Content uploaded by Dan-Adrian Corfar
Author content
All content in this area was uploaded by Dan-Adrian Corfar on Jun 29, 2024
Content may be subject to copyright.
EXPERIMENTAL TEST ON A HYBRID INTER-MODULE CONNECTION FOR
STEEL MODULAR BUILDING SYSTEMS
D.-A. Corfar1 & K. D. Tsavdaridis2
1
Doctoral Researcher, Department of Engineering, School of Science and Technology, City, University of
London, London, United Kingdom,
2 Professor of Structural Engineering, Department of Engineering, School of Science and Technology, City,
University of London, London, United Kingdom, Konstantinos.Tsavdaridis@city.ac.uk
Abstract: The context of the ongoing climate emergency has highlighted the pressing challenge of achieving
a resilient and sustainable built environment. Steel modular buildings have gained a lot of traction due to
noteworthy opportunities for disassembly and reuse harnessed by prefabrication, standardisation, and the use
of demountable connections, yet research has shown that the seismic resilience of steel modular buildings
remains problematic. In this regard, this study investigated the in-plane structural response to cyclic lateral
loading of an inter-module joint (IMJ) prototype equipped with a hybrid inter-module connection (IMC)
comprising of a high-damping rubber bearing and a high-strength steel bolt. The FEMA-SAC cyclic loading
sequence was conducted in a quasi-static fashion on a meso-scale IMJ subassemblage to understand the
effect of the high-damping rubber bearing on the seismic performance of the joint, by revealing its lateral load
capacity, deformation modes, and energy dissipation capacity. The re-centring effect of the joint due to the
high-damping rubber bearing was confirmed by the large recoverable displacement identified from the force-
displacement curve, which came at the expense of lower energy dissipation capacity. The IMJ prototype
reached the end of the loading programme with 85% of the peak strength and limited plastic damage sustained
by the frame members, meeting the qualifying drift angle capacity at strength degradation (4% storey drift
angle) for a special moment frame. Overall, it was shown that the hybrid IMC fulfilled its passive damage
control role by delaying the full contribution of the IMJ frame members until large storey drift amplitudes were
reached, effectively improving the reclaim and reuse opportunities in the event of an earthquake for the next-
generation, sustainable and resilient steel modular buildings.
1 Introduction
In the context of the current climate emergency, steel Modular Building Systems (MBSs) have emerged as a
sustainable modern method of construction (MMC) with compelling disassembly and reuse prospects
(Gunawardena and Mendis, 2022; Nguyen et al., 2023), yet natural hazards such as earthquakes can damage
the structural frame beyond recovery without due consideration of structural resilience. Recent studies have
emphasised the interdependence between sustainability and resilience in structural steel design (McConnell
and Fahnestock, 2015; Fujita et al., 2023; Grigorian, Sedighi and Mohammadi, 2023; Hao et al., 2023),
highlighting the increasing interest in damage control design principles. In line with the United Nations’
Sustainable Development Goals (in particular Goals 9 and 11) (United Nations, Department of Economic and
Social Affairs, 2015), it is worthwhile to further develop the technology of steel MBSs by mitigating the impact
of extreme events on the functionality, repairability, demountability, and continued service of these structures.
WCEE2024 Corfar & Tsavdaridis
2
The ever-increasing height of self-standing steel MBSs has drawn attention to the connections between
modules (i.e., inter-module connections or IMCs) and the critical influence of their mechanical behaviour on
the global structural response of sway modular structural systems to lateral load (Lacey et al., 2020; Farajian
et al., 2022; Wang and Tsavdaridis, 2022; Wang et al., 2023). So far, studies have investigated the
performance of IMCs based on capacity design criteria, in order to dissipate energy through the formation of
plastic hinges in predetermined locations within the structural system. Although this method is effective in
preventing collapse by controlling the global damage mechanism, the permanent damage sustained by
sacrificial structural elements hinders functionality and retrofitting prospects (McCormick et al., 2008).
