Content uploaded by Dan-Adrian Corfar
Author content
All content in this area was uploaded by Dan-Adrian Corfar on May 17, 2024
Content may be subject to copyright.
© 2022. This manuscript version is made available under the CC-BY-NC-ND 4.0
license https://creativecommons.org/licenses/by-nc-nd/4.0/
Proof-of-concept finite element analysis of a novel hybrid
inter-module connection
D-A Corfar1, K D Tsavdaridis2
1 City, University of London, Northampton Square, London, EC1V 0EH, UK
2 City, University of London, Northampton Square, London, EC1V 0EH, UK
E-mail: Konstantinos.Tsavdaridis@city.ac.uk
Abstract. Inter-module connections (IMCs) play a crucial role in the structural behaviour of
steel Modular Building Systems (MBSs) by ensuring the vertical and horizontal load-transfer
paths between modules, yet existing designs display limited disassembly opportunities and lack
damage control features. This study introduces a novel, hybrid demountable IMC comprising
bespoke corner fittings, a resilient high-damping rubber core and a shape-memory alloy (SMA)
bolt. Proof-of-concept connection tests have been carried out using validated, continuum finite
element analysis (FEA) to determine the mechanical behaviour of the proposed IMC with respect
to the main deformation modes expected to occur in the joints of tall steel MBSs under the
combined effect of vertical and horizontal loading. Main findings show that both the HDR core
and the SMA bolt contribute effectively to the overall hybrid response of the IMC under tension
and combined compression and shear loading, preventing the formation of significant plastic
damage in the MBS’s corner fittings to facilitate reusability of modules.
1. Introduction
Steel Modular Building Systems (MBSs) are a modern method of construction (MMC) that has gained
a lot of traction due to well-established benefits such as halved construction times [1], lower capital
costs [2], reduced on-site labour [3], less construction waste [4] and safer work sites [5]. Moreover,
technological advancements achieved in Off-Site Manufacturing (OSM), Off-Site Construction (OSC)
and Building Information Modelling (BIM) have accelerated the expansion of steel MBSs within the
high-rise construction sector, where pre-finished corner-supported volumetric modules are typically
assembled into self-standing or hybrid MBSs [6,7] by means of inter-module connections (IMCs). A
particularly challenging scenario is met in tall self-standing steel MBSs, where connections between
modules are part of the building’s lateral load resisting systems (LLRS) and are subjected to more
complex loading scenarios due to height-induced amplification of lateral load effects [8–10].
Hence, IMCs play a crucial role in the structural behaviour of MBSs by ensuring the vertical and
horizontal load-transfer paths between modules, while also controlling the buildability of MBSs by
embracing principles of design-for-assembly/-disassembly (DfA/DfD). Given that existing IMCs
display limited disassembly opportunities and lack damage control features [11], there is reasonable
scope for developing a new generation of hybrid, demountable IMCs that contribute effectively to the
global damage distribution mechanism to improve the reuse chances of structural modules after
exposure to demanding lateral loads such as strong winds or earthquakes.
In this regard the authors introduce a novel, hybrid IMC comprising bespoke corner fittings, a
resilient high-damping rubber core and a pseudoelastic smart alloy bolt. The present study focuses on
the development of a high-fidelity proof-of-concept finite element analysis (FEA) model used to identify
the desired working mechanism and to characterise the mechanical behaviour of the proposed IMC
subjected to tension and combined compression and shear loading. An extensive study done by the
authors [12] provides further details regarding the ductility, energy dissipation and damping of the
proposed connection, as well as a parametric FEA focused on the variation of bolt preload, endplate
thickness, axial load magnitude and the vertical layout of the HDR core.
© 2022. This manuscript version is made available under the CC-BY-NC-ND 4.0
license https://creativecommons.org/licenses/by-nc-nd/4.0/
2. The hybrid inter-module connection (hybrid IMC)
The connection configuration illustrated in figure 1 consists of a high-damping rubber (HDR) core
secured between the corner fittings of volumetric modules by means of a preloaded pseudoelastic shape-
memory alloy (SMA) bolt.
