materials

Article
Mechanical Behavior of Dowel-Type Joints Made
of Wood Scrimber Composite
Minjuan He 1 , Duo Tao 1 , Zheng Li 1, * and Maolin Li 2
1 Department of Structural Engineering, Tongji University, Shanghai 200092, China;
hemj@tongji.edu.cn (M.H.); 91taoduo@tongji.edu.cn (D.T.)
2 Tianjin Hualin (Group) Co., Ltd., 88 Nanhuan Road, Tianjin 300350, China; benilitj@gmail.com
* Correspondence: zhengli@tongji.edu.cn; Tel.: +86-21-6598-1582

Academic Editor: George Papanicolaou

Received: 3 June 2016; Accepted: 12 July 2016; Published: 15 July 2016

Abstract: As a renewable building material with low embodied energy characteristics, wood has
gained more and more attention in the green and sustainable building industry. In terms of material
resource and physical properties, scrimber composite not only makes full use of fast-growing wood
species, but also has better mechanical performance and less inherent variability than natural wood
material. In this study, the mechanical behavior of bolted beam-to-column joints built with a kind of
scrimber composite was investigated both experimentally and numerically. Two groups of specimens
were tested under monotonic and low frequency cyclic loading protocols. The experimental results
showed that the bolted joints built with scrimber composite performed well in initial stiffness,
ductility, and energy dissipation. A three-dimensional (3D) non-linear finite element model (FEM)
for the bolted beam-to-column joints was then developed and validated by experimental results.
The validated model was further used to investigate the failure mechanism of the bolted joints through
stress analysis. This study can contribute to the application of the proposed scrimber composite in
structural engineering, and the developed FEM can serve as a useful tool to evaluate the mechanical
behavior of such bolted beam-to-column joints with different configurations in future research.

Keywords: scrimber composite; bolted joints; mechanical behavior; finite element model;
failure mechanism

1. Introduction
With low embodied energy and carbon storage characteristics, wood has gained more and
more attention in the green and sustainable building industry in recent years. However, wood
normally has larger variabilities in mechanical properties compared to concrete or steel. Therefore,
different kinds of engineered wood-based composite materials have been developed. Compared to
solid sawn lumber, the variability in mechanical properties of wood-based composite is significantly
reduced since natural defects in wood are dispersed. The decrease in non-uniformity leads to efficient
utilization of fiber resources, and better mechanical performance can be obtained from engineered
wood-based composites.
In the past two decades, researches on the structural application of wood-based composites have
been conducted. The performance of glulam timber columns reinforced by FRP (Fiber Reinforced
Plastic) sheets were investigated by Taheri et al. [1], and experimental results and computational
modelling showed that the stiffness and strength could be significantly improved with FRP that
were externally bonded. O’Loinsigh et al. [2] conducted experimental and numerical investigations
on multi-layered wooden beams with welded-through wood dowels, showing that a desirable
bending stiffness is attainable with a reasonable combination of material and geometric parameters.
Quiroga et al. [3] investigated the feasibility of rapid housing construction with wood-cement

Materials 2016, 9, 581; doi:10.3390/ma9070581 www.mdpi.com/journal/materials

Materials 2016, 9, 581 2 of 16
Materials 2016, 9, 581 2 of 16

et al. [3] investigated the feasibility of rapid housing construction with wood-cement composites. It
was shown that
composites. It wassuch
showncomposites
that suchare able to compete
composites are ablewith existing
to compete building
with materials
existing buildinginmaterials
structural
in
durability and production
structural durability cost. Incost.
and production addition, a large
In addition, amount
a large amount of researches
of researchesononthe the mechanical
mechanical
properties
properties ofofinnovative
innovative wood-based
wood-based composite
composite havehave
been been conducted
conducted [4–14].[4–14]. These researches
These researches showed
showed
that wood-based composites could meet both structural and non-structural demands demands
that wood-based composites could meet both structural and non-structural with
with desirable
desirable
mechanical mechanical
performance performance
and betterand usebetter
of wooduseresources.
of wood resources.
The
The authors have proposed a kind of veneer-based scrimber
authors have proposed a kind of veneer-based scrimber composite
composite made
made fromfrom fast-growing
fast-growing
wood. The desirable appearance makes such scrimber composite
wood. The desirable appearance makes such scrimber composite a potential material a potential material choicechoice
for post-
for
and-beam timber constructions. The mechanical behavior of beam-to-column
post-and-beam timber constructions. The mechanical behavior of beam-to-column joints has a joints has a significant
influence
significanton the seismic
influence on theperformance of a post-and-beam
seismic performance structural
of a post-and-beam system.
structural In thisInstudy,
system. the
this study,
mechanical
the mechanical behavior
behaviorandand
failure mode
failure mode of of
bolted
boltedbeam-to-column
beam-to-columnjoints jointsbuilt
builtwith
with the proposed
the proposed
scrimber composite were investigated by monotonic and low frequency cyclic
scrimber composite were investigated by monotonic and low frequency cyclic tests. A 3D non-linear tests. A 3D non-linear
finite
finite element
element model
model (FEM)
(FEM) was then developed
was then developed and and validated
validated by by experimental
experimental results,
results, with
with which
which
the mechanical behavior
the mechanical behavior andand failure
failure mechanism
mechanism of of the bolted beam-to-column
the bolted beam-to-column joints joints waswas further
further
investigated.
investigated. Moreover,
Moreover, to to evaluate
evaluate thethe feasibility
feasibility of
of using
using thethe scrimber
scrimber composite
composite in in post-and-beam
post-and-beam
buildings,
buildings, the rotational performance of the scrimber joints was compared with that of
the rotational performance of the scrimber joints was compared with that of glulam
glulam joints,
joints,
since glulam is considered as the most commonly used material for existing post-and-beam
since glulam is considered as the most commonly used material for existing post-and-beam buildings. buildings.
The presented experimental
The presented experimental andand numerical
numerical analyses
analyses can
can serve
serve asas aa fundamental
fundamental basis basis for
for promoting
promoting
more application of
more application of such
such scrimber
scrimber composite
composite material
material inin modern
modern timber
timber structures.
structures.

2. Rotational
RotationalPerformance
Performanceof
ofScrimber
ScrimberBeam-to-Column
Beam-to-ColumnJoints
Joints
The production
production process
processand andtypical
typicalcross
crosssection
sectionofofthe
theproposed
proposed scrimber
scrimber composite
composite is shown
is shown in
in Figure
Figure 1. The
1. The logslogs are firstly
are firstly slicedsliced into veneer
into veneer lumber lumber
with a with a thickness
thickness of only 2ofmm.
onlyThen
2 mm.the Then the
adhesive
adhesive impregnation
impregnation process isprocess
conducted is conducted on the
on the dried dried
veneer veneerafter
lumber, lumber,
which after
thewhich
veneers theare
veneers are
squeezed
squeezed
and molded andunder
moldedhighunder high temperature
temperature and pressure. andThepressure. The cross
cross section of thesection
scrimberof the scrimber
composite is
composite is characterized by a curved veneer, and this is quite different from
characterized by a curved veneer, and this is quite different from the existing wood-based composites the existing wood-
based
that arecomposites
engineered. that
Thearemechanical
engineered. The mechanical
properties of such properties of such scrimber
scrimber composite have been composite have
investigated
been
by Li investigated by Li
et al. [15]. With et al. [15].strength
a desirable With a and
desirable
modulusstrength and modulus
of elasticity (MOE),of it elasticity (MOE), that
is recommended it is
recommended that the scrimber composite can be used for structural beams
the scrimber composite can be used for structural beams and columns for a post-and-beam timber and columns for a post-
and-beam timber
system. In this system.
study, In this study,
both monotonic and low both monotonic
frequency cyclicand
testslow
werefrequency
conductedcyclic tests were
to investigate the
conducted to investigateofthe
rotational performance rotational
scrimber performance joints.
beam-to-column of scrimber beam-to-column joints.

Figure
Figure 1.
1. Production
Production and
and cross
cross section
section of
of the
the scrimber
scrimber composite.
composite.

2.1. Experimental Description

2.1.1. Specimens

The configuration
configurationofofthe
thespecimen
specimen is is
shown in Figure
shown 2. The
in Figure beam
2. The member
beam was 178
member wasmm 178 ˆ
mm150× mm
150
in cross
mm section
in cross and 900
section andmm
900 in
mmlength, and the
in length, andcolumn member
the column had the
member hadsame cross section
the same as theas
cross section beam
the
member, but it was 1100 mm in length. A slot with a thickness of 11 mm was cut for the 10 mm-thick
Materials 2016,9,
Materials 2016,
2016, 9,581
9, 581
581 33 of
of 16
16