In this regard, several review articles (Deng et al., 2020; Chen et al., 2021; Corfar and Tsavdaridis, 2022) have
emphasised the potential of employing replaceable energy dissipating components in IMCs to enhance their
mechanical response, yet accounts of this topic remain limited (Sultana and Youssef, 2018; Wu et al., 2019;
Jing et al., 2021; Sendanayake et al., 2021; Zhang, Xu and Li, 2022). In a recent paper (Corfar and Tsavdaridis,
2023), the authors have introduced a hybrid demountable IMC with high-damping rubber (HDR) and shape-
memory alloy (SMA) components, and characterised the axial tension and combined compression and shear
behaviours through validated finite element analysis (FEA). The proof-of-concept FEA study realised the first
stage of the investigation, showcasing the cyclic performance of the connection at small-scale micro-level, yet
further research is necessary in order to determine the large-scale, macro-level stress evolution and force-
transfer mechanisms between the interfaces of modular structural systems equipped with the hybrid IMC.
Hence, this paper introduces the second stage of the investigation, which aims to study the in-plane structural
response to cyclic lateral loading of an inter-module joint (IMJ) prototype equipped with a hybrid inter-module
connection (IMC) comprising of a high-damping rubber (HDR) bearing and a high-strength steel (HSS) bolt.
The experimental test has been carried out on a meso-scale joint sub-assemblage made of full-scale structural
members to avoid potential issues pertaining to scaling effects. The IMJ prototype has been subjected to cyclic
lateral loading using the standard FEMA-SAC loading sequence to reveal its aseismic behaviour and
understand the effect of the HDR bearing on a bolted inter-module connection, as well as the contribution of
the hybrid IMC to the overall structural performance of the IMJ assembly.
2 The hybrid inter-module connection
The assembly illustrated in Figure 1 depicts a corner IMJ assembled with a laminated elastomeric bearing
(LEB) as a core clamped between the steel box corner fittings of volumetric modules by means of a bolt
assembly. The size of the box corners was restricted to that of the connecting frame members with a flush
finish to allow members to be closely fitted together at external or internal IMCs where more than two modules
are connected, without increasing the horizontal gaps between modules which would act adversely to the
stiffness of the structure. The connection has been designed to fulfil the essential functions of vertical and
horizontal connectivity between modules, while the centred alignment of the member cross-sections to the box
corners eliminates the unfavourable effect of eccentric loads caused by offsets. Axial compression is
transferred between the corner posts through the laminated elastomeric bearing made with steel reinforcing
plates to control the level of vertical displacement, whereas tensile axial force is resisted by the bolt assembly.
Horizontal shear forces are transferred through a combined mechanism of friction between the faying steel
surfaces and the superposition of shear resistances of the rubber layers and bolt rod, while the interlocking
pins prevent accidental sliding.
To improve the energy dissipation capacity of the connection, the bearing has been fabricated with filled (high-
damping) rubber instead of low-damping (unfilled) rubber, as the presence of high percentages of carbon black
filler in HDR results in a much more pronounced hysteresis at the lower working shear strain levels (in the
range of 50%-100%) due to the breakdown of carbon filler chain networks (Lindley and Gough, 2015). By
comparison, unfilled (low-damping) rubber typically exhibits hysteresis due to strain-crystallisation at
extensions greater than 200% strain. Moreover, filled (high-damping) rubber shows a higher initial shear
stiffness, ensuring that the IMC is not easily excited during more frequent low-magnitude earthquakes or
common low-intensity wind loads, reserving the available supplemental damping for stronger lateral loads.
WCEE2024 Corfar & Tsavdaridis
3
Figure 1. Configuration of the hybrid inter-module connection.
3 Experimental programme
3.1 Details of the meso-scale test arrangement
The meso-scale inter-module joint (IMJ) prototypes (Figure 2) are based on the beam-column subassemblage
with column loading (J/C) recommended by Lacey et al. (2022). The IMJ prototype consists of top and bottom
steel beam-column subassemblages connected by means of the hybrid inter-module connection (IMC). The
choice for this type of frame has been attributed to its ability to realistically replicate the anticipated points of
inflection in the deformed shape of unbraced modular frames subjected to earthquake loading, including the
contribution of the connecting framing members and that of the beam-column joints.
Figure 2. Schematic of meso-scale IMJ prototype.
3.2 Dimensions and material properties of the IMJ prototype components
Steel beam-column sub-assemblages
The beam-column frames are made of standard hollow sections joined to box corner fittings by complete joint
penetration (CJP) groove welds, while cap plates are attached to each member’s end through all around fillet
welds. The role of the endplates is to help realise the bolted connections between the samples and test rig.