Figure 1. Configuration of the novel
hybrid inter-module connection for a
corner joint.
a) exploded view
b) assembled joint
The system has been designed to fulfil the essential function of vertical and horizontal connectivity
between modules, while also acting as a self-centring, energy dissipating component under uplift and
combined compression and shear loading. The desired working mechanism of the novel IMC relies on
a hybrid mechanical behaviour that fosters the unique material properties of its constituent components.
The HDR core consists of a laminated elastomeric bearing (LEB) designed to withstand the
compressive stresses between the corners of modules by controlling the level of vertical displacement.
Due to the innate high shear strength of engineering rubber, the bearing accommodates large inter-storey
shear forces without failing, while the reduced shear stiffness alleviates the stress concentrations in the
vicinity of the joint. The lateral displacement in the joint is limited above certain sway thresholds by the
stiffening exhibited at high shear strains (S-shaped stress-strain curve), while the addition of carbon
black fillers in the rubber mix (filled rubber) supplies the enhanced hysteresis (i.e., energy dissipation
and damping) that characterises high-damping rubber (HDR) [13].
The shape-memory alloy (SMA) bolt complements the shear strength of the IMC by resisting the
inter-storey drifts through a combined bending and shear action, yet the main advantage lies in the
pseudoelasticity effect of SMAs which allows the bolt to develop large elastic strains during loading
that are largely recoverable upon unloading. SMAs are a class of smart materials with physical properties
similar to structural steel, capable of ‘memorising’ their shape due to the reversible transitions between
its two main phases (i.e., martensite and austenite) caused by a shear lattice distortion mechanism [14].
SMA components in austenitic phase manifest a superelastic (or pseudoelastic) effect, characterised by
a flag-shaped stress-strain curve with high load-unload stiffnesses and reduced permanent strains [15].
This effect equates to improved IMC resilience and demountability even after damage-inducing
loading scenarios, as opposed to conventional high-strength steel bolts which either fail or become
jammed due to permanent deformations. In addition, energy dissipation and damping are also supplied
through hysteresis during the cyclic transition between austenite-martensite phases of the smart alloy.
The plug-in lugs on the outer plates of the bearing act as last-resort stoppers during late loading stages
in the event of bolt fracture, ensuring a limited level of joint connectivity until the commencement of
retrofitting works.
SMA
bolt
HDR
core
Upper
module
Lower
module
© 2022. This manuscript version is made available under the CC-BY-NC-ND 4.0
license https://creativecommons.org/licenses/by-nc-nd/4.0/
3. Methodology
3.1. Framework
The proof-of-concept connection tests have been carried out using validated, continuum finite element
analysis (FEA) to determine the mechanical behaviour of the proposed IMC with respect to the main
deformation modes expected to occur in the joints of tall steel MBSs under the combined effect of
vertical and horizontal loading.
To calibrate the advanced material models required for capturing the viscoplastic behaviour of high-
damping rubber (HDR) and the pseudoelastic (superelastic) behaviour of austenitic shape-memory
alloys (SMAs), both materials were characterised through stress-strain and load-displacement curves
extracted from specific experimental testing detailed below.
The ultra-high damping rubber compound from Tun Abdul Razak Research Centre (TARRC -
http://tarrc.co.uk/) is adopted for the elastomeric layers of the LEB. Standard double-bonded shear tests
have been carried out at TARRC on cylindrical test pieces made of two rubber layers hot-bonded to
metal plates. The material model characterisation tests have been done at ambient temperature and the
test pieces have been subjected to triangular waveforms up to +/-200% shear strain for 6 cycles at 10%
strain rate (0.1/s) with a 2-minute wait before being subjected to triangular waveforms up to +/-200%
shear strain for 1 cycle over an amplitude range of 25% shear strain with stress relaxation segments of
2 minutes inserted between each step change. The 6 consecutive cycles of +/- 200% shear strain have
been done to account for the cumulative damage (also known as Mullins damage), such that the
calibration was done on an already stabilised stress-strain behaviour of the HDR. The load-relax-load
incremental steps are necessary for the calibration of the hysteresis parameters of the Bergström-Boyce
model, which requires test data for at least two different strain rates to accurately predict the desired
mechanical behaviour.
The SMA adopted is a Nickel-rich Nickel-Titanium alloy (Ni51-Ti49 % at.) supplied by the company
2SMArtEST Srl (https://2smartest.com/). The test samples have been heat treated at 500 °C for 30 min
to adjust the final transformation temperatures (austenite finish temperature ) and give proper
superelastic behaviour at a practical range of temperatures for real-life applications.
