S03, the maximum lateral load resistance was equal tions highlighted, in general, the importance of friction phenomena
to V max ¼ 28.56 kN, and the collapse occurred because of the in the definition of the global stiffness of shear log walls subjected
cracking of a top joint (Fig. 10). to in-plane lateral loads.
Comparison with the Behavior of Traditional Timber Walls. In this work, further experimental tests have been performed on
Blockhaus shear walls exhibited a dissipative capacity comparable simple specimens with the same type of contact surface between
with that of traditional timber-framed shear walls with brick ma- logs used in the tested walls to estimate a reasonable static friction
sonry infill (ν eq ≈ 0.15), recently tested under cyclic in-plane loads coefficient to be taken into account in a subsequent numerical mod-
by Vasconcelos et al. (2013). Good correlation of dissipative capac- eling phase. A setup similar to a double shear connector test was
ity was found also with results of full-scale cyclic experiments per- used [Fig. 11(a)]. The typical specimen (24 in total) consisted of
formed by Vogt et al. (2012) on modern light-frame timber shear three 0.60-m long timber logs subjected to a horizontal precom-
walls (ν eq ≈ 0.14). However, the flexibility of Blockhaus shear pression V L , introduced by means of four pretensioned steel bolts.
walls was larger than that of masonry filled timber-framed and During the tests, the effects of various parameters were taken
light-frame shear walls, which failed, respectively, at a maximum into account in the estimation of the average static friction coeffi-
lateral displacement of approximately Hw =45 and H w =31. cient μ. A compression load V L of 10 and 20 kN, respectively, was
separately applied to each specimen by balanced tightening of four
Φ16 steel rods. This compression load was then monitored during
Experimental Investigation on Blockhaus Components
the friction test by means of a load cell [Fig. 11(a)].
Based on the full-scale test results discussed in the section “Dis- At the same time, different geometries of timber logs were in-
cussion of Full-Scale Cyclic and Monotonic Test Results,” further vestigated. As result, according to Figs. 11(a) and 12 tests have

Fig. 10. Deformed shape of specimen S03 at the end of monotonic test (failure condition)

© ASCE 04014118-6 J. Struct. Eng.

J. Struct. Eng., 2015, 141(4): -1--1

Downloaded from ascelibrary.org by Politehnica University of Timisoara on 05/22/17. Copyright ASCE. For personal use only; all rights reserved.

Fig. 11. (a) Experimental setup for static frictional tests performed on Fig. 12. Compressive behavior of standard joints: (a) test setup
simple Blockhaus specimens (front view); (b) cross section of Schweiz (nominal standard joint dimensions); (b) failure mechanism
log; (c) detail of contact surfaces (Tirol specimen) (measures in mm)

a standard joint (l ¼ 0.06 m × w ¼ 0.06 m × h ¼ 0.076 m) typi-

cally used in Blockhaus structural systems (e.g., specimen S01)
been performed on Tirol logs (cross section A-A’ given in Fig. 1), in accordance with the setup displayed in Fig. 12(a). Based on
whereas 12 additional tests have been performed on 0.13 × 0.16-m the results of the full-scale shear wall tests previously discussed,
Schweiz timber logs, consisting of three 45-mm thick-lamellas the adopted setup should ideally reproduce the interaction between
of C24 spruces glued together [cross section A-A’ given in a main timber log under in-plane lateral loads and the orthogonal
Fig. 11(b)]. logs in contact with it. The small stiffening contribution provided
In each test, after the introduction of the compressive load V L , by the orthogonal walls, neglected in the used setup, was confirmed
the middle log was subjected to a vertical increasing force P by numerical investigations. Each test was performed by applying a
(Pmax;avg ¼ 10.56 kN and 20.45 kN for Tirol specimens, with compressive force to the 0.06 × 0.076 m2 -lateral resisting surface
V L ¼ 10 and 20 kN, respectively; Pmax;avg ¼ 13.13 and 26.94 kN of a single standard joint, thus leading the joint to collapse, and
for Schweiz specimens, with V L ¼ 10 and 20 kN), and the corre- monitoring the corresponding compression C-vertical displacement
sponding vertical slip was measured. As expected, friction tests d relationship.
resulted in an extremely variable static coefficient, being defined Because the observed failure mechanism resulted almost iden-
as a function of the imposed precompression and the cross-section tical in each specimen, only three tests were performed. Experi-
geometry of logs. The average static friction coefficient resulted, ments resulted in a high ultimate strength f u;90 of timber when
respectively, equal to μ ¼ 0.520 for Tirol timber logs (CoV ¼ compressed perpendicular to the grain and, with the exception
11.24%) and μ ¼ 0.665 for Schweiz timber logs (CoV ¼ 10.24%). of test 01, in a large ultimate displacement at collapse du (Table 2).
In particular, experimental measurements were affected not only
by the level of precompression V L or the geometry of logs, but also
by small contact defects between adjacent logs [Fig. 11(c)]. Table 2. Results of Compressive Tests on Standard Joints

