Engineering Structures 303 (2024) 117473

Contents lists available at ScienceDirect

Engineering Structures
journal homepage: www.elsevier.com/locate/engstruct

Seismic performance of RC frame structures with infill CLT shear walls

Richard Yip Je Too a, *, Masanobu Sakashita b, Hiroshi Isoda c, Yasuhiro Araki d
a
Tsukuba Research Institute, Sumitomo Forestry Co., Ltd, Japan
b
Building Research Institute, Japan
c
Research Institute for Sustainable Humanosphere, Kyoto University, Japan
d
National Institute for Land and Infrastructure Management, Japan

A R T I C L E I N F O A B S T R A C T

Keywords: This paper focuses on the newly proposed RC-CLT hybrid wall systems used in residential buildings. Cyclic
Hybrid structures loading tests were conducted using 2-storey specimens, to confirm the seismic performance and behavior of the
Earthquake engineering hybrid system and also to investigate the behavior of the RC beams in mid-level and the transmission of shear
CLT panels
stress by the CLT wall from upper to lower level. The experimental results showed that CLT could potentially be
Reinforced concrete structures
Seismic retrofit
more superior than RC wall infill in terms of ductility, due to CLT exhibiting higher inter-story drift during shear
failure. Furthermore, this study proposes numerical models to simulate the experimental results of the hybrid
systems. The RC structural members were modeled using linear elements with MS model inserted at the ends to
consider for the N-M interactions. As for the CLT wall, it was modeled using several braced models. The nu­
merical results generally show good agreements with the experimental results. As a result, the numerical model
proposed in this study can contribute to better predict the structural behavior of CLT shear wall infilled RC frame
with concealed steel plate and drift pin connections.

1. Introduction substantial amount of greenhouse gases, when compared to steel and

concrete [2]. Among the available engineered wood products (EWP), the
In many countries, reinforced concrete (RC) frame structure with RC advent of Cross Laminated Timber (CLT) has been described as a sig­
load-bearing wall infill is a typical structure that is widely applied in nificant development, and recent studies have enabled the construction
buildings. However, due to the recent growing concern on the climate of tall CLT buildings, even in earthquake-prone areas [3,4].
change, there is a global imperative to use eco-friendly materials in the In Japan, achieving a pure mid-to-high rise timber structure is
construction industry to mitigate the effects of global warming and difficult, due to the strict restrictions of fire safety implemented in the
provide economical and sustainable housing. In this case, the utilization building code. Therefore, in order to achieve greater building heights
of wood can fulfill such requirements. One crucial factor that makes while also enabling the utilization of wood, the application of hybrid
wood a viable option for construction is its environmental impact structures is a potential way to solve such problem. A secondary benefit
throughout the life cycle of a building. When it comes to the argument of using hybrid wood structures in seismic design is their reduced
about climate change, the energy spent by the Heating, Ventilation and weight. As base shear is proportional to weight, they attract lower
Air Conditioning (HVAC) systems should not be overlooked. Smiroldo seismic forces compared to concrete structures. Furthermore, the
et al. [1] proposed an integrated intervention approach that aims to wooden components in hybrid systems can be easily prefabricated.
improve both the seismic and thermal behaviors of existing RC frames Depending on the simplicity of mechanical connections, it can lead to a
retrofitted by inserting CLT panels. The authors understand the impor­ reduction in construction cost and time [5]. CLT is also known for its
tance of considering the energetic aspect and believe that it should be high strength and stiffness. Stazi et al. analytically investigated the
treated as a separate research topic. Therefore, it should be noted that in-plane behavior and performance of CLT [6]. The numerical results
the aim of this paper is to investigate the structural aspect of the pro­ highlighted that CLT has high potential to be used as shear walls.
posed structure. In recent years, extensive research has been conducted on the RC
Studies have shown that applying wood in structures can decrease frames with CLT shear wall infill hybridizations (RC+CLT). In Europe,

* Corresponding author.
E-mail address: RICHARDYIP_jetoo@star.sfc.co.jp (R.Y.J. Too).

https://doi.org/10.1016/j.engstruct.2024.117473
Received 9 October 2023; Received in revised form 6 December 2023; Accepted 2 January 2024
Available online 16 January 2024
0141-0296/© 2024 Published by Elsevier Ltd.
R.Y.J. Too et al. Engineering Structures 303 (2024) 117473

various researchers have suggested an approach that involves posi­ frame with infill CLT shear walls in the transverse direction. In 2019,
tioning CLT panels in addition to the already existed shear walls to research was conducted for the CLT wing wall type hybridization and
reduce seismic vulnerability [7,8]. The CLT in this proposed solution is the contents were summarized in [14–16]. This paper focuses on the
connected to the RC beams using innovative dissipative friction transverse direction of the prototype building. It is noteworthy that
dampers. Moreover, Smiroldo et al. [9] conducted an analytical inves­ Fig. 1 is a schematic drawing and limited to the structural aspect of the
tigation on a newly proposed seismic retrofit intervention that is prototype building. Other fundamental non-structural elements (e.g.,
applicable to existing RC frame structures, replacing the original ma­ thermal insulations) that form a wall should be included in actual
sonry infill with timber structural panels. The results demonstrated that building constructions when necessary.
the CLT infilled hybrid systems could potentially produce higher resis­ Into the RC frame, CLT shear walls are infilled as seismic resisting
tance and ductile responses, compared to the brittle masonry walls. In element that is expected to significantly increase the strength and
Japan, a method which involves a series of Laminated Veneer Lumber stiffness of the structure against seismic load. Comparing the existing RC
(LVL) inserted side by side and firmly secured in place in the RC frames shear wall structures with RC frame with CLT shear wall structures,
using epoxy resin adhesives was proposed [10,11]. Similar method was there are advantages and disadvantages to both cases. In the case of RC
also being investigated using CLT by Haba et al. [12]. Both systems show shear wall structure, the initial stiffness is usually very high and is able
similar results in terms of ultimate strength. However, slip failure along to keep the deformation at minimum during earthquakes. But once the
the adhesives were observed at small deformations, which led to damage, such as crack has been developed, the damage will be likely to
degradation in stiffness as well as the strength. remain. On top of that, CLT shear walls have greater flexibility in de­
This paper proposes a CLT infill method into RC frames that is mainly formations, that allows failures to occur at larger drift ratios. Therefore,
intended for new constructions. Thus, the sections of RC members are utilizing CLT shear walls might be advantageous in terms of damage
designed not to be damaged by additional lateral and vertical forces control.
from the CLT panels and to withstand dead and live load in the building
even without the CLT infill. In order to apply such method as retrofits, 3. Details of test specimens and description of testing and
further research has to be conducted to investigate the case of adopting specimen design method
weaker RC frames. As a result, the authors believe that this should be a
subject in another paper that is dedicated to retrofitting interventions, 3.1. Details of test specimens
for instance the paper written by Smiroldo et al. [13]. In the proposed
method, the CLT panels are connected to the RC frame using ductile Fig. 2 shows the overview of the specimens. A total of three 2/3 scale,
mechanical connections called drift pins (DP), instead of using adhe­ two-story, one bay frame specimens, were prepared to represent the
sives. The aim of this paper is to conduct experimental and analytical bottom two-story of the prototype, in the transverse direction. The
studies to investigate the seismic behavior and performance of the newly components in the specimens (e.g the cross-sectional dimensions of RC
proposed RC frame with CLT shear wall infill system. After the experi­ members, CLT panels and mechanical connections) were carefully
ments, several novel numerical models were developed and validated by scaled so that the behavior of the scaled specimens could represent the
comparing the numerical results with the test results. Specifically, the behavior of a full-sized structure. Regarding the RC members, the ratios
research findings of this study will enhance the accuracy of seismic of dimensions and reinforcements were adjusted, so that it is the same as
design and evaluation method in practical design work and future those of the actual sized prototype building. Generally, RC specimens
investigations. are made at 1/4 to 1/3 scale, due to the capacity of the testing in­
struments and budgets. However, the RC frame in this test was made at
2. Description of the proposed system larger scale of 2/3. Also, the RC frame was designed so that the flexural
behavior in the RC members would be dominant, by making sure that
Fig. 1 shows the typical floor plan of the prototype building. The sufficient amount of shear reinforcements was provided (i.e. sudden
prototype building is a 6-storey apartment, which is made up of RC strength degradation would not occur). It was already mentioned that if
frame with infill CLT wing walls in the longitudinal direction, and RC proper scaling of the shear reinforcements and the cross-sectional

Fig. 1. Typical floor plan of the prototype building (unit: mm).

