Damage assessment of cross laminated timber connections subjected to simulated earthquake loads

Abstract

Wood-frame is the most common construction type for residential buildings in North America. However, there is a limit to the height of the building using a traditional wood-frame structure. Cross-laminated timber (CLT) provides possible solutions to mid-rise and high-rise wood buildings. CLT offers many advantages such as improved dimensional stability, a quicker erection time and good performance in case of fire. In order to introduce the cross-laminated timber products to the North American market, it is important to gain a comprehensive understanding of its structural properties. This paper focuses on the seismic performance of CLT connections. Over the last few years FPInnovations of Canada has conducted a test program to determine the structural properties of CLT panels and its application in shear walls. The test program comprised of more than 100 connection tests which followed the loading procedures of CUREE and ISO test protocols as specified in ASTM Standards ASTM E 2126-09 (2009). These tests were performed parallel and perpendicular to the grain of the outer layer, respectively. The impact of different connections on the seismic performance of CLT walls was investigated in a second phase on full size shearwall. CLT panels are relatively stiff and thus energy dissipation must be accomplished through the ductile behaviour of connections between different shear wall elements and the connections to the story below. A literature review on previous research work related to damage prediction and assessment for wood frame structures was performed. Different approaches for damage indices were compared and discussed. This paper describes how the energy-based cumulative damage assessment model was calibrated to the CLT connection and shear wall test data in order to investigate the damage under monotonic and cyclic loading. Comparison of different wall setup provided a deeper insight into the damage estimation of CLT shear walls and determination of the key parameters in the damage formulation. This represents a first published attempt to apply the damage indices to estimate the seismic behaviour of CLT shear walls.

Full text

DAMAGE ASSESSMENT OF CROSS LAMINATED TIMBER
CONNECTIONS SUBJECTED TO SIMULATED EARTHQUAKE
LOADS
Johannes Schneider1, Siegfried F. Stiemer2, Solomon Tesfamariam3,
Erol Karacabeyli4, Marjan Popovski5
ABSTRACT: Wood-frame is the most common construction type for residential buildings in North America. However,
there is a limit to the height of the building using a traditional wood-frame structure. Cross-laminated timber (CLT)
provides possible solutions to mid-rise and high-rise wood buildings. CLT offers many advantages such as improved
dimensional stability, a quicker erection time and good performance in case of fire. In order to introduce the cross-
laminated timber products to the North American market, it is important to gain a comprehensive understanding of its
structural properties. This paper focuses on the seismic performance of CLT connections. Over the last few years
FPInnovations of Canada has conducted a test program to determine the structural properties of CLT panels and its
application in shear walls. The test program comprised of more than 100 connection tests which followed the loading
procedures of CUREE and ISO test protocols as specified in ASTM Standards ASTM E 2126-09 (2009). These tests
were performed parallel and perpendicular to the grain of the outer layer, respectively. The impact of different
connections on the seismic performance of CLT walls was investigated in a second phase on full size shearwall. CLT
panels are relatively stiff and thus energy dissipation must be accomplished through the ductile behaviour of
connections between different shear wall elements and the connections to the story below. A literature review on
previous research work related to damage prediction and assessment for wood frame structures was performed.
Different approaches for damage indices were compared and discussed. This paper describes how the energy-based
cumulative damage assessment model was calibrated to the CLT connection and shear wall test data in order to
investigate the damage under monotonic and cyclic loading. Comparison of different wall set-up provided a deeper
insight into the damage estimation of CLT shear walls and determination of the key parameters in the damage
formulation. This represents a first published attempt to apply the damage indices to estimate the seismic behaviour of
CLT shear walls.
KEYWORDS: Cross laminated timber, Connections, , Seismic performance, Damage index, Damage prediction,
Shearwall
1 INTRODUCTION
123
Throughout North America, timber framing is used for
most residential buildings. Recently a new material
called Cross-laminated timber (CLT) was developed in
1
Ph.D. student, School of Engineering, University of British
Columbia, Tel: (250)-807-8185, E-mail:
jonny.schneider@gmx.net
2
Professor, Dept. of Civil Engineering, University of British
Columbia, Tel: (604) 822-6301,
E-mail: sigi@civil.ubc.ca
3
Assistant Professor, School of Engineering, University of
British Columbia, Tel: (250)-807-8185,
E-mail: Solomon.Tesfamariam@ubc.ca
4 Adjunct Professor, Dept. of civil Engineering, University of
British Columbia and Manager of FPInnovations’ Wood
Products division, Canada, V6T 1W5, E-mail:
erol.karacabeyli@fpinnovations.ca,
5 Principal Scientist, FPInnovations, Tel: 1-604-222-5739;
E-mail: marjan.popovski@fpinnovations.ca
Austria and Germany. CLT found various applications in
residential and non-residential buildings. It can be
produced in sizes with maximum dimensions of 4.8m x
20m. Panels can be pre-cut and prepared with all
openings on CNC- machines. Floors and walls can be
prefabricated and erected within small tolerances in a
short time. Benefiting from the solid structure of such a
wood structure, good thermal insulation values as well as
fire performance can be achieved. European architects
and engineers have designed mid-rise buildings in CLT
up to 9 storeys height with very good performance. The
North American market could benefit from this
development and implement this product into their code
as prospective alternative to traditional timber-framing.
FPInnovations has started several testing series to
investigate the material properties of CLT. The
researchers focussed on the connection to the foundation,
connection between wall and deck as well as wall to