Compressive Behavior of Standard Joints Ultimate Ultimate Ultimate

The shear strength of the studied walls, as discussed in the section Specimen load displacement strength
number Cu (kN) du (mm) f u;90 (N=mm2 )
“Specimen S01,” depends on the number and geometrical configu-
ration of joints [EN 1995-1-1:2009 (CEN 2009a)]. Because of 1 52.24 6.5 11.46
the in-plane lateral loads and the interaction between main logs 2 62.47 18 13.70
and orthogonal log walls, the joints used are primarily subjected 3 71.55 15 15.70
to compressive forces perpendicular to timber grains. Therefore, Mean 62.08 13.17 13.62
CoV 15.56 45.31 15.57
tests have been performed on small specimens representative of

© ASCE 04014118-7 J. Struct. Eng.

J. Struct. Eng., 2015, 141(4): -1--1

 Advanced Numerical Modeling of Blockhaus Shear density was assumed equal to ρ ¼ 420 kg=m3 , whereas the
Walls under In-Plane Lateral Loads corresponding mechanical properties were described by means of
ABAQUS/Standard engineering constants (E0;mean ¼ 11,000 MPa,
Geometry Description E90;mean ¼ 370 MPa, and Gmean ¼ 690 MPa). For steel, an iso-
tropic, linear elastic mechanical behavior was used {ρs ¼
To investigate in detail the typical lateral response of the tested full- 7,500 kg=m3 , Es ¼ 130 GPa, and ν s ¼ 0.32 [EN 1993-1-1:2004
scale Blockhaus shear walls, a numerical model was implemented (CEN 2004a)]}. The vertical precompression (10 kN=m) was
in the software package ABAQUS/Standard (Simulia 2012). described by means of a uniformly distributed, constant vertical
The lateral response of specimen S01 under in-plane monotonic pressure qv applied to the upper surface of the top log. Similarly,
loads was analyzed, and a finite element (FE)-model with the nomi- the horizontal in-plane lateral load was applied to the edge surface
nal geometry of the tested wall was taken into account. Careful con- of the top timber log as a uniformly distributed, quasi-static, linearly
sideration was given to the geometrical description of standard joints. increasing time-varying pressure qh . Finally, gravity loads were
Downloaded from ascelibrary.org by Politehnica University of Timisoara on 05/22/17. Copyright ASCE. For personal use only; all rights reserved.