Isothermal mechanical tests for material characterisation consisting of pseudo-static cyclic tensile
tests with a monotonic strain controlled loading (
ε
" = 3·
10!"
s!#
) up to a given strain,
ε$%$
, followed by
complete load-controlled unloading (
σ
" = 12 MPa·
s!#
) have been performed at University of Calabria
on standard dog-bone samples in fully austenitic conditions (T = 20 °C). The tests have been carried out
using a universal testing machine (Instron E10000, USA) and strains have been measured by an
extensometer (accuracy class 0.5 ISO) with a gauge length of 10 mm.
3.2. Finite Element Models
The computational analyses in the present study have been conducted using the standard (implicit) solver
of commercial FEA package Abaqus® [16] including geometric, material and boundary nonlinearities.
The steel connection components are based on typical sizes of modular framing members reported
in the literature [17], while the design of the HDR core and SMA bolt has been based on provisions for
elastomeric bearings as per EN 1337-3 [18] and recommendations from experimental observations on
threaded SMA bars [14] respectively. The detailed drawings of all connection parts are illustrated in
figure 2.
© 2022. This manuscript version is made available under the CC-BY-NC-ND 4.0
license https://creativecommons.org/licenses/by-nc-nd/4.0/
Ceiling corner fitting
Floor corner fitting
HDR core
SMA bolt
Figure 2. Detailed drawings of connection components.
For contact interactions, normal behaviour is defined using the nonlinear, penalty-enforced, “Hard”
pressure-overclosure relationship, while for tangential direction, the default penalty friction formulation
based on the basic Coulomb friction model is adopted. The friction coefficient is taken as
µ = 0.3
,
assuming steel-to-steel interfaces that have been cleaned by wire-brushing or flame cleaning, with loose
rust removed as per EN-1090-2 [19].
General static steps are defined for each loading stage. The initial bolt preload is applied during the
first step in both tension and shear FEA. To provide initial stiffness without affecting the ductility and
re-centring capability, the bolt pre-stress is limited to 50% of the forward transformation start stress,
σ&'
, resulting in a preload force
F()*
= 80 kN. For the tension FEA setup, a displacement-controlled
vertical load is applied during the second step, while for the shear FEA a compressive axial load is
applied in the second step to simulate the effect of permanent gravitational loading, followed by a
displacement-controlled horizontal shear loading during the third step. The compressive axial load has
been taken as
N+,
= 97.5 kN, equal to 5% of the compressive yield capacity (
N-).,
) of the column’s
cross-section. Loading and boundary conditions are applied using reference points coupled to cross-
sections with distributed continuum coupling constraints, while the bolt load is defined through an
orthogonal surface passing through the middle of the bolt’s shank. The cyclic loading histories in figure
3 are based on the EN 15129 [20] load protocol using the maximum displacement as the design
displacement, corresponding to the displacement at which the stress in the bolt’s shank reaches the finish
transformation stress,
σ&/
. The maximum displacement,
Δ012
, has been determined through an initial
monotonic load sequence up to
Δ012
for each load case considered.
© 2022. This manuscript version is made available under the CC-BY-NC-ND 4.0
license https://creativecommons.org/licenses/by-nc-nd/4.0/
Tension load case
Shear load case
Figure 3. Cyclic loading histories
The steel and SMA parts have been meshed with first-order 8-node linear brick (3D solid) elements
(C3D8R) with reduced integration, while the steel plates in the HDR core have been meshed with the
equivalent hybrid elements (C3D8RH). The rubber layers have been meshed with fully integrated first-
order 8-node hybrid linear brick elements (C3D8H). The adequate mesh refinement has been explored
in terms of accuracy of results versus analysis efficiency through a rigorous sensitivity analysis. An
overview of the final FEA model setup and mesh details is illustrated in figure 4.
(a) tension load case
(b) combined compression and
shear load case
(c) mesh details
Figure 4. FEA model setup.
3.3. Material models
The FEA model includes three different materials, namely structural steel, high-damping rubber and
pseudoelastic SMA, each presenting specific characteristics in their constitutive models.