wall, and the behaviour under seismic loading. Using the
data from individual connection and shear wall tests, in
this paper, Kratzig’s damage accumulation principles are
utilized [7]. The calculated damage stages will be
compared with visual observations.
2 PREVIOUS RESEARCH
Traditional timber-frame structures used numerous
connections between studs and sheathing that are ductile
and provide a variety of load paths [1]. Ductility in CLT
shear walls is achieved through attachments to the
foundation or floors. The CLT panel itself is very stiff.
European research facilities have conducted CLT tests
with various objectives. Tests with different set-ups were
conducted at the Tree and Timber Institute of Italy
(IVALSA) [2] at University of Ljublijana, Slovenia [3],
and at University of Karlsruhe [4]. Blass and Uibel
considered connections in the narrow side as well as the
face side of the CLT panel [4], calibrated and modified
Johansen’s theory [5] accordingly. At University of
Ljubljana, the main focus of all the tests was on the
connector performance under monotonic and seismic
load. Previous tests in Ljubljana and at IVALSA showed
a number of different failure modes in the connection.
Failure occurred in the bracket, in the fasteners and in
the wood. At University of Karlsruhe a test series was
conducted with individual and groups of fasteners under
various loading conditions [4]. In nearly all test
configurations commercial brackets and fasteners were
used in the tests. The Trees and Timber Institute tested a
custom made hold-down [2]. When a wall was
connected to foundation the failure occurred in the CLT
panel, in the bracket, and in the fasteners. The most
common failure mode was the pull-out of fasteners
which is the desired failure mode as the most energy can
be dissipated by forming plastic hinges in the fasteners.
The research projects in Italy [2] and Ljubijana [3] were
interested in the entire behaviour of the shear walls less
on individual bracket connections.
3 DAMAGE INDICES
Damage accumulation index is a measure to evaluate
damage to elements or buildings after earthquakes or
other impact loads [6,7]. Cosenza and Manfredi [6],
Williams and Sexsmith [7] have discussed various global
and local damage indices. The local and global damage
indices can be categorized into the following four
categories [7]: i) non-cumulative indices; ii)
deformation-based cumulative indices; iii) energy-based
cumulative indices, and iv) combined indices. Non-
cumulative indices neglect the effect of repeating
loading cycles. Based on simplicity and ease of
interpretation, however, non-cumulative indices are
widely used [7]. Ductility and inter-storey drift are
widely used to assess structures with non-cumulative
indices [7]. Newmark and Rosenblueth [8] have
proposed ductility ratio, defined as the ratio between
maximum displacement and the yield displacement as a
structural damage measurement. Park and Ang’s model
is the most widely accepted cumulative damage indices
for reinforced concrete. Their model includes maximum
deformation as well as the influence of repeated cyclic
loadings [7]. Park and van de Lindt [9] has extended the
Park and Ang index model for isolated shear walls based
on sheathing perimeter nail spacing. In the past, rare
attempts have been made to evaluate wood structures by
applying damage indices [9,10]. The maximum
displacement and the hysteretic energy absorption are
the required parameters for the modified function by van
de Lindt and Gupta [10]. The analyzed connection and
shear walls tests in this paper will be focused on the
development of the damage accumulation over time and
will be validated with the on-going dynamic loading
protocol and observed damage of the test specimen. In
this paper only the energy-based cumulative index will
be applied for the analysis.
3.1 ENERGY-BASED CUMULATIVE INDEX
Gosain et al. [14] defined energy absorption as a
measure of damage.
; / 0.75
ii
e i y
iyy
F
D F F
F