“Exploded” view of IMC components Assembled IMC
Top module
(bottom corner)
High-damping
rubber bearing
Bolt
assembly
Bottom module
(top corner)
Top module
Bottom module
Overall dimensionsIMJ subassemblage of a full-scale modular frame
Ceiling beam
Top post
Floor beam
Hybrid IMC
Bottom post
WCEE2024 Corfar & Tsavdaridis
4
The box corner fittings have been fabricated from 15-mm-thick steel plates joined together by CJP groove
welding. Detailed drawings of the steel frames are illustrated in Figure 3, while material properties of the frame
members based on mill test certificates are provided in Table 1.
Figure 3. Details of the steel beam-column frames (units: millimetres).
Table 1. Material properties of steel.
Frame element
Sectional properties
(mm x mm x mm)
Steel
grade
Yield
Strength
(
𝐍/𝐦𝐦!
)
Ultimate
Strength
(𝐍/𝐦𝐦!)
Elongation
(%)
Top and bottom posts
150 x 150 x 10
S355J2H
367
514
30
Ceiling beam
150 x 150 x 8
S355J2H
506
539
33.4
Floor beam
150 x 200 x 8
S355J2H
363
519
29
High-damping rubber core
The laminated elastomeric bearing has been fabricated using a proprietary high damping rubber (HDR)
compound. Double-bonded shear (DBS) testing has been carried out to characterise the material properties
of the rubber. Sinusoidal waveforms of 0.5 Hz frequency at ±1, ±2, ±5, ±10, ±20, ±50, ±100, ±150, ±200 %
shear strains have been applied to the test pieces at a constant temperature of 23 +/-2 °C. Six cycles have
been performed at each strain amplitude to capture the stress-softening behaviour (Mullins effect (Mullins,
1969)) typically exhibited in rubber during cyclic loading at a constant amplitude. As the Mullins effect is most
pronounced during the first two cycles and becomes negligible after six-to-ten cycles (Burtscher and Dorfmann,
2004), the data extracted for material characterisation is taken from the sixth loading cycle. Figure 4 (a) shows
the results for the cyclic DBS test used to determine the representative material properties listed in Table 2.
The highly nonlinear load-deflection behaviour characteristic for HDR is unmistakeable. The shear modulus,
G, the effective shear stiffness,
k"##$%
, and the effective damping ratio,
ξ"##$%
, have been determined as per EN
Top BC frame
Bottom BC frame Ceiling corner fitting
Floor corner fitting
WCEE2024 Corfar & Tsavdaridis
5
15129 (BSI, 2018) using equations (1) – (3), where
τ&'(
is the maximum nominal shear stress,
γ&'(
is the
maximum nominal shear strain,
F)
and
F*
are the maximum and minimum shear force values in the chosen
cycle with
d)
and
d*
as the corresponding displacements, and
H
is the energy dissipated per cycle.
G = τ!"#/γ!"#
(1)
k$%%.' = (F(− F))/(d(− d))
(2)
ξ$%%.' =2H
πk$%%.' (d(− d))*
(3)
Table 2. Mechanical properties of the high damping rubber.
Property
Value
Hardnessa (IRHDb)
86
Shear modulus, Gc (MPa)
0.61
Effective damping ratio,
ξ"##$%
c (%)
18.46
a based on shear modulus at 5% shear strain
b International rubber hardness degree
c at 100% shear strain
The variation of shear stiffness and damping properties of the HDR with strain amplitude is provided in Figure
4 (b). It can be seen that the shear modulus assumes values between 16.97 MPa and 0.44 MPa, falling abruptly
for strains up to 10% and stabilising after about 50% shear strain. This is in good agreement with the well-
known phenomenon of decreasing modulus with increasing deformation due to the breakage of weak filler
clusters within the filled rubber’s microstructure, known as the Payne effect (Payne, 1962). On the other hand,
the effective damping ratio’s dependence on the shear strain magnitude is a lot less prominent, ranging
between 17.22% and 22.06% for the strain range considered herein.
Figure 4. Results of the DBS tests: (a) hysteretic curves with sixth cycle highlighted, (b) variation of secant
shear modulus and equivalent viscous damping ratio with strain amplitude.