2
R.Y.J. Too et al. Engineering Structures 303 (2024) 117473

Fig. 2. Overview of specimens (unit: mm).

dimensions are taken into account, the scale effect should be negligible is rational and should not have significant effect on the obtained results.
[17]. Due to these reasonings, it can be said that the scale effect of the RC Specimen C is the RC bare frame specimen without any shear wall
frame is insignificant. About the CLT panels in the specimens, the scale infill. Two types of hybrid specimens, named WA and WB, were pre­
of cross-sectional dimension was adjusted so that it matches to those of pared with different shear force transmission in the CLT shear wall.
the RC members. This is to make sure that the punching shear effect and Specimen WA is expected to be the higher strength and stiffness option
the impact on axial forces in the RC columns are independent from the due to the installation of the metal shear keys. On the other hand,
scale effect. The only difference that should be given attention is the specimen WB mainly transfer shear forces through the compression strut
number of laminas, which is smaller than those of the actual sized formed in the CLT shear wall. The differences between specimen WA and
prototype building. Although this will more or less affect the shear WB, are the installation of metal shear keys and the grain direction of the
strength and buckling behavior of the CLT panels, it should not be sig­ outer layer of the CLT shear wall. The purpose of the metal shear keys is
nificant. Based on the above statements, the scaling of the hybrid system to increase the flow of lateral force being transmitted from the RC beam

Table 1
Details of the specimens.

3
R.Y.J. Too et al. Engineering Structures 303 (2024) 117473

to the CLT panel. By doing so, several compression struts would be Table 2
formed in specimen WA instead of the single compression strut formed Material properties of concrete.
in WB. This will result in specimen WA demonstrating high load capacity Specimen Floor Compressive strength σc Elastic modulus E
in smaller deformations and higher stiffness as the expected advantage (N/mm2) (kN/mm2)
of the metal shear keys. However, it should be taken into considerations WA 1st floor 38.9 28.8
that there is a possibility where the yielding of the RC beams at the shear 2nd floor 38.3 25.9
attached position would occur, when substantial amount of shear force 3rd floor 34.5 29.2
is being borne by the metal shear keys. The length between the center­ WB 1st floor 40.4 29.2
2nd floor 36.7 28.1
lines of the RC columns is 3500 mm. The floor height is 2000 mm. The 3rd floor 38.3 30.2
configuration of CLT utilized is graded S60 with 4 layers. The CLT in C 1st floor 36.3 32.0
specimen WA consists of layers with grain direction horizontally on the 2nd and 3rd 35.8 32.0
outer layer. Whereas the CLT in specimen WB consists of layers with floor
grain direction vertically on the outer layer. The reason for the change of
grain direction was to also investigate the possible occurrence of buck­
ling of CLT in the failure mechanisms. Table 1 summarizes the details of Table 3
the specimens. Fig. 3 illustrates the details of the reinforcement in the RC Material properties of mortar.
frame, the shear wall infill, vertical and horizontal connections of the Specimen Compressive strength σc Elastic modulus E
cross-section lines shown in Fig. 2(b). Only the details of specimen WA (N/mm2) (kN/mm2)
are shown here as they are similar in other specimens. The dimensions of Non-shrink WA 73.3 23.0
the RC columns and beams are, 400 mm × 400 mm and 300 mm × mortar WB 76.4 23.6
400 mm, respectively. In the CLT infilled specimens, non-shrink mortar Fiber reinforced WA 34.7 18.3
was infilled in between the CLT shear wall and the RC frame, to absorb mortar

construction errors and also to integrate CLT shear wall and RC frame
when resisting lateral load. Vertical connections were also installed at
the ends of the CLT shear wall, which was fixed to the RC columns by a Table 4
series of bolts. The connection consists of a T-shaped metal concealed Material properties of steel reinforcements, DP and bolts.
plate inserted into the slit of the CLT panel, and 25-∅12 drift pin con­ Specimen Yield Tensile Elastic
nectors. In specimen WA, metal shear keys were installed, which were strength σy strength σt modulus E
fixed to the RC beams using bolts. In between the CLT panel and the (N/mm2) (N/mm2) (kN/mm2)

metal shear keys, fiber reinforced mortar was infilled, instead of non- Transverse WA, WB 356 500 182
shrink mortar. The reason is to prevent cracks from propagating into reinforcement C 337 447 179
(D10)
the non-shrink mortar under cyclic loading.
Longitudinal WA, WB 374 556 177
reinforcement C 382 564 170
(D16)
3.2. Materials properties DP (φ12) WA, WB 438aa 528 200
Bolts (M16) WA, WB 687aa 829 177
After the structural experiments, a series of element tests were con­ Bolts (M20) WA 580aa 790 182

ducted to obtain the material properties of each structural component. a

0.2% offset strength
Tables 2–5 show the material properties of concrete, mortar, steel and
CLT shear wall. Standard cylindrical compression test was conducted to were also conducted. However, in this study, tests on the connection
obtain the material properties of concrete and non-shrink mortar. As for considering the concealed plate were not conducted. Therefore, the
the steel reinforcements, DP and bolts, tensile tests were conducted. In material properties of concealed plate with one drift pin connection
addition, both compression and shear tests were conducted to obtain the were obtained by numerical analysis, as shown in Section 3.4.
material properties of CLT shear wall. Bearing tests of CLT with drift pins

Fig. 3. Details of the reinforcement in the RC frame, the shear wall infill, horizontal and vertical connections (Specimen WA, unit: mm).

4
R.Y.J. Too et al. Engineering Structures 303 (2024) 117473

Table 5
Material properties of CLT shear wall.
Direction of outer layer Compressive Elastic Shear strength
strength σc (N/ modulus E σs (N/mm2)
mm2) (kN/mm2)

Vertical axial (parallel 17.9 4.16 3.17

to loading direction)
Horizontal axial 17.8 4.47
(transverse to
loading direction)

