Soil Dynamics and Earthquake Engineering 149 (2021) 106802

Contents lists available at ScienceDirect

Soil Dynamics and Earthquake Engineering

journal homepage: www.elsevier.com/locate/soildyn

An upgrade of existing practice-oriented FE design models for the seismic

analysis of CLT buildings
Vincenzo Rinaldi a, *, Daniele Casagrande b, Catia Cimini a, Maurizio Follesa c,
Massimo Fragiacomo a
a
Department of Civil, Architecture and Building and Environmental Engineering, University of L’Aquila, Via Giovanni Gronchi 18, 67100, L’Aquila, Italy
b
Institute of Bioeconomy, National Research Council of Italy CNR-IBE, via Biasi, 75, 38010, San Michele all’Adige, Trento, Italy
c
dedaLEGNO, Via degli Artisti 6, 50132, Firenze, Italy

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

Keywords: Practice-oriented finite element (FE) modelling strategies represent a fundamental tool for the seismic analysis
CLT buildings and design of Cross Laminated Timber (CLT) structures. Although substantial research has been undertaken
Seismic analysis concerning the seismic behaviour of CLT buildings, practice-oriented FE modelling strategies are still at an early
Practice-oriented
stage. This paper presents an upgrade of an existing practice-oriented FE design model for the seismic design of
Finite element
CLT structures. The upgrade is supported through the same modelling strategy presented by Christovasilis et al.
Numerical modelling
Initial configuration (2020) adding some features of analytical equations models presented by Casagrande et al. (2016).
2D elements are used both for the modelling of CLT panels and for mechanical connections, which are rep­
resented by a horizontal strip with a height smaller than 7% of the height of the panel. Analytical equations for
determining the elastic modulus of elasticity and the shear modulus of the horizontal strip are reported ac­
counting for both rocking and sliding behaviour for single- and multi-panel CLT shearwalls, including the effect
of the vertical loads by using a secant stiffness. The validation of the proposal is carried out in terms of the
shearwall lateral stiffness through the results of experimental tests on full-scale shearwalls published by other
authors and through a validation/comparison between a detailed non-linear model and the Christovasilis et al.
(2020) strategy on two multi-storey seismic-resistant lateral systems configurations with different amount of
vertical loads. Finally, a proposal for the initial layout of connections to reduce the iterative seismic design and
analysis process for CLT buildings is presented through a case study.

1. Introduction of angle-brackets for wall-to-foundation connections, Izzi et al. [7]

analysed the behaviour of annular-ringed shank nails used for CLT
In recent years, Cross-Laminated-Timber (CLT) buildings have been connections whereas, more recently, Dong et al. [8] performed an
characterised by a fast-growing in the market of wood-based structural extensive experimental campaign on hold-downs. D’Arenzo et al. [9]
systems ([1,2]). The lightness of structural panels and the capacity of investigated innovative CLT connections with high shear/tension ca­
mechanical connections to dissipate seismic energy make CLT buildings pacity, compared with [5], available in the market, whereas Marchi
a valuable alternative to other traditional structural types in seismic et al. [10] analysed a unique ductile steel solution (wall-to-wall,
prone areas ([3,4]). wall-to-floor, wall-to-foundation) developed for uni/bi-directional ac­
The current knowledge of the seismic behaviour of CLT structures is tions. At wall level, static and cyclic experimental tests on full-scale
the result of investigations carried out at three different levels: i) shearwalls were performed by Popovski et al. [11], Gravric et al. [12],
connection-, ii) wall-, and iii) building-level. Several researches have Flatscher et at [13]. and Hummel et al. [14] by varying the number of
been conducted on the characterisation of CLT connections in last ten connections and the amount of vertical loads, while Hristovski et al. [15]
years. Gavric et al. [5] investigated the behaviour of hold-downs and conducted shaking-table dynamic test on single-storey CLT shearwalls.
angle-bracket for wall-to-foundation/floor connections via static and At building level, shaking-table tests were carried out within the SOFIE
cyclic tests, Tomasi et al. [6] studied the monotonic and cyclic behaviour project [16] on a three-storey full-scale building with three

* Corresponding author.
E-mail address: [email protected] (V. Rinaldi).

https://doi.org/10.1016/j.soildyn.2021.106802
Received 20 November 2020; Received in revised form 18 March 2021; Accepted 30 April 2021
Available online 7 July 2021
0267-7261/© 2021 Elsevier Ltd. All rights reserved.
V. Rinaldi et al. Soil Dynamics and Earthquake Engineering 149 (2021) 106802

wooden elements. The major deformation contributions of single-panel

shearwalls with no openings are commonly due to the rocking and
sliding behaviour. A significant contribution of the CLT panel flexibility
may be conversely observed in case of single-panel shearwalls with
either door or window openings.
Two main rocking kinematic modes are observed in multi-panel
shearwalls [22,23]. In case of relative flexible vertical joints, a
coupled-panel behaviour with a significant slip along the vertical joint is
generally achieved (see Fig. 2a), whereas a single-wall behaviour with a
single centre of rotation is expected in case of relative stiff vertical joints
(see Fig. 2b).
Several analytical equations have been proposed to determine the
Fig. 1. (a) Sliding and (b) rocking deformation contribution in a single-panel lateral stiffness and the load-carrying capacity of single-panel CLT
CLT shearwall.
shearwalls [24]. The main differences among the proposed equations
regard the calculation of the panel length in contact with the foundation
and the behaviour of mechanical anchors (hold-down and angle
bracket), which may be characterised by either an uni- or a
bi-directional behaviour. Analytical models for multi-panel CLT shear­
walls were proposed by Gavric et al. [25], Tamagnone et al. [26],
Casagrande et al. [23], while Popovski et al. [27] presented an analytical
model for multi-panel balloon CLT shearwalls.

1.1. Practice-oriented modelling

Although substantial research has been undertaken dealing with the

seismic behaviour of CLT buildings, practice-oriented finite element
(FE) modelling strategies for the seismic analysis of CLT structures are
Fig. 2. Rocking kinematic modes in a multi-panel CLT shearwall: (a) coupled- still at an early stage. If a general consensus has been achieved on the
panel (CP), (b) single-wall (SW). modelling of CLT panels, their interaction (e.g. contact between panels,
contact with foundation, friction, non-linear behaviour of connections),
configurations and different openings size [17] and on a seven-storey still represents a key and challenging aspect in the use of practice-
full-scale building [16]. Flatscher et al. [13] performed a oriented linear-elastic FE models for the seismic analysis of CLT
shaking-table test on a three storeys full-scale building as well within the buildings.
SERIES project. More recently, van de Lindt et al. [18] performed a A first modelling strategy for the seismic analysis of CLT buildings
shaking-table tests on a two-storey CLT platform building. Cycle tests on was developed within the SOFIE project (2006–2007 [16]), with the aim
two-storey full-scale buildings were carried out by Yasumura et al. [19], to predict the dynamic response of a 3- and a 7-storey CLT building in a
Popovski et al. [20] and Matos et al. [21]. shaking-table test. These models were adopted to perform non-linear
The studies conducted at the different levels on CLT structures dynamic analyses, taking into account the non-linear hysteretic behav­
showed that mechanical connections are designed to dissipate the iour of mechanical connections and the non-linear mutual interaction
seismic energy while CLT wooden panels behave almost elastically. It between the panels and between the panels and the foundation. Other
was observed that three main deformation contributions characterise authors proposed similar strategies for non-linear analyses of
the in-plane behaviour of single-and multi-panel CLT shearwalls, light-frame timber buildings [28]. However, it is noteworthy to mention
namely rocking, sliding and panel deformation. Sliding (Fig. 1a) and that the complex implementation of these non-linear procedures and the
rocking (Fig. 1b) contributions are directly related to: i) the flexibility of need of input of several mechanical parameters have made these
mechanical connections (e.g. hold-down and angle-brackets) used to modelling strategies more suitable for research purposes rather than
anchor the CLT panel to the foundation or the floor below, ii) the panel design practice [3].
aspect ratio and iii) the amount of vertical loads; the panel deformation Depending on how CLT panels are schematised, the practice-oriented
conversely takes into account the shear and bending flexibility of the FE modelling strategies of CLT shearwalls can be divided into two main
categories, see Fig. 3. In the “Frame model” approach, two vertical and

Fig. 3. Modelling strategies for CLT panels and connections.