Recent contributions of literature highlighted the validity of 3D automatically computed by ABAQUS/Standard for the whole model
solid models for the structural analysis of carpentry connections components.
(Pang et al. 2011; Tannert et al. 2010). In this work, to reduce An important role was assigned to the mechanical interaction
the number of mechanical interactions among multiple surfaces between timber surfaces in contact. Using a formulation available
in contact, thus the computational cost of simulations, the 0.01 × in ABAQUS/Standard library, surface-to-surface interactions were
0.01 m2 -notches and protrusions characterizing the top and bottom automatically detected between overlapping logs, along their entire
surfaces of timber logs were neglected. In accordance with Fig. 1, length, and between main logs and orthogonal logs based on the
each Tirol log was modeled, with the exception of its ends, as a nominal geometry of adopted joints. For all of them (430 contact
beam having a regular 0.09 × 0.16 m2 -cross section and consisting interactions, in total), appropriate mechanical behaviors were speci-
of three-dimensional elements. Based on preliminary sensitivity fied, to correctly reproduce possible relative slidings between two
studies, an extremely fine mesh composed of eight-node linear adjacent surfaces (tangential behavior) or interactions because of
brick (C3D8R) and six-node linear prism (C3D6R) 3D-stress solid loads perpendicular to surfaces in contact (normal behavior). In
elements was used. Combination of eight-node and six-node ele- the first case (tangential behavior), the main input parameters is
ments allowed the user to correctly describe the nominal geometry the static friction coefficient μ. Based on experimental results
of logs, both along their main body and near the joints and to ensure discussed in the section “Prediction of the Static Friction Coeffi-
accuracy of multiple interaction formulations (Fig. 13). cient,” μ was preliminarily set to 0.52. At the same time, careful
The vertical cables were modeled using truss elements (T32DH consideration was given to the calibration of a further nondimen-
type). Their geometry and position was described according to sional parameter (slip tolerance Ff ). Slip tolerance Ff , defined in
Fig. 4. The connection between each cable and the wall top log ABAQUS/Standard as a ratio of the surface mesh size lmesh , is
was then modeled by means of a rigid link, fixed at the top timber representative of possible elastic reversible tangential slidings γ f ¼
log (ux ¼ uy ¼ uz ¼ 0 and rx ¼ ry ¼ rz ¼ 0, ui and ri signify- Ff × lmesh between two surfaces in contact, before the occurrence
ing the displacement along and rotation around the axis i, with of irreversible sliding motion (namely, for shear stresses τ lower
i ¼ x; y; z), and hinged to the cable end (see detail of Fig. 4). than the critical shear stress τ crit ). If Ff → 0, no elastic reversible
At the same time, both the cables and the bottom timber log were motion occurs. In this work, the default value Ff ¼ 0.005 was pre-
restrained at the base (ux ¼ uy ¼ uz ¼ 0). The timber was defined liminarily used (γ f ≈ 0.5%lmesh ). Concerning the normal behavior,
as a linear elastic, orthotropic material. Based on values provided again a default mechanical interaction was used (hard contact pres-
by standards for C24 spruce [EN 338:2009 (CEN 2009b)], its sure-overclosure). As a result, when compressive contact pressures
p are applied, the interpenetration between two surfaces in contact
is avoided. At the same time, if p → 0, the two surfaces in contact
can separate during the analysis.

Preliminary Numerical Considerations

A static incremental, geometric nonlinear simulation was per-
formed on the preliminary FE-model (M01) described in the sec-
tion “Geometry Description.” Detailed investigation of numerical
results, although not discussed in this paper, highlighted that
M01 FE-model is not able to correctly reproduce the structural
behavior of specimen S01 under in-plane lateral loads (Bedon et al.
2013). In particular, no relative sliding occurred between overlap-
ping logs. The numerical wall, because of the applied lateral load,
behaved as a rigid body overturning around the base. In addition,
the use of default slip tolerance (Ff ¼ 0.005) resulted in a strong
underestimation (≈ 60%) of the experimental elastic stiffness for
the same specimen. In any case, sensitivity studies highlighted that
the static friction coefficient μ (set to 0.52, 0.511, respectively, from
experiments, and also 0.6, 0.5, and 0.4) does not significantly affect
numerical results, thus a final value μ ¼ 0.5 was used in further
parametric simulations. At the same time, the slip tolerance Ff
was set to 0.0005, so that possible reversible sliding between over-
Fig. 13. Mesh adopted in the FE analysis of specimen S01 (ABAQUS/
lapping logs could be avoided (γ f ≈ 0.05%lmesh ). This improve-
Standard)
ment resulted in increase of the shear wall rigidity, thus in an

© ASCE 04014118-8 J. Struct. Eng.

J. Struct. Eng., 2015, 141(4): -1--1

 overall good correspondence between experimental and numerical (Fig. 16, tension side). Good correlation is also found, for the same
elastic stiffnesses of the examined specimen. loading phase, in terms of lateral displacement at the top log
(Fig. 15). Nevertheless, in this latter case numerical results are more
Parametric Numerical Study and FE-Model sensitive to the amplitude of geometrical gaps.
Optimization Qualitative good correlation is also found, finally, for the global
behavior of the shear wall under in-plane lateral loads, as for
In the next modeling phase, some mechanical and geometrical example showed in Fig. 17.
parameters were varied to improve the quality of preliminary Although the geometrical parameters considered for M03 FE-
numerical predictions. model led to a better correspondence between experimental and
Geometrical Defects numerical load V-displacement δ curves, up to approximately
First, geometrical imperfections were introduced in ABAQUS 50 mm of lateral displacement, further extended series of simula-
tions have been performed to improve the accuracy of numerical
Downloaded from ascelibrary.org by Politehnica University of Timisoara on 05/22/17. Copyright ASCE. For personal use only; all rights reserved.