3.3. Experimental procedure

Fig. 4 shows a diagram of the test settings. Two 700 kN center-hole

jacks were used to apply constant axial load of 535 kN (N/N0 = 0.1,
where N is the exerted axial load, and N0 is the compressive strength of Fig. 5. Loading schedule of the cyclic test for specimen WA and WB.
the RC columns). The lateral load exerted was controlled to be equal for
the hydraulic jacks from both sides (north and south). The specimens initial stiffness was set to 0 for the case of major direction, according to
were subjected to static cyclic loading, and was controlled by drift ratio previous literature [18]. For the case of minor direction, the ratio
(%), which is δ/3800 mm. Here, δ is the averaged lateral displacement (=0.14) was determined from the results of the bearing element test.
measured on both sides of the specimens at the center of 3 F RC beam Since the obtained load-displacement curve from the analysis was for
and 3800 mm is the height from the surface of the RC stub to the center one DP, the result was multiplied by 25 to account for the vertical
of 3 F RC beam. The loading cycles were repeated once at R= ± 0.125%, connections in the experiments. However, it is reported in the Japanese
twice at R= ± 0.25%, ± 0.5%, ± 0.75%, ± 1%, ± 2%, and ± 3%. For Standard [19] that when multiple DPs are inserted, the load capacity
specimen WA and WB, the final pushover was applied up to R= +3.4%, will decrease according to the number of DPs in a row. The reduction
as shown in Fig. 5. For specimen C, the cycles were repeated twice at R= factor was calculated using Eq. (2), proposed in the Japanese Standard.
± 4% and the final pushover was applied up to R= +5%. The reduction factor for the load capacity was calculated to be 0.81,
based on the maximum number of drift pins in a row (n = 13). The
3.4. Numerical model for drift pin joint load-displacement relationships and the characteristic values of the
concealed plate with drift pin connections are shown in Fig. 8 and
The load-deformation of the concealed steel plate drift pin joint was Table 6, respectively.
obtained using a numerical model shown in Fig. 6. The drift pins were
Kn = ni − 0.081
(2)
imitated by the beam elements and the bearing capacity of laminas with
different grain directions was modeled using uniaxial springs with a
spacing of 3 mm. The bending moment acting on the drift pins was 3.5. Design procedure of horizontal and vertical joints
assumed to reach its plastic moment, calculated using Eq. (1), due to the
large bending deformations observed in the experiments. In most previously reported structural experiments [20], in which
CLT shear walls were installed in RC frame, focused on a single story.
F × d3 Because of that, stress transfer mechanisms between the upper and
Mp = (1)
6 lower stories were seldom being considered. Therefore, in this paper,
two-story one-span partial frame specimens, extracted from the lower
Where Mp = plastic bending moment (N•mm), F = yield strength of DP
level of the prototype building, were prepared to verify the path of force
from tensile element test (= 438 N/mm2), d = diameter of drift pin
transfer from the 2nd floor to the 1st floor. It should be noted that the
(=12 mm).
specimens can only verify the stress transfer under the conditions of the
No uniaxial spring was provided in the 12 mm range in the center at
experiment, and is not suitable for considering other cases such as stress
the slit position. Two analysis cases were considered, in which the
transfer in mid-story (e.g.: 3rd floor to 2nd floor). In this experiment,
laminas of the outer layer of the CLT are in major (case 1) and minor
same lateral loads were applied to the columns by hydraulic jacks. As
direction (case 2), respectively. The uniaxial springs were modeled as
shown in Fig. 9, when the lateral shear force is transmitted through the
bilinear model, of which the bearing capacity was obtained from the
compression strut formed in the diagonal direction of the CLT shear
bearing element test results, as shown in Fig. 7. The ratio of secondary to
wall, axial tensile force will be exerted onto the RC beam in the middle
story in proportion to the amount of lateral shear force, which may cause
early tensile yielding of the RC beam reinforcements and may also limit
the shear force transmitted by the CLT shear wall. In consideration of
workability and cost, the connection between the RC frame and the CLT
shear wall should be simplified as much as possible. However, to resist
the reaction force from the compression strut, the vertical joints (con­
cealed steel plate with drift pin joints) were designed and installed.
Lateral shear capacity of CLT wall can be roughly calculated to be
818 kN by multiplying the section size (120 mm × 3100 mm) to the CLT
design shear strength (2.2 N/mm2). By formation of diagonal compres­
sion strut, half of lateral shear capacity of CLT wall (409 kN) may be
transferred as lateral tensile force in the RC beam. As shown in Table 6,
part of the lateral tensile force can be transferred through the vertical
joints, to decrease the lateral tensile force borne by the RC beams. Yield
strength of the vertical joints is 218 kN for specimen WA (Case 1) and
Fig. 4. Diagram of the test settings (unit: mm). 357 kN for specimen WB (Case 2).

5
R.Y.J. Too et al. Engineering Structures 303 (2024) 117473

Fig. 6. Numerical model of concealed plate with drift pin joint (unit: mm).

Table 6
Characteristic values of the concealed plate with drift pin connections.
Layers Number Reduction Load
of DP Factor (kN)

Horizontal Initial Case 1 25 0.81 168.9

direction of Stiffness Outer:
WA and (kN/mm) Major
Vertical Secondary Inner: 11.7
direction of Stiffness Minor
WB (kN/mm)
Yield 218
Strength
(kN)
Maximum 436
Strength
Fig. 7. Model of the uniaxial springs. (kN)

Vertical Initial Case 2 25 0.81 321.6

direction of Stiffness Outer:
WA and (kN/mm) Minor
Horizontal Secondary Inner: 7.5
direction of Stiffness Major
WB (kN/mm)
Yield 357
Strength
(kN)
Maximum 499
Strength
(kN)

(◆), 1 F RC column (■) and 2 F RC column (•), while the symbol (×)
Fig. 8. Load-displacement relationship of concealed plate with drift represents the maximum load. The observed behavior of each specimen
pin connection. is summarized and discussed in their respective sections below.

In specimen WA, metal shear keys were installed, which were fixed 4.1.1. Specimen C (RC bare frame)
to the RC beams using bolts similar to the vertical connections. Lateral At 0.125%, flexural cracks were observed at both ends of 2 F and 3 F
shear capacity of metal shear keys can be calculated to be 322 kN and RC beam. As the drift ratio increased to 0.5%, flexural cracks were
178 kN, respectively. The former is given by the vertical section size of significant at all ends of the RC beams and the stiffness degraded. At 1%
metal shear keys (120 mm × 150 mm) and compression strength of CLT drift when the lateral force was 218 kN, additional flexural cracks
(17.9 N/mm2) shown in Table 5. The latter is given by the horizontal appeared at the bottom of 1 F RC column, forming a yielding mechanism
section size (120 mm × 675 mm) and design shear strength of CLT of whole structure later at 2% drift that could provide high energy
(2.2 N/mm2). According to the calculations, the shear keys of specimen dissipation, as seen in Fig. 10(a). The result shows a typical hysteresis of
WA can resist 60% of lateral shear capacity of CLT wall (=500 kN) RC framed structure, without any slip deformation. Also, after the
without the vertical joints. overall structure failure mechanism has been formed (Fig. 13(a)), the
strength further increased with the increase of the drift ratio. The
4. Experiment results maximum lateral force increased up to 275 kN at 4% drift, before it
degraded due to the spalling of concrete cover and the buckling of
4.1. Load-drift ratio relationship and failure mechanisms longitudinal reinforcement in the beams. The shear reinforcement at the
north end of 3 F beam also yielded during the final pushover.
Fig. 10 and Fig. 11 shows the hysteresis curves and the crack pattern
of the RC frame of all 3 specimens. The crack pattern in the positive and 4.1.2. Specimen WA (with CLT wall and metal shear keys)
negative loading directions are shown in red and blue lines, respectively. The stiffness and strength of specimen WA increased significantly
The shaped symbols in the hysteresis curves represent the drift ratio that compared to that of the bare frame specimen C. Flexural cracks firstly
the yielding of steel reinforcements in 3 F RC beam (▴), 2 F RC beam appeared on the 3 F and 2 F RC beams during the respective positive and

6
R.Y.J. Too et al. Engineering Structures 303 (2024) 117473

Fig. 9. Shear force transmission in the CLT shear wall.