Structural steel has been modelled using the standardised bi-linear plus nonlinear isotropic hardening
model proposed by Yun and Gardner [21] due to its practicality. The model accurately depicts the stress-
strain response of steel based just on the three basic material parameters readily available in design
codes, namely the modulus of elasticity E, the yield strength
f3
, and the ultimate strength
f4
. Material
parameters used to characterise the elasto-plastic behaviour of steel grade S355 (table 1) have been taken
from Eurocode 3 [22] based on nominal material properties and nominal element thickness
t ≤ 40
mm.
ZSYMM
Fixed
Tension
Bolt Load
RP-2
RP-1
Compression
Fixed
RP-1
RP-2
RP-3
Shear
ZSYMM
Bolt Load
C3D8R
C3D8R
C3D8R
Stress
measurement
location
σ')$
σ')'
C3D8RH
C3D8H
C3D8RH
C3D8H
C3D8RH
© 2022. This manuscript version is made available under the CC-BY-NC-ND 4.0
license https://creativecommons.org/licenses/by-nc-nd/4.0/
Table 1. Nominal material properties for steel grade S355.
Steel
grade
ρa
E
ν
b
f3
f4
ε4c
(kg m5
⁄)
(Nmm6
⁄)
(-)
(Nmm6
⁄)
(Nmm6
⁄)
(%)
S355
7,850
210,000
0.3
355
490
16.53
a Density of steel
b Poisson’s ratio
c Ultimate strain
To capture the non-linear viscoelastic (non-LVE) response of high-damping rubber, the Bergström-
Boyce (BB) model based on Yeoh hyperelasticity has been used. The calibration of the material model
has been performed using the commercial material parameter extraction tool MCalibration [23], with
the final material parameters summarised in table 2.
Table 2. Material properties for high-damping rubber.
Model
Hyperelasticity coefficients
Non-linear viscoelasticity
(Bergström-Boyce)
NMADg
Fitness
SFc
Ad
me
Cf
Yeoh
C!"a
0.139
C#"
a
-0.014
C$"a
0.003
D!b
0
D#b
0
D$b
0
24.7
141.4
10.85
-0.99
8.2 %
a Stiffness coefficients
b Compressibility coefficients taken as 0 to approximate fully incompressible behaviour of elastomers
c Stress scaling factor
d Creep parameter
e Effective stress exponent
f Creep strain exponent
g Normalized Mean Absolute Difference/Error
The typical flag-shaped constitutive behaviour of pseudoelastic SMA under isothermal conditions
has been modelled through the combination of elasticity (for the austenite phase) and superelasticity (for
the martensite phase) using a built-in material model based Auricchio and Taylor’s work [24,25]. This
model has been widely adopted in the literature and has accurately reflected the mechanical behaviour
of Ni-Ti based SMA bars [26–28]. The material parameters used to model the superelastic SMA have
been extracted from the first load-unload cycle of the experimental stress-strain curve and are presented
in table 3.
Table 3. Material properties for pseudoelastic SMA.
E%a
(MPa)
E&b
(MPa)
ν%
c
(-)
ν&d
(-)
σ&!
e
(MPa)
σ&"
f
(MPa)
σ%!
g
(MPa)
σ%"
h
(MPa)
ε'i
(%)
87500
38438
0.33
0.33
525
615
180
90
3.64
a Young’s modulus of austenite.
b Young’s modulus of martensite.
c Poisson’s ratio of austenite.
d Poisson’s ratio of martensite.
e Forward (load) transformation start stress.
f Forward (load) transformation finish stress.
g Reverse (unload) transformation start stress.
i Reverse (unload) transformation finish stress.
h Total transformation strain.
© 2022. This manuscript version is made available under the CC-BY-NC-ND 4.0
license https://creativecommons.org/licenses/by-nc-nd/4.0/
3.4. FEA Validation
The FEA modelling techniques have been validated by simulating experimental tests on three separate
IMCs studied by Chen et al [29] in figure 5 (a), Lacey et al. [30] in figure 5 (b), and Chen et al [31] in
figure 5 (c) and the laminated elastomeric bearing (LEB) studied by Rahnavard and Thomas [32] in
figure 5 (d). In addition, the validation of the calibrated material models presented in the previous section
is also completed by simulating the experimental characterisation tests for rubber in figure 5 (e) and
SMA in figure 5 (f). Overall, the validation results presented herein demonstrate the high accuracy of
the present FEA modelling techniques, further supporting their use for predicting the mechanical
behaviour of the proposed hybrid IMC.