To fabricate the bearing, steel shims and rubber sheets have been alternated in a custom mould, where the
rubber vulcanisation (curing) and bonding between rubber and metal surfaces occur simultaneously at
elevated temperatures under pressure (Lindley and Gough, 2015). For the rubber to metal bonding process,
a primer layer and proprietary bonding adhesive are applied to the degreased and sand-blasted surfaces of
the steel plates. During vulcanisation, strong crosslinks are formed between the molecules of the viscous liquid
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
-20 -10 010 20
Shear force Fs(kN)
Shear displacement ds(mm)
0
8
16
24
32
40
0
5
10
15
20
25
1 2 5 10 20 50 100 150 200
Effective damping ratio ξeff.b (%)
Shear modulus G (MPa)
Maximum shear strain γmax (%)
Shear modulus
Effective damping ratio
(a) (b)
WCEE2024 Corfar & Tsavdaridis
6
elastomer, turning the compound into the solid polymer with high elasticity and strength required for
engineering applications (Gent, 2012).
The bearing consists of two outer steel plates (150 mm x 150 mm x 15 mm), four rubber layers (150 mm x 150
mm x 4 mm) and three steel shims (150 mm x 150 mm x 3 mm), designed to achieve a shape factor of S =
8.85 to limit the vertical displacement. A central hole (40 mm diameter) has been cut through the plates to
accommodate the bolt assembly. The hole diameter was purposefully made larger than the hole diameter in
the corner fittings to ensure that the bolt does not interfere with the inner surface of the bearing during the
lateral cyclic loading imposed on the IMJ prototype. Blind holes (15 mm diameter x 10 mm deep) have been
milled on one side of the outer plates to push-fit the four cylindrical steel lugs. Loctite has been used to ensure
the lugs on the bottom surface do not detach and fall during transportation and testing. The design details of
the fabricated bearings are illustrated in Figure 5.
Figure 5. Details of the HDR core (units: millimetres).
Bolt assembly
A standard M24 x 150mm full-thread hexagon head bolt made of class 8.8 high-strength steel has been used
to connect the box corner fittings. The choice for a typical structural bolt has facilitated the focus of this research
on the effect of the rubber bearing over the mechanical behaviour of the tested IMJ prototype.
3.3 Details of the test apparatus and loading sequence
The test rig illustrated in Figure 6 has been designed to accommodate the bi-axial loading of the IMJ sub-
assemblage. A test commences with the application of a force-controlled compressive axial load of 100 kN
over 1 minute through the vertical hydraulic jack to simulate the gravitational load from the self-weight of upper
building levels. Based on a realistic on-site installation scenario of a modular building, a snug-tight condition
is achieved in the hybrid IMC before initiating the axial load step. The axial load is then maintained for 10
minutes to account for the potential loss of bolt pre-tension due the compression set in the rubber bearing.
The applied axial load is equivalent to ca. 5% of the compressive yield capacity of the column’s cross-section,
N+$,-
, as defined in Eurocode 3, Part 1 (BSI, 2015) for members not susceptible to local buckling failure. The
axial load is maintained during subsequent test stages, while the sample is subjected to cyclic lateral load
Bearing mould Fabricated bearing
Dimensions of the bearing
WCEE2024 Corfar & Tsavdaridis
7
using the horizontal actuators. Following the prequalification and cyclic qualification testing provisions as per
ANSI/AISC 341-22 (AISC, 2022), the displacement-controlled standard FEMA/SAC loading sequence (SAC
Joint Venture, 2000) given in Table 3 has been adopted in a quasi-static cyclic fashion at a rate of 10mm/min.
The equivalent horizontal displacement applied at the top of the column,
Δ.
, has been determined based on
the storey drift angle (also known as chord rotation), θ (rad) using equation (4). Due to constructional limitations
of the test rig, the maximum lateral drift level that could be applied was that of θ = ±0.04 rad.
𝜃 = Δ.𝐻/01234
⁄
(4)
Figure 6. Schematic of the IMJ prototype and its loading system.