beammember,
beam member,but butititwas
was1100
1100mm mmin inlength.
length.A Aslot
slotwith
withaathickness
thicknessof of11
11mm
mmwas wascutcutfor
forthe
the1010mm-
mm-
steel
thick connection
thicksteel
steelconnection plate,
connectionplate, and
plate,andbolt holes
andbolt
boltholeswith
holeswith diameter
withdiameter of
diameterof 18
of18 mm
18mm were
mmwere pre-drilled
werepre-drilled
pre-drilledforfor the
forthe M16
theM16 bolts
M16bolts
bolts
(Ningbo
(Ningbo Xinderui
(Ningbo XinderuiStandard
Xinderui StandardParts
Standard Co.,
Parts
Parts Ltd.,
Co.,
Co., Ningbo,
Ltd.,
Ltd., Ningbo,
Ningbo,China) (i.e.,(i.e.,
China)
China) boltsbolts
(i.e., with with
bolts diameter of 16 mm).
with diameter
diameter 16Q235B
of 16
of mm).
mm).
steel
Q235B
Q235B plate
steel(manufactured
steel plate by Qingdao
plate (manufactured
(manufactured by East Steel
by Qingdao
Qingdao Tower
East
East Stock
Steel
Steel Tower
Tower Co.,Stock
Ltd., Co.,
Stock Qindao,
Ltd.,China)
Co., Ltd., Qindao,
Qindao, with a nominal
China)
China) with
with
yielding
aanominal strength
nominalyielding of 235
yieldingstrength MPa,
strengthof as
of235 specified
235MPa,
MPa,as in the
asspecified Chinese
specifiedininthe code
theChinese of
Chinesecode notations
codeof for
ofnotations designations
notationsfor of
fordesignationsiron
designations
and
of steel
of iron
iron and [16],
and steel was used
steel [16],
[16], wasfor
was used the connection
used for
for the plate.
the connection Bolts
connection plate. with the
plate. Bolts
Bolts with grade
with the of 6.8,
the grade conforming
grade of of 6.8, to Chinese
6.8, conforming
conforming to to
Standard
Chinese GB/T
Standard 1231-2006
GB/T [17]
1231-2006were used
[17] as
were fasteners
used as for the
fasteners specimens.
for the The nominal
specimens.
Chinese Standard GB/T 1231-2006 [17] were used as fasteners for the specimens. The nominal yielding The yielding
nominal strength
yielding
of the bolts
strength
strength of wasbolts
of the
the 480 MPa.
bolts was 480
was A total
480 MPa.ofA
MPa. A10total
specimens
total of 10
of were prepared
10 specimens
specimens for the tests.
were prepared
were prepared One
for the
for group
the tests.
tests. One
Onewith five
group
group
duplicates
with five
with was testedwas
five duplicates
duplicates under
was testedmonotonic
tested loading, and
under monotonic
under monotonic the other
loading,
loading, andgroup
and withgroup
the other
the other five duplicates
group with five
with was tested
five duplicates
duplicates
under
was low
was tested frequency
tested under
under low cyclic
lowfrequency loading.
frequency cyclic
cyclicloading.
loading.

Figure2.
Figure
Figure 2.Specimen
2. Specimen configuration(all
Specimenconfiguration
configuration (alldimensions
(all dimensionsare
dimensions arein
are inmm).
in mm).
mm).

2.1.2. Test
2.1.2.
2.1.2. TestDesign
Test Designand
Design andData
Data Measurement
Measurement
Thespecimens
The
The specimenswerewererotated
rotated90
rotated 90degrees
90 degreesfor
degrees forease
for easeofof
ease of loading
loading
loading in
inin the
the
the laboratory.
laboratory.
laboratory. As
AsAs shown
shown
shown inFigure
in in Figure
Figure 3a,
3a, the
3a,
the the column
columncolumn was horizontally
was was horizontally
horizontally fixed
fixedfixed
on theonon the ground
the
ground ground
by anchorby anchor
by anchorand
bolts, bolts,
bolts, and the
the and
beam the
wasbeam
beam
placed was
was placed
placed
vertically
vertically
vertically
with with
with
its top its
endits topend
top endconnected
connected connected tothe
to theelectro-hydraulic
electro-hydraulic
to the electro-hydraulic servoactuator.
servo
servo actuator. actuator.
Figure Figure
3bFigure
shows3b3b shows
theshows thetest
the
test set-up test
of
set-up
set-up of the specimen
of the specimen
the specimen in the laboratory.
in the laboratory.
in the laboratory. The actuator
The actuator
The actuator has a maximal
has a maximal
has a maximal loading capacity
loading capacity
loading capacity of 300 kN of 300
ofand
300 kNkN and
and
a stroke
a stroke
stroke
arange range
ofrange
˘250 ofof±250
mm.±250 mm.
mm.

Figure3.
Figure 3. Testdesign:
design: (a)test
test layout(all
(all dimensionsare
are inmm);
mm); (b)test
test set-upin
in thelaboratory.
laboratory.
Figure 3.Test
Test design:(a)
(a) testlayout
layout (alldimensions
dimensions are in
in mm); (b)
(b) test set-up
set-up in the
the laboratory.

The rotationof
The of thejoint
joint wasmeasured
measured bysevenseven linearvoltage
voltage displacementtransducers
transducers (LVDTs)
The rotation
rotation of the
the jointwas
was measured by by sevenlinear
linear voltagedisplacement
displacement transducers (LVDTs)
(LVDTs)
(manufactured by
(manufactured by Liyang
Liyang City
City Instrument
Instrument andand Meter
Meter Plant,
Plant, Liyang,
Liyang, China),
China), which
which were
were distributed
(manufactured by Liyang City Instrument and Meter Plant, Liyang, China), which were distributed distributed
along the height
along height of specimen
specimen as illustrated
illustrated in Figure
Figure 3a. LVDT
LVDT 1 was usedused to record
record the horizontal
horizontal
along the
the height ofof specimen asas illustrated inin Figure 3a.
3a. LVDT 11 was
was used to
to record thethe horizontal
displacement of the
displacement the beam’s free
free end. LVDT
LVDT 2, LVDTLVDT 3, and
and LVDT 44 were
were used to to measure the
the beam
displacementof of thebeam’s
beam’s freeend.
end. LVDT2, 2, LVDT 3,3, and LVDT
LVDT 4 were usedused to measure
measure the beam
beam
rotation.LVDT
rotation. LVDT22and andLVDT
LVDT33were
wereinstalled
installedon
onboth
bothsides
sidesof
ofthe
thesteel
steelconnection
connectionplate
platesymmetrically.
symmetrically.
rotation. LVDT 2 and LVDT 3 were installed on both sides of the steel connection plate symmetrically.
LVDT 5 andLVDT
LVDT LVDT 6 wereconnected
connected tothethe steelconnection
connection platewith
with a distanceof of 50mm
mm tomeasure
measure
LVDT 55 and
and LVDT 66 were
were connected to to the steel
steel connection plate
plate with aa distance
distance of 5050 mm to
to measure
Materials 2016, 9, 581 4 of 16
Materials 2016, 9, 581 4 of 16
the rotation of the steel connection plate. LVDT 7 was installed at the end of the column to record the
horizontal rigid body movement of the entire joint.
the rotation of the the
Considering steel connection
centroid plate.
of the groupLVDT 7 was
of bolts installed
as the at the
rotation endofofthe
center thejoint,
columnthe to recordand
moment the
horizontal rigid body movement of the entire joint.
corresponding rotation of the joint can be calculated by Equations (1)–(4).
Considering the centroid of the group of bolts as the rotation center of the joint, the moment and
M  FH
corresponding rotation of the joint can be calculated by Equations (1)–(4).
(1)

   BS M
 “BCFˆ SC
H (2)
(1)

θ“θ “θ ´θ (2)
 „   S 2 +SBC
S 4 BS 3  /2 
SC
180
 BC   arctan ´ pS2 `S3 q/2
S4 120 

180 (3)
θ BC “ arctan  ˆ (3)
120 π
  S180
S SS56 ´
„ 
5 6 180
 SC
θSC= “
arctan
arctan   ˆ (4)
(4)
 50 50  π

where FF is
where is the
the lateral
lateral force
force applied
applied onon the
the specimen
specimen byby the
the actuator,
actuator, andand H H is
is the
the vertical
vertical distance
distance
between
between the rotation center and the loading point. As shown in Figure 4, θ is the rotation of theof
the rotation center and the loading point. As shown in Figure 4, θ is the rotation the
beam
beam relative to the rotation center, θ is the rotation of the beam relative to the steel
relative to the rotation center, θBS is theBSrotation of the beam relative to the steel connection plate, θBC connection
plate, θ BC is the
is the rotation of rotation
the beamofrelative
the beam
to therelative to the
column, andcolumn,
θSC is theand θ SC isofthe
rotation therotation
columnof the column
relative to the
relative to the steel connection plate. S2, S3, S4, S5, and S6 are the displacements
steel connection plate. S2, S3, S4, S5, and S6 are the displacements measured by LVDT 2–LVDT 6, measured by LVDT
2–LVDT 6, respectively.
respectively.

Figure 4.
Figure 4. Calculation
Calculation of
of relative
relative rotation.
rotation.

2.1.3. Loading
2.1.3. Loading Protocol
Protocol
Displacement-control protocol
Displacement-control protocol was
was adopted
adopted for for both
both monotonic
monotonic and and lowlow frequency
frequency cyclic
cyclic tests.
tests.
In accordance with American standard ASTM D1761-12 [18], the monotonic
In accordance with American standard ASTM D1761-12 [18], the monotonic loading was applied loading was applied atata
a constant rate of 5 mm/min until the failure of specimen occurred (i.e., the loading
constant rate of 5 mm/min until the failure of specimen occurred (i.e., the loading dropped by more dropped by more
than 20%
than 20% ofof the
the peak
peak value).
value). For
For the
the low
low frequency
frequency cyclic
cyclic tests,
tests, the
the Consortium
Consortium of of Universities
Universities for for
Research in Earthquake Engineering (CUREE) protocol was adopted according
Research in Earthquake Engineering (CUREE) protocol was adopted according to American standard to American standard
ASTM E2126-11
ASTM E2126-11[19].
[19].As
Asshown
shownininFigure
Figure5,5,the
thereference
reference displacement
displacement of of
thethe CUREE
CUREE protocol
protocol (i.e.,
(i.e., ∆)
Δ) was determined as 60% of the maximal displacement obtained by
was determined as 60% of the maximal displacement obtained by monotonic tests. The CUREE monotonic tests. The CUREE
protocolincludes
protocol includesinitiation
initiationcycles,
cycles, primary
primary cycles,
cycles, andand trailing
trailing cycles.
cycles. The The loading
loading protocol
protocol begins begins
with
with six initiation cycles with an amplitude of 0.05Δ, followed by the first primary
six initiation cycles with an amplitude of 0.05∆, followed by the first primary cycle with an amplitude cycle with an
amplitude of 0.075Δ, and then a rise up to the final primary cycle with an amplitude
of 0.075∆, and then a rise up to the final primary cycle with an amplitude of 2.0∆. Each primary cycle of 2.0Δ. Each
primary
is followedcycle
by is followed
several by several
trailing trailing
cycles with 75%cycles
of thewith 75% of of
amplitude thethe
amplitude
primary of the primary cycle.
cycle.
Materials 2016, 9, 581 5 of 16
Materials 2016, 9, 581 5 of 16
Materials 2016, 9, 581 5 of 16