(1)
where De = energy-related damage index, Fi = force in i-
th cycle, δi = displacement in i-th cycle, Fy = force at
yielding, δy= displacement at yielding.
Hysteresis loops when dropping below 75% of the
yielding value after reaching yielding were considered to
be negligible for the remaining capacity of the member.
Kratzig et al. [15] developed a more complex energy
formulation. Kratzig’s formulation is based on following
half-cycles. The first half-cycle of loading at certain
amplitude is called primary half-cycle (PHC). The
subsequent part of the cycle after peak load is called
follower half-cycle (FHC). There are different
cumulative formulations for the positive and negative
part of the hysteresis. For the positive part of the
response, the damage parameter is defined as:
,p i i
fi
EE
DEE







(2)
where Ep,i = energy in a PHC, Ei = energy in a FHC, Ef =
energy absorbed in a monotonic test to failure.
The parameters for the negative deformation have to be
calculated accordingly. The overall damage index is
defined as:
D D D D D
   
  
(3)
where D+ = damage in positive cycle, D- = damage in
negative cycle, D+D- = interaction of D+ and D-
The inclusion of the FHC energy in the numerator as
well as in the denominator limits the influence on a low
level compared to the primary term. Both deformation
and fatigue-type damage are taken into account. A
similar high value of the damage index can be achieved
by a single high-amplitude cycle or by repeating cycles
at lower amplitude.

Figure 1: (PHC) and follower (FHC) half-cycles [7]
4 EXPERIMENTAL STUDIES
In the research reported here, individual CLT connection
tests were carried out, as well as the shearwall tests using
the same CLT connections.
4.1 CONNECTION TEST
The connection tests were conducted separately for
loading parallel to grain and perpendicular to grain. The
CLT samples were prepared to 7 × 14 inches. For this
test series two commercial brackets in size 90 × 48 × 3.0
× 116 mm (Bracket A) and 105 × 105 × 3.0 × 90 mm
(Bracket B) (Fig. 3) and five different fasteners (Fig. 4)
were combined and tested. The table below shows the
combinations and number of tests for each group.
Figure 2: Bracket A (90x48x3.0x116) and Bracket B
(105x105x3.0x90)
Figure 3: Fasteners as used in the testing program
Table 1: Connection tests
Combination
No. of
Tests
Bracket A
18 Spiral Nail 16d × 3 1/2"
19
12 Ring Shank Nail 10d × 3"
16
12 Ring Shank Nail 16d ×
60mm
16
18 Screw 4 × 70mm
17
9 Screw 5 × 90mm
16
Bracket B
10 Spiral Nail 16d × 3 1/2"
16
A universal testing machine was used to carry out the
connection test, and parallel to grain and perpendicular
to grain loading were set up differently.
4.2 SHEAR WALL TEST
All walls were tested on the same test set-up with minor
changes for each different test. Each wall sat on a solid
base beam, which provided holes for all wall set-ups.
Figure 5 illustrates schematic of the test set-up. The wall
length varied between 2.3m and 3.45m.. The brackets
were connected to the base beam with bolts and washers
either 1/2" (12.7 mm) or 3/8" (9.52 mm) depending on
the bracket used, and connected with different fasteners
to the CLT wall panel. The vertical load was applied to
the wall panel by a top beam that was attached to the top
edge side of the wall panel. The vertical load was chosen
to 20 kN/m. This value is given by the Canadian code
and represents the load on the bottom wall of a four-
story building. Depending on length either two or four
vertical actuators equipped with a load cell applied a
constant force between 23 kN and 69 kN. To allow
maximum freedom of rotation of the panel, the
horizontal force was induced through a pinned
connection on each end of the wall.In the starting
position of the test, the centre of the actuator lines up
with the top edge of the wall. The large actuator was
connected to a tall braced tower. To prevent tipping out
of plane of the wall panel, there were two rollers on
either side of the top beam to provide lateral support at
the top end of the wall.
Figure 4: 3D sketch of wall test set-up
Table 2 gives an overview of the tested walls which will
be considered for the damage accumulation indices.