Table 3. Loading sequence
Load step
Story drift angle, θ (rad)
Equivalent lateral displacement,
𝚫5 (mm)
Number of
cycles
1
0.00375
10.3125
6
2
0.005
13.75
6
3
0.0075
20.625
6
4
0.01
27.5
4
5
0.015
41.25
2
6
0.02
55
2
7
0.03
82.5
2
8
0.04
110
2
3.4 Instrumentation layout
The in-plane lateral displacement applied at the top of the sample is recorded by the built-in linear variable
displacement transformers (LVDTs) of the servo-hydraulic actuators (T05) and is used to obtain the load-
displacement hysteresis loops. Two linear potentiometers (T01-02) have been arranged horizontally along the
centrelines of the floor and ceiling beams to record the relative displacement between the floor and the ceiling
beams. Additionally, two vertical potentiometers have been installed between the floor and ceiling corner
-Dh
+ Dh
10
1. IMJ prototype
2. Pinned support
3. Hinged roller supports
4. Loading crosshead beam
5. Lateral actuators
6. Hydraulic jack
7. Load cell
8. Roller support
9. Out-of-plane restraint
10. Self-reacting test frame
7
6
4
1
2
9
3
5
8
WCEE2024 Corfar & Tsavdaridis
8
fittings (T03) and at the top jack (T04) to measure the vertical displacement of the rubber bearing and the total
vertical displacement of the specimen respectively. A total of 10 strain gauges have been installed on the
beam and posts near the beam-column joint zone to monitor the evolution of strains in the regions expected
to develop the highest stress levels. Details of the arrangement of potentiometers and the locations of strain
gauges are illustrated in Figure 7.
Figure 7. Instrumentation plan.
4 Results of the experimental tests
4.1 Hysteretic performance
At the end of the 10 minute wait step with 100 kN applied vertical load on the IMJ test frame, the linear
potentiometer T03 recorded a compressive set in the rubber bearing of -0.54 mm. This effect generated a
partial loss of tension in the original snug-tight condition of the bolt assembly, preventing the full involvement
of the bolt at small horizontal displacements. As the quasi-static cyclic loading sequence commenced, the 3
mm bolt hole clearance cumulated with the partially loose condition of the bolt assembly ensured that the initial
non-linear force-displacement relationship was almost entirely due to the shear behaviour of the rubber
bearing, characterised by a high initial stiffness with quick softening followed by increased stiffness at larger
strains.
The graph of the applied lateral load versus the displacement at the tip of the top post recorded during the
quasi-static cyclic loading test is plotted in Figure 8 (a). An initial qualitative appraisal of the hysteretic loop
reveals a stable performance, with no obvious signs of strength or stiffness degradation and a pronounced S-
shaped path during the small-deformation cycles, typical for filled rubber under shear loading conditions. The
curve shows very limited residual displacement up to 3% story drift angle, attributed to the re-centring capability
of the rubber bearing, while the increasing permanent displacement at 4% story drift angle indicates the onset
of a more significant contribution of the steel frame members and the beam-column joint to the overall response
of the IMJ to cyclic lateral loading. The rubber bearing’s effective re-centring effect (represented by 68%
recoverable displacement in the positive loading direction and 55% recoverable displacement in the negative
loading direction) has come at the expense of a fat loop like those exhibited by metallic yielding-based systems,
suggesting that the hysteresis demonstrated by the hybrid IMJ is mostly a result of the work done by the high-
damping rubber bearing during the cyclic shear straining.
Transducers layout Strain gauge locations
WCEE2024 Corfar & Tsavdaridis
9
Figure 8. Hysteretic curve (a) and weld crack at the ceiling beam-column joint (b).
The structural behaviour of the IMJ prototype has been characterised by three main force-displacement stages,
each identifiable by a change in stiffness attributed to events observed during the test. During Stage 1 (0 mm
to ±20.625 mm lateral displacement), the IMJ prototype displayed a lower stiffness as lateral load was mostly
resisted by a combination of the friction between the box corner endplates and rubber bearing outer plates,
the shear stiffness of the rubber and slipping of the bolt assembly. As the IMJ prototype entered Stage 2
(limited by ±82.5 mm lateral displacement), the bolt rod experienced a combined shear-bending action as it
was fully engaged between the top and bottom box corner plates leading to an increased overall stiffness of
the IMJ prototype due to the additional contribution of the flexural stiffness of its members. The third and final
stage displayed a lower stiffness, explained by the increasing bending deformation of the bolt rod which allows
further extension to occur in the rubber layers of the bearing. Overall, the IMJ prototype displayed stiffer
behaviour in the negative loading direction due to the contribution of the intra-module connections represented
by the rigid beam-to-column joints which stiffened that side of the assembly. The values of the lateral load
capacity (Pi), lateral load displacement (di), and stiffness (ki) representative for each stage have been reported
in Table 4, using data from the first cycle of the respective load step.
Table 4. Mechanical parameters of the IMJ prototype.