Figure 5. Consortium of Universities for Research in Earthquake Engineering (CUREE) protocol for
Figure
the low5.frequency
Consortium of tests.
cyclic
Consortium Universities
Universities for Research in Earthquake Engineering
Engineering (CUREE)
(CUREE) protocol
protocol for
the low frequency cyclic tests.
2.2. Failure Modes
2.2. Failure
Failure Modes
In the beginning stage of loading, the rotation of joints was almost free due to the initial clearance
In
between the
the beginning
beginning
bolts and wood stage of loading,
loading, the
components. Soonrotation
a stop of
rotation of joints was almost
almost
the potential free
free due
rotation due to
was to the
the initial
initial
observed clearance
clearance
with partial
between bolts and
contact between wood
beam components.
and column on Soon a stop
stop of
the compressionof the potential
theside.
potential rotation
rotation waswas observed
The moment-resisting observed with
withofpartial
capacity partial
joints
contact between beam and column on the compression side. The moment-resisting
was mainly dependent on the extrusion between beam and column before the clearance was capacity of joints
was mainlydependent
mainly
eliminated. dependent
With ononthe the
the increase ofextrusion
extrusion between
between
moment, thebeam beam
localand and failure
column
crushing column
before the before the
clearance
appeared clearance
was
with was
eliminated.
slight crack
eliminated.
With
development With
the increase ofthe
on the increase
moment,
beam theoflocal
as shown moment,
Figurethe
incrushing local
6,failure
and crushing
appeared
then failure
with appeared
slight
the moment-resisting with of
crackcapacity slight
development joints crack
on the
was
development
beam on the beam as shown in Figure 6, and then the moment-resisting
mainly dependent on the interaction between steel components and wood components. It on
as shown in Figure 6, and then the moment-resisting capacity of joints was capacity
mainly of joints
dependent wasis
mainly
the
suggested dependent
interaction between
the progresson the
steel
of theinteraction
components
rotation ofbetween
and
jointswoodsteel
could components
components.
be It isand
influenced by wood
local components.
suggested
the the progress
crushing ofItthe
between is
suggested
elements.of the
rotation progress
joints could beofinfluenced
the rotation byof
thejoints
localcould be influenced
crushing by the local crushing between
between elements.
elements.

Figure 6.
Figure 6. Local
Local crushing between beam
crushing between beam and
and column.
column.
Figure 6. Local crushing between beam and column.
The failure modes of the bolted scrimber joints from monotonic tests are shown in Figure 7a,b.
The failure modes of the bolted scrimber joints from monotonic tests are shown in Figure 7a,b.
WithThethe failure
increasemodes of the bolted
of rotations, scrimber
splitting alwaysjoints fromtowards
appeared monotonicthe tests
bottomareof
shown
beam inonFigure 7a,b.
the tension
With the increase of rotations, splitting always appeared towards the bottom of beam on the
With the increase of rotations, splitting always appeared towards the bottom of beam
side and then developed upwards. At larger rotations, the joint eventually failed with the appearance on the tension
tension side and then developed upwards. At larger rotations, the joint eventually failed with the
side and then developed
of run-through crack alongupwards.
the beam,At larger rotations,
and severe plug the joint
shear eventuallythe
underneath failed
boltswith
wasthe appearance
also observed.
appearance of run-through crack along the beam, and severe plug shear underneath the bolts was also
of run-through
After crackand
disassembly along the beam, and
examination severe
of the plugspecimens,
failed shear underneath the boltswas
bolt yielding wasobserved
also observed.
with
observed. After disassembly and examination of the failed specimens, bolt yielding was observed with
After disassembly and examination of the failed specimens, bolt yielding was
embedment deformation of bolt holes. As shown in Figure 7c,d, the failure modes from cyclic tests observed with
embedment deformation of bolt holes. As shown in Figure 7c,d, the failure modes from cyclic tests
embedment
were similardeformation
to, but moreof bolt holes.
severe As shown
than, those from in Figure 7c,d,
monotonic theThis
tests. failure
wasmodes
due tofrom cyclic
the fact thattests
the
were similar to, but more severe than, those from monotonic tests. This was due to the fact that the
were
cyclicsimilar
loadingto,
in but moreand
positive severe than,directions
negative those from monotonic
added tests. This
to the damage was due to of
accumulation thethe
fact that the
specimens,
cyclic loading in positive and negative directions added to the damage accumulation of the specimens,
cyclic
and theloading in positive
maximal and negative
displacement of the directions added
actuator from to the
cyclic damage
tests accumulation
(i.e., 2.0Δ) was largerofthan
the specimens,
that from
and the maximal displacement of the actuator from cyclic tests (i.e., 2.0∆) was larger than that from
and the maximal
monotonic displacement
tests (i.e., of the actuator from cyclic tests (i.e., 2.0Δ) was larger than that from
about 1.7Δ).
monotonic tests (i.e., about 1.7∆).
monotonic tests (i.e., about 1.7Δ).
Materials 2016, 9, 581 6 of 16
Materials
Materials2016,
2016,9,9,581
581 66of
of16
16

Figure
Figure7.
Figure 7.Failure
7. Failuremodes:
Failure modes:(a)
modes: (a)external
(a) externalobservation
external observationfrom
observation frommonotonic
from monotonictest;
monotonic test;(b)
test; (b)internal
(b) internalobservation
internal observationfrom
observation from
from
monotonic
monotonic test; (c)
test; (c)
monotonic test; external
(c) external observation
external observation from
observation from cyclic
from cyclic test;
cyclic test; (d)
test; (d) internal
(d) internal observation
internal observation from
observation from cyclic
from cyclic test.
cyclic test.
test.

2.3.
2.3.Results
ResultsofofMonotonic
MonotonicTests
Tests
Figure
Figure88shows showsthe themoment-rotation
moment-rotation(M-θ) (M-θ)curvescurves fromfrom monotonic
monotonic tests.tests. TheThefirst
firstbranch
branchof ofthe
the
curve isisclose
curve is close
closetoto
to the
thethe xx axis,
axis,
x axis, indicating
indicating
indicating low
low moment-resisting
moment-resisting
low moment-resisting capacity capacity
capacity
of joints of joints
ofcaused
jointsbycaused
initialby
caused by initial
initial
clearance
clearance
clearance
between bolts between
betweenand bolts
bolts and
wood and wood
wood components.
components. When the When
components. When the
rotation of rotation
the rotation
beam of
of beam
relative beam relative to
to column
relativeincreased
to column column to
increased
almost 1.5to
increased to almost
almostthe
degrees, 1.5 degrees,
1.5 stiffness
degrees,of the
the stiffness
stiffness
joints of
of joints
significantly significantly
jointsincreased
significantlywith increased
sufficientwith
increased with sufficient
sufficient
contact between contact
contact
bolts
between
between
and wood bolts
bolts and
andwood
components. woodItcomponents.
is noted thatItItthe
components. isisnoted
noted
five M-θ that
that the
thefive
curves areM-θ
five M-θ curves
curves
similar witharea similar
are similarwith
variation with aavariation
variation
coefficient less
coefficient
coefficient
than 10%, thus less than
less the
than 10%, thus
10%, can
results the
thusbethe results
results can
represented can be
bybe represented
therepresented
average curve by the
by the average
average
as shown curve
curve
in the as shown
as shown
figure. in
in the
The average the
figure.
curve isThe
figure. The average
almostaverage
linearcurve
withisis
curve almost
jointslinear
almost
the in thewith
linear with
elastic the
the joints
jointsin
deformation inthe
the elastic
elastic
stage. deformation
Due to the loadstage.
deformation stage. Due
Dueto
distribution the
toafter
the
load
load distribution
distribution after
after plug
plug shear
shear failure,
failure, aa short
short period
period ofof volatility
volatility
plug shear failure, a short period of volatility is observed with unstable moment-resisting behavior. is
is observed
observed with
with unstable
unstable
moment-resisting
moment-resisting
Afterwards, therebehavior.
behavior.
is a yieldingAfterwards,
plateau inthere
Afterwards, there isisaayielding
the curve yielding plateau
plateauof
with yielding in the
thecurve
inbolts,curve
showingwith
withyielding
yielding
that of
ofbolts,
bolts,
the scrimber
showing
showing that
thatthe
joints perform the
wellscrimber
scrimber joints
jointsperform
in ductility. performwell wellin inductility.
ductility.

Figure
Figure 8. Moment-rotation (M-θ) curves and the average M-θ curve
curveof
ofmonotonic
monotonictests.
Figure 8.
8. Moment-rotation
Moment-rotation (M-θ)
(M-θ) curves
curves and
and the
the average
average M-θ
M-θ curve of monotonic tests.
tests.