Table 2: Test matrix of shear walls
Wall
Configuration
Test
No.
Connections
Loading
Protocol
1
00C
4 Bracket A
SN 16d, n=18
CUREE
01M
Ramp
01C
CUREE
03
CUREE
04
RN 10d, n=12
CUREE
05
S1 n=18
CUREE
06
S2 n=9
CUREE
20
7 Bracket A
SN 16d, n=18
CUREE
2
07M
3 Bracket A
2 Hold-down
SN 16d, n=18
Ramp
07C
CUREE
08
CUREE
3
09M
4 Bracket C
TR 65, n=10
Ramp
09C
CUREE
10
CUREE
4
11
4 Bracket A
SN 16d, n=18
in step joint WT-
T, n=12
CUREE
12
CUREE
12ISO
ISO
5
13M
9 Bracket B
SN 16d, n=10
Ramp
13C
ISO
6
15M
9 Bracket B
SN 16d, n=10
in step joint
SFS 1, n=8
Ramp
15C
ISO
19
7 Bracket B
SN 16d, n=10
ISO
7
24M
4 Bracket D
TR 65, n=40
Ramp
24C
ISO
26
TR 90, n=40
ISO
27
TR 90, n=20
ISO
8
28M
4 Bracket A
/storey
SN 16d, n=6
Ramp
28C
ISO
29
SN 16d, n=8
ISO
SN = Spiral nail 3.9 × 89mm, RN = Ring shank nail 2.4
× 76mm, S1 = Screw 4 × 70mm, S2 = Screw 5 × 90mm,
TR65 = Timber rivet 65mm, TR90 = Timber rivet
90mm.
4.3 LOADING PROTOCOL
The connection tests and shear wall tests were conducted
under monotonic and cyclic loading. For the cyclic tests
CUREE protocol as well as ISO was applied.
4.3.1 Loading protocol for connection tests
For the first set of connection tests the monotonic
program was done with a unidirectional and downward
loading at a rate of 6.35 mm (1/4") per minute. All tests
were displacement-controlled. The machine stopped
loading when 50% of the peak load was reached after
passing peak load. In different sets of tests the load was
applied parallel to grain and perpendicular to grain
respectively. For cyclic loading tests CUREE and ISO
protocol was followed ASTM E 2126 Standard [18].
Both protocols are displacement-controlled loading
procedures. Each primary cycle in the CUREE protocol
is followed by two cycles at 80% of the first one. ISO
load protocol has three cycles of same magnitude in each
loading level. Two different schedules were used to
address the two different situations of displacement in
the negative direction. In the first case the wall sample
sat on the foundation and did not allow any displacement
in the negative direction. The second case addressed the
horizontal movement (perpendicular to grain). The two
loading schedules are shown in Figure 6.
Figure 5: Cyclic loading protocols for connection test
parallel and perpendicular to grain
4.3.2 Loading protocol for shearwall tests
For the shearwall tests the same type loading protocols
as for the connection tests were used. However, in the
full size wall test, some walls were tested with the ISO
protocol, which is a method, accepted internationally and
also included in the ASTM E 2126 Standard (2009).
The monotonic test was loaded horizontally
unidirectional and in plane at a loading rate of 0.4"
(10.16 mm) per minute. The displacement-controlled
loading stopped at 60% of the peak load after the peak
load was reached. Using the obtained data, the
displacement at 80% of the peak load is calculated which
is necessary for the cyclic tests. For the CUREE protocol
a displacement rate of 0.2" per second (5.08 mm/s) was
chosen. Figure 7 depicts an appropriate plot time vs.
displacement of the cyclic loading. The ISO protocol
schedule was followed with a displacement rate of 1
mm/sec. The amplitudes of the reversed cycles are a
-160
-120
-80
-40
0
40
80
120
160
0 200 400 600 800 1000
Displacement (% of Max)
Time (seconds)
Perpendicular to grain