numerical model, in the form of small gaps representative of pos-

sible production tolerances or geometrical defects affecting the predictions.
standard joints. Nominal log dimensions were modified, according
to Fig. 14, by introducing a vertical gap dgap between each main log
and its orthogonal interceptions. The gap amplitude was assumed
equal to dgap ¼ 0.5 mm (M02), based on production tolerances,
and successively increased up to dgap ¼ 1 mm (M03), as suggested
by results of additional numerical analyses.
As shown in Figs. 15 and 16, the presence of small vertical gaps
provides close agreement between the numerical predictions and
the test results. This finding can be noticed especially for the first
loading phase (V max ≈ 20–25 kN), in terms of vertical uplift

Fig. 16. Monotonic test (S01) and ABAQUS/Standard results (M02);

load V-vertical uplift δV curves

Fig. 14. Introduction of a small gap between main timber logs and
perpendicular walls (ABAQUS/Standard, cross section); nominal di-
mensions in mm

Fig. 15. Monotonic test (S01) and ABAQUS/Standard results (M02); Fig. 17. (a) Photo; (b) ABAQUS/Standard prediction (M03) of the
load V-lateral displacement δ curve deformed shape of specimen S01 on the tension side

© ASCE 04014118-9 J. Struct. Eng.

J. Struct. Eng., 2015, 141(4): -1--1

 Timber Compressive Strength the yielding strength able to provide a good agreement between
In a second modeling phase, to take into account the occurrence experiments and model results is very close to the average compres-
of inelastic phenomena in joints, the mechanical behavior of spruce sive strength perpendicular to the grain estimated in the section
was described in M04 FE-model by means of an elastoplastic curve “Compressive Behavior of Standard Joints.”
(with dgap ¼ 1 mm, Ff ¼ 0.0005, μ ¼ 0.5). Conversely, if compared with standard recommendations for
Several values of yielding strength were taken into account C24 spruce [EN 338:2009 (CEN 2009b)], the characteristic value
when performing large parametric numerical studies (2.5 MPa ≤ of the compression strength (fc;90;k ¼ 2.5 MPa) approximately
f y ≤ 16 MPa), to highlight the possible effects of timber mechani- corresponds to a mean strength value fc;90;mean ≈ fc;90;k =0.7 ¼
cal behavior on the overall response of the examined shear wall. 3.57 MPa, which is significantly lower than the experimentally
The analyses, partly discussed in this paper, demonstrated that a measured value fy ¼ 13.62 MPa (section “Compressive Behavior
good correspondence between test results and numerical calcula- of Standard Joints”). This difference could be attributed to the
tions can be obtained if the yielding strength of spruce is assumed anisotropy of timber or to the interaction between multiple logs.
Downloaded from ascelibrary.org by Politehnica University of Timisoara on 05/22/17. Copyright ASCE. For personal use only; all rights reserved.

equal to fy ¼ 11.5 MPa, as shown in Figs. 18 and 19. In these fig- Possible defects and imperfections were in fact taken into account
ures, specifically, test predictions are compared with three numeri- in ABAQUS model in the form of a geometrical gap of constant
cal plots (fy ¼ 10, 11.5 and 13.62 MPa). The corresponding values amplitude dgap along the elevation of the shear wall, and this
of yielding strength, respectively, derive from numerical calibration assumption obviously represents an idealization.
(fy ¼ 10 MPa) and compressive experiments performed on stan- To assess the validity of numerical models, further comparisons
dard joints (section “Compressive Behavior of ‘Standard’ Joints,” have then been performed between test results obtained for a mon-
with f y ¼ 11.5 MPa the minimum experimental value and fy ¼ otonic test performed on a single standard joint [Fig. 20(a); Tomasi
13.62 MPa the average test result). A significant improvement et al. 2012] and the corresponding ABAQUS/Standard numerical
of the experimental-numerical comparison can be found especially model. Experiments have been carried out on single standard joints,
after the attainment of large lateral displacements (δ > 60 mm ≈ subjected to a vertical precompression (10 kN/m) and to a linear
H c =50, Fig. 18), in which major discrepancies between parametric increasing monotonic in-plane lateral load (Tomasi et al. 2012).
numerical results occur. In this context, it should be noticed that Again, mechanical and geometrical assumptions discussed for
the full-scale S01 specimen (M04), provide good correspondence
between a standard joint test and the corresponding ABAQUS
model [MST01, Fig. 20(b)], with f y ¼ 13.62 MPa (section “Com-
pressive Behavior of Standard Joints”).