Glulam
Glulam isis
Glulam is normally
normally used
normally used in
used in timber
in timber constructions.
timber constructions. In
constructions. Inorder
In orderto
order to evaluate
to evaluate the
evaluate the feasibility
the feasibility of
feasibility of using
of using
using
scrimber
scrimber joints
scrimber joints in
joints in structural engineering,
structural engineering,
in structural comparison
engineering, comparison between
comparison between
between the the scrimber
the scrimber joints
scrimber joints and
joints and glulam joints
glulam joints
and glulam joints
was
was also
also performed
performed in
in this
this study.
study. Based
Based on
on the
the same
same design
design target,
target, including
including
was also performed in this study. Based on the same design target, including stiffness and moment stiffness
stiffness and
and moment
moment
resisting
resisting capacity,
resisting capacity, the
capacity, the dimension
the dimension of
dimension of scrimber
of scrimber joints
scrimber joints (i.e.,
joints (i.e., the
(i.e., the cross
the cross sections
cross sections of
sections of scrimber
of scrimber beam
scrimber beam and
beam and
and
column
column were
column wereboth
were both178
both 178mm
178 mm׈
mm ×150150mm)
150 mm)was
mm) waseventually
was eventuallysmaller
eventually smallerthan
smaller thanthat
than thatof
that ofglulam
of glulamjoints
glulam joints(i.e.,
joints (i.e.,the
(i.e., thecross
the cross
cross
sections
sections of
of glulam
glulam beam
beam and
and column
column were
were 300
300 mm
mm ×
× 200
200 mmmm and
and
sections of glulam beam and column were 300 mm ˆ 200 mm and 300 mm ˆ 250 mm, respectively). 300
300 mmmm ×× 250
250 mm,
mm, respectively).
respectively).
The
Theexperimental
The experimentalresults
experimental resultsof
results ofthe
of theglulam
the glulam
glulam joints
joints
joints were
were
werereported
reported
reported by
byby He and
HeHe and Liu
and [20].
LiuLiu
[20]. For
Forcomparison,
[20]. comparison,
For comparison, the
the
CEN
CEN (European
the CEN(European Committee
Committee
(European Committeefor
forStandardisation)
Standardisation)
for Standardisation) method
method was
wasadopted
method adopted to
tocalculate
was adopted calculate the
theinitial
initial
to calculate thestiffness
stiffness
initial
and
andductility
ductilityratio
ratio[21].
[21].TheTheinitial
initialstiffness
stiffnessisisdefined
definedby bythetheslope
slopeof ofaasecant
secantlinelinefrom
from10%10%to to40%
40%of of
the peak moment. The yield point is determined as the intersection of
the peak moment. The yield point is determined as the intersection of the secant line and a tangentthe secant line and a tangent
Materials 2016, 9, 581 7 of 16

stiffness and ductility ratio [21]. The initial stiffness is defined by the slope of a secant line from
10% to 40% of the peak moment. The yield point is determined as the intersection of the secant line
Materials 2016, 9, 581 7 of 16
and a tangent line with a slope equal to one sixth of the slope of the secant line. The ductility ratio
is line
determined
with a slope as the ratio
equal of the
to one sixthmaximal rotation
of the slope of thetosecant
the yield
line.rotation corresponding
The ductility to the yield
ratio is determined as
point. These key points, used to calculate various performance parameters
the ratio of the maximal rotation to the yield rotation corresponding to the yield point. These of joints, are not always
key
located
points,inused
the to
first branchvarious
calculate of the curve, so the adverse
performance parameters impact on analysis
of joints, results caused
are not always located by erection
in the first
error and initial clearance is effectively eliminated. The comparisons of initial
branch of the curve, so the adverse impact on analysis results caused by erection error and initial stiffness, ultimate
moment,
clearance failure rotation,eliminated.
is effectively and ductility Theratio are given in
comparisons of Table
initial1.stiffness,
Except the aforementioned
ultimate design
moment, failure
performance,
rotation, anditductility
is notedratio
thatare
thegiven
ductility ratio
in Table 1. of scrimber
Except joints is 2.28. This
the aforementioned is mainly
design due toit the
performance, is
fact thatthat
noted improvements
the ductilityof theof
ratio material
scrimberhardness
joints is and
2.28.tensile
This is strength
mainly due perpendicular
to the fact thatto improvements
the grain of the
scrimber composite
of the material enhance
hardness andthetensile
interaction
strengthbetween wood and
perpendicular to steel members,
the grain of theleading
scrimber to composite
yielding of
enhance
bolts and thethe delay
interaction betweenHowever,
of cracking. wood andthe steel members,
failure modeleading
of glulamto yielding of bolts
joints was andsince
brittle, the delay
wood
of cracking.
splitting However,
appeared almostthe failure mode of
simultaneously glulam
with joints was
the yielding of brittle,
bolts. since wood splitting appeared
almost simultaneously with the yielding of bolts.
Table 1. Comparison of glulam joints and scrimber joints.
Table 1. Comparison of glulam joints and scrimber joints.
Initial Stiffness Ultimate Moment Failure Rotation Ductility Ratio
Joint Type Initial Stiffness Ultimate Moment Failure Rotation Ductility Ratio
Joint Type Kαα Mpeak (KN¨m) θθpeak (˝ ) µμ
K M peak (KN∙m) peak (°)

Glulam joint
Glulam joint [17]
[17] 4.49 17.94
17.94 6.71
6.71 ––**
Scrimber joint
Scrimber joint 4.79
4.79 15.62
15.62 6.97
6.97 2.28
2.28
* The
* The ductility
ductility ratioofofglulam
ratio glulam joints
joints is
is not
notconsidered
considereddue
dueto to
its its
brittle failure
brittle mode.
failure mode.

2.4.
2.4.Results
ResultsofofCyclic
CyclicTests
Tests

2.4.1.
2.4.1.Hysteretic
HystereticLoop
Loopand
andBackbone
BackboneCurve
Curve
The
Thetypical
typicalhysteretic
hystereticloop
loopandand corresponding
corresponding backbone curve from
backbone curve from cyclic
cyclictests
testsare
areshown
showninin
Figure
Figure9. The typical
9. The hysteretic
typical loop shows
hysteretic a reverse
loop shows a “S” shape“S”
reverse withshape
significant
with pinching phenomenon,
significant pinching
which is mainly caused by the unrecoverable plastic deformation of wood components.
phenomenon, which is mainly caused by the unrecoverable plastic deformation of wood components. With the
increase
With theof rotation, therotation,
increase of pinchingthe phenomenon becomes more
pinching phenomenon severe more
becomes due tosevere
the damage
due to accumulation
the damage
inaccumulation
the specimens. Thespecimens.
in the backbone curve is the envelope
The backbone curve isobtained by connecting
the envelope obtained bythe peak point
connecting theinpeak
each
primary cycle of the hysteretic loop. It is noted that the backbone curve is nonlinear
point in each primary cycle of the hysteretic loop. It is noted that the backbone curve is nonlinearwith stiffness
degradation
with stiffnesscaused by woodcaused
degradation fracture
by and
wood bolt bending.
fracture andThe
boltbackbone
bending. curve enters a softening
The backbone curve entersstage
a
insoftening
the last primary
stage in cycle, showing
the last primary a decrease of moment-resisting
cycle, showing capacity with severe
a decrease of moment-resisting damage
capacity withof
severe
the damage of the specimen.
specimen.

Figure9.9.Analyses
Figure Analysesof
ofcyclic
cyclictests:
tests: (a)
(a) typical
typical hysteresis
hysteresis loop;
loop; (b)
(b) backbone
backbonecurve.
curve.

2.4.2.Stiffness
2.4.2. StiffnessDegradation
Degradation
Thestiffness
The stiffnessofofthe
thejoint,
joint,which
whichisiscrucial
crucial for
for the
the lateral
lateral deformability
deformabilityofofaapost-and-beam
post-and-beamtimber
timber
system, would decline due to the unrecoverable damage such as
system, would decline due to the unrecoverable damage such as fracture and embedment fracture and embedment
deformation
deformation of wood components. To evaluate the degree of damage accumulation, the secant
stiffness of each primary cycle is calculated by:
Materials 2016, 9, 581 8 of 16

Materials 2016, 9, 581 8 of 16

of wood components. To evaluate the degree of damage accumulation, the secant stiffness of each
primary cycle is calculated by:
 M i  M i
K i  |`M | ` |´Mi | (5)
Ki “  i  i  i (5)
|`θi | ` |´θi |
where+M
where +Mi i and
and ´M −Miiare
arethe
thepositive
positiveandandnegative
negative peak
peak moments
moments of the the i-th
i-th primary
primary cycle, respectively;
respectively;
+θi iand
+θ and´θ−θii are the rotations
rotations corresponding
corresponding to to +M
+Mii and −M
´Mi,i ,respectively.
respectively.
The comparison
The comparison of stiffness degradation
degradation between
betweenscrimber
scrimberjointsjointsand
andglulam
glulamjoints
jointsis is
shown
shown in
Figure
in Figure 10.10.
It isItsuggested
is suggestedthatthat
the the
secant stiffness
secant of scrimber
stiffness of scrimberjoints is larger
joints at loading
is larger stages
at loading from
stages
0.075Δ
from to 0.7Δ,
0.075∆ and the
to 0.7∆, andstiffness of glulam
the stiffness jointsjoints
of glulam decreased faster faster
decreased than thethanstiffness of scrimber
the stiffness joints.
of scrimber
This isThis
joints. dueistodue thetofact thatthat
the fact the the
failure of of
failure scrimber
scrimberjoints
jointswaswasmainly
mainlycaused
caused byby gradual damagedamage
accumulation,while
accumulation, whilethethefailure
failureofofglulam
glulamjoints
jointswas
wascaused
causedby by the
the sudden
sudden appearance
appearance of of run-through
run-through
crackin
crack inthe
the wood.
wood. Moreover,
Moreover,the themaximal
maximalsecant
secantstiffness
stiffnessof
ofscrimber
scrimberjoints
jointsisisabout
about30%30% larger
largerthan
than
thatof
that ofglulam
glulamjoints.
joints.

Figure 10.
Figure 10. Comparison
Comparison of
of average
average stiffness
stiffnessdegradation
degradationbetween
betweenscrimber
scrimberjoint
jointand
andglulam.
glulam.