function of the mean value of the ultimate slip (νu)
obtained in the monotonic test.
Figure 6: Cyclic loading protocols for shear wall test
CUREE (top) and ISO (bottom)
5 EXPERIMENTAL RESULTS AND
CALIBRATION OF DAMAGE
INDICES
5.1 FAILURE MODES
Connection tests as well as full size shearwall tests
showed similar failure modes. The failure occurred at the
connections between the CLT panel and the base. Seven
failure modes were observed within the test series: Pull-
out failure of the fasteners, shear failure of the fasteners,
edge break out of the CLT panel, extended wood
crushing, CLT delamination, net tension failure of the
bracket, and group tear-out of fasteners (Fig. 8).
Pull-out Failure
Shear failure
Edge break out
Wood crushing
Net tension failure
Group tear-out
Figure 7: Failure modes in connection and wall tests
5.2 CONNECTION TEST ANALYSIS WITH THE
ENERGY-BASED APPROACH
The damage evaluation approach using the energy-based
method of Kratzig [7] does not require a user defined
factor. The function is only related to the absorbed
energy of the monotonic and cyclic test. Figure 9 shows
a typical load displacement curve of a conducted
connection test.
The analysis was done separately for each connection
and direction. Figure 10 shows test results of the group
Bracket A with 18 Spiral nails 3.9 × 89mm. It can be
shown that the loading direction (parallel or
perpendicular) does not affect the slope of the curve
which represents the damage increasing rate. Up to D =
0.75 the values of the individual test results are in a
narrow range. Table 3 considers collapse at D = 0.75
which was proven in an earlier publication of the author.
In all connection combinations the damage index D
follows a nearly linear function until collapse is reached.
In Figure 11 the average damage accumulation curves
for the individual combinations are plotted. A lower
slope will represent a more ductile behaviour of the
connection since the CLT specimen can be considered as
stiff.
Figure 8: Cyclic response for the connection test
perpendicular to grain with 9 Ring shank nails
-160
-120
-80
-40
0
40
80
120
160
0 200 400 600 800 1000 1200
Displacement (% of Max)
Time (seconds)
Shearwall test - CUREE Protocol
-160
-120
-80
-40
0
40
80
120
160
0 200 400 600 800 1000 1200
Displacement (% of Max)
Time (seconds)
Shearwall test - ISO Protocol
-60
-30
0
30
60
-60 -40 -20 0 20 40 60
Load [kN]
Deflection [mm]
Load - Deflection Curve
A-R-P-C8

Figure 9: Damage index for Bracket A with 18 Spiral
nails
Figure 10: Average values of all connection types
5.3 SHEAR WALL ANALYSIS
The shear walls tests are divided into eight groups based
on size and connections according to Table 2. The
damage assessment was carried out based on the cyclic
data sets for 22 walls. The damage-time curves of a wall
with the size of 2.3m × 2.3m and various connectors are
compared in Figure 12. The validation of the calculated
value was done by visual observation of the test walls.
The first group of wall tests shows a similar behaviour
like the connection tests. The curves for walls with spiral
nails can be found at the lower end of the curves (Figure
12; curve 00, 01C, 03).
It is interesting that in group 3 both tests performed
identical. The timber rivets of wall named 09C separated
gradually with the increasing loading protocol compared
to test #10 (Fig. 13). This softer behaviour can be seen at
the slope difference of the 2 tests.
Figure 11: Wall results for group 1 (2.3m x 2.3)
Figure 12: Wall results for group 3 (2.3m x2.3m)
Figure 13: Wall results for group 6 (3.45m x 2.3m)
A comparison of group #1 with test #10 shows a
significant difference in the damage accumulation where
D = 0.75 was reached with timber rivets quite faster.
The results from group 6 show the influence of the wall
panel on the damage accumulation (Fig. 14). Wall #19
consists of 3 separate panels connected with screws in
the step joints. Even with a higher number of brackets
the slope of the curve is very low. Wall 15C consisting
of one panel with 2 openings loaded with the same
0.00
0.25
0.50
0.75
1.00
0 100 200 300 400 500
Damage factor [-]
Time [s]
A-N-L-C1
A-N-L-C2
A-N-L-C3
A-N-L-C4
A-N-L-C5
A-N-L-C6
A-N-L-C7
A-N-P-C2
A-N-P-C3
A-N-P-C4
A-N-P-C5
A-N-P-C6
0.00
0.25
0.50
0.75
1.00
0 100 200 300 400 500
Damage factor [-]
Time [s]
Bracket B, Spiral
nails
Bracket A, Ring
shank nails
Bracket A, Ring
shank nails(short)
Bracket A, Screw 4
x 70mm
Bracket A, Screw 5
x 90mm
Bracket A, Spiral
nail
0.00
0.25
0.50
0.75
1.00
0 100 200 300 400 500
Damage Factor [-]
Time [s]
01C
00
03
04
05
06
20
0.00
0.25
0.50
0.75
1.00
0 100 200 300 400 500
Damage Factor [-]
Time [s]
09C
10
0.00
0.25
0.50
0.75
1.00
0 100 200 300 400 500
Damage Factor [-]
Time [s]
15C
19