2
V. Rinaldi et al. Soil Dynamics and Earthquake Engineering 149 (2021) 106802

two horizontal frame elements are used to create a hinged frame with rocking may represent a major contribution in the lateral deflection of
the same width and height of the CLT panel, whereas a diagonal truss (or multi-storey CLT shearwalls system.
link) element is adopted to represent the lateral stiffness due to the The upgrade is supported through the same modelling strategy pre­
bending and shear panel deformation [29–31]. According to a “2D sented by Christovasilis et al. [36] adding some of the main features of
model” approach either multi-layer or homogenous equivalent ortho­ the FE models presented by Casagrande et al. [49]. The strategy pre­
tropic 2D elements are conversely used to represent the CLT panels [19, sented in this paper adopts: i) 2D elements for the modelling of both CLT
32–40]. For both the two approaches, the elastic properties (i.e. modulus panels and connections, ii) analytical formulations proposed by Casa­
of elasticity, shear modulus and Poisson ratio) of frame elements and 2D grande et al. [23,49] to properly take into account the rocking defor­
elements can be determined according to the analytical equations pro­ mation contribution in the determination of the effective modulus of
posed by Refs. [41–43]. elasticity and shear modulus of the 2D elements.
Three different approaches are commonly adopted for the modelling Since an iterative process might be required in seismic analyses to
of connections. According to a “Component level – CL” approach, a link define a proper initial layout of connections, a proposal for the deter­
(spring) element is used to represent each single fastener (e.g. screw or mination of the initial layout of connections in the practice-oriented
nail in vertical joints) or mechanical anchor (hold-downs and angle- modelling process is presented as well. The layout and typology of
brackets), see Refs. [18,19,30,33,34,37–40]. A “Semi-component Level connections may in fact strongly influence the CLT building dynamic
– SCL” approach is adopted when a single link element is used to model behaviour, to which, in turn, the magnitude of seismic loads is related
multiple connections or multiple mechanical anchors ([17,31,35,36]). too.
No spring elements are conversely adopted in the models where a More specifically, the objectives of the study are to:
“Phenomenological – PH” approach is used ([29,44]): the deformation
contribution of connections and mechanical anchors is directly englobed • include the rocking contribution due to mechanical anchors and the
in the elastic properties of either frame or 2D elements used for the CLT stabilising effect of the vertical load in the FE modelling strategy
panels. presented by Christovasilis et al. [36];
A “Frame model” was used by Ceccotti and Follesa ([17,45]) to • validate the upgraded modelling strategy by means of full-scale
model the CLT panels while spring elements were adopted to model the experimental tests on CLT shearwalls published by others and
interaction between the panels according to a SCL approach. An available in literature;
extensive comparison between the CL and PH approaches was presented • validate the upgraded modelling strategy presented in this study
by Pozza et al. [46], showing how the uplift forces on hold-downs are through a comparative study with a detailed modelling approach;
generally overestimated in a PH than in a CL approach. Sustersic et al. • demonstrate the improvement of the proposed upgrade model by
[44] introduced a detailed numerical model based on 2D elements for means of a comparative study with the practice-oriented modelling
CLT panels and discrete springs with either uni- or bi-directional strategy proposed by Christovasilis et al. [36];
behaviour for connections and mechanical anchors (CL approach). A • present a simplified process for determining the initial layout of
similar strategy was adopted by Rinaldin et al. [39] while more recently, connections in a FE model to be used in the seismic analysis of CLT
Mestar et al. [30] proposed an equivalent frame model (EFM) for CLT buildings.
walls with openings where the CL approach was used for modelling
connections and mechanical anchors. Schickhofer and Ringhofer [47] 2. An upgraded practice-oriented modelling strategy
presented a frame model in which an equivalent two-degree of freedom
spring (shear/rotation) is lumped at the shearwall base. 2.1. Description of the modelling strategy and of the kinematic modes
Follesa et al. [35] developed a 3D building model for the seismic
analysis of the 3-storey CLT building tested in the SOFIE project [16]; 2D Single- and multi-panel CLT shearwall are modelled using only
elements were used to schematise CLT panels while symmetrical truss orthotropic homogeneous 2D elements, see Fig. 4. The CLT shearwall of
elements were used for angle-brackets at the bottom of shearwalls and total height H is divided into two parts: i) a horizontal strip of height h
the vertical joints between the CLT according to an SCL. Floor and roof placed at the bottom of the shearwall reproducing the properties of the
panels were not explicitly modelled using a rigid in-plane constraint to connections and ii) an effective CLT panel dimension of height equal to
simulate the diaphragm behaviour of the floor and roof elements. (H-h) accounting for its deformability. All 2D elements have the same
Christovasilis et al. [36] adopted 2D elements to simulate both the thickness t as the CLT panel.
flexibility of CLT panels and the deformation contribution due to Effective moduli of elasticity along the vertical and horizontal di­
angle-brackets and vertical joints between the panels according to a PH rection, Eef,2D,v and Eef,2D,h are used to take into account the in-plane
approach. Both the models presented by Follesa et al. [35] and Chris­ axial deformation contribution of the CLT panel, whereas an effective
tovasilis et al. [36] were validated through the dynamic behaviour of in-plane shear modulus G2D is defined to consider the influence of
low-amplitude shaking table tests [48], neglecting the deformation torsional deformation of laminations on the in-plane shear deformability
contribution due to the rocking behaviour of the CLT shearwalls. of the panel. The analytical equations reported in Refs. [42] or [41,43]
The stabilising effect of the vertical load on the activation of the can be adopted for the determination of Eef,2D,v, Eef,2D,h and G2D
rocking behaviour of CLT shearwalls was taken into account in the
“Frame model” presented by Rossi et al. [31] and Casagrande et al. [49].
The elastic properties of the frame models were established adopting
analytical equations taking into account panel deformation, sliding and
rocking behaviour through an iterative approach.

1.2. Contribution

The work presented in this paper aims to present an upgrade of the

practice-oriented FE modelling strategy presented by Christovasilis et al.
[36] in order to properly consider the rocking deformation contribution,
given by the mechanical anchors and the vertical joints, and take into
account the stabilising effect of the vertical load both for single- and Fig. 4. Modelling of CLT shearwalls by means of 2D elements and horizontal
multi-panel shearwalls. Several research studies showed in fact, that area strips.

3
V. Rinaldi et al. Soil Dynamics and Earthquake Engineering 149 (2021) 106802

depending on the panel layout. A null value of Poisson ratios is adopted deformation of the multi-panel CLT shearwalls. In Ref. [36], in fact, the
according to Refs. [36,50] disaggregating the stress/deformation in the contribution of vertical joints was only related to the shear deformation
two orthogonal directions. of the panels.
The equivalent modulus of elasticity along the vertical direction Estrip In the upgraded model, the friction contribution is neglected as done
and shear modulus Gstrip of the horizontal bottom strip take into account in common practice and specified in Ref. [3] when the Force-Based
the deformation contribution due to rocking and sliding behaviour of Design (FBD) approach is adopted. In Refs. [35,44] an additional stiff­
each shearwall, see Fig. 5. The height h of the horizontal bottom strip is ness provided by the friction effect is included in the total stiffness of the
assumed to equal to or smaller than 0.07H. Small values of h lead to angle bracket considering the yield displacement of the weak fastener of
account for the effective deformability of the CLT panel of height H, and the shear-transferring connections and the total friction force assuming a
to adopt the analytical equations, later reported, which are valid for h friction coefficient equal to 0.4 for each wall.
close to zero. The equivalent modulus of elasticity along the vertical
direction Estrip and shear modulus Gstrip of the bottom horizontal strip are 2.2. Elastic lateral stiffness of single- and multi-panel CLT shearwalls
calculated to take into account the deformation contribution due to
rocking and sliding behaviour of each shearwall. The estimation of Estrip The estimation of the elastic stiffness associated with the sliding and
and Gstrip is carried out via analytical equations specific for the shearwall rocking mechanisms for single- and multi-panel CLT walls is based on
typology. the mechanical elastic behaviour under in-plane lateral loads.
The floor is modelled as a 2D element with its effective thickness s The rocking stiffness Kr is evaluated as the ratio of the 40% of the
placed at the top of the shearwall. As for the CLT wall, the thickness of overturning moment strength capacity MHD,y and the corresponding
2D element is assumed to equal to t, simulating the effective contact area relative lateral displacement d0.4HD,y according to ASTM 2126 [52]. The
between the floor and the CLT wall. The vertical elastic modulus is taken overturning moment strength capacity MHD,y is estimated considering
equal to the out-of plane stiffness of the floor (e.g. the modulus of the yield of the hold-downs for both single- and multi-panel CLT
elasticity perpendicular to grain Ec,90). shearwalls. In multi-panel shearwalls, the yield of the vertical joint
For single-panel shearwalls, a rigid rotation of shearwalls due to the should follow the yield of the hold-down to ensure the CP behaviour, as
rocking mechanism, which causes a linear increase of the lateral explain in Ref. [51].
displacement of the upper stories (i.e. cantilever behaviour) is supposed,
see Fig. 6a. Due to the smaller size of panels in a multi- than a single- Kr =
M0.4HD,y
(1)
panel wall, it is conversely assumed that in multi-panel shearwalls the d0.4HD,y ⋅H
panels do not significantly contribute to the out-of-plane bending of the Since a secant rather than the initial lateral stiffness is adopted, a
floor panels, which, for this reason, do not rotate (ϕ = 0, see Fig. 6b). No linear analysis can be conducted without considering the activation
rotation accumulation along the height of the buildings is supposed; force of rocking through an iterative approach. The contribution of the
therefore no influence of the rocking behaviour at the upper stories is activation force is automatically included in the determination of the
thus considered, according to a multi-panel shear-type behaviour. rocking capacity at MHD,y as shown in Fig. 7 for both single- and multi-
In order to ensure an high level of ductility and energy dissipation in panel shearwalls.
multi-panel shearwalls, a coupled-panel behaviour (CP) with the The effect of the activation/non-activation of the rocking mechanism
yielding of the vertical joints is assumed at each storey. Therefore, the is considered by comparing the stabilising moment contribution Mstab
modelling strategy presented in this section should be adopted only if due to the vertical loads and the overturning moment M0.4HD,y. The
the multi-panel shearwalls are designed according to a capacity design stabilising moment contribution is directly defined once the amount of
approach such that the CP behaviour is ensured and the SW behaviour is the vertical load per unit length w and the total length of the wall L in
avoided. Analytical expressions and design provisions for designing a single-panel shearwalls, and the length/number bclt/m of each single-
multi-panel CLT shearwall characterised by a CP behaviour where all panel in multi-panel shearwalls are known. The overturning moment
vertical joints yield, are available in Ref. [51]. capacity at the yielding of the hold-downs MHD,y is estimated via kine­
It is noteworthy to mention that vertical strips made with equivalent matic relationships specific for the shearwall type.
2D elements were used by Christovasilis et al. [36] to properly model the If M0.4HD,y is greater than Mstab, the secant stiffness Kr denes the
deformation contribution due to vertical joints in multi-panels shear­ rocking stiffness, on the contrary, the rocking is considered inhibited.
walls. In the upgraded model presented in this paper, the deformation This approach allows accounting for the only sliding deformation
contribution of the vertical joint is conversely englobed in the bottom contribution into the Gstrip if the vertical loads inhibit the rocking
horizontal strip. This choice was made to make the deformation mechanism. As basic assumption for both shearwall types, the angle-
contribution of vertical joint properly taken into account in the rocking brackets are assumed to work elastically along the tension direction.