2.4.3. Energy
2.4.3. Energy Dissipation
Dissipation
The energy
The energy dissipation
dissipation capacity
capacity of
of different
different joints
joints can
can be
be compared
compared by
bythe
theequivalent
equivalent viscous
viscous
dampingratio
damping ratio(EVDR),
(EVDR),which
whichisiscalculated
calculatedby:
by:
1 E
he he “ 1 dˆ Ed (6)
(6)
2 2πE p E p

As
Asillustrated
illustratedin
inFigure
Figure11a,11a,EEdd is
is the
the energy
energy dissipated
dissipated inin one
one cycle
cycle equal
equal toto the
the enclosed
enclosed area
areaofof
the
the hysteretic
hystereticloop, andEEpp is
loop,and is the
the nominal
nominal elastic
elastic potential
potential energy
energy equal
equal to
to the
the sum
sum of of the
the product
product ofof
the
thepeak
peakmoment
moment andand the
the corresponding
corresponding rotation
rotation inin the
the positive
positive direction
direction and
and the negative direction.
The
Thecomparison
comparisonof ofEVDR
EVDRfrom from0.2∆
0.2Δtoto2.0∆
2.0Δisisshown
shownin inFigure
Figure11b.
11b. The
The EDVR
EDVRof of scrimber
scrimberjoints
joints
is
is about
about twice
twice that
thatof
ofglulam
glulamjoints
jointsatateach
eachprimary
primarycycle,
cycle,showing
showingthatthatthe
thescrimber
scrimberjoints
jointsperform
perform
well
wellininenergy
energydissipation.
dissipation.With
Withbending
bendingdeformation
deformationof ofbolts,
bolts,the
theEDVR
EDVRof ofscrimber
scrimberjoints
jointsincreases
increases
rapidly
rapidly from 0.4Δ to 1.6Δ. Since bolt yielding has become the major energy dissipation mechanismat
from 0.4∆ to 1.6∆. Since bolt yielding has become the major energy dissipation mechanism at
the
thelast
lastprimary
primarycycle,
cycle,the
theEDVR
EDVRof ofscrimber
scrimberjoints
jointsalmost
almostremains
remainsunchanged.
unchanged.
Materials 2016, 9, 581 9 of 16
Materials 2016, 9, 581 9 of 16

Figure 11. Calculation of equivalent viscous damping ratio (EVDR): (a) mathematical meaning of Ed
Figure 11. Calculation of equivalent viscous damping ratio (EVDR): (a) mathematical meaning of Ed
and Ep; (b) comparison between scrimber joint and glulam joint.
and Ep ; (b) comparison between scrimber joint and glulam joint.

3. Numerical Analysis
3. Numerical Analysis
Numericalmodelling
Numerical modellingisisanan effective
effective wayway of investigating
of investigating the mechanical
the mechanical behavior
behavior of bolted
of bolted joints.
joints. During the past decade, some numerical research for bolted joints based on two-dimensional
During the past decade, some numerical research for bolted joints based on two-dimensional (2D) FEM
(2D) FEM was conducted [22,23]. It was suggested that 2D FEM was only appropriate for some
was conducted [22,23]. It was suggested that 2D FEM was only appropriate for some specific situations
specific situations such as for very thin or very thick wood components. Therefore, a 3D FEM
such as for very thin or very thick wood components. Therefore, a 3D FEM considering the anisotropic
considering the anisotropic elastoplastic behavior of scrimber composite was developed and
elastoplastic behavior of scrimber composite was developed and validated by experimental results.
validated by experimental results. The validated model was further used to investigate the
The validated model was further used to investigate the mechanical behavior and failure mechanism
mechanical behavior and failure mechanism of bolted scrimber joints through stress analyses
of bolted scrimber joints through stress analyses combined with parametric analyses.
combined with parametric analyses.
3.1. Finite Element Model
3.1. Finite Element Model
3.1.1. Material Properties
3.1.1. Material Properties
The steel is regarded was isotropic with Young’s modulus equal to 210 GPa and Poisson’s ratio
equalThe steelThe
to 0.3. is regarded was isotropic
stress-strain propertieswith Young’s
of bolts andmodulus
steel plateequal
weretoboth
210 GPa and Poisson’s
assumed ratio
as a bi-linear
equal to 0.3.
hardening The stress-strain
model. The yielding properties
strength and of ultimate
bolts and steel plate
strength were
of steel both assumed
connection as a 270
plate were bi-linear
MPa
hardening
and 550 MPa, model. The yielding
respectively, withstrength
ultimateandstrain
ultimate
equalstrength
to 0.2.ofThe
steelyielding
connection plate were
strength 270 MPa
and ultimate
and 550 of
strength MPa,boltsrespectively,
were 480 MPa withandultimate
600 MPa,strain equal to with
respectively, 0.2. The yielding
ultimate strength
strain equal to and
0.1.ultimate
All the
strength of bolts
stress-strain were 480
parameters wereMPa and 600 MPa,
determined respectively,
by tension tests. with ultimate strain equal to 0.1. All the
stress-strain
The scrimberparameters
compositewere determined
was modelled asbytransverse
tension tests.
isotropy material with identical characteristics
Theand
in radial scrimber
tangentialcomposite
directions.was modelled
For bolted timberas joints,
transverse isotropy
the bearing material
capacity with governed
was mainly identical
characteristics
by the embedment in radial and tangential
properties of wooddirections.
and bending For resistance
bolted timber joints,Hong
of bolts. the bearing capacity
et al. [24,25] was
found
mainly governed by the embedment properties of wood and bending resistance
the MOE obtained by embedment tests was quite different from that obtained by compression tests. of bolts. Hong et al.
[24,25] found
Therefore, the MOE
a so-called obtained by
embedment embedment
region around the tests
boltwas quite
holes wasdifferent
defined fromwith athat
crossobtained
section by of
compression
2.5 d. ˆ 2.5 d.tests.
(d. isTherefore,
the diameter a so-called
of bolts).embedment
The MOE ofregion around theregion
the embedment bolt holes was defined
was taken with
by reducing
a cross
the value section of 2.5 d. ×tests.
from material 2.5 d.The
(d.reduction
is the diameter
factor ofwasbolts).
takenThe MOE
as 0.2 in of
thethe embedment region
parallel-to-grain was
direction
taken
and 0.7byinreducing the value from material
the perpendicular-to-grain tests.based
direction The reduction
on previous factor was taken as 0.2 in the parallel-
experience.
to-grain
As thedirection and 0.7
generalized in theofperpendicular-to-grain
version Von-Mises yield criterion, direction
the Hillbased on previous
criterion experience.
was adopted to consider
As the generalized
the orthotropy version
of the scrimber of Von-Mises
composite [26]. Theyield criterion,
equivalent stressthe
canHill criterion by
be expressed was adopted(7):to
Equation
consider the orthotropy of the scrimber composite [26]. The equivalent stress can be expressed by
ı1{2 ?
Equation (7):
” ` ˘2 ˘2
σ “ a1 σy ´ σz ` a2 pσz ´ σx q2 ` a3 σx ´ σy ` 3a4 τzx 2 2 2
`
` 3a5 τyz ` 3a6 τxy { 2 (7)
1/ 2
   a1  y   z   a 2  z   x   a 3  x   y   3 a 4 zx2  3 a 5 yz2  3 a 6 xy2 
2 2 2
/ 2 (7)
where a1 “ 2 1
´ 1
, a2 “ a3 “ 2
, a4 “ a5 “ a6 “ 
; σi and τ ij are the normal stress and shear
2
f c,90 2
f c,0 2
f c,0 3 f v2
2 1 1 and2 fc,90
stress
whereof athe
1 
scrimber
2
 2 composite, , a 4  a 5  fac,0
, a 2  a 3  respectively; 6 
; σare theτcompressive
i and strengths
ij are the normal stressparallel and
and shear
perpendicularf c ,to f c , 0 and fv is the
90 grain, f c2, 0 shear strength of 2
3 f the
v scrimber composite.
stress of the scrimber composite, respectively; fc,0 and fc,90 are the compressive strengths parallel and
perpendicular to grain, and fv is the shear strength of the scrimber composite.
Materials 2016, 9, 581 10 of 16
Materials 2016, 9, 581 10 of 16

The stress-strain
The stress-strain properties
properties of the composite
of the scrimber scrimber incomposite in parallel-to-grain
parallel-to-grain and
and perpendicular-to-
perpendicular-to-grain directions are defined by two different multi-linear kinematic
grain directions are defined by two different multi-linear kinematic hardening models (i.e., KINH hardening
models in
models (i.e., KINHasmodels
ANSYS) shown inin ANSYS)
Figure 12,aswhich
shownareincombined
Figure 12, which
with are combined
the Hill criterion towith the Hill
simulate the
criterion to simulate
elastic-plastic behaviorthe
of elastic-plastic behavior
the material [27]. of the
Based on the material tests,
[27]. the
Based on the material
mechanical tests,
properties of
the scrimber
mechanical properties of the scrimber composite used
composite used in the model are given in Table 2. in the model are given in Table 2.

Figure 12. Stress-strain propertiesofofscrimber

Stress-strain properties scrimber composite:
composite: (a) (a) parallel-to-grain;
parallel-to-grain; (b) perpendicular-to-
(b) perpendicular-to-grain.
grain.
Table 2. Mechanical properties of scrimber composite used in the model.
Table 2. Mechanical properties of scrimber composite used in the model.
Modulus of Elasticity Modulus of Rigidity Strengths Poisson Ratio Fracture Energies
Modulus of Elasticity Modulus of Rigidity Strengths Poisson Ratio Fracture Energies
(N/mm2 ) (N/mm2 ) (N/mm2 ) – (N/mm2 )
(N/mm2) (N/mm2) (N/mm2) – (N/mm2)
ft,90 = 3.4
ft,90==84.0
3.4 µLR = 0.028
EL = 16,800 fc,0 GcI I= 0.56
EL E=R16,800 GLR = 1120 f = 84.0 µμRT
LR = 0.028
= 0.050
= 1300 fc,90 = 16.0
c,0 GG
II = 0.56
c c = 1.32
GLR = 1120 μLT
µ = 0.028
RT = 0.050 II
ER = 1300 fc,90
fv ==6.3
16.0 Gc = 1.32
μLT = 0.028
fv = 6.3
3.1.2. Crack Simulation
3.1.2. Crack Simulation
To simulate the failure process of bolted joints explicitly, the cohesive zone material (CZM) was
usedTo to simulate
model the thecrack
failure process
[28]. of bolted
Contact joints
elements explicitly,
were set on the
bothcohesive
sides ofzone material (CZM)
the designed was
crack path
used to model
as shown the crack
in Figure [28].element
13. The Contact pairs
elements
wouldwere beset on both according
separated sides of thetodesigned
the stresscrack path as
distribution
shown in Figure
of the defined 13. The
CZM, andelement pairscrack
then a new would be separated
surface according
was created. to the stress
The available CZMdistribution of the
in the software
defined
ANSYS is CZM,basedand
on then a new crack
the models surface
proposed was created.
by Alfano The available
and Crisfield CZMincludes
[29], which in the software ANSYS
Mode I, Mode II,
is based on the models proposed by Alfano and Crisfield [29], which includes
and the mixed mode. Considering the combined effect of perpendicular-to-grain tensile stress and Mode I, Mode II, and
the mixed mode.
parallel-to-grain shearConsidering
stress, the the combined
mixed model iseffect
chosenoftoperpendicular-to-grain tensile
realize the fracture criteria stress and
as illustrated in
parallel-to-grain
Equation (8): shear stress, the mixed model is chosen to realize the fracture criteria as illustrated
in Equation (8):
GI GII
G I G IG`II G I I “ 1 (8)
I
+c II =c1 (8)
Gc Gc
where G I and GcI are the fracture energy and the critical fracture energy considering normal contact
I
where andGcI IGare
stress; GGI II and c are the fracture
the fracture energyenergy
and the and the critical
critical fracturefracture energy considering
energy considering tangentialnormal
contact
stress, respectively. I II
II Since it IIis difficult to determine the values of Gc and Gc through material tests,
contact stress; G and Gc are the fracture energy and the critical fracture energy considering
in this study, GcI and GcI I were calibrated versus the joint test results. As shown in Table 2,I the critical
tangential contactI stress, Irespectively. Since it is difficult to determine the values of Gc and GcII
fracture energy Gc and Gc I are taken as 0.56 N/mm and 1.32 N/mm, respectively.
through material tests, in this study, GcI and GcII were calibrated versus the joint test results. As
shown in Table 2, the critical fracture energy GcI and GcII are taken as 0.56 N/mm and 1.32 N/mm,
respectively.
Materials 2016, 9, 581 11 of 16
Materials 2016, 9, 581 11 of 16