protocol shows a significant higher increase of the
damage accumulation. It shows the necessity of having
the same boundary conditions to be able to compare the
damage results.
The visual observation of test 24C showed a group tear-
out failure of the end brackets in an early stage of the
test. This can be seen in the flat slope of the damage
curve. In test #27 the wall was performing very strong
until 3 brackets failed in tension in a proceeded stage of
the loading as seen in the top end of the curve (Fig. 15).
Figure 14: Wall results for group 7 (4.8m x 2.3m)
5.4 COMPARISON OF CONNECTION AND
FULL SIZE WALL TEST RESULTS
For a representative comparison between connection and
wall tests #7 walls were picked. Six walls were of size
2.3m x 2.3m. One wall has the dimensions of 3.45m x
2.3m. To compare the results the average values of the
connection tests were used. All walls were subjected to a
vertical load of 20 kN/m. The Comparison of Figures 16
to 20 shows that the wall results do never exceed the
damage level of the connection tests before D= 0.75 is
reached. The connection test set-up restricts movements
to the particular direction which will be tested.
Movements out of plane as well as bending of test
samples are prevented. Compared to the connection tests
the test walls were only guided at the top end to keep
them in plane. The movement at the bracket was free in
all directions. The walls for this comparison were
generally connected with 4 brackets to the foundation.
Through rocking of the wall panel these brackets
experienced a considerably bigger displacement than the
inner brackets. However, the calculated damage index
considers the entire wall. This is the reason of the
different slopes of the damage index curves between
connection and wall test.
Walls in Figures 16 to 19 had a size of 2.3m x 2.3m and
were connected to the foundation with 4 brackets. Wall
13C (Fig. 20) had a size of 3.45m x 2.3m and was hold
in place with 9 brackets. Wall 13C shows a good
correlation between connection and wall test. However,
walls with a different size and number of connectors
cannot reach this correlation although Figure 16 shows a
close result.
The comparison of the energy-based damage
accumulation index of connection and wall test indicate
that size and number of brackets have a significant
influence on the result. For all conducted wall tests the
damage index stayed at any point below the damage
index of the isolated connection test.
Figure 15: Comparison between connection and wall
test, Bracket A with Spiral nail 3.8 x 89mm
Figure 16: Comparison between connection and wall
test, Bracket A with Ring shank nail 3.4 x 76mm
0.00
0.25
0.50
0.75
1.00
0 200 400 600 800 1000
Damage Factor [-]
Time [s]
24C
26
27
0.00
0.25
0.50
0.75
1.00
0 100 200 300 400 500
Damage factor [-]
Time [s]
Bracket A, Spiral nail
Wall 01C, 4 Bracket A @ 18 Spiral nails, 2.3 x 2.3m
Wall 00, 4 Bracket A @ 18 Spiral nails, 2.3 x 2.3m
Wall 03, 4 Bracket A @ 18 Spiral nails, 2.3 x 2.3m
0.00
0.25
0.50
0.75
1.00
0 100 200 300 400 500
Damage factor [-]
Time [s]
Bracket A, Ring shank nails
Wall 04, 4 Bracket A @ 12 Ring shank nails, 2.3 x 2.3m