Fig. 18. Monotonic test (S01) and ABAQUS/Standard results; load

V-lateral displacement δ curve

Fig. 19. Monotonic test (S01) and ABAQUS/Standard results; uplift Fig. 20. Monotonic test on single standard joint: (a) test setup;
δ V -time t curve; tension side (b) experimental-numerical comparison (ABAQUS)

© ASCE 04014118-10 J. Struct. Eng.

J. Struct. Eng., 2015, 141(4): -1--1

 Conclusions in-plane cyclic loads.” Proc., XV ANIDIS Conf. (CD-ROM), Padua,
Italy.
In the paper, experimental and numerical results have been pre- Branco, J., and Araújo, J. P. (2010). “Lateral resistance of log timber walls
sented for full-scale Blockhaus shear walls subjected to in-plane subjected to horizontal loads.” Proc., of WCTE 2010 (CD-ROM).
lateral loads. As shown, these structural systems are character- Branco, J., and Araújo, J. P. (2012). “Structural behaviour of log timber
ized by: walls under lateral in-plane loads.” Eng. Struct., 40, 371–382.
Ceccotti, A., Sandhaas, C., Okabe, M., Yasumura, M., Minowa, C., and
• High flexibility: cyclic tests performed on Blockhaus shear
Kawai, N. (2013). “SOFIE project—3D shaking table test on a
walls with three different types of joints were stopped at the seven-storey full-scale cross-laminated timber building.” Earthquake
attainment of a maximum lateral displacement of the top log Eng. Struct. Dyn., 42(13), 2003–2021.
equal to 80 mm (≈ H w =36), and no damage was observed in European Committee for Standardization (CEN). (2004a). “Eurocode
the deformed specimens. Monotonic tests performed on the 3—Design of steel structures—Part 1-1: General rules and rules for
same specimens confirmed their high flexibility by attainment buildings.” EN 1993-1-1:2004, Brussels, Belgium.
Downloaded from ascelibrary.org by Politehnica University of Timisoara on 05/22/17. Copyright ASCE. For personal use only; all rights reserved.