Figure 13.
Figure 13. The
The 3D
3D finite
finite element
element model
model (FEM).
(FEM).

3.1.3. Contact Simulation

3.1.3. Simulation
Contacts exist in
Contacts in the
the interface
interfacebetween:
between:thethe scrimber
scrimber members
members andandsteel plate,
steel the bolts
plate, and
the bolts
scrimber members, the bolts and steel plate, and the scrimber beam and
and scrimber members, the bolts and steel plate, and the scrimber beam and column members. column members. The
interaction
The between
interaction different
between components
different was simulated
components was simulatedby surface-to-surface contactcontact
by surface-to-surface pairs, where
pairs,
the surface with higher rigidity was defined as the target surface, and the surface with
where the surface with higher rigidity was defined as the target surface, and the surface with lower lower rigidity
was defined
rigidity was as the contact
defined as thesurface.
contactThe normal
surface. The pressure
normaland tangential
pressure friction force
and tangential were transferred
friction force were
between the
transferred contactthepairs
between without
contact considering
pairs without penetration.
considering The initial
penetration. normal
The initial stiffness
normal was
stiffness
dependent on the material characteristics of the softer surface. The friction
was dependent on the material characteristics of the softer surface. The friction was evaluated was evaluated by the
the
Coulomb stick-skip
Coulomb stick-skip criterion, which allows no relative movement until the friction force is reached.
Accordingtotothe
According theresearch
researchonon bolted
bolted timber
timber joints
joints [30],
[30], thethe friction
friction coefficient
coefficient between
between scrimber
scrimber and
and bolt,
bolt, and scrimber and steel connection plate were set as 0.6 and 0.1, respectively.
and scrimber and steel connection plate were set as 0.6 and 0.1, respectively. The friction coefficient The friction
coefficient
between between
bolt bolt connection
and steel and steel connection
plate, and plate, and scrimber
scrimber beam and beam and column
column were taken
were taken as 0.2asand
0.2
and 0.5, respectively.
0.5, respectively.

3.1.4.
3.1.4. Meshing and
and Boundary
Boundary Conditions
Conditions
The
The 8-node
8-nodehexahedral
hexahedral elements
elements with reduced
with integration,
reduced which
integration, existsexists
which in ANSYS elementelement
in ANSYS library,
were used
library, were to model
used tosteel and wood
model components.
steel and Each node of
wood components. the node
Each elements haselements
of the three translational
has three
degrees of freedom
translational degreesin directions
of freedom of x,iny,directions
and z, respectively. To track
of x, y, and the whole deformation
z, respectively. To track the process
wholeof
the joints, the Newton-Raphson method was chosen as the incremental
deformation process of the joints, the Newton-Raphson method was chosen as the incremental scheme. Considering the high
MOE
scheme.of scrimber
Considering composite,
the high theMOE
reduced integration
of scrimber could prevent
composite, nearlyintegration
the reduced incompressible couldelements
prevent
from
nearlyvolume locking. However,
incompressible elementsthe artificial
from volumestrain energy,
locking. which was
However, theintroduced to control
artificial strain the energy
energy, which
spurious mode, may
was introduced have an
to control theeffect
energyon spurious
the calculation
mode,accuracy.
may have With the ratio
an effect of artificial
on the strain
calculation energy
accuracy.
to the the
With total strain
ratio energy being
of artificial strainless than to
energy 5%,
thethe calculation
total accuracy
strain energy being was almost
less within
than 5%, the the limit of
calculation
error. On was
accuracy the contrary, the elements
almost within the limit should be remeshed
of error. to increase
On the contrary, the number
the elements of elements
should be remeshedand theto
mesh density
increase of someofirregular
the number elements areas.
and theMoreover, some other
mesh density techniques
of some irregular inareas.
modeling could be
Moreover, someadopted
other
to effectively
techniques in solve the problems
modeling appeared
could be adopted toin meshingsolve
effectively [31–33]. Two typesappeared
the problems of boundary conditions
in meshing [31–
were considered,
33]. Two types of including
boundary those due towere
conditions the loading and including
considered, those duethose
to theduedisplacements
to the loading as and
shown in
those
Figure
due to 13.
the displacements as shown in Figure 13.

3.2.
3.2. Validation
Validation of
of the
the Numerical
Numerical Model
Model
Figure 13 shows
Figure 13 showsthe
the3D3D FEM
FEM established
established by software
by the the software package
package ANSYS.
ANSYS. The numerical
The numerical failure
failure
modes modes are given
are given in Figure
in Figure 14 with
14 with thethe comparativeexperimental
comparative experimentalphenomena.
phenomena.Similar
Similar to
to the
the
experimental phenomena of monotonic tests, varying degrees of crack development were observed
on the tension and compression sides with embedment deformation of bolt holes. The stress
Materials 2016, 9, 581 12 of 16

Materials 2016, 9, 581 12 of 16

Materials 2016, 9, 581 12 of 16
experimental phenomena of monotonic tests, varying degrees of crack development were observed on
nephogram
the tension and shows that the sides
compression stress at mid-span of deformation
bolts had reached the yielding strength when
nephogram shows that the stresswith embedment
at mid-span of bolts of boltthe
had reached holes. The stress
yielding nephogram
strength when
scrimber
shows joints
that the failed,
stress leading
at mid-span to a “one
of bolts hinge”
had reached yielding mode
the yielding of the bolts. The comparison of
scrimber joints failed, leading to a “one hinge” yielding mode strength whenThe
of the bolts. scrimber joints failed,
comparison of
numerical
leading and
to a “one experimental
hinge” yielding M-θ curves
mode is
of the shown
bolts. The in Figure 15,
comparison and the results
of numerical of comparing
numerical and experimental M-θ curves is shown in Figure 15, and the resultsand experimental
of comparing
rotational
M-θ curves performance
is shown in are given
Figure 15, in Table
and the 3. It isofsuggested
results comparing that the numerical
rotational and experimental
performance are given in
rotational performance are given in Table 3. It is suggested that the numerical and experimental
results
Table are
3. Itare
results in good
is suggested agreement with relative
that the numerical
in good agreement errors
anderrors
with relative less
experimentalthan 10%,
less than results thus the
are the
10%, thus developed
in good agreement
developed 3D
3D FEM FEM
with can
be be
relative
can
an effective
errors less than
an effective tool for
10%,
tool further investigations.
thus theinvestigations.
for further developed 3D FEM can be an effective tool for further investigations.

Figure 14. Numerical failure modes: (a) cracking of beam component; (b) yielding of (all
bolts (all
Figure 14. Numerical
Figure 14. Numericalfailure modes:
failure (a) (a)
modes: cracking of beam
cracking component;
of beam (b) yielding
component; of bolts
(b) yielding numbers
of bolts (all
in numbers
the figurein the
are figure
in MPa).are in MPa).
numbers in the figure are in MPa).

Figure 15. Comparison of experimental and numerical M-θ curves.