Fig. 10. Hysteresis curves of the tested specimens.

Fig. 11. Crack distribution of the RC frame.

negative cycle of 0.125% drift at around the same lateral load of 264 kN. which involved the horizontal misalignments of the CLT lamina
At the subsequent cycle of 0.25% drift, cracks were developed in the occurred (Fig. 12(a)), causing the strength degradation after reaching its
fiber reinforced mortar adjacent to the metal shear keys. Flexural cracks maximum lateral load of 825 kN in the first cycle. Because of this, the
were also observed at the bottom of 1 F RC columns in the same loading story deformation was concentrated on 2 F and the longitudinal rein­
cycle. The longitudinal reinforcement in the 3 F RC beam began to yield forcement at the bottom of 2 F RC columns yielded, which led to the
at 0.25% drift, while the longitudinal reinforcement in 2 F RC beam formation of local failure mechanism instead of the overall structure
yielded later at 0.6% drift. As the drift ratio progressed to 1%, the drift failure mechanism occurred in specimen C, as shown in Fig. 13(b). No
pins in the vertical connection began to show obvious bending de­ yielding occurred in the shear reinforcement of the RC beams. None­
formations. During the second cycle of the same drift ratio, shear failure theless, the increase of the lateral load and the stiffness were 3.0 and 3.3

7
R.Y.J. Too et al. Engineering Structures 303 (2024) 117473

Fig. 13. Yielding conditions of rebars at 1/50 rad (only positive loading).

Fig. 12. Failure observes at the CLT shear wall.

times, respectively, when compared to specimen C. became obvious, which led to the separation between the T-shaped
concealed plate and the CLT, and thus the bending deformation of the
4.1.3. Specimen WB (with only CLT wall) drift pins. Regardless, the lateral load continued to increase until 2%
In specimen WB, the out-of-plane deformation of CLT was investi­ drift, where the shear failure of the CLT shear wall occurred (Fig. 12(b)),
gated as buckling of the inner layer laminas was expected to occur. causing the degradation in strength after it reached the maximum lateral
However, the experimental results showed that buckling behavior could load of 767 kN. In the same cycle, yielding occurred at bottom and top of
be neglected in condition that both ends of CLT panels were fixed with the 1 F RC column, forming a failure mechanism. As can be observed
drift pin connections. The overall behavior of specimen WB was almost from Fig. 13(c), the longitudinal reinforcement of the columns yielded in
the same as specimen WA until 1% drift. Also, the initial stiffness was tension at both top and bottom of the columns. This indicates that there
similar between the 2 specimens. During the first cycle of 0.125% drift, was no antisymmetric distribution of bending moment in the columns.
flexural cracks were initially observed at the ends of 3 F RC beam at However, similar to specimen WA, inter-story deformation concentrated
lateral load of 180 kN, then subsequently at the ends of 2 F RC beam at on 2 F due to the shear failure of the 2 F CLT shear wall. Because of this,
lateral load of 208 kN. At 0.25% drift, flexural cracks started to develop at 3% drift, the longitudinal reinforcement yielded at the bottom of 2 F
at the bottom of 1 F RC columns. As the drift ratio progressed to 1% drift, RC column, causing a shift in the formation of failure mechanism from
where specimen WA reached its maximum lateral load of 825 kN, the overall to local structure. Towards the cycle of final pushover, buckling
lateral load of specimen WB was only 621 kN. This suggests that the of longitudinal reinforcement and spalling of cover concrete were
installation of metal shear keys can contribute to the early development observed. Also, severe damages were also observed on the CLT shear
of strength. At the same cycle, shear deformation of the CLT shear wall wall, which involved out-of-plane deformation of laminas at the outer

8
R.Y.J. Too et al. Engineering Structures 303 (2024) 117473

layer, shown in Fig. 12(c). This was caused by the strong horizontal of the minimum principal strain tend to point towards the metal shear
compression force exerted by the vertical connection across the CLT keys, demonstrating that the location of the metal shear keys were the
shear wall, which eventually led to the separation of the laminated layer. starting points for the formation of diagonal compression struts.
At the vertical connection on 2 F, cracks in between the drift pin holes Therefore, compared to specimen WB, the directions were not as uni­
were also observed (Fig. 12(d)). Similar to specimen WA, no yielding form throughout the CLT infill. Also, the difference in the magnitude of
occurred in the transverse reinforcements of the RC beams. the minimum principal strain between 1 F and 2 F was smaller
The increase of the lateral load and initial stiffness were 2.8 and 3.5 compared to specimen WB. Conversely, in specimen WB, the magnitude
times, respectively, when compared to specimen C. of the minimum principal strain was large at the corners. Together, the
directions were oriented to the corners of the CLT shear wall. These
4.2. Curvature of the RC beams suggest that the formation of compression strut may have taken place
diagonally from corner to corner of the CLT infill.
Fig. 14 shows the curvature profiles along the 2 F and 3 F RC beams
of all specimens. The curvatures were calculated by using the recorded 4.4. Maximum shear strain of CLT shear walls
strain values by the strain gauges on the longitudinal reinforcement in
the beams, as shown in the figure. It should be noted that the position of Fig. 18 show the maximum shear strain on CLT shear wall of each
the strain gauges in specimen C was slightly different. The yield cur­ specimen, also from R = 0.25% to 1% drift for WA and from R = 0.25%
vatures shown in dotted lines were calculated with reference to Fig. 15. to 2% drift for WB specimen. It was also evaluated from the recordings of
The yield curvatures were calculated considering the equilibrium of the the 12 rosette gauges, using Eq. (5).
compression-tensile stress and strain relationship, when the RC beam is √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
{ }̅
under bending deformation. γ max = 2 (εh − εd )2 + (εd − εv )2 (5)
The curvature profiles of specimen WB and C were similar, except
As can be observed from Fig. 18, there was an obvious difference in
that the yielding of north end of the 3 F RC beam in specimen WB
the trend of shear force transmission in the CLT infill between both
occurred at smaller drift ratio, due to the additional exertion of force
specimens. Due to the effect of the metal shear keys, the infill of spec­
through the formation of compression strut from the CLT shear wall. On
imen WA show large maximum shear strain at top and bottom rows near
the other hand, there were some irregularities in the curvature profile of
the position of metal shear keys, whereas in specimen WB, the maximum
specimen WA, displaying a sudden change in curvature near the position
shear strain is particularly large in the center row. The maximum shear
of metal shear keys. This was affected by the bending of the studs in the
strain recorded at maximum lateral load was 4462μ and 5611μ for
RC beams, when the metal shear keys transmit the lateral forces. Due to
specimen WA and WB, respectively. It can be said that the CLT shear
the exertion of bending force from the studs, discontinuity in internal
wall reached its maximum capacity in certain areas, as these shear
bending moment of the RC beams was observed, therefore affecting the
strains were approximately the standard shear strain (=2.2 N/mm2/
curvature profile of specimen WA.
500 N/mm2 × 106 =4460μ), which was calculated using the design
From the curvature profiles, the length of plastic zone can also be
shear strength and shear modulus from the Japanese CLT Manual [21].
approximated by observing the position where the curvatures reached or
exceeded the yield curvature at which the longitudinal reinforcement
5. Numerical model
yielded. As a result from the observation, the length of plastic zone was
almost the depth of the RC beam (=400 mm).
5.1. Common details of the numerical model