Figure 17: Comparison between connection and wall
test, Bracket A with Screw 4 x 70mm
Figure 18: Wall Comparison between connection and
wall test, Bracket A with Screw 5 x 90mm
Figure 19: Comparison between connection and wall
test, Bracket B with Spiral nails 3.8 x 89mm
5.5 DAMAGE PREDICTION
In this paper the connection and shear wall tests were
analyzed using Kratzig’s index (energy-based index). In
order to make the proposed damage model useful for
predicting and evaluating the damage state of CLT
connections, the relationship between calculated damage
index and observed damage needs to be established.
Table 3: Classification of Damage for CLT connections
Degree of
damage
Damage description
Damage index
scale
None
No visible damage observed
D < 0.20
Minor
Minor pull-out of fasteners;
light plastic deformation of
bracket; minor repairs are
required
0.20 ≤ D < 0.35
Moderate
Visual permanent deflections
of bracket; shear failure of
small number of fasteners;
extensive pull-out of
fasteners;
0.35 ≤ D < 0.65
Severe
Major or complete failure of
fasteners; severe crack in
bracket; separation of bracket
from CLT panel; requires
replacement of bracket in
different position at CLT wall
to be serviceable again;
severe wood crushing in outer
layer of CLT
0.65 ≤ D < 0.75
Collapse
Total or partial collapse of
connection
D > 0.75
Obtained hysteresis graphs (Fig. 15) and visual
observations (Fig. 20) were used to calibrate the applied
damage index. The classification consists of five damage
limit states. To address the observed damage in the
conducted tests damage states for CLT connections were
defined accordingly. The results from the visual
observations were validated with the calculated damage
indices. Previous research considered the final stage after
completing the loading protocol for the damage indices.
The focus for CLT connections was placed on the
development of the damage accumulation over time. The
results were validated with the destructivity of the
connection. The limit states were determined as None,
Minor, Moderate, Severe, and Collapse. The relationship
between damage index and observed damage is shown in
Table 3
6 CONCLUSIONS
To use CLT in regions with a high earthquake hazard, it
is necessary to evaluate the damage of such structures
which may be subjected to ground motions due to a
major earthquake. In this paper, Kratzig’s energy-based
damage accumulation index was utilized to evaluate
connections and entire CLT walls subjected to cyclic
loading.
0.00
0.25
0.50
0.75
1.00
0 100 200 300 400 500
Damage factor [-]
Time [s]
Bracket A, Screw 4 x 70mm
Wall 05, 4 Bracket A @ 18 Screws 4 x 70mm, 2.3 x2.3m
0.00
0.25
0.50
0.75
1.00
0 100 200 300 400 500
Damage factor [-]
Time [s]
Bracket A, Screw 5 x 90mm
Wall 06, 4 Bracket A @ 9 Screws 5 x 90mm, 2.3 x 2.3m
0.00
0.25
0.50
0.75
1.00
0 100 200 300 400 500
Damage factor [-]
Time [s]
Bracket B, Spiral nails
Wall 13C, 9 Bracket B @ 10 Spiral nails, 3.45 x 2.3m