Figure 15. Comparison
Comparison of experimental and numerical M-θ curves.
M-θ curves.
Figure 15. 3.
Table Comparisonofofexperimental and
numerical and numerical results.
experimental
Table
Table 3. Comparison
Initial
3. Comparison of numerical
numerical
Stiffness Ultimate
of and experimental
Moment
and experimental results.
Failure Rotation Ductile Coefficient
results.
Analysis Type
Kα Mpeak (KN∙m) θpeak (°) μ
Initial Stiffness Ultimate Moment Failure Rotation Ductile Coefficient
Analysis Type
Experimental resultInitial Stiffness
4.79 15.62
Ultimate Moment 6.97
Failure 2.28
Kα Mpeak (KN∙m) θpeakRotation
(°) DuctileμCoefficient
Analysis Typevalue
Numerical 4.56 16.49 7.12 2.12
Experimental result 4.79
K−4.8% 15.62
Mpeak5.6%
(KN¨m) 6.97( )
θpeak ˝ 2.28µ
Relative error α 2.2% −7.1%
Numerical value 4.56 16.49 7.12 2.12
Experimental result 4.79 15.62 6.97 2.28
3.3. Relative
Numerical error
value
Stress Distribution −4.8%
4.56 5.6%
16.49 2.2%
7.12 −7.1%
2.12
Relative error ´4.8% 5.6% 2.2% ´7.1%
Figure
3.3. Stress 16a,b show the perpendicular-to-grain stress distribution and shear stress distribution
Distribution
when initial splitting appears, respectively. It is noted from Figure 16a that high-stressed regions are
3.3. Stress
FigureDistribution
16a,bthe
show
observed along rowthe perpendicular-to-grain
of bolts on the tension side and stress distribution
near the contact and shearbetween
segment stress distribution
scrimber
when
beam initial
and splitting
column, appears,
where respectively.
the initial It
cracks is noted
would from
probablyFigure 16a
appear that high-stressed
according
Figure 16a,b show the perpendicular-to-grain stress distribution and shear stress distribution to regions are
experimental
observed along
phenomenon.
when the row
With the
initial splitting of bolts on
development
appears, the tension
of cracks,
respectively. side
It is theand near
region
noted the contact
fromofFigure 16a that high-stressedscrimber
segment between
high perpendicular-to-grain stress
regions
beam
moves
are andup column,
observed onalong where
the tension
the side
row the
andinitial
of bolts cracks
gradually
on the would
expands
tension probably
to the
side upper appear
and bolt
near theaccording
holes on to experimental
both sides.
contact Moreover,
segment between
phenomenon.
the maximal
scrimber With the development
beamperpendicular-to-grain
and column, where thetensileof cracks,
initial stress the
cracks(i.e., region
3.8
would MPa) ofis high
probably higher perpendicular-to-grain
appear than stress
the perpendicular-to-
according to experimental
grainup
moves tensile
on thestrength
tension (i.e.,
side and3.4 MPa).
graduallyFigure
expands16b shows that the
to the upper boltshear
holesstress is symmetrically
on both sides. Moreover,
the maximal perpendicular-to-grain tensile stress (i.e., 3.8 MPa) is higher than the perpendicular-to-
grain tensile strength (i.e., 3.4 MPa). Figure 16b shows that the shear stress is symmetrically
Materials 2016, 9, 581 13 of 16

distributed on both sides of the steel connection plate. However, the maximal shear stress around
bolt holes
Materials on9, the
2016, 581 tension side (i.e., 4.96 MPa) is only 58% of the shear strength (i.e., 8.6 MPa). It 13
canof be
16
concluded that the combination of perpendicular-to-grain tensile stress and shear stress governs the
failure mode of bolted scrimber joint.
phenomenon. With the development of cracks, the region of high perpendicular-to-grain stress
moves up on the tension side and gradually expands to the upper bolt holes on both sides.
Moreover,
Materials 2016, the
9, 581maximal perpendicular-to-grain tensile stress (i.e., 3.8 MPa) is higher than 13 ofthe
16
perpendicular-to-grain tensile strength (i.e., 3.4 MPa). Figure 16b shows that the shear stress is
distributed
symmetrically on both sides ofonthe
distributed bothsteel connection
sides plate.
of the steel However,
connection the maximal
plate. However,shear stress around
the maximal shear
bolt
stressholes on the
around tension
bolt holesside
on (i.e.,
the 4.96 MPa)
tension is only
side (i.e.,58%
4.96ofMPa)
the shear strength
is only 58% (i.e.,
of the 8.6 MPa).
shear It can be
strength
concluded that the
(i.e., 8.6 MPa). combination
It can be concludedof perpendicular-to-grain
that the combinationtensile stress and shear stresstensile
of perpendicular-to-grain governs the
stress
failure mode of bolted scrimber joint.
and shear stress governs the failure mode of bolted scrimber joint.

Figure 16. Stress distribution when initial splitting appears: (a) perpendicular-to-grain stress;
(b) shear stress (all numbers in the figure are in MPa).

3.4. Parametric Analyses

In this section, parametric analyses were conducted to investigate the influence of two different
parameters. The scrimber joints with five different bolts with grades ranging from 4.6 to 8.8 have
been modeled, and the rotational performance indexes with various yielding strength of bolts are
shown in Figure
Figure
17a. The momentwhen
16. Stress
resisting capacity is appears:
proportionally increased with the increase of
Figure 16. Stress distribution
distribution when initial initial splitting
splitting appears: (a)
(a) perpendicular-to-grain
perpendicular-to-grain stress;
stress;
bolts(b)
grade. However, the growth thetendency in initial stiffness becomes slower with the increase of
(b) shear
shear stress (all numbers
stress (all numbers in the
in figure are
figure are in
in MPa).
MPa).
bolts grade. It is suggested that the bolts in grade of 6.8 and 8.8, with the yielding strength of 480 MPa
and 640 MPa respectively,
3.4. are the proper choice in consideration of performance.
3.4. Parametric
Parametric Analyses
Analyses
The mechanical properties of the scrimber composite can be modified through the impregnation
In
In this
this section,
section, parametric
parametric analyses
analyses were
were conducted
conducted to
to investigate
investigate the
the influence
influence of
of two different
twocomposite
different
and molding process. To investigate the influence of mechanical properties of the scrimber
parameters.
parameters. The scrimber
The scrimber joints
joints with
with fivefive different
different bolts
bolts with grades
with grades ranging
ranging from
fromwere 4.6 to 8.8 have
4.6 tomultiplied
8.8 have been
on the joint behavior, all the strength indexes of the scrimber used in this study by a
been modeled,
modeled, and and
the the rotational
rotational performance
performance indexes indexes
with with various
various yielding yielding
strengthstrength
of boltsofare
bolts are
shown
factor in a reasonable and achievable range from 0.7 to 1.1. Thus, the strength modification factor is
shown
in in Figure
Figure 17a. 17a.
The momentThe moment
resistingresisting
capacityAscapacity
is shown is proportionally
proportionally increased with the increase of
considered as another influential parameter. in Figureincreased with themoment,
17b, the ultimate increase yielding
of bolts
bolts
grade. grade.
However,However, the
thestiffness growth
growth are tendency
tendency in initial
in initial stiffness stiffness
becomesbecomesslower slower
with ofwith the
thescrimber increase
increasestrength. of
of bolts
moment, and initial all proportionally increased with the increase
bolts
grade. grade. It is suggestedthe
It is suggested thatbolts
the bolts in grade 6.8ofand
6.8 8.8,
andwith
8.8, with the yielding strength of 480 MPa
Compared to the boltthat grade, improving in grade theofproperties of wood theis yielding strength
a more economical of 480
andMPa and
effective
and
640 640
wayMPa MPa respectively,
respectively,
for enhancing theare are the
the proper
rotational proper choice in consideration
choice in consideration
performance of the bolted timber of performance.
of performance.
joints.
The mechanical properties of the scrimber composite can be modified through the impregnation
and molding process. To investigate the influence of mechanical properties of the scrimber composite
on the joint behavior, all the strength indexes of the scrimber used in this study were multiplied by a
factor in a reasonable and achievable range from 0.7 to 1.1. Thus, the strength modification factor is
considered as another influential parameter. As shown in Figure 17b, the ultimate moment, yielding
moment, and initial stiffness are all proportionally increased with the increase of scrimber strength.
Compared to the bolt grade, improving the properties of wood is a more economical and effective
way for enhancing the rotational performance of the bolted timber joints.

Figure 17.
Figure 17. Parametric
Parametric analyses:
analyses: (a)
(a) various
various yielding
yielding strength
strength of
of bolts;
bolts; (b)
(b) various
various strength
strength factor
factor of
of
scrimber composite.
scrimber composite.

The mechanical properties of the scrimber composite can be modified through the impregnation
and molding process. To investigate the influence of mechanical properties of the scrimber composite

Figure 17. Parametric analyses: (a) various yielding strength of bolts; (b) various strength factor of
scrimber composite.
Materials 2016, 9, 581 14 of 16

on the joint behavior, all the strength indexes of the scrimber used in this study were multiplied by a
factor in a reasonable and achievable range from 0.7 to 1.1. Thus, the strength modification factor is
considered as another influential parameter. As shown in Figure 17b, the ultimate moment, yielding
moment, and initial stiffness are all proportionally increased with the increase of scrimber strength.
Compared to the bolt grade, improving the properties of wood is a more economical and effective way
for enhancing the rotational performance of the bolted timber joints.

3.5. Comparison between Glulam and Scrimber Joints with the Same Geometrical Dimensions
In this section, the numerical results of scrimber joints and the experimental results of glulam
joints with the same geometrical dimensions were compared. The scrimber joint with the beam section
of 300 mm ˆ 200 mm and the column section of 300 mm ˆ 250 mm was analyzed. The initial stiffness,
ultimate moment, and failure rotation of the scrimber joint were 38.8%, 63.4% and 8.5% larger than
those of glulam joint, respectively. Moreover, scrimber joints still perform well in ductility with the
ductility ratio equal to 2.16. It is noted that the increase of cross sectional size of the scrimber joints can
produce far larger increase of stiffness and load carrying capacity with a smaller decrease of ductility.

4. Discussion
The development of scrimber composite can not only increase the value of fast-growing
wood, which is widely available in China, but also provide an alternative material for wood-based
constructions. A series of experiments were conducted to evaluate the practical applications of
such material, and a 3D model was then developed to investigate the stress distribution and failure
mechanism. It was noted that the bolted joints built with the scrimber composite performed much
better than commonly used glulam joints in initial stiffness, ductility, and energy dissipation. However,
when subjected to lateral load, the crack development of scrimber joints was similar to that of
glulam joints. Previous researches have shown that the failure of glulam joints is governed by the
perpendicular-to-grain tensile stress, which was verified as the key factor for failure of scrimber joints
through numerical analysis. These findings suggest that the bolted joints built with scrimber composite
have certain similar attributes to bolted glulam joints, which contributed to developing the design
procedure of scrimber composite with appropriate reliability levels. Due to the novel manufacturing
process, scrimber joints also have some unique attributes such as complicated interactions among
the curved layers. The numerical model is suitable for further research on other applications of such
material such as floor and shear wall.