4.3. Magnitude and angle of the minimum principal strain in the CLT In this section, the models that were developed to accurately predict
shear walls the pushover results of the structural experiment are presented and
described. The models for specimen C and WA are shown in Fig. 19. As
To investigate the directions and magnitudes of the compression for the model of specimen WB, it was identical to that of specimen WA,
struts formed in the CLT shear walls, the values and angles of the min­ only without the shear springs which imitated the metal shear keys.
imum principal strain in the CLT shear walls were evaluated from the Similar to the structural experiments, the models were subjected to
recordings of the rosette gauges provided on the CLT shear wall, using constant axial load onto both columns when the lateral loads were
Eqs. (3) and (4), respectively. A total of 12 rosette gauges were pasted on exerted from both sides. However, in this study, pushover analyses were
the surface of the outer lamina in the fashion shown in Fig. 16. performed for numerical verification of the tests because pushover
{ √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
{ }̅
} analysis is becoming a popular method for evaluating the seismic de­
εmin = εh + εv − 2 (εh − εd )2 + (εd − εv )2 (3) mands and performance of new and existing structures in performance-
based seismic design methodologies [22]. A commercial nonlinear
1
(
2εd − εh − εv
) analysis software SNAP [23] was used for the analyses.
θ = tan− 1
(4) To give a short description for the common modeling method for all
2 εh − εv
specimens, the models were constructed so that the story height and
where, εh , εv and εd are the horizontal, vertical and diagonal (45-degree) span lengths are the same as those of the specimens. In addition, rigid
strain of the CLT shear wall. supports were inserted at the bottom of 1 F columns. Pin supports were
Fig. 17 shows the magnitude and direction of the minimum principal also inserted as boundary conditions for axial springs, imitating the CLT
strain of the CLT shear wall infilled specimens (from R = 0.25% to 1% compression of the infill. The detailed modeling method for the speci­
drift for WA and from R = 0.25% to 2% drift for specimen WB). The mens will be described in the following sections. Generally, the char­
length of the arrow in the figure indicates the magnitude of the mini­ acteristic values obtained from the element tests were used as the
mum principal strain. In both specimens, generally the magnitudes of parameter values for the models.
the minimum principal strain increased with the increase of drift ratio.
The directions were also considerably consistent, before the shear failure 5.2. Model of specimen C
occurred in the CLT infill of specimen WA and WB at R = 1% and 2%
drift, respectively. In both specimens, the force transmission through the 5.2.1. Modeling of RC column
compression struts was observed to be taken place diagonally from a The RC columns were modeled using software defined column linear
corner to another in the CLT shear wall. In specimen WA, the directions elements. In the center of the linear elements, shear springs were

9
R.Y.J. Too et al. Engineering Structures 303 (2024) 117473

Fig. 14. Curvature profiles along the RC beams.

10
R.Y.J. Too et al. Engineering Structures 303 (2024) 117473

Fig. 15. Calculation method of yield curvature of the RC beams.

the RC columns. Here, the plastic length of the MS model was set to be
the column’s depth (=400 mm). The MS model consisted of several
uniaxial steel and concrete springs to represent the inelastic flexural
rotation and the N-M interaction at the ends. Fig. 20(a) shows the
modeling of the column cross section for the MS model. In the MS
models, 3 different materials were defined, which were the cover con­
crete, confined concrete and the longitudinal reinforcement. Fig. 20(b)
and (c) illustrates the stress-strain relationship of concrete and steel
reinforcements for the MS model. NewRC model [25], was adopted for
the envelope compressive stress-strain relationship for both cover and
confined concrete including the ascending and descending branches. It
has been reported that the NewRC model is able to accurately capture
Fig. 16. Location of the strain gauges on the CLT shear wall of specimen WA the confined effects by transverse reinforcements. It is noteworthy that
(unit: mm). the tensile behavior of the concrete was not considered. As for the lon­
gitudinal reinforcement, bilinear model was used to model its
inserted to imitate the shear behavior of the RC columns. At both ends of stress-strain relationship.
the columns, a model called the multi-spring model (MS model) pro­ For the model of shear springs, shown in Fig. 21, trilinear model
posed by Lai et al. [24], was used to simulate the inelastic behavior of which consisted of the cracking and ultimate shear strength was

Fig. 17. Magnitude and direction of the minimum principal strain on the CLT shear wall in positive direction.

11
R.Y.J. Too et al. Engineering Structures 303 (2024) 117473

Fig. 18. maximum shear strain on the CLT shear wall（unit: μ）.

adopted. The cracking shear strength, ultimate shear strength and ulti­ taken into considerations in the model. Therefore, similar to RC col­
mate shear strain were calculated using Eqs. (6)–(8), proposed in [26]. umns, MS model was also applied at the ends of RC beams to account for
However, according to the validation of the equations where the the N-M interactions. Fig. 22(a) shows the modeling of beam cross
calculated values were compared to a large database obtained from a section for the MS model. The plastic length of the MS model is set to be
series of conducted RC shear element tests [27], the results showed that the beam’s depth (=400 mm), which was defined from the curvature
there is a possibility that the calculated values obtained using Eq. (6) profiles of the beam. The stress-strain relationship of the longitudinal
would underestimate the experiment values. Since the purpose of the rebars in the MS model was the same as RC column. However, section
numerical analysis is to verify the structural experiment results, a shape and transverse reinforcement were slightly different, therefore
modified factor of 1.4 was multiplied, as suggested by the reported resulted in the change of model of the confined concrete, shown in
validation Fig. 22(b). The RC beam model also consisted of a shear spring in the
√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅ / center to model the shear characteristics. Here, the model of the shear
Qsc = σ 2cr + σ 0 σ cr bD ks (6) spring was the same as the RC column. Similar to the RC column, the
calculated values were subjected to the modified factor of 1.4, due to the
√̅̅̅̅̅
where, σcr is tensile strength of concrete (=0.33 Fc ) (N/mm2), Fc is possibility of underestimation of numerical results when compared to
2
the compressive strength of concrete (N/mm ), σ0 is axial stress in the the experimental results.
columns (=N/bD) (N/mm2), b is the width of RC column (mm), D is the
depth of RC column (mm), ks is shape factor (=1.5) 5.3. Model of CLT infilled specimens (WA and WB)
⎧ ⎫
⎪
⎪ ⎪
⎪
⎨0.068p (F + 18)
√̅̅̅̅̅̅̅̅̅̅̅̅
⎬ 5.3.1. Modeling of RC frame
(7)
t c
Qsu =
⎪ M
+ 0.12
+ 0.85 pw σwy + 0.1σ0 bj
⎪ The model of the RC columns in specimens WA and WB was divided
⎪ ⎪
⎩ ⎭ into 3 portions, due to the extra nodes for the insertion of axial-shear
Qd

springs, at the location of vertical connections. On the other hand, the

where, pt is the ratio of tensile steel rebars (%), M/Qd is the shear span RC beams were divided into a total of 8 equal portions with a length of
ratio, pw is ratio of transverse reinforcement in the RC column, σwy is 387.5 mm, also for the insertion of shear and axial springs, to imitate the
yield strength of transverse reinforcement (N/mm2), j is length between metal shear keys and CLT compression, respectively. The number of
the compressive and tensile resultant forces (=7d/8) (mm), d is length divisions was decided so that the length was almost equivalent to the
between surface of the cover concrete and the position of tensile steel plastic length of the RC beams at the ends (=400 mm), as observed in
reinforcements (=358 mm). Section 4.2. The MS models with the plastic length of 387.5 mm, were
applied into the divided sections in between, which was the length of
Qu • ke
γ su = 0.004 − (8) each portion. However, inserting MS models with plastic length as long
Gc • bc • Dc
as the length of the element itself may cause poor simulation in initial
where, Qu is ultimate shear strength of RC column (N), ke is shape factor stiffness (e.g: lower initial stiffness in analysis results). Because of this,
according to the energy law (=1.2), Gc is shear elastic modulus of con­ such effect was further investigated by comparing the results of analysis
crete (N/mm2), bc is width of RC column (mm), Dc is depth of RC column models with and without MS models. The results showed that while
(mm). there was difference in initial stiffness (below 1 mm), inserting plastic
length of 387.5 mm in the MS models of RC beams seemed to be
5.2.2. Modeling of RC beam acceptable. In the center of each section in both RC columns and beams,
Due to the stress transfer that takes place between the upper and shear springs were also inserted. Other details regarding the modeling of
lower stories of the specimen, the transmission of axial force should be RC structural members were similar to the mentioned descriptions
above in Section 5.2.