Fig. 5. Multi-storey deformation: (a) sliding deformation, (b) rocking deformation.

4
V. Rinaldi et al. Soil Dynamics and Earthquake Engineering 149 (2021) 106802

M0.4HD,y − 0.5⋅w⋅L2
d0.4HD,y = ∑ ⋅H (3)
L ⋅kHD,up + kAB,up ⋅ n0AB x2i
2

where:

M0.4HD,y is the 40% of the overturning moment capacity at the hold-

down yielding
d0.5HD,y is the lateral top displacement related to M0.4HD,y
xi is the distance of the angle bracket i-th from the centre of rotation
Fy,HD,up is the hold-down yield tensile strength
kHD,up is the hold-down stiffness along vertical (tensile) direction
kAB,up is the angle-bracket stiffness along vertical (tensile) direction
uHD,y is the yield displacement of hold-down along vertical (tensile)
direction
nAB is the number of angle-brackets per each shearwall
w is the vertical load per unit length

According to Ref. [53] and as proposed in the new draft of Eurocode

8 [54], in multi-panel CLT shearwall, all vertical joints should yield
before hold-down and angle-brackets maximise the ductility and energy
Fig. 6. Kinematic modes for multi-storey (a) single and (b) multi-
panel shearwalls. dissipation as exhibit in a couple-panel behaviour, see Fig. 8b. Fixing the
rotation θHD,y,m as the ratio of the yield tensile deformation of the
hold-down and the length of each CLT panel bCLT, the overturning ca­
In a single-panel shearwall, the overturning moment MHD,y related to pacity at the hold-down yielding can be evaluated through Equation (4).
the yielding of the hold-downs (Equation (2)) is calculated by fixing the The multi-panel shearwall lateral displacement d0.4HD,y is estimated with
rotation θHD,y equal to the ratio between the yield displacement of hold- Equation (5).
down along the vertical (tensile) direction and the distance between the
centre of rotation and the distance of the hold-down, approximately equal

[ ( ) ]
1 kAB,up ∑
nAB
w⋅m⋅bCLT
MHD,y = bCLT ⋅ Fy,HD,up ⋅ 1 + x2 + (m − 1) ⋅ nSC ⋅ Fy,SC + (4)
b2CLT kHD,up 0 i 2

to L, from the centre of rotation. The shearwall lateral displacement d0.4HD,y

corresponding to M0.4HD,y can be evaluated through Equation (3).
M0.4HD,y − 0.5⋅w⋅b2CLT ⋅m
[ ] d0.4HD,y = [ ]⋅H (5)
uHD,y ⋅kAB,up ∑
nAB
w⋅ L kAB,up ∑nAB 2
MHD,y = L⋅ Fy,HD,up + x2
i + (2) b2CLT ⋅ kHD,up + (m − 1)⋅nSC ⋅kSC + ⋅ 0 xi
L2 2 b2CLT
0

Fig. 7. Increase of the yield capacity and of the corresponding secant lateral stiffness due to the stabilising effect of the vertical load on (a) a single and (b) multi-
panel CLT shearwalls (general behaviour).

5
V. Rinaldi et al. Soil Dynamics and Earthquake Engineering 149 (2021) 106802

Fig. 8. Rocking kinematic configurations used to determine the equivalent lateral stiffness of shearwalls at the yield condition of (a) single and (b) multi-
panel shearwalls.

where: h
Gstrip = Ks ⋅ (7)
L⋅t
bCLT is the length of each CLT panel ⎧
xi is the distance of the angle bracket i-th from the centre of rotation ⎪
⎪ 12⋅H2 ⋅h w⋅L2
⎪
⎨ Kr ⋅ 3 ≤ Eef,2D,v if Mstab = < M0.4HD,y
of the CLT panel L⋅t 2
Estrip = (8)
m is the number of panels ⎪
⎪ w⋅L2
⎪
nSC is the number of fasteners along the vertical joint ⎩ Eef,2D,v if Mstab =
2
≥ M0.4HD,y
Fy,SC is the shear strength of each fastener
kSC is the stiffness of each fastener along the vertical joint For multi-panel shearwalls:
⎧( )− 1
⎪ 1 1 h h m⋅w⋅b2clt
The lateral stiffness related to sliding mechanism Ks, for both ⎪
⎪
⎨ Kr Ks+ ⋅ ≤ ⋅Ks if Mstab = < M0.4HD,y
L⋅ t L⋅t 2
shearwall types, is calculated according to Equation (6) considering the Gstrip =
⎪
stiffness of angle-brackets and hold-down along the horizontal direction. ⎪
⎪
⎩K ⋅ h m⋅w⋅b2clt
s if Mstab = ≥ M0.4HD,y
L⋅t 2
Ks = nAB ⋅kAB,s + nHD ⋅kHD,s (6)
(9)
where:
Estrip = Eef,2D,v (10)

nHD is the number of hold-downs where:

kAB,s is the angle bracket stiffness along horizontal (shear) direction
kHD,s is the hold-down stiffness along horizontal (shear) direction H is the height of the shearwall panel
L is the length of the shearwall
2.3. Analytical equations for the elastic properties of bottom horizontal h is the height of the horizontal bottom strip
strip 2D elements t is the thickness of the CLT panel

The analytical equations adopted to define the modulus of elasticity Estrip and Gstrip are calculated at different levels and, depending on the
Estrip and shear modulus Gstrip of 2D elements representing the defor­ shearwall’s mechanical properties, can assume different values at
mation contribution of mechanical anchors and vertical joints in the different storeys.
bottom horizontal strip are presented.
For each single-panel wall, Estrip and Gstrip are calculated by consid­ 3. Validation of the proposed modelling strategy
ering the rocking and sliding mechanical behaviour, respectively. For
multi-panel walls, since the accumulation of rotations along the height 3.1. Validation of experimental tests on single-storey CLT shearwalls
of the multi-storey shearwall system is neglected and a shear-type published by others
behaviour is assumed, both rocking and sliding behaviour are eng­
lobed in the shear modulus of the strip Gstrip. The shear deformation of The validation of the modelling strategy presented in the previous
the horizontal strip does not in fact contribute to the rotation at the top section was carried out by comparing the values of the lateral stiffness of
of the shearwall, and for this reason, the lateral displacement of the single- and multi-panel CLT shearwalls obtained from the experimental
upper storeys is not influenced by the accumulation of rotations by the tests presented in Ref. [12] with those from finite element models.
lower storeys. Three wall panel configurations with dimensions equal to 2.95 × 2.95
The values of modulus of elasticity and shear modulus of 2D ele­ m were considered: single-panel walls (Test I – Fig. 9a), two-panel walls
ments of the strip are determined in order that the CLT shearwalls are with a half-lap vertical joint (Test II) and two-panel walls with spline ver­
characterised by the same rocking Kr and sliding Ks equivalent lateral tical joint (Test III), see Fig. 9b. All wall specimens were made of 5-layer
elastic stiffness of the shearwall according to Equations (7)–(10). CLT panels with a total thickness of 85 mm (17-17-17-17-17). The me­
For single-panel shearwalls: chanical properties in terms of elastic stiffness Kel, yield strength Fy, and