of maximum lateral displacements up to ≈ Hw =15. In this con- European Committee for Standardization (CEN). (2004b). “Eurocode
text, it is interesting to notice that this deformation limit largely 8—Design of structures for earthquake resistance—Part 1: General
exceeds the ultimate lateral displacement of light-frame shear rules, seismic actions and rules for buildings.” EN 1998-1:2004,
walls or masonry filled timber-framed walls under in-plane Brussels, Belgium.
European Committee for Standardization (CEN). (2006a). “Adhesives,
lateral loads.
phenolic and aminoplastic, for load-bearing timber structures—
• High strength: although directly dependent on the geometry of Classification and performance requirements.” EN 301:2006, Brussels,
the adopted carpentry joints, the tested walls generally showed Belgium.
high shear resistance to in-plane seismic loads. European Committee for Standardization (CEN). (2006b). “Timber struc-
• Large dissipative capacities: cyclic tests highlighted, for each tures—Test methods—Cyclic testing of joints made with mechanical
specimen, significant energy dissipation of Blockhaus shear fasteners.” EN 12512:2006, Brussels, Belgium.
walls (equivalent damping ratio in the range 10 to 20% depend- European Committee for Standardization (CEN). (2009a). “Eurocode
ing on the type of carpentry timber joint), which was similar, in 5—Design of timber structures—Part 1-1: General-common rules
this case, to that of ancient brick masonry filled timber-framed and rules for buildings.” EN 1995-1-1:2009, Brussels, Belgium.
walls and modern light-frame shear walls. European Committee for Standardization (CEN). (2009b). “Structural
timber-strength classes.” EN 338:2009, CEN European Committee
Because of the influence of several geometrical and mechanical
for Standardization, Brussels, Belgium.
parameters, predicting the seismic response of Blockhaus shear Gavric, I., Fragiacomo, M., Popovski, M., and Ceccotti, A. (2014). “Behav-
walls is a complex task. Despite that, as shown in this paper, iour of cross-laminated timber panels under cyclic loads.” Mater. Joints
the use of sophisticated numerical FE-models effectively calibrated Timber Struct., 9, 689–702.
can allow the user to reproduce the results of monotonic experimen- Hirai, T., Kimura, T., Yanaga, K., Sasaki, Y., and Koizumi, A. (2004).
tal tests with acceptable accuracy. Phenomena such as friction “Lateral resistance of log constructions.” Proc. of WCTE 2004
along the contact surfaces of the logs, geometrical imperfections (CD-ROM), Vol. 1.
in the form of small gaps in the carpentry joints, and timber plas- Pang, S. J., Ok, J. K., Park, J. S., Park, C. Y., and Lee, J. J. (2011).
ticization in compression were found to significantly affect wall “Moment-carrying capacity of dovetailed mortise and tenon joints with
behavior. Each phenomenon, consequently, had to be properly or without beam shoulder.” J. Struct. Eng., 10.1061/(ASCE)ST.1943-
541X.0000323, 785–789.
modeled to obtain good correspondence with experimental results.
Parida, G., Fragiacomo, M., and Johnsson, H. (2013). “Prefabricated timber
walls anchored with glued-in rod connections: Racking tests and pre-
liminary design.” Eur. J. Wood Wood Prod., 71(5), 635–646.
Acknowledgments Simulia ABAQUS v.6.12 [Computer software]. Providence, RI, Dassault
Systems.
Rubner Haus AG SpA is gratefully acknowledged for funding Tannert, T., Lam, F., and Vallée, T. (2010). “Strength prediction for rounded
the research project. Dr. Annalisa Battisti is also acknowledged dovetail connections considering size effects.” J. Eng. Mech., 10.1061/
for her cooperation during the experimental and modeling phase. (ASCE)0733-9399(2010)136:3(358), 358–366.
Dr. Andrea Polastri [CNR IVALSA Trees and Timber Institute, Tomasi, R., Sartori, T., Grossi, P., and Wenzel, L. (2012). “Sistema cost-
San Michele all’Adige (Italy)] is acknowledged for the technical ruttivo Blockhaus: prove su giunzioni d’angolo.” Internal Rep., Univ. of
support provided during the full-scale experiments performed on Trento, Italy (in italian).
Blockhaus shear walls. Dr. Franco Trevisan [University of Trieste, Van de Lindt, J. W. (2004). “Evolution of wood shear wall testing, model-
ing and reliability analysis: Bibliography.” Pract. Period. Struct. Des.
DICAR—Department of Engineering and Architecture, Trieste
Constr., 10.1061/(ASCE)1084-0680(2004)9:1(44), 44–53.
(Italy)] is finally acknowledged for his technical support for Van de Lindt, J. W., Pei, S., and Pryor, S. E. (2011). “Construction and
the friction and compressive experiments performed on small experimental seismic performance of a full-scale six-story light-frame
Blockhaus components. wood building.” Procedia Eng., 14, 1599–1605.
Vasconcelos, G., Poletti, E., Salavessa, E., Jesus, A. M. P., Lourenço, P. B.,
Pilaon, P. (2013). “In-plane shear behaviour of traditional timber walls.”
References Eng. Struct., 56, 1028–1048.
Vogt, T., Hummel, J., and Seim, W. (2012). “Timber framed wall elements
Bedon, C., Fragiacomo, M., Amadio, C., and Sadoch, C. (2013). “Exper- under cyclic loading.” Proc. WCTE-World Conf. on Timber Engineer-
imental and numerical investigation of ’Blockhaus’ shear walls under ing (CD-ROM), Auckland.

© ASCE 04014118-11 J. Struct. Eng.

J. Struct. Eng., 2015, 141(4): -1--1