12
R.Y.J. Too et al. Engineering Structures 303 (2024) 117473

Fig. 19. Details of the numerical models.

Fig. 20. Modeling of MS model and stress-strain relationship for concrete and steel reinforcement (RC Column).

13
R.Y.J. Too et al. Engineering Structures 303 (2024) 117473

Eq. (9) was used to calculate the equivalent cross-sectional area, which
could be determined by assuming the cases where lateral deformations
per unit force in both infilled wall and braced wall are equivalent. As for
the strength of the truss element, Eq. (11) was applied to convert the
lateral shear force Q (Eq. (10)) to the axial force, with reference to
Fig. 23(c).
(√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅ )3
G•t• L2 + H 2
Ab = (9)
2•E•L•H

Fig. 21. Modeling of MS model and shear springs for the RC columns. where, G is the shear elastic modulus of CLT= (500 N/mm2), t is the
width of CLT (mm), L is length of the braced model (mm), H is height of
the braced model (mm), E is Young’s modulus of CLT (N/mm2).
5.3.2. Modeling of CLT shear wall
In the numerical model, the shear resistance of the CLT shear wall Q=τ•L•t (10)
was modeled using braced model, illustrated in Fig. 23(a). The braced
model consisted of 4 elastic linear elements and 2 inelastic truss ele­ where, τ is the shear stress of CLT (N/mm2)
ments, which were all connected by pins. Multiple braced models were √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
Q L2 + H 2
applied and placed appropriately according to the positions of springs TCLT orCclt = • (11)
which imitated the connections around the CLT infill. The advantage of 2 L
adopting the braced model is that it is able to take into account of not Further, the stiffness of the vertical and horizontal elements was also
only the shear stiffness, but also axial stiffness of a structural member. required to be properly determined, if the shear and flexural deforma­
The modeling method of the braced model was in accordance with the tion of the braced model were to be accurately reproduced. Here, the
Japanese guidelines [28], where the equivalent stiffness of the elements required stiffness for both vertical and horizontal elements were deter­
in the braced model was evaluated so that it was similar to the CLT shear mined using Eqs. (12) and (13), respectively. The first term of the
wall. The load capacity of the hybrid systems was observed to get equation calculates the total stiffness of the CLT shear wall in each di­
affected by the shear capacity of the CLT infill, therefore an inelastic rection, which would be subtracted by the total axial stiffness compo­
bilinear model shown in Fig. 23(b), was adopted for the truss elements in nent of the truss element in the respective directions. Then, the resulted
the analysis. value would be divided by the number of sections in each direction to
Fig. 23(c) shows the schematics for the calculations required for the determine the required stiffness for each vertical and horizontal
braced model. To appropriately define the stiffness of the truss element, element.

Fig. 22. Modeling of MS model and stress-strain relationship for concrete (RC Beam).

Fig. 23. Modeling of shear resistance of CLT shear wall.

14
R.Y.J. Too et al. Engineering Structures 303 (2024) 117473

/
t • 3100 • EMA would likely to experience shear failure. On the contrary, when the
K(V)req =( − 16 • KCLT(V) ) 9 (12)
H length is longer (i.e: 1390 mm), compression failure would be more
likely occur. To consider these failing scenarios in the numerical model,
/
t • 1560 • EMI Eq. (18) was used to calculate the bearing strengths with different shear
K(H)req = ( − 6 • KCLT(H) ) 4 (13)
L cross-sectional area A1 and A2 .

where, EMA and EMI are Young’s modulus of CLT in major and minor Ks = 15.6 • t • dm (17)
direction, respectively (N/mm2), KCLT(V) and KCLT(H) are vertical and ( )
τ • (A1 or A2 ) t • dm • σc
horizontal stiffness component of the truss element, respectively. Fs = min , (18)
1000 1000
The compressive resistance incurred at the contact surface between
the CLT wall and the RC frame was modeled using bearing spring ele­ Here, t is width of CLT (mm), dm is bearing depth of CLT on metal
ments, placed at the corners of the braced model. In order to reproduce shear keys (=150 mm), τ is design shear strength of CLT (=2.2 N/mm2),
the separation between CLT wall and RC frame, the bearing springs were σc is compressive strength of CLT (N/mm2), A1 =635mm × 120mm,
adjusted to resist only in compression. Each bearing spring was modeled A2 =1390mm × 120mm.
with an elastic-plastic transition, as shown in Fig. 24(a). The values of
bearing stiffness Kc and compressive strength Fc were determined using 5.3.4. Modeling of vertical connection (concealed steel plate with DP
Eqs. (14)–(16). For the bearing springs inserted at the edges of vertical connection)
direction, it was required to consider for the effect of the notches below In the CLT shear wall infilled specimens, concealed steel plate with
the metal shear keys. As shown in Fig. 24(b), the blue area of CLT be­ DP connections were installed at the sides as vertical connections to
tween the RC frame and the metal shear key (L=635 mm) may experi­ transmit vertical and horizontal shear forces from the CLT to the RC
ence either shear or compression failure when shear force Q is exerted. columns. Fig. 26(a) illustrates the behavior of the vertical connection
This was taken into account in the first term of Eq. (16). Each bearing during the experiment. It was observed that the vertical connection
spring was assumed to bear the compression load by a certain behaved such that, compression of CLT occurred at the top half and the
compression area. Here, the second term considered the remaining tensile of DP occurred at the bottom half of the left vertical connection.
compression area that was required to be borne by the bearing springs at While on the right, the vertical connection behaved in the opposite
the edges of vertical direction. fashion.
In the numerical model, axial and shear springs were used to imitate
E•t•d
Kc = (14) the behavior of vertical connection. The springs were inserted in the
h
center of the top and bottom half of grouped drift pins. The restoring
Here, E is the Young’s modulus of CLT (N/mm2), t is width of CLT force model for the CLT compression was the same to that shown in
(mm), d is distance between the bearing springs (mm), h is distance Fig. 25(a). On the other hand, the restoring force model for the tensile
between the local strain readings in the compression element test (mm) and shear springs is as shown in Fig. 26(b). It is a trilinear model tracing
Fc(H) and Fc(V) = σc • t • d (15) the values of the load-displacement curves obtained from the numerical
analysis of DP connection, shown in Fig. 8. Here, Fy and δy are the 5%
(
τ • A t • ln • σ c
)
t • (lb − ln ) • σc offset value yield strength and yield displacement, and, Fu and δu are
Fc(Vedge ) = min , + (16) the ultimate strength and ultimate displacement at 20 mm in the
1000 1000 1000
restoring force model. K was determined from the slope of the line
Here, σ c is compressive stress if CLT (N/mm2), t is width of CLT connecting the points at 0.1 and 0.4 times of Fu .
(mm), d is distance between the bearing springs (mm), τ is design shear
strength of CLT (=2.2 N/mm2), A is shear cross-sectional area (mm2), ln 6. Simulation results
is height of notches (=175 mm), lb is bearing length of axial spring
(=260 mm). 6.1. Load-drift ratio relationship