(a) IMC validation 1
(b) IMC validation 2
(c) IMC validation 3
(d) LEB model
(e) rubber DBST
(f) SMA tensile test
Figure 5. FEA validation studies.
4. FEA Results
4.1. Tension behaviour
The main deformation modes and stress states of the hybrid IMC under cyclic tension loading are
illustrated in figure 6.
The stress contour in figure 6 (a) shows the Mises stress in the bolt’s shank, corresponding to the
start of “yield”-like point, just as the stress overcomes the forward transformation start stress,
σ&' =
525:MPa
. As the applied tensile displacement increases (figure 6 (b)), the inelastic stress concentrates
uniformly in the bolt’s shank, up to the ultimate/maximum point corresponding to the stress contours in
the SMA bolt’s shank reaching the forward transformation finish stress
σ&/ =615:MPa
. Figure 6 (c)
shows the Mises plots limited by the yield strength of S355 steel,
f3=355:MPa
, indicating stress
concentrations near bolt holes in the steel endplates of corner fittings, while the actual values barely
exceed the yield limit. These results indicate that the hybrid IMC achieves the desired working
mechanism in tension, as the tensile axial loading is mainly resisted by the SMA bolt, while the
contribution of the endplates bending remains negligible.
This observation is further supported by the equivalent plastic strain (PEEQ) contours in figure 6 (d),
in which the non-zero values represent very limited regions where the steel has sustained plastic
© 2022. This manuscript version is made available under the CC-BY-NC-ND 4.0
license https://creativecommons.org/licenses/by-nc-nd/4.0/
deformation. The figures reveal that the plastic regions at the end of the third cycle of maximum tensile
loading are localised towards the inner sides of the IMC, at interface between endplates and the outer
plates of the HDR core, caused by the sudden section change as part of the endplates near bolt hole rims
cantilever over the larger hole in the HDR core.
(a) +δ$)3 = 1.11 mm
(b) +δ$)012 = 2.48 mm
(c) +δ$)012 = 2.48 mm
(d) 3rd +δ δ012
⁄ = 1
Figure 6. Mises stress plots and PEEQ contours under cyclic tension loading.
The hysteresis loop corresponding to the stress states and deformation modes under tension loading
is illustrated in figure 7 (a), while the main mechanical parameters are extracted in table 4. The loop’s
flag-shape is a good qualitative indication of the connection’s good energy dissipation capacity due to
the transformation-induced pseudoelastic deformations in the SMA bolt, while a self-centring effect
with limited residual displacements is also present. The blue dashed curve represents the initial
monotonic loading performed to determine the maximum displacement used for tailoring the cyclic
loading protocols for each of the tested specimens. The generic force-displacement curve is illustrated
in figure 7 (b) and comprises four characteristic stages governed by key changes in the stress-strain
response of the SMA bolt. Since stress state beyond the martensite transformation point equates to the
development of permanent inelastic strains in the SMA, it is desired to keep the mechanical response of
the connection within the first three stages to maintain the re-centring effect under tension loading and
the post-load demountability of the hybrid IMC. The overall shape and stages that characterise the force-
displacement behaviour of the hybrid IMC in tension are in good agreement with previous findings
reported in the literature [33–35].
Table 4. Mechanical parameters extracted from tension FEA.
F()"
a
δ()"
b
k*
c
F()+
d
δ()+
e
k**
f
F(),-.
g
δ(),-.
h
k***
i
(kN)
(mm)
(kN/mm)
(kN)
(mm)
(kN/mm)
(kN)
(mm)
(kN/mm)
44.27
0.08
610.44
168.05
1.11
119.80
181.07
2.48
9.50
a Initial tension strength.
b Initial displacement.
c First stage stiffness.
d Yield-like tension strength.
e Yield-like displacement.
f Second stage stiffness.
g Maximum tension strength.
i Maximum displacement.
h Third stage stiffness.
© 2022. This manuscript version is made available under the CC-BY-NC-ND 4.0
license https://creativecommons.org/licenses/by-nc-nd/4.0/
(a) Hysteresis loop
(b) Idealised F-d curve
Figure 7. Load-displacement response under cyclic tension loading.