The connection tests with nails and screws in CLT have
showed adequate seismic performance. It is of interest to
see that the direction of the applied load does not change
the slope of the curve of an energy-based damage index.
By comparing the calculated damage index over time
with the visual observation of the individual connection
tests it could be shown a good correlation with the
proposed connection damage scale. The calculated
damage indices for the full size shearwalls were
validated according to the same principle as the
connection tests. By analyzing the results it can be
proven that the proposed connection damage scale can
represent the wall behaviour satisfactorily.
The CLT wall panel in the conducted tests can be
assumed as stiff. Thus, the ductility of the wall system
comes from the connections. Among these different
types of connections, bracket A and B with spiral nails
agreed with the wall behaviour the best.
It can be summarized that the slope of the damage
accumulation allows determining the stiffness of the
connection and the degree of damage which will be
reached at certain times. The calculated damage curves
can be compared in a clear way with other calculated
curves. For this comparison it is required that the CLT
panels have the same size as well as the same loading
protocol. For a comparison of individual connection and
entire wall test result additional research has to be
undertaken to determine the influence of size, vertical
loading and number of connectors on the damage index.
The energy-based damage index is a reasonable
alternative in evaluation of results from cyclic testing of
connections and walls.
7 ACKNOWLEDGEMENT
This research was supported through funding to the
NSERC Strategic Network on Innovative Wood
Products and Building Systems (NEWBuildS).
REFERENCES
[1] Rainer, & E.Karacabeyli. Wood-frame construction
- meeting the challenge of earthquakes. Canadian
Wood Council, 1999.
[2] Ceccotti A., Follesa M., Lauriola M.P., Sandhaas C.
Sofie Project – Test Results on the Lateral
Resistance of Cross-Laminated Wooden Panels.
Proceedings of the First European Conference on
Earthquake Engineering and Seismicity, Geneva,
Switzerland, 2006.
[3] Dujic B., Pucelj J., Zarnic R. Testing of Racking
Behavior of Massive Wooden Wall Panels.
Proceedings of the 37th CIB-W18 Meeting, paper 37-
15-2, Edinburgh, Scotland, 2004.
[4] Blass H.J.& Uibel T. Tragfaehigkeit von
stiftfoermigen Verbindungsmitteln in Brettsperrholz.
Karlsruhe: Universitaetsverlag Karlsruhe.2007.
[5] Johansen, K.W. Theory of timber connections.
International Association for Bridge and Structural
Engineering, Vol.9, 249-262.1949
[6] Cosenza E.& Manfredi G. Damage indices and
damage measures. Pro.Struct.Engng Mater., 2:50-
59.2000
[7] Williams M. S. & Sexsmith, R.G. Seismic Damage
Indices for Concrete Structures: A State-of-the-Art
Review. Earthquake Spectra, Vol.11, No.2, 319-
349. 1995.
[8] Newmark N.M. & Rosenblueth E. Fundamental of
earthquake engineering Prentice-Hall, Englewood
Cliffs, N.J. 1974.
[9] Park, S., & Lindt, J. W. Formulation of Seismic
Fragilities for a Wood-Frame Building Based on
Visually Determined Damage Indexes. Journal of
Performance of Constructed Facilities, 346-352.
2009.
[10] Lindt, J. W., & Gupta, R. Damage and Damage
Prediction for Wood Shearwalls Subjected to
Simulated Earthquake Loads. Journal of
Performance of Constructed Facilities, 176-184.
2006.
[11] Liang, H., Wen, Y.-K., & Foliente, G. c. Damage
Modeling and Damage Limit State Criterion for
Wood-Frame Buildings Subjected to Seismic Loads.
Journal of Structural Engineering, 41-48. 2011.
[12] Banon H., Biggs J.M., Irvine H.M. Seismic damage
in reinforced concrete frames, Journal of Structural
Engineering, ASCE, Vol.107, No.9, 1713-1729.
1981.
[13] Wang M.L., Shah SL.P. Reinforced concrete
hysteresis model based in the damage concept.
Earthquake Engineering and Structural Dynamics,
Vol.15,No.8,993-1003. 1987
[14] Gosain N.K., Brown R.H., Jirsa J.O. Shear
requirements for load reversals on RC members.
Journal of Structural Engineering, ASCE ,
Vol.103,No.7,1461-1476. 1977.
[15] Kratzig W.B., Meyer I.F., Meskouris K. Damage
evolution in reinforced concrete members under
cyclic loading. Proceedings of the 5th Int.Conf.on
Structural Safety and Reliability, San Francisco.
Vol.II 795-802. 1989
[16] Lindt, J. W. Damaged-based seismic reliability
concept for woodframe structures. Journal of
Structural Engineering, 131(4),668-675. 2005.
[17] Park Y.J., Ang A.H.-S., Wen Y.K. Damage-limiting
aseismic design of buildings. Earthquake Spectra,
3(1), 1-26. 1987.
[18] ASTM E 2126. Standard Test Methods for Cyclic
(Reversed) Load Test for Shear Resistance of
Vertical Elements of the Lateral Load Resisting
Systems for Buildings. ASTM International, West
Conshohocken, PA, USA. 2009.