5.3.3. Modeling of metal shear keys in specimen WA Fig. 27. shows the comparison of load-drift ratio relationship be­
Shear springs were inserted to reproduce the shear resistance of tween test and analysis results for all specimens. For the RC bare frame C
metal shear keys in the numerical model. Fig. 25(a) shows the restoring specimen, the stiffness from the analysis result could be observed to be
force model of shear springs, which yields at the compressive strength of in good agreement with the test result. As for the CLT infilled WA and
CLT to the shear key. The stiffness Ks was calculated using Eq. (17), in WB specimens, the analysis result could only predict up until the shear
which the bearing stiffness obtained from the Japanese CLT Manual [21] failure of CLT in the analysis, as the numerical model did not consider
equivalent to 15.6 N/mm3 was applied. Under lateral load, 2 areas with for strength degradation due to the shear failure of CLT shear wall. In
different lengths were assumed to be compressed, as illustrated in this case, the CLT shear wall in the analysis model was assumed to
Fig. 25(b). Here, when the length is shorter (i.e: 635 mm), the effective encounter shear failure, when any one of the 3 rows of the braced model
shear cross-sectional area is smaller, therefore the compressed area was completely yielded.

Fig. 24. Modeling of compressive resistance of CLT shear wall.

15
R.Y.J. Too et al. Engineering Structures 303 (2024) 117473

Fig. 25. Modeling of the metal shear keys in specimen WA.

Fig. 26. Modeling of the vertical connections.

Fig. 27. Comparison of load-displacement relationship between test and analysis.

In terms of the yielding deformations of the steel reinforcements, previously.

good agreements between the experiment and numerical results could In specimen C, the moment diagram from the analysis shows typical
be observed as well. As noted above, the simulated load-displacement moment diagram of an RC frame. Also, similar to the experiment result,
curves of the specimens were in close agreement with the experiment an overall structure yielding mechanism which involves the formation of
results, indicating that the proposed analytical models may be able to plastic hinges at the ends of RC beams and at the bottom of 1 F RC
contribute to the seismic design calculations for RC frame with CLT columns was achieved in the numerical model. In specimens WA and
shear wall infill. However, the proposed model is still complicated and WB, the RC columns were not subjected to antisymmetric moments, as
should be simplified in order to be put to practical use in structural seen on the 2 F of specimen C. This shows that the RC columns did not
design. mainly bear the shear forces, as they are being transmitted by several
other structural components, for instance the vertical connections and
the CLT shear wall. Furthermore, as described above, the compression
6.2. Moment diagram and yielding condition
strut from corner to corner in the CLT shear wall was dominant in
transmitting the lateral force in specimen WB. As a result, it can be seen
Fig. 28 shows the moment diagrams and yielding conditions for all
that there is a reverse shear force acting near the center of the RC col­
specimens obtained from the analyses, each at the angular displacement
umns of specimen WB; a different result when compared to specimen
of shear failure of CLT, when any one of the 3 rows of the braced model
WA.
was completely yielded (in the results, the middle row was completely
As described in Sections 4.1.2 and 4.1.3, local failure mechanism
yielded). Among the figures, the moment distribution in the RC beam of
occurred in specimen WA, of which only the longitudinal reinforcement
specimen WA was discontinuous at the shear key location, due to the
yielded at the ends of the RC beams, when the shear failure of CLT
transmission of shear forces through the metal shear keys, as described

16
R.Y.J. Too et al. Engineering Structures 303 (2024) 117473

Fig. 28. Bending moment diagrams and yielding conditions from the analyses.

occurred at 1%. On the other hand, specimen WB primarily showed 1. Based on the observed results, it can be said that both specimen WA
formation of overall structure failure mechanism until when the shear and WB are suitable to be utilized in seismic-resistant structures,
failure of CLT occurred at 2%, after which the vertical reinforcement of considering the adequacy of load capacity and stiffness. Nonetheless,
1 F RC column yielded. Here, the analysis results show correspondence they both have pros and cons respectively. For example,
to such yielding conditions and also the structural behavior where the The pros of specimen WA: Suitable for structures that demand high
shear failure of CLT occurred at smaller (=0.64%) and larger (=1.45%) load capacity and stiffness demands that focuses on strength rather
drift ratio in WA and WB specimens respectively in the analysis. than ductility in structural design.
Fig. 29 shows the maximum shear strain of the CLT shear wall The cons of specimen WA: Additional costs and construction time
recorded by the rosette gauges for each specimen, at 1% and 2% drift, due to the installation of the metal shear keys
respectively. As already been described in section 5.4, the pattern of the The pros of specimen WB: Has better performance in terms of
compression strut and shear force transmissions in the CLT shear wall ductility. Since the strength degradation occurred at relatively
are obviously different in specimens WA and WB. When comparing the higher drift ratio, it is suitable to be applied together with framed
results to the analysis results shown in Fig. 27, similar trend in shear structures which also have high ductility.
force transmissions can also be observed. The cons of specimen WB: The transmission paths of the forces are
limited, therefore structural designers have to pay special attention
7. Conclusions to the localization of shear force acting on the RC columns, and also
the tensile axial force acting on the RC beams.
In this paper, with the aim of developing new structural design 2. The load-relationship results indicated that by infilling CLT shear
method for RC and CLT hybridization, the findings obtained from the wall into an RC frame, significant increase in the maximum capacity
cyclic loading tests conducted on two RC frame specimens with infilled (up to 3.0 times) and initial stiffness (up to 3.5 times) can be ach­
CLT shear walls and analytically investigated results by newly proposed ieved. Also, the ductility responses were relatively better, given that
numerical models were presented and discussed. The major findings are the strength degradation occurred at 1% and 2% drift in specimen
summarized as follows: WA and WB, respectively, compared to that of 0.4% which is
considered as the standard drift for shear failure for RC walls in
practical structural design in Japan [29].

Fig. 29. Maximum shear strain of the CLT shear wall obtained from the test (Values: μ).