4.2. Combined compression shear behaviour
The main deformation modes and stress states occurring in the IMC during combined compression-
shear loading are illustrated in figure 8. The contours plotted in figure 8 (a) show the Mises stress in the
hybrid IMC when the shear displacement reaches the onset of slipping,
δ')%'
. At this point, the initial
gap of 1.5 mm due to the 3 mm bolt hole clearance provided in the corner fittings is still unchanged,
while the two diagonally opposite shank regions which exceeded the forward transformation start stress,
σ&' =525:MPa
, are a sign of the bolt’s main deformation mode defined as combined bending and
shear. As the shear displacement increases and the bolt begins to slide, closing the clearance gap (figure
8 (b)), the stress in the bolt shank decreases below
σ&'
. At this stage the end of slipping displacement,
δ')7'
, is reached as the bolt shank has come into full contact with the bolt hole walls in the corner fitting.
Further loading beyond this point resumes the development of stress in the tension side of the shank up
until the forward transformation finish stress,
σ&/
, is eventually exceeded at the peak shear
displacement,
δ')012
(figure 8 (c)).
As explained in the case of pure tension loading, the hybrid IMC is intended to work only within the
pseudoelastic limits of the SMA bolt, hence
σ&/
governs the ultimate limit of the mechanical behaviour
during combined compression-shear loading as well. Overall, the stress evolution indicates that the
hybrid IMC achieves the desired working mechanism in combined compression-shear, as the stress
mainly develops in the SMA pin by virtue of the HDR core’s flexibility in shear which prevents the
direct bearing of corner fittings, keeping the stress levels in the module members negligible.
The equivalent plastic strain (PEEQ) contours illustrated in figure 8 (d) reveal that the plastic regions
at the end of the third cycle of peak shear displacement are localised around the rims of the bolt holes
in the corner fitting endplates and are caused by the high contact pressure that acts repetitively upon
these surfaces as the SMA pin slides from side to side during cyclic shear loading.
0
50
100
150
200
250
0 1 2 3
Tension Strength Ft(kN)
Displacement δt(mm)
Tension Strength Ft(kN)
Displacement δt(mm)
on-set of "yielding"
!!.# > !$!
!!.# > !$%
Stage II Stage III Stage IV
Stage I
##.& ##.' ##.()*
$
#.&
$
#.'
$
#.()*
Decompression
© 2022. This manuscript version is made available under the CC-BY-NC-ND 4.0
license https://creativecommons.org/licenses/by-nc-nd/4.0/
(a) +
δ')%'
= 1.00 mm
(b) +
δ')7'
= 3.75 mm
(c) +
δ')012
= 7.25 mm
(d) 3rd +
δ δ012
⁄ = 1
Figure 8. Mises stress plots and PEEQ contours under combined compression and cyclic shear.
To facilitate the qualitative and quantitative assessment of the cyclic shear response of the hybrid
IMC, the hysteresis loop is plotted in figure 9 (a), while corresponding mechanical parameters are
extracted in table 5. The generic force-displacement curve is illustrated in figure 9 (b) and comprises
four characteristic stages governed by the IMC slip behaviour and key changes in the stress-strain
response of the SMA bolt. As defined for the tension behaviour, Stage IV reaches beyond the upper
limit of the pseudoelastic response in the SMA pin and is considered outside of the safe working
mechanism intended for the proposed design.
The curves in figure 9 (c) show the force contribution of the SMA bolt and HDR core when loaded
up to a shear displacement of + 8 mm, while the corresponding maximum shear strengths of each
component,
F8&9
= 45.11 kN,
F:;.
= 12.43 kN, are in good agreement with the total
F<&*
= 57.02 kN
(error < 1%). From the graph, it is confirmed that in the current design the SMA pin provides the major
contribution to the total shear response of the hybrid IMC (i.e., 79.1%) while the HDR core plays a
secondary role.
Table 5. Mechanical parameters extracted from combined compression and shear FEA.