5. Conclusions
A series of monotonic and low frequency cyclic tests were conducted to investigate the rotational
performance of bolted joints built with a wood-based scrimber composite material. Due to the initial
clearance between bolts and wood components, the moment-resisting capacity of joints was relatively
low at the beginning of loading. With local crushing of the beam on the compression side, the
moment-resisting capacity of joints significantly increased after the bolts were sufficiently contacted
to the wood components. At large rotations, splitting cracks of the beam developed upwards on
the tension side, and the joints almost accessed a plastic deformation phase with the yielding of
bolts. The joints eventually failed when the run-through crack appeared along the beam with obvious
embedment deformation of bolt holes. Ductile failure with yielding of bolts was observed as the
main failure mode for the specimens. These results are due to the production process of the wood
scrimber composite bring a considerable increase in its mechanical property (i.e., strength and stiffness),
and bolt yielding is therefore obtained as the main failure model and cracking of wood is delayed at
the same time. Ductile failure with bolts yielding is therefore obtained as the main failure model of
the joints. For the results of cyclic tests, the maximum secant stiffness and energy dissipation capacity
were 30% and 90% larger than those of glulam joints, respectively. Considering the dimension of
scrimber joints is eventually smaller than that of glulam joints, scrimber joints performed better in both
Materials 2016, 9, 581 15 of 16

ductility and energy dissipation capacity, which makes the scrimber composite a promising material
for timber structures.
A 3D numerical model was then developed and validated by comparisons of M-θ curves and
failure modes. The proposed FEM was able to capture the bending deformation of bolts and the
deformation of wood embedment near the bolt holes. The development of cracks in the high-stress
region is also captured by the FEM. Through analysis of different stress distribution combined
with crack development, it is suggested the tensile stress perpendicular-to-grain governs the failure
mode of the scrimber joints. Results suggested that the initial stiffness, ultimate moment resisting
capacity, and failure rotation of the scrimber joints were 38.8%, 63.4%, and 8.5% larger than those of
glulam joints with the same member sizes, which further showed the superiority of such scrimber
composite material. The presented experimental results and validated numerical model can serve
as a fundamental basis and useful tool for evaluating the mechanical behavior of scrimber joints
with different configurations in future research, thus supporting more applications of such composite
material in structural engineering.

Acknowledgments: The authors gratefully acknowledge the National Natural Science Foundation of China for
supporting this work with a research grant (Grant No. 51378382).
Author Contributions: Minjuan He and Zheng Li conceived and designed the experiments; Duo Tao performed
the experiments and analyzed the data; MaoLin Li contributed materials; Duo Tao and Zheng Li wrote the paper.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Taheri, F.; Nagaraj, M.; Khosravi, P. Buckling response of glue-laminated columns reinforced with
fiber-reinforced plastic sheets. Compos. Struct. 2009, 88, 481–490. [CrossRef]
2. O’Loinsigh, C.; Oudjene, M.; Shotton, E.; Pizzi, A.; Fanning, P. Mechanical behaviour and 3d stress analysis
of multi-layered wooden beams made with welded-through wood dowels. Compos. Struct. 2012, 94, 313–321.
[CrossRef]
3. Quiroga, A.; Marzocchi, V.; Rintoul, I. Influence of wood treatments on mechanical properties of
wood–cement composites and of populus euroamericana wood fibers. Compos. Part B Eng. 2016, 84,
25–32. [CrossRef]
4. Cabrero, J.M.; Heiduschke, A.; Haller, P. Analytical assessment of the load-carrying capacity of axially loaded
wooden reinforced tubes. Compos. Struct. 2010, 92, 2955–2965. [CrossRef]
5. Borri, A.; Corradi, M.; Speranzini, E. Reinforcement of wood with natural fibers. Compos. Part B Eng. 2013,
53, 1–8. [CrossRef]
6. Oudjene, M.; Meghlat, E.M.; Ait-Aider, H.; Batoz, J.L. Non-linear finite element modelling of the structural
behaviour of screwed timber-to-concrete composite connections. Compos. Struct. 2013, 102, 20–28. [CrossRef]
7. Candan, Z.; Akbulut, T. Nano-engineered plywood panels: Performance properties. Compos. Part B Eng.
2014, 64, 155–161. [CrossRef]
8. Khelifa, M.; Celzard, A. Numerical analysis of flexural strengthening of timber beams reinforced with CFRP
strips. Compos. Struct. 2014, 111, 393–400. [CrossRef]
9. Tran, V. D.; Oudjene, M.; Méausoone, P.J. Experimental and numerical analyses of the structural response of
adhesively reconstituted beech timber beams. Compos. Struct. 2015, 119, 206–217. [CrossRef]
10. Jin, M.; Hu, Y.; Wang, B. Compressive and bending behaviours of wood-based two-dimensional lattice truss
core sandwich structures. Compos. Struct. 2015, 124, 337–344. [CrossRef]
11. Bru, D.; Baeza, F.J.; Varona, F.B.; García Barba, J.; Ivorra, S. Static and dynamic properties of retrofitted timber
beams using glass fiber reinforced polymers. Mater. Struct. 2016, 49, 181–191. [CrossRef]
12. Alaee, S.A.M.; Sullivan, T.; Rogers, C.A.; Nascimbene, R. Semi-empirical method to predict the displacement
capacity and resistance of cold formed steel frame wood-panel shear walls. In Proceedings of the 12th World
Conference on Timber Engineering, Auckland, New Zealand, 15–19 July 2012; Volume 5, pp. 450–455.
13. Sebastian, W.; Piazza, M.; Tomasi, R. State-of-the-art in timber materials and structures research.
Constr. Build. Mater. 2016, 102, 961–962. [CrossRef]
Materials 2016, 9, 581 16 of 16

14. Feroldi, F.; Russo, S. Structural Behavior of All-FRP Beam-Column Plate-Bolted Joints. J. Compos. Constr.
2016. [CrossRef]
15. Li, Z.; He, M.; Tao, D.; Li, M. Experimental buckling performance of scrimber composite columns under
axial compression. Compos. Part B Eng. 2016, 86, 203–213. [CrossRef]
16. Notations for Designations of Iron and Steel; GB/T 221-2008, China’ National Standard; Standardization
Administration of China (SAC): Beijing, China, 2008.
17. Specifications of High Strength Bolts with Large Hexagon Head, Large Hexagon Nuts, Plain Washers for Steel
Structures; GB/T 1231-2006, China’ National Standard; Standardization Administration of China (SAC):
Beijing, China, 2006.
18. Standard Test Methods for mechanical Fasteners in Wood; ASTM D1761-12, ASTM Standard Test Methods;
American Society for Testing Material Standard (ASTM): West Conshohocken, PA, USA, 2012.
19. Standard Test Methods for Cyclic (Reversed) Load Test for Shear Resistance of Vertical Elements of the Lateral Force
Resisting Systems for Buildings; ASTM E2126-11, ASTM Standard Test Methods; American Society for Testing
Material Standard (ASTM): West Conshohocken, PA, USA, 2012.
20. He, M.J.; Liu, H.F. Comparison of glulam post-to-beam connections reinforced by two different dowel-type
fasteners. Constr. Build. Mater. 2015, 99, 99–108. [CrossRef]
21. Muñoz, W.; Mohammad, M.; Salenikovich, A.; Quenneville, P. Determination of yield point and ductility of
timber assemblies: In search for a harmonised approach. In Proceedings of the 10th World Conference on
Timber Engineering, Miyazaki, Japan, 2–5 June 2008; Volume 3, pp. 1064–1072.
22. Chui, Y.H.; Li, Y. Modeling timber moment connection under reversed cyclic loading. J. Struct. Eng. 2005,
131, 1757–1763. [CrossRef]
23. Bouchaïr, A.; Racher, P.; Bocquet, J.F. Analysis of dowelled timber to timber moment-resisting joints.
Mater. Struct. 2007, 40, 1127–1141. [CrossRef]
24. Hong, J.P.; Barrett, J.D. Wood material parameters of numerical model for bolted connections-compression
properties and embedment properties. In Proceedings of the 10th World Conference on Timber Engineering,
Miyazaki, Japan, 2–5 June 2008; Volume 2, pp. 1018–1026.
25. Hong, J.P.; Barrett, J.D.; Lam, F. Three-dimensional finite element analysis of the japanese traditional
post-and-beam connection. J. Wood Sci. 2011, 57, 119–125. [CrossRef]
26. Audebert, M.; Dhima, D.; Taazount, M.; Bouchaïr, A. Numerical investigations on the thermo-mechanical
behavior of steel-to-timber joints exposed to fire. Eng. Struct. 2011, 33, 3257–3268. [CrossRef]
27. ANSYS, Inc. Release 15.0 Documentation for ANSYS: ANSYS Mechanical APDL Theory Reference; ANSYS, Inc.:
Canonsburg, PA, USA, 2013.
28. Franke, B.; Quenneville, P. Numerical modeling of the failure behavior of dowel connections in wood.
J. Eng. Mech. 2011, 137, 186–195. [CrossRef]
29. Alfano, G.; Crisfield, M.A. Finite element interface models for the delamination analysis of laminated
composites mechanical and computational issues. Int. J. Numer. Methods Eng. 2001, 50, 1701–1736. [CrossRef]
30. Sjödin, J.; Serrano, E.; Enquist, B. An experimental and numerical study of the effect of friction in single
dowel joints. Holz als Roh-und Werkstoff 2008, 66, 363–372. [CrossRef]
31. Belytschko, T.; Wong, B.L.; Chiang, H-Y. Advances in one-point quadrature shell elements. Comput. Methods
Appl. Mech. Eng. 1992, 96, 93–107. [CrossRef]
32. Nascimbene, R.; Venini, P. A new locking-free equilibrium mixed element for plane elasticity with continuous
displacement interpolation. Comput. Methods Appl. Mech. Eng. 2002, 191, 1843–1860. [CrossRef]
33. Nascimbene, R. Towards Non–Standard Numerical Modeling of Thin-Shell Structures: Geometrically Linear
Formulation. Int. J. Comput. Methods Eng. Sci. Mech. 2014, 15, 126–141. [CrossRef]

© 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC-BY) license (http://creativecommons.org/licenses/by/4.0/).