17
R.Y.J. Too et al. Engineering Structures 303 (2024) 117473

3. The installation of metal shear keys in the CLT walls contributed to [3] Tannert T, Follesa M, Fragiacomo M, Conzalez P, Isoda H, Moroder D. Seismic
design of cross-laminated timber buildings. Wood Fiber Sci 2018:3–26.
the early display of maximum capacity. Because of this, strength
[4] Brandner R, Flatscher G, Ringhofer A, Shickhofer G, Thiel A. Cross Laminated
degradation due to the shear failure of CLT occurred earlier at 1% Timber (CLT): overview and development. Eur J Wood Wood Prod 2016;74:
drift, leading to the formation of local failure mechanism on the 2nd 331–51.
floor. On the other hand, the strength was maintained for specimen [5] Abrahamsen R.B., Malo K.A. Structural design and assembly of“ Treet”–A 14-storey
timber residential building in Norway. In: World Conference on Timber
WB up to 2% drift, after which the shift of failure mechanism from Engineering. Quebec City, Canada; 2014.
overall structure to local structure occurred. Significant difference in [6] Stazi F, Serpilli M, Maracchini G, Pavone A. An experimental and numerical study
the shear force transmission and the formation of compression struts on CLT panels used as infill shear walls for RC buildings retrofit. Constr Build Mater
2019;211:605–16.
were also observed in both specimens. [7] Aloisio A, Boggian F, Tomasi R. Design of a novel seismic retrofitting system for RC
4. By utilizing the proposed numerical models, the experimental structures based on asymmetric friction connections and CLT panels. Eng Struct
behavior of the specimens could be well evaluated. Specifically, the 2022;254:113807.
[8] Margani G, Evola G, Tardo C, Marino EM. Energy, seismic, and architectural
stiffness, strength, yielding conditions of the reinforcements were renovation of RC framed buildings with prefabricated timber panels. Sustainability
well captured. 2020;12:4845. https://doi.org/10.3390/su12124845.
5. A number of braced models were adopted to simulate the shear [9] Smiroldo F, Giongo I, Piazza M. Use of timber panels to reduce the seismic
vulnerability of concrete frame structures. Eng Struct 2021;244:112797.
behavior of the CLT shear wall of the hybrid systems. The proposed [10] Suga J., Ono M., Aoki K., Fukuhara T., Kurihara T., Maeda T. Timber Shear Walls
model was confirmed to agree well with the trend of the shear force for Seismic Retrofit of Reinforced Concrete Buildings. In: World Conference on
transmission in the CLT walls of the experiments. Timber Engineering. Seoul, Republic of Korea, 20–23 August; 2018.
[11] Kurihara, T., Fukuhara, T., Usami, T., Suga, J., Ishikawa, Y., Isoda, H., Mori, T. and
Kitamori, A. Seismic Retrofit of RC Frames Using Timber Shear Panels (Part 1:
The above research outcomes show that the numerical model pro­ Research Programs and Experimental Plans of LVL Wall). Summaries of Technical
posed in this study can contribute to more precise seismic design cal­ Papers of Annual Meeting Architectural Institute of Japan (in Japanese),
culations for CLT shear wall infilled RC frame with concealed steel plate Architectural Institute of Japan, pp. 687–688, 2015.
[12] Haba R, Kitamori A, Mori T, Fukuhara T. Development of CLT panels bond-in
and drift pin connections. However, further verifications are recom­ method for seismic retrofitting of RC frame structure. J Struct Constr Eng 2016;81:
mended because the findings are based on limited number of specimens, 1299–308.
given the fact that there are not many 2-storey structural experiments [13] Smiroldo F, Kallioras S, Sommacal G, Bournas D, Piazza M, Giongo I. Full-scale
testing of masonry-infilled RC frames retrofitted with cross-laminated timber
conducted previously. Further studies like developing a simpler nu­ panels. Eng Struct 2023;v. 294:116789. https://doi.org/10.1016/j.
merical model with fewer number of elements and conducting seismic engstruct.2023.116789.
analyses to evaluate its ability to perform under large deformations and [14] Sakashita, M., Maida, Y., Taki, Y., Araki, Y., Fukuyama, H., Nakagawa, T. and
Isoda, H. Experimental Study on Structural Performance of RC Frame
damping performances, should be conducted in order to fully promote Subassemblies with CLT Wing Walls (Part 1. Outline of test) Summaries of
such hybridization into practical use. Technical Papers of Annual Meeting Architectural Institute of Japan (in Japanese),
Architectural Institute of Japan, pp. 423–424, 2020.
[15] Maida, Y., Sakashita, M., Taki, Y., Araki, Y., Fukuyama, H., Nakagawa, T. and
CRediT authorship contribution statement Isoda, H. Experimental Study on Structural Performance of RC Frame
Subassemblies with CLT Wing Walls (Part 2. Force-deformation relationship and
Araki Yasuhiro: Supervision, Resources. Sakashita Masanobu: damage condition) Summaries of Technical Papers of Annual Meeting
Architectural Institute of Japan (in Japanese), Architectural Institute of Japan, pp.
Writing – review & editing, Project administration, Methodology,
425–426, 2020.
Investigation. Richard Yip Je Too: Writing – original draft, Methodol­ [16] Taki, Y., Sakashita, M., Maida, Y., Araki, Y., Fukuyama, H., Nakagawa, T. and
ogy, Investigation, Conceptualization. Isoda Hiroshi: Writing – review Isoda, H. Experimental Study on Structural Performance of RC Frame
& editing, Methodology. Subassemblies with CLT Wing Walls (Part 3. Strain of each component) Summaries
of Technical Papers of Annual Meeting Architectural Institute of Japan (in
Japanese), Architectural Institute of Japan, pp. 427–428, 2020.
Declaration of Competing Interest [17] Park R, Paulay T. Reinforced Concrete Structures, A Wiley-Interscience
Publication. John Wiley & Sons, Inc,; 1975. p. 287–8.
[18] AIJ (Architectural Institute of Japan), Design Method for Engineered Timber
The authors declare that they have no known competing financial Joints, 2017. (in Japanese).
interests or personal relationships that could have appeared to influence [19] AIJ (Architectural Institute of Japan), AIJ Standard for Structural Design of Timber
the work reported in this paper. Structures, 2006. (in Japanese)
[20] Kurihara, T., Fukuhara, T., Usami, T., Suga, J., Ishikawa, Y., Isoda, H., Mori, T. and
Kitamori, A. Seismic Retrofit of RC Frames Using Timber Shear Panels (Part 5–7).
Data availability Summaries of Technical Papers of Annual Meeting Architectural Institute of Japan
(in Japanese), Architectural Institute of Japan, pp. 193–198, 2016.
[21] Comittee of CLT Manual, Design and construction manual for CLT Buildings.,
Data will be made available on request. Tokyo: Japan Housing and Wood Technology Center, 2016. (in Japanese).
[22] ASCE/SEI 41–06, Seismic rehabilitation of existing buildings, American Society of
Acknowledgements Civil Engineers, 2007.
[23] SNAP Ver.8, “Finite Element Structural Analysis Software.,” KOZO SYSTEM.
[24] Lai S, Will GT, Otani S. Model for inelastic biaxial bending of concrete members.
This study is conducted by General Technological Development J Struct Eng 1984;110(11):2563–84.
Project, “Development of Design and Construction Technology for [25] Komuro D. Study on the Compressive Properties of Reinforced Concrete Columns
Mixed Structures Utilizing Newly Developed Wood-Based Materials” Using Ultra-High Strength Concrete (PhD Dissertation). Faculty of Engineering,
Kyoto University,; 2007. p. 67–77 (PhD Dissertation).
by National Institute for Land and Infrastructure Management, Min­ [26] AIJ (Architectural Institute of Japan), AIJ Standard for Lateral Load-carrying
istry of Land, Infrastructure and Tourism, Japan. Capacity Calculation of Reinforced Concrete Structures, 2018. (in Japanese).
[27] Mukai Tomohisa, Validation of equations for evaluating structural properties of
reinforced concrete members using experimental database”, No.175, Building
References Research Materials, Building Research Institute, 2016. (in Japanese).
[28] AIJ (Architectural Institute of Japan), AIJ Standard for Structural Design of
[1] Smiroldo F, Paviani I, Giongo I, Zanon S, Albatici R, Piazza M. An integrated Reinforced Concrete Boxed-Shaped Wall Structures, 2015. (in Japanese).
approach to improve seismic and energetic behaviour of RC framed buildings using [29] Ministry of Land, Infrastructure and Tourism, 2016, “Commentary on the
timber panels. Sustainability 2021;13:11304. https://doi.org/10.3390/ structural calculation based on the revised enforcement order, Building Standard
su132011304. Law.” Building Center of Japan, Ministry of Land, Infrastructure and Tourism,
[2] Isoda H., Okada H., Kawai N., Yamaguchi N. Research and development programs Housing Bureau and Building Research Institute, Tsukuba (in Japanese).
on timber structures in Japan. NIST Special Publication SP; 2002: p. 83–90.

18