F/)0/
a
δ/)0/b
k*c
F/)1/d
δ/)1/e
k**f
F/),-.g
δ/),-.
h
k***i
(kN)
(mm)
(kN/mm)
(kN)
(mm)
(kN/mm)
(kN)
(mm)
(kN/mm)
24.25
1.00
24.25
26.27
3.75
0.73
54.43
7.25
8.05
a Onset of slipping strength.
b Onset of slipping displacement.
c First stage stiffness.
d End of slipping strength.
e End of slipping displacement.
f Second stage stiffness.
g Maximum shear strength.
i Maximum displacement.
h Third stage stiffness.
© 2022. This manuscript version is made available under the CC-BY-NC-ND 4.0
license https://creativecommons.org/licenses/by-nc-nd/4.0/
(a) Hysteresis loop
(b) Idealised F-d curve
(c) Individual contribution
Figure 9. Load-displacement response under combined compression and shear loading.
5. Concluding remarks
In the present study a novel hybrid IMC has been proposed and its tension and combined compression
and shear behaviours are studied and quantified by means of validated high-fidelity FEA. The key points
are summarised as follows.
The hybrid IMC’s tension behaviour is governed by the stress-strain response of the SMA pin,
demonstrating a robust cyclic performance with effective re-centring effects and is divided into four
separate load stages corresponding to transformation-induced stress changes suffered by the bolt.
The combined compression-shear behaviour of the proposed IMC relies on the effectiveness of the
resilient HDR core to undergo large shear straining without failing, while the main resistance is provided
by a combined shear-bending deformation state in the SMA pin’s shank.
The total shear response has been successfully decoupled, revealing the major contribution of the
SMA bolt regarding recoverable deformation, while the HDR core plays a supporting role.
Due to the effective contribution of each component to the combined hybrid response, the connection
succeeds in preventing the formation of significant plastic damage in the MBS’s corner fittings to
facilitate reusability of modules, while the resilience and demountability of the IMC are ensured
provided that the SMA pin is designed to work within the pseudoelastic domain.
Acknowledgements
The authors would like to thank Mr Carmine Maletta, Professor of Machine Design at University of
Calabria and CEO of 2SMArteST, and Mr Fabrizio Niccoli, Researcher at University of Calabria and
CTO of 2SMArtEST for their very valuable and continuous technical support. Their experience and
expertise have been instrumental to the material characterisation and fabrication of the pseudoelastic
shape-memory alloy bolts. The authors would also like to thank Mr Hamid Ahmadi, Head of Division
at TARRC, for his technical support with regards to the material characterisation and fabrication of high-
damping rubber cores.
6. References
[1] Gorgolewski M T, Grubb P J and Lawson R M 2001 Modular Construction using Light Steel
Framing: Design of Residential Buildings (Ascot: SCI)
[2] Ahn Y H and Kim K-T 2015 Sustainability in modular design and construction: a case study of
‘The Stack.’ Int. J. Sustain. Build. Technol. Urban Dev. 5 250–9
[3] Jabar I I, Ismail F, Mustafa A A 2013 Issues in Managing Construction Phase of IBS Projects
Procedia Soc. Behav. Sci. 101 81–9
[4] Nahmens I and Ikuma L H 2012 Effects of Lean Construction on Sustainability of Modular
Homebuilding J. Archit. Eng. 18 155–63
[5] Kamali M and Hewage K 2016 Life cycle performance of modular buildings: A critical review
Renewable Sustainable Energy Rev. 62 1171–83
[6] Hough M J and Lawson R M 2019 Design and construction of high-rise modular buildings
based on recent projects. Proc. Inst. Civ. Eng.: Civ. Eng. 172 37–44
© 2022. This manuscript version is made available under the CC-BY-NC-ND 4.0
license https://creativecommons.org/licenses/by-nc-nd/4.0/
[7] Lawson R M and Richards J 2010 Modular design for high-rise buildings. Proc. Inst. Civ. Eng.:
Struct. Build. 163 151–64
[8] Lacey A W, Chen W, Hao H and Bi K 2013 Effect of inter-module connection stiffness on
structural response of a modular steel building subjected to wind and earthquake load Eng.
Struct. 213 110628
[9] Wang Z and Tsavdaridis K D 2022 Optimality criteria-based minimum-weight design method
for modular building systems subjected to generalised stiffness constraints: A comparative study
Eng. Struct. 251 113472
[10] Farajian M, Sharafi P, Eslamnia H, Kildashti K and Bai Y 2022 Classification of inter-modular
connections for stiffness and strength in sway corner-supported steel modular frames J. Constr.