6
V. Rinaldi et al. Soil Dynamics and Earthquake Engineering 149 (2021) 106802

Fig. 9. Experimental shearwall tests presented in Ref. [12] (in meters): (a) single and (b) multi-panel shearwalls.

yield displacement uy of connections used in the experimental tests are

reported in Refs. [5,55] and summarised in Table 1. For angle-brackets a
Table 1
Mechanical parameters of hold-down, angle-brackets and screws in vertical bi-directional behaviour along vertical (tensile) and horizontal (shear) di­
joints ([5,55]). rection was assumed, while uni-directional behaviour along the direction
parallel to the vertical joints was adopted for half-lap/spline joints and
Hold-Down Angle bracket Screws Half-lap Screws Spline
along vertical (tensile) direction for hold-downs. The shear stiffness of the
Tension Tension Shear Shear Shear hold-downs was neglected.
Fy [kN] 40.44 19.24 23.01 3.24 4.85 The number of hold-down, angle-brackets, screws in the vertical
Fmax [kN] 48.33 23.47 26.85 5.25 7.33 joints and vertical load are reported in Table 2 for the validation tests.
Fult [kN] 38.79 18.74 21.48 4.20 5.86
Test II and III adopt partially threaded HBS screws.
Kel [kN/mm] 4.59 2.65 1.96 1.27 0.85
Kp1 [kN/mm] 0.68 0.41 0.23 0.10 0.09 The effective elastic moduli along the vertical Eef,2D,v and horizontal
Kp2 [kN/mm] − 2.77 − 0.86 − 1.60 − 0.13 − 0.45 Eef,2D,h direction as well as the effective shear modulus G2D of the 2D ele­
uy [mm] 8.81 7.26 11.74 2.55 5.70 ments representing the CLT panels in the FE models were equal to 6748
umax [mm] 20.30 17.69 28.51 23.50 34.37 MPa, 4622 MPa and 530 MPa respectively and were calculated according
uult [mm] 23.75 23.19 31.86 31.55 37.66
to equations proposed in Refs. [41,42], assuming an elastic modulus of
elasticity E0 and shear modulus Gmean of wooden boards equal to 11000
MPa and 690 MPa, respectively. The modulus of elasticity Estrip and the
Table 2
Number of hold-down, angle-brackets, screws in the vertical joints and vertical shear modulus Gstrip of the bottom horizontal strip were calculated ac­
load [12]. cording to the analytical equations reported in the previous section, as
shown in Table 3.
Shear wall Test ID Num. HD Num. AB Num. SC Vert. Load [kN/m]
The values of the lateral stiffness obtained from the experimental
Single-panel I.1 2 2 – 18.5 tests Kexp and from the finite element models Kproposal are reported in
I.2 2 4 18.5
Table 3. The values of Kproposal were assessed by means of a linear-elastic
–
I.3 2 4 – 9.25
analysis, as the ratio of a top horizontal force of 10 kN and the corre­
Multi-panel Half-Lap II.3 2 4 10 18.5
sponding lateral displacement of the shearwall. The scatter of the
Multi-panel Spline III.2 2 4 2× 10 18.5 experimental results of tests III.2-4-5 shows the intrinsic randomness of
III.4 2 4 2× 10 18.5 the experimental values, responsible for the same discrepancy of tests
III.5 2 4 2× 10 18.5
III.6 2 4 2× 10 0.0
I.1–2. Therefore, the results show a reasonable accuracy of the model­
ling strategy, especially for the multi-panel configurations.

Table 3
Comparison in terms of lateral stiffness between experimental tests and FE
models.
Test Gstrip Estrip Kproposal [kN/ Kexp [kN/ Var.
ID [MPa] [MPa] mm] mm] [%]

I.1 3.13 371.92 3.49 4.69 − 25.62

I.2 6.25 253.61 5.49 4.78 +14.95
I.3 6.25 128.70 4.62 4.97 − 6.98
II.3 3.66 – 4.36 4.48 − 2.66
III.2 3.10 – 3.77 3.5 +7.83
III.4 3.10 – 3.77 3.13 +20.57
III.5 3.10 – 3.77 3.77 +0.10
III.6 2.11 – 2.67 2.82 − 5.29

Fig. 10. MSS configuration 1 (L/H > 1) (in meters): (a) single and (b) multi-
panels shearwalls.

7
V. Rinaldi et al. Soil Dynamics and Earthquake Engineering 149 (2021) 106802

Table 5
MSS configuration 2 (L/H ≤ 1): layout of anchors. (*with double
strength and stiffness of the standard values).
Single-panel model Multi-panel model

Storey shearwall-3m shearwall-2.4m shearwall-2.8m shearwall-2.4m

1 2 HD 2 HD 2 HD 2 HD
2 AB 1 AB 3 AB* 5 AB
5 SC 5 SC

2 2 HD 2 HD 2 HD 2 HD
2 AB 1 AB 5 AB 4 AB
5 SC 5 SC

3 2 HD 2 HD 2 HD 2 HD
1 AB 1 AB 2 AB 2 AB
5 SC 5 SC

Fig. 11. MSS configuration 2 (L/H ≤ 1) (in meters): (a) single and (b) multi-
(shearwall-2.8 m 1:1 and shearwall-2.4 m 0.86:1). The use of different wall
panels shearwalls.
aspect ratios was followed to investigate configurations in which the rocking
mechanisms could be limited (L/H > 1) or substantial (L/H ≤ 1).
3.2. Validation and comparison
The two configurations of MSS were designed as single-panel and
multi-panels shearwall according to the provisions reported in the tim­
This section aims to compare the results obtained from the Upgraded
ber section of the draft document for the revision process of Eurocode 8
Proposal modelling strategy (UP) presented in this study with those ob­
[54]. Each configuration was analysed with and without vertical loads
tained from detailed non-linear model called Target Model (TM) and show
to examine different scenarios.
the upgrade obtained from the UP modelling strategy in comparison with
The two shearwalls were made with 5-layer CLT panel with a total
the original proposal made by Christovasilis et al. [36] (CH). Two in-plane
thickness of 100 mm. The CLT panel height was 2.8 m with a floor thick­
structural configurations made in CLT single-/multi-panel shearwalls are
ness of 200 mm for a total inter-storey height of 3.0 m. The same me­
taken into account.
chanical anchors (hold-down -HD and angle-brackets -AB) and screws SC
The structural configurations are composed of multi-storey shearwall
along the half-lap vertical joints presented in the previous section for the
system (MSS) with three storeys and two shearwalls. The first configuration
validation with the experimental tests were adopted. The number and
(see Fig. 10) has shearwalls with aspect ratio L/H, considering H as the
layout of hold-down, angle-brackets and screws at each storey are reported
effective wall-height, greater than one (shearwall-6m, 2.14:1, and
in Tables 4 and 5.
shearwall-4m, 1.43:1) whereas the second configuration (see Fig. 11) has
A uniform distributed vertical loads equal to 18 kN/m and 6 kN/m was
shearwalls with aspect ratio L/H either equal to one and smaller than one
applied at the inter-storeys and roof level, respectively. The lateral design
loads were distributed over the building’s height either according to the
Table 4 seismic masses and stories height.
MSS configuration 1 (L/H > 1): layout of anchors. (*with double
The CLT panels of three modelling strategies are modelled using 2D
strength and stiffness of the standard values).
elements with the values of effective elastic moduli Eef,2D,v, Eef,2D,h, and
Single-panel model Multi-panel model the effective shear modulus G2D reported in section 3.1.
In the TMs (Fig. 12a–c, Fig. 13a–c), the mechanical anchors (hold-
Storey shearwall-6m shearwall-4m shearwall-6m shearwall-4m down, angle-bracket and screws) were modelled by means of non-linear
1 2 HD 2 HD 2 HD 2 HD links (spring) whose mechanical behaviour were represented by the tri-
5 AB 3 AB 7 AB* 9 AB linear curve reported in Table 1. An uni-directional behaviour along the
5 SC 5 SC vertical–tensile direction was assumed for the hold-downs, while for angle-
2 2 HD 2 HD 2 HD 2 HD brackets a bi-directional tensile-shear behaviour was used. For the sake of
4 AB 2 AB 5 AB* 7 AB simplicity, equivalents angle-brackets are lumped in the middle of each
5 SC 5 SC
panel. The floor’s compressibility was accounted with gap-elements
3 2 HD 2 HD 2 HD 2 HD assuming an elastic modulus Ec,90 equal to 370 MPa and calibrated with
2 AB 1 AB 4 AB 3 AB the mesh-size. The base contacts were assumed rigid. Rigid horizontal
5 SC 5 SC
frames connect the two shearwalls in multi-panel configurations. Each
single-panel is hinged at the top centre with the rigid frame, while non-

Fig. 12. FE model layout, configuration 1 (L/H > 1): (a) Target Model for single-panel, (b) Upgraded Proposal and Christovasilis et al. [36] for single-panel, (c)
Target Model for multi-panel and (d) Christovasilis et al. [36] for multi-panel.

8
V. Rinaldi et al. Soil Dynamics and Earthquake Engineering 149 (2021) 106802

Fig. 13. FE model layout, configuration 2 (L/H ≤ 1): (a) Target Model for single-panel, (b) Upgraded Proposal and Christovasilis et al. [36] for single-panel, (c)
Target Model for multi-panel and (d) Christovasilis et al. [36] for multi-panel.

Table 6
MSS configuration 1 (L/H > 1): Elastic and shear moduli of the horizontal stripes (with vertical loads “W”, without vertical loads “W/O”).
Christovasilis et al. Upgraded Proposal

Single-panel model Multi-panel model Single-panel model Multi-panel model

Storey Modulus shearwall-6m shearwall-4m shearwall-6m shearwall-4m Modulus shearwall-6m shearwall-4m shearwall-6m shearwall-4m

1 Gstrip [MPa] 3.27 2.94 9.15 8.82 Gstrip [MPa] 3.27 2.94 W-9.15 W-8.82
W/O-3.02 W/O-2.90
Estrip [MPa] 6748 6748 6748 6748 Estrip [MPa] W-6748 W-6748 – –
W/O-34.55 W/O-41.45
Gsc [MPa] – – 1.22 1.22 – – – – –

2 Gstrip [MPa] 2.61 1.96 6.53 6.86 Gstrip [MPa] 2.61 1.96 W-5.52 W − 5.27
W/O-2.51 W/O-2.52
Estrip [MPa] 6748 6748 6748 6748 Estrip [MPa] W-6748 W-6748 – –
W/O-31.08 W/O-36.37
Gsc [MPa] – – 1.22 1.22 – – – – –

3 Gstrip [MPa] 1.31 0.98 2.61 2.94 Gstrip [MPa] 1.31 0.98 W-1.78 W-1.86
W/O-1.48 W/O-1.55
Estrip [MPa] 6748 6748 6748 6748 Estrip [MPa] W-65.57 W-64.95 – –
W/O-4.25 W/O-31.52
Gsc [MPa] – – 1.22 1.22 – – – – –

Table 7
MSS configuration 2 (L/H ≤ 1): Elastic and shear moduli of the horizontal stripes (with vertical loads “W”, without vertical loads “W/O”).
Christovasilis et al. Upgraded Proposal

Single-panel model Multi-panel model Single-panel model Multi-panel model

Storey Modulus shearwall-2.8m shearwall-2.4m shearwall-2.8m shearwall-2.4m Modulus shearwall-2.8m shearwall-2.4m shearwall-2.8m shearwall-2.4m

1 Gstrip [MPa] 2.80 1.63 8.40 8.17 Gstrip [MPa] 2.80 1.63 W-7.83 W-6.38
W/O-2.02 W/O-1.72
Estrip [MPa] 6748 6748 6748 6748 Estrip [MPa] W-6748 W-6748 – –
W/O-51.96 W/O-52.53
Gsc [MPa] – – 1.22 1.22 – – – – –

2 Gstrip [MPa] 2.80 1.63 7.00 6.53 Gstrip [MPa] 2.80 1.63 W-3.86 W-3.17
W/O-1.87 W/O-1.58
Estrip [MPa] 6748 6748 6748 6748 Estrip [MPa] W-1501.93 W-1281.73 – –
W/O-51.96 W/O-52.53
Gsc [MPa] – – 1.22 1.22 – – – – –

3 Gstrip [MPa] 1.40 1.63 2.80 3.27 Gstrip [MPa] 1.40 1.63 W-1.45 W-1.41
W/O-1.23 W/O-1.19
Estrip [MPa] 6748 6748 6748 6748 Estrip [MPa] W-73.10 W-79.17 – –
W/O-45.02 W/O-52.53
Gsc [MPa] – – 1.22 1.22 – – – – –

linear springs and contacts connect the downer frame to the panel bottom. shearwall. The vertical joint for the CH models was reproduced with
In single-panel configurations, kinematic constraints connect the shear­ area elements of height 50 mm and length 2.6 m, and shear modulus Gsc.
walls while floors were not explicitly modelled. The floors were modelled in both models with an elastic modulus
The material properties of the area elements for the UP and CH Ec,90 in the vertical direction equal to 370 MPa whereas the remaining
models (Fig. 12b–d, Fig. 13b–d) are reported in Tables 6–7. The elastic moduli were equal to 110 GPa.
moduli Estrip and the shear modulus Gstrip of the horizontal bottom strips The validation/comparison between the modelling strategies is car­
were calibrated for a height of 200 mm and length equal to the ried out by considering the lateral stiffness of the MSS of the models. Due

9
V. Rinaldi et al. Soil Dynamics and Earthquake Engineering 149 (2021) 106802

Table 8
Comparison of the lateral stiffness of the MSSs.
Configuration Shearwalls Model With vertical loads “W” Without vertical loads “W/O”

Ke [kN/m] Var. [%] Ke [kN/m] Var. [%]

“1′′ Single-panel Target Model 5110.56 - 2490.39 -

L/H > 1 Christovasilis et al. 5736.11 þ12.24 5736.11 þ130.33
Upgraded Proposal 5631.21 þ10.19 2739.00 þ9.98

Multi-panel Target Model 8258.39 - 4843.28 -

Christovasilis et al. 11698.29 þ41.65 11698.29 þ141.54
Upgraded Proposal 10181.16 þ23.28 6098.52 þ25.92

“2′′ Single-panel Target Model 1923.37 - 671.23 -

L/H ≤ 1 Christovasilis et al. 2345.18 þ21.93 2345.18 þ249.39
Upgraded Proposal 2201.22 þ14.45 717.40 þ6.88

Multi-panel Target Model 2658.01 - 1734.64 -

Christovasilis et al. 4292.07 þ61.48 4292.07 þ147.43
Upgraded Proposal 3130.83 þ17.79 1983.54 þ14.35

to the multi-degree of freedom nature of the physical problem, the MSS strong influence on the whole dynamic behaviour to which, in turn, the
lateral stiffness Ke of UP and CH was determined through a modal magnitude of seismic loads is related. The mechanical connections are
analysis by using the values of the modal mass M1 and period T1 cor­ designed to resist uplift and sliding actions obtained through a response
responding to the first mode of vibration of the MSS, see Eq. (11). spectrum linear analysis depending on the fundamental building period,
function of the shearwall rocking, and sliding stiffness, to which the
4⋅π⋅M1
Ke = (11) connections themselves contribute too. The typology, number, and
T12
layout of mechanical connections have to be defined based on an iter­
The lateral stiffness for TM was estimated via pushover analysis. A ative process so that the stiffness and strength capacities are consistent
secant lateral stiffness at 40% of the maximum load was used due to the with the seismic demand at connection-, wall- and building-level.
presence of non-linear elements (connections and contact). The pushover The definition of a proper initial layout of connections is essential in
analysis was carried out using a load profile derived from the UP’s modal implementing a FE model to limit the number of iterations in the anal­
analysis. An elastoplastic curve was drawn following the equivalent energy ysis and design process. In fact, mechanical connections strongly
elastic-plastic (EEEP) method. contribute to the response of the models much more than CLT panels to
The results are summarised and compared in Table 8, whereas Fig. 14 which they are connected.
display the elastic lateral stiffness of the MSSs and the elasto-plastic curves In this section, a proposal for determining the initial layout of con­
of the TMs. The values of the lateral stiffness of the TMs were assumed as nections in the practice-oriented modelling process is presented. The
reference values. proposal is based on a distinction of different groups of CLT multi-/
The scatters between TM and UP models show that the UP over­ single-shearwall systems characterised, in each direction, by a similar
estimates the lateral stiffness for all conditions in a range of +7% to +25%. amount of vertical load related to the orientation (perpendicular or
The reasons are associated with the accuracy given by a non-practice- parallel) and the tributary length of the supported floor (Fig. 15).
oriented approach which guarantees a more realistic force/deformation For each of these groups, a “reference Multi-/Single- Shearwall
distribution. However, based on linear-elastic laws, the proposed practice- system (rMSS)” is defined to represent all other multi-/single-shearwalls
oriented model highlights the ability to provide reasonable lateral stiffness of the same group called “non-reference Multi-/Single Shearwall system
for shearwall systems with different aspect ratio and presence/absence of (nrMSS)”. The connections of each rMSS are designed based on seismic
the vertical loads. actions obtained from a Lateral Force Method (LFM) where the funda­
The comparison between the UP and CH models shows that the CH mental period is calculated according to simplified equations reported in
strategy, due to the neglected effect of the rocking mechanism, predicts the standards as a function of either the building’s height or the number
better the configurations with vertical loads and shearwalls with L/H > of stories.
1. For each rMSS the values of Estrip and Gstrip are calculated from the
Minor errors are observable for single-panel shearwalls instead of analytical equations reported in Section 2 on the base of connections
multi-panels shearwalls. The variation with respect to TM is highlighted, (hold-down, angle-brackets and screws) obtained from the preliminary
especially for configurations with L/H < 1 without vertical loads with a LFM analyses.
maximum error of +249.39%, while UP reports a scatter of þ6.88%. The number and typology of the connections of each “reference
Hence, the proposed model can be considered acceptable independently shearwall” composing the rMSS are used for the design of the shearwalls
from the aspect ratio of the shearwalls and the amount of the vertical of the nrMSS at the same storey. The values of Estrip and Gstrip of the
loads. nrMSS shearwalls are estimated by assuming a linear variation of the
It is noteworthy to mention that, in seismic design, the benefits of an number of the angle-brackets with the total length L as reported in
accurate estimation of the elastic stiffness can be observed for structures Equations (12) and (13). In this way, the distance between the angle-
with a fundamental period located into the constant-velocity branch of brackets is kept constant. The typology of hold-down, screws in the
the pseudo-elastic spectrum (over the corner period) thanks to lower vertical joints and the vertical load are conversely kept constant at the
design accelerations. However, the use of the UP allows verifying this storey level. A FE model is then implemented. It is noteworthy to
condition directly by applying the proposed method. mention that Estrip and Gstrip are determined for each group of shearwall
systems at each storey in both directions applying Equations (7)–(10).
4. The initial layout of connections for an iterative seismic Fig. 16 and Fig. 17 display the general trends of the elastic moduli
analysis of CLT buildings Estrip and the shear moduli Gstrip varying the total length of the shearwall
for single-panel and multi-panel shearwalls, respectively.
The seismic design process of CLT structural systems is based on an The limit (L/H)lim defines the critical length of the wall for which the
iterative process. The layout and typology of connections may have rocking is active or inactive. The limit is evaluated by imposing equality

10
V. Rinaldi et al. Soil Dynamics and Earthquake Engineering 149 (2021) 106802

Fig. 14. Graphical comparison of the lateral stiffness of the MSS: elasto-plastic curves of the Target Model and the lateral stiffness of the Christovasilis strategy and
the Upgraded Proposal.

Fig. 15. Identification of the shearwalls system groups and the reference shearwall systems.

11
V. Rinaldi et al. Soil Dynamics and Earthquake Engineering 149 (2021) 106802

Fig. 16. Single-panel: General trend of the modulus of elasticity Estrip and shear modulus Gstrip of each horizontal bottom strip.

Fig. 17. Multi-panel: General trend of the modulus of the shear modulus Gstrip of each horizontal bottom strip.

in Equation (8) for single-panel shearwalls. For multi-panels, a proper nAB(L), nAB(m) is the integer number of angle-brackets for the
design of multi-panel shearwalls should ensure the couple-panel shearwall with increased or decreased length
behaviour thus, the activation of the rocking mechanism should be L is the increased or decreased length of the shearwall
attained in all conditions. m is the increased or decreased number of the panels of the shearwall
For single-panel shearwalls: nAB,ref is the number of angle-brackets of the “reference shearwall”
Lref is the length of the “reference shearwall”
nAB,ref ⋅L
nAB (L) = (12) bCLT,ref is the length of the single CLT panel of the “reference
Lref
shearwall”
For multi-panel shearwalls: mref is the number of the CLT panels of the “reference shearwall”
nAB,ref ⋅m
nAB (m) = (13)
mref

where:

12
V. Rinaldi et al. Soil Dynamics and Earthquake Engineering 149 (2021) 106802

Fig. 18. Case-study (in centimetres): Group H in red, Group M in green and Group L blue. (For interpretation of the references to colour in this figure legend, the
reader is referred to the Web version of this article.)

code period formulation of 0.19 s and a behaviour factor of 2.0.

Table 9 The sum of the lengths of the MSS is similar in both directions: 34.4
Properties and actions of the reference walls composing the rMSS. m along the x-direction and 34.0 m in the y-direction. Three main groups
Group Storey rMSS Lref [m] nAB nHD NEd [kN/m] VEd [kN] MEd [kNm] of MSS are defined, namely Group H, Group M and Group L corre­
H 2 X7 4.0 1 1 18 12.4 37.2
sponding to a high, medium and low amount of vertical load for each
1 X7 4.0 2 1 36 29.1 124.5 storey, respectively.
As shown in Fig. 18, Group H includes walls X7-X9, Group M includes
M 2 X11 3.75 1 1 9 10.9 32.7
1 X11 3.75 2 1 18 25.6 109.4 walls X1-X2-X3-X4-X5-X6-X7-X8-X10-X11-X12-X13, while walls
Y1–Y2–Y3–Y4–Y5–Y6–Y7–Y8–Y9–Y10–Y11–Y12 belong to Group L. The
L 2 Y2 3.5 1 1 0 10.4 31.2
1 Y2 3.5 2 1 0 24.3 104.0
identification of the rMSS for each group is based on a mean value of the
wall length, namely wall X7 for Group H, wall X11 for Group M and wall Y2
for Group L.
Only the rMSS X7, X11 and Y2 are designed by means of an LFM
analysis assuming a distribution proportional to the square length.
4.1. A case-study
Length Lref, number of angle-bracket nAB, number of hold-down at each
corner nHD, axial load NEd, shear load VEd and bending moment MEd
A case-study on a two-storey CLT building made with single-panel is
obtained from the LFM analysis are reported in Table 9 for each refer­
analysed in this section.
ence wall at each storey.
The building seismic mass is 50.64 ton for the first storey (3.6 kN/
Hold-downs and angle-brackets are designed respectively for the
m2) and 18.9 ton for the roof (1.2 kN/m2). The walls, with an inter-
overturning and the sliding mechanisms. The mechanical properties
storey height equal to 3.0 m, are made of 5-layer (30-20-20-20-30)
(strength Ryd, elastic stiffness Kser and yield slip uy of hold-down and
CLT panels (C24) with a total thickness of 120 mm. The inter-storey floor
angle-brackets) used in the case study are reported in Table 10. No yield
has a thickness equal to 200 mm. The structure, located in L’Aquila
slip of angle-brackets is considered since the tension behaviour is
(Italy), is designed for a seismic base-shear of 206 kN assessed with the
neglected. The strengths and the elastic stiffness are assessed within the
formulation provided in Eurocode 5 [56]. It should be noted that the
Table 10 values of elastic stiffness assessed with Eurocode 5 [56] are generally
Connections properties. much higher than the ones derived from experimental results, as
Anchors ID Storey Ryd[kN] Kser [kN/mm] uy [mm] confirmed in Ref. [5]. However, as reported in Ref. [35] they provide a
AB TTF200 2x14Ø4x60 2 16.6 8.70 – good estimation of the horizontal stiffness of the building when
TCF200 14Ø4x60 1 16.6 17.40 – compared with test results on whole buildings.
HD WHT340 14Ø4x40 2 24.2 24.35 0.99
Fig. 19 show the variation of the equivalent modulus of elasticity
WHT340 14Ø4x60 1 29.7 24.35 1.22 Estrip and shear modulus Gstrip with a parametric analysis by changing the
length of the shearwalls. The length L of the shearwall was varied

13
V. Rinaldi et al. Soil Dynamics and Earthquake Engineering 149 (2021) 106802

Fig. 19. Parametric curves through the variation of the shearwall length.

Table 11 between two values of Lmin and Lmax with an incremental step of length
Parametric analysis. ΔL, see Table 11, and consistently increasing the number of angle-
Group Storey rMSS Lmin [m] Lmax [m] ΔL [m] brackets (see Equation (13)). The red crosses in Fig. 19 show the posi­
tion of the reference shearwalls.
H 2 X7 3.0 5.0 0.25
1 X7 3.0 5.0 0.25
The parametric curves evidence the variation of the Estrip modulus,
which depends on the vertical loads. In shearwalls of Group H, the
M 2 X11 1.0 4.5 0.25
rocking mechanism is always inhibited due to high vertical loads instead
1 X11 1.0 4.5 0.25
of Group L. Group M, as aspected, could have shearwalls with the
L 2 Y2 1.5 4.0 0.25
1 Y2 1.5 4.0 0.25

14
V. Rinaldi et al. Soil Dynamics and Earthquake Engineering 149 (2021) 106802

Fig. 20. Modal Analysis: (a) T1 = 0.130 s, (b) T2 = 0.123 s, (c) T3 = 0.106 s.

rocking mechanism active or inactive depending on the length of the springs simplifying the modelling process, ii) deals with single- and
shearwall. multi-panels shearwalls, iii) includes the effect of the vertical loads into
The Gstrip modulus is independent of the presence of vertical loads the overturning mechanisms in linear-elastic modelling without itera­
(friction contribution neglected). The singularity point observable in tive steps, iv) shows an acceptable discrepancy between a detailed non-
walls X11-Y2 at the first storeys is related to the effective number of linear model in terms of global lateral stiffness for different scenarios in
angle-brackets since rounded integer values are adopted following terms of shearwall size and vertical loads and v) provides a simplified
Equation (12). approach for the definition of the initial layout of the mechanical con­
According to the proposal presented in this paper, a FE model was nections in seismic analysis.
implemented with SAP2000 v.14. The elastic properties of all CLT
panels were estimated with [41,42] finding the vertical, the horizontal Funding
and the shear moduli equal to 7457 MPa, 3913 MPa and 530 MPa,
respectively. The overall mass is considered lumped at each floor, while The authors received no financial support for the research, author­
the 2D elements have a null mass. The floor is modelled with a strip of ship, and/or publication of this article.
height equal to 200 mm with a Young modulus Ec90 in the gravity di­
rection equal to 370 MPa while the remaining moduli are kept with high Declaration of competing interest
values equal to 100 GPa, neglecting the deformation contribution in the
remaining directions. Rigid diaphragms connect all MSS at the floor The authors declare that they have no known competing financial
levels. For each shearwall, the values of Estrip and Gstrip were calculated interests or personal relationships that could have appeared to influence
according to Equations (7) and (8). the work reported in this paper.
Fig. 20 shows the vibration periods and the deformed shape of the 3D
structure. Three vibration periods can represent the global dynamic Acknowledgements
behaviour. Due to the structural regularity, about 88% of the participant
mass is associated with the two primary directions X and Y. The lower The authors would like to thank the anonymous referees for their
values of the periods calculated depend on the stiffness of the anchors valuable comments.
assessed with the EC5 formulations [56]. The first iteration allows
estimating for each wall the corresponding demand in terms of vertical Author statement
compression, shear and overturning based on the effective wall stiffness
and plan shape. This model represents the initial configuration for the Vincenzo Rinaldi: Conceptualisation, Methodology, Formal analysis,
following steps in the seismic analysis of multi-storey CLT building. Writing – original draft, Writing – review & editing. Daniele Casagrande:
Methodology, Writing – original draft, Writing – review & editing. Catia
5. Conclusions Cimini: Formal analysis, Writing – original draft, Maurizio Follesa:
Writing – original draft, Writing – review & editing, Supervision, Mas­
An upgrade of an existing practice-oriented FE design model for the simo Fragiacomo: Supervision.
seismic analysis of CLT buildings made in single- and multi-panel walls
is presented in this paper. Christovasilis et al. [36] model strategy References
combined with analytical equations presented by Casagrande et al. [49]
are implemented to propose a novel FE modelling strategy. [1] Hildebrandt J, Hagemann N, Thrän D. The contribution of wood-based
The proposal was validated through both experimental tests pub­ construction materials for leveraging a low carbon building sector in europe.
Sustainable Cities and Society 2017;34:405–18. https://doi.org/10.1016/j.
lished by others and a non-linear detailed finite element model on two
scs.2017.06.013.
in-plane configurations with different shearwalls length and vertical [2] Muszynski L, Larasatie P, Guerrero JE, Albee R. Global CLT industry in 2020:
loads. The same configurations were used to demonstrate the upgrade growth beyond the alpine region. In: Conference: Society of wood Science and
technology (SWST) international ConventionAt: virtual conference; 2020.
compared to Christovasilis et al. [36] model strategy. The outcomes, in
[3] Izzi M, Casagrande D, Bezzi S, Pasca D, Follesa M, Tomasi R. Seismic behaviour of
terms of global lateral stiffness, show scatters between 7% and 25% Cross-Laminated Timber structures: a state-of-the-art review. Eng Struct 2018;170:
compared with a detailed non-linear model and a significant reduction 42–52. https://doi.org/10.1016/j.engstruct.2018.05.060.
against Christovasilis et al. [49] strategy especially for multi-panel [4] Pei S, van de Lindt JW, Popovski M, Berman JW, Dolan JD, Ricles J, et al. Cross-
Laminated timber for seismic regions: progress and challenges for research and
shearwalls without vertical loads. A proposal for determining the implementation. J Struct Eng 2016;142(4). https://doi.org/10.1061/(asce)
initial layout of connections in the practice-oriented modelling process st.1943-541x.0001192.
is presented providing an approach to limit the number of iterations in [5] Gavric I, Fragiacomo M, Ceccotti A. Cyclic behaviour of typical metal connectors
for cross-laminated (CLT) structures. Materials and Structures/Materiaux et
the seismic analysis process. Constructions 2015;48(6):1841–57. https://doi.org/10.1617/s11527-014-0278-7.
The proposal strategy demonstrates its applicability as a practice- [6] Tomasi R, Smith I. Experimental characterization of monotonic and cyclic loading
oriented method since: i) adopts 2D elements instead of discrete responses of CLT panel-to-foundation angle bracket connections. J Mater Civ Eng
2015;27(6):04014189. https://doi.org/10.1061/(ASCE)MT.1943-5533.0001144.

15
V. Rinaldi et al. Soil Dynamics and Earthquake Engineering 149 (2021) 106802

[7] Izzi M, Flatscher G, Fragiacomo M, Schickhofer G. Experimental investigations and [31] Rossi S, Casagrande D, Tomasi R, Piazza M. Seismic elastic analysis of light timber-
design provisions of steel-to-timber joints with annular-ringed shank nails for frame multi-storey buildings: proposal of an iterative approach. Construct Build
Cross-Laminated Timber structures. Construct Build Mater 2016;122:446–57. Mater 2016;102-2:1154–67. https://doi.org/10.1016/j.conbuildmat.2015.09.037.
https://doi.org/10.1016/j.conbuildmat.2016.06.072. [32] Rinaldin G, Fragiacomo M. Non-linear simulation of shaking-table tests on 3- and
[8] Dong W, Li M, Ottenhaus LM, Lim H. Ductility and overstrength of nailed CLT hold- 7-storey X-Lam timber buildings. Eng Struct 2016;113:133–48. https://doi.org/
down connections. Eng Struct 2020;215. https://doi.org/10.1016/j. 10.1016/j.engstruct.2016.01.055.
engstruct.2020.110667. 110667. [33] Sustersic I, Fragiacomo M, Dujic B. Seismic analysis of cross-laminated multistory
[9] D’Arenzo G, Rinaldin G, Fossetti M, Fragiacomo M. An innovative shear-tension timber buildings using code-prescribed methods: influence of panel size,
angle bracket for Cross-Laminated Timber structures: experimental tests and connection ductility, and schematization. J. Struct. Eng. 2016;142(4):E4015012.
numerical modelling. Eng Struct 2019;197. https://doi.org/10.1016/j. https://doi.org/10.1061/(ASCE)ST.1943-541X.0001344 (United States).
engstruct.2019.109434. 109434. [34] Hummel J, Seim W. Displacement-based design approach to evaluate the
[10] Marchi L, Trutalli D, Scotta R, Pozza L. Macro-element modelling of a dissipative behaviour factor for multi-storey CLT buildings. Eng Struct 2019;201. https://doi.
connection for CLT structures26. Structures; 2020. p. 582–600. org/10.1016/j.engstruct.2019.109711. 109711.
[11] Popovski M, Schneider J, Schweinsteiger M. Lateral load resistance of cross- [35] Follesa M, Christovasilis IP, Vassallo D, Fragiacomo M, Ceccotti A. Seismic design
laminated wood panels. In: 11th world Conference on timber engineering 2010. of multi-storey CLT buildings according to Eurocode 8. Ingegneria Sismica 2013;4
WCTE; 2010. p. 20–4. (Special Issue on Timber Structures, 30).
[12] Gavric I, Fragiacomo M, Ceccotti A. Cyclic behavior of CLT wall systems: [36] Christovasilis IP, Riparbelli L, Rinaldin G, Tamagnone G. Methods for practice-
experimental tests and analytical prediction models. J Struct Eng 2015;141(11): oriented linear analysis in seismic design of Cross Laminated Timber buildings. Soil
04015034. https://doi.org/10.1061/(ASCE)ST.1943-541X.0001246. Dynam Earthq Eng 2020;128. https://doi.org/10.1016/j.soildyn.2019.105869.
[13] Flatscher G, Bratulic K, Schickhofer G. Experimental tests on cross-laminated 105869.
timber joints and walls. In: Proceedings of the Institution of civil engineers - [37] Demirci C, Málaga-Chuquitaype C, Macorini L. Seismic drift demands in multi-
Structures and buildings. 11th168; 2015. storey cross-laminated timber buildings. Earthq Eng Struct Dynam 2018;47(4):
[14] Hummel J, Flatscher G, Seim W, Schickhofer G. CLT wall elements under cyclic 1014–31. https://doi.org/10.1002/eqe.3003.
loading – details for anchorage and connection. In: Focus solid timber solutions - [38] Franco L, Pozza L, Saetta A, Savoia M, Talledo D. Strategies for structural
European Conference on Cross laminated timber. CLT; 2014. p. 152–65. modelling of CLT panels under cyclic loading conditions. Eng Struct 2019;198.
[15] Hristovski V, Dujic B, Stojmanovska M, Mircevska V. Full-scale shaking-table tests https://doi.org/10.1016/j.engstruct.2019.109476. 109476.
of XLam panel systems and numerical verification: specimen 1. J. Struct. Eng. [39] Rinaldin G, Amadio C, Fragiacomo M. A Component approach for the hysteretic
2013;139(11):2010–8. https://doi.org/10.1061/(asce)st.1943-541x.0000754. behaviour of connections in cross-laminated wooden structures. Earthq Eng Struct
[16] Ceccotti A, Sandhaas C, Okabe M, Yasumura M, Minowa C, Kawai N. SOFIE project Dynam 2013;42(13):2023–42. https://doi.org/10.1002/eqe.2310.
- 3D shaking table test on a seven-storey full-scale cross-laminated timber building. [40] Polastri A, Izzi M, Pozza L, Loss C, Smith I. Seismic analysis of multi-storey timber
Earthq Eng Struct Dynam 2013;42(13):2003–21. https://doi.org/10.1002/ buildings braced with a CLT core and perimeter shear-walls. Bull Earthq Eng 2019;
eqe.2309. 17(2):1009–28. https://doi.org/10.1007/s10518-018-0467-9.
[17] Ceccotti A. New technologies for construction of medium-rise buildings in seismic [41] Bogensperger T, Moosbrugger T, Silly G. Verification of CLT-plates under loads in
regions: the XLAM case. In: Structural engineering international: Journal of the plane. In: 11th world Conference on timber engineering 2010. WCTE; 2010. p. 1–9.
international Association for Bridge and structural engineering. IABSE; 2008. [42] Blass H, Fellmoser P. Design of solid wood panels with cross layers. In: 8th world
p. 156–65. Conference on timber engineering; 2004.
[18] van de Lindt JW, Furley J, Amini MO, Pei S, Tamagnone G, Barbosa AR, et al. [43] Flaig M, Blass HJ. Shear strength and shear stiffness of CLT-beams loaded in plane.
Experimental seismic behavior of a two-story CLT platform building. Eng Struct In: PROC-CIB; 2013.
2019;183:408–22. https://doi.org/10.1016/j.engstruct.2018.12.079. [44] Sustersic I, Dujic B. Simplified cross-laminated timber wall modelling for linear-
[19] Yasumura M, Kobayashi K, Okabe M, Miyake T, Matsumoto K. Full-scale tests and elastic seismic analysis. In: International Council for Research and Innovation in
numerical analysis of low-rise CLT structures under lateral loading. J Struct Eng Building and construction. CIB-W18; 2012.
2016;142(4):E4015007. https://doi.org/10.1061/(asce)st.1943-541x.0001348. [45] Ceccotti A, Follesa M, Lauriola MP. Quale fattore di struttura per gli edifici
[20] Popovski M, Gavric I. Performance of a 2-story CLT house subjected to lateral multipiano a struttura di legno? (in Italian). In: XII Convegno ANIDIS L’ingegneria
loads. J Struct Eng 2016;142. https://doi.org/10.1061/(asce)st.1943- sismica in Italia; 2007. Pisa.
541x.0001315. [46] Pozza L, Savoia M, Franco L, Saetta A, Talledo D. Effect of different modelling
[21] Matos FT, Branco JM, Rocha P, Demschner T, Lourenço PB. Quasi-static tests on a approaches on the prediction of the seismic response of multi-storey CLT buildings.
two-story CLT building. Eng Struct 2019;201. https://doi.org/10.1016/j. Int. J. Comput. Methods Exp. Measur. 2017;5(6):953–65. https://doi.org/10.2495/
engstruct.2019.109806. 109806. CMEM-V5-N6-953-965.
[22] Nolet V, Casagrande D, Doudak G. Multipanel CLT shearwalls: an analytical [47] Schickhofer G, Ringhofer A. The seismic behaviour of buildings erected in Solid
methodology to predict the elastic-plastic behaviour. Eng Struct 2019;179:640–54. Timber Construction - seismic design according to EN 1998 for a 5-storey reference
https://doi.org/10.1016/j.engstruct.2018.11.017. building in CLT, vol. 1. Austria: TU Graz; 2012.
[23] Casagrande D, Doudak G, Mauro L, Polastri A. Analytical approach to establishing [48] Ceccotti A, Follesa M. Seismic behaviour of multi-storey X-lam buildings. In: COST
the elastic behavior of multipanel CLT shear walls subjected to lateral loads. E29 Workshop on earthquake Engineering on timber structures. Coimbra; 2006.
J Struct Eng 2018;144. https://doi.org/10.1061/(ASCE)ST.1943-541X.0001948. [49] Casagrande D, Rossi S, Sartori T, Tomasi R. Proposal of an analytical procedure and
United States. a simplified numerical model for elastic response of single-storey timber shear-
[24] Lukacs I, Björnfot A, Tomasi R. Strength and stiffness of cross-laminated timber walls. Construct Build Mater 2016;102-2:1101–12. https://doi.org/10.1016/j.
(CLT) shear walls: state-of-the-art of analytical approaches. Eng Struct 2019;178: conbuildmat.2014.12.114.
136–47. https://doi.org/10.1016/j.engstruct.2018.05.126. [50] Turesson J, Berg S, Ekevad M. Impact of board width on in-plane shear stiffness of
[25] Gavric I, Popovski M. Design models for CLT shearwalls and assemblies based on cross-laminated timber. Eng Struct 2019;196. https://doi.org/10.1016/j.
connection properties. In: Proceedings of meeting, vol. 47. INTER- International engstruct.2019.05.090. 109249.
Network on Timber Engineering Research; 2014. [51] Casagrande D, Doudak G, Polastri A. A proposal for the capacity-design at wall-and
[26] Tamagnone G, Fragiacomo M. On the rocking behavior of CLT wall assemblies. In: building-level in light-frame and cross-laminated timber buildings. Bull Earthq Eng
WCTE 2018 - world Conference on timber engineering; 2018. 2019;17:3139–67. https://doi.org/10.1007/s10518-019-00578-4.
[27] Chen Z, Popovski M. Mechanics-based analytical models for balloon-type cross- [52] ASTM-E2126-19. Standard test methods for cyclic (reversed) load test for shear
laminated timber (CLT) shear walls under lateral loads. Eng Struct 2020;208. resistance of vertical elements of the lateral force resisting systems for buildings.
https://doi.org/10.1016/j.engstruct.2019.109916. 109916. West Conshohocken: ASTM International; 2019.
[28] Follesa M, Ni C, Popovski M, Karacabeyli E. Blind prediction of the seismic [53] Follesa M, Fragiacomo M, Casagrande D, Tomasi R, Piazza M, Vassallo D, et al. The
response of the neeswood capstone building. In: Proceedings of 10th world new provisions for the seismic design of timber buildings in Europe. Eng Struct
Conference on Timber engineering WTCE. Riva del Garda; 2010. p. 543–8. 2018;168:736–47. https://doi.org/10.1016/j.engstruct.2018.04.090.
[29] Pozza L, Scotta R, Trutalli D, Polastri A. Behaviour factor for innovative massive [54] CEN/TC250/SC8. prEN 1998-1-2:20XX: Eurocode 8: Design of structures for
timber shear walls. Bull Earthq Eng 2015;13:3449–69. https://doi.org/10.1007/ earthquake resistance. Part 1-2: Rules for new buildings. Draft document 30/04/
s10518-015-9765-7. 20. Brussels: Comité Européen de Normalisation; 2020.
[30] Mestar M, Doudak G, Caola M, Casagrande D. Equivalent-frame model for elastic [55] Gavric I, Fragiacomo M, Ceccotti A. Cyclic behavior of typical screwed connections
behaviour of cross-laminated timber walls with openings. In: Proceedings of the for cross-laminated (CLT) structures. European Journal of Wood and Wood
Institution of Civil Engineers - Structures and Buildings. 5th173; 2020. p. 363–78. Products 2015;73:179–91. https://doi.org/10.1007/s00107-014-0877-6.
[56] CEN. Eurocode 5: design of timber structure - Part 1-1: general – Common rules
and rules for buildings. 1995. Brussels.

16