Steel Res. 197 107458
[11] Corfar D-A and Tsavdaridis K D 2022 A comprehensive review and classification of inter-
module connections for hot-rolled steel modular building systems J. Build. Eng. 50 104006
[12] Corfar D-A and Tsavdaridis K D 2023 A Hybrid Inter-Module Connection for Steel Modular
Building Systems with SMA and High-Damping Rubber Components Eng. Struct. submitted
[13] Lindley P B 2015 Engineering Design with Rubber (Hertford: TARRC)
[14] Fang C and Wang W 2020 Shape Memory Alloys for Seismic Resilience (Singapore: Springer)
[15] Abavisani I, Rezaifar O and Kheyroddin A 2021 Multifunctional properties of shape memory
materials in civil engineering applications: A state-of-the-art review. J. Build. Eng. 44 102657
[16] Dassault Systèmes SIMULIA Corp 2020 Abaqus CAE
[17] Lawson R M, Ogden R G and Goodier C 2014 Design in Modular Construction (Oxon: CRC
Press)
[18] BSI 2006 Structural bearings — Part 3: Elastomeric bearings (London: BSI)
[19] BSI 2018 Execution of steel structures and aluminium structures - Part 2: Technical
requirements for steel structures (London: BSI)
[20] BSI 2018 Anti-seismic devices (London: BSI)
[21] Yun X and Gardner L 2017 Stress-strain curves for hot-rolled steels J. Constr. Steel Res. 133
36–46
[22] BSI 2015 Eurocode 3: Design of steel structures - Part 1-1: General rules and rules for
buildings (London: BSI)
[23] PolymerFEM 2023 MCalibration
[24] Auricchio F and Taylor R L 1997 Shape-memory alloys: modelling and numerical simulations
of the finite-strain superelastic behavior Comput. Methods Appl. Mech. Eng. 143 175–94
[25] Auricchio F, Taylor R L and Lubliner J 1997 Shape-memory alloys: macromodelling and
numerical simulations of the superelastic behavior Computer Methods in Applied Mechanics
and Engineering 146 281–312
[26] Askariani S S, Garivani S, Hajirasouliha I and Soleimanian N 2022 Innovative self-centering
systems using shape memory alloy bolts and energy dissipating devices J. Constr. Steel. Res.
190 107127
[27] Qiu C, Liu J and Du X 2022 Cyclic behavior of SMA slip friction damper Eng. Struct. 250
113407
[28] Chen J, Wang W and Fang C 2022 Manufacturing, testing and simulation of novel SMA-based
variable friction dampers with enhanced deformability J. Build. Eng. 45 103513
[29] Chen Z, Wang J, Liu J and Cong Z 2020 Tensile and shear performance of rotary inter-module
connection for modular steel buildings J. Constr. Steel Res. 175 106367
[30] Lacey A W, Chen W, Hao H, Bi K and Tallowin F J 2019 Shear behaviour of post-tensioned
inter-module connection for modular steel buildings J. Constr. Steel Res. 162 105707
[31] Chen Z, Liu J, Yu Y, Zhou C and Yan R 2017 Experimental study of an innovative modular
steel building connection J. Constr. Steel Res. 139 69–82
[32] Rahnavard R and Thomas R J 2019 Numerical evaluation of steel-rubber isolator with single
and multiple rubber cores Eng. Struct. 198 109532
© 2022. This manuscript version is made available under the CC-BY-NC-ND 4.0
license https://creativecommons.org/licenses/by-nc-nd/4.0/
[33] Fang C, Wang W, He C and Chen Y 2017 Self-centring behaviour of steel and steel-concrete
composite connections equipped with NiTi SMA bolts Eng. Struct. 150 390–408
[34] Wang B, Nishiyama M, Zhu S, Tani M and Jiang H 2020 Development of novel self-centering
steel coupling beams without beam elongation for earthquake resilience Eng. Struct. 232
111827
[35] Wang B, Jiang H and Wang J 2019 Numerical simulation and behavior insights of steel
columns with SMA bolts towards earthquake resilience J. Constr. Steel Res. 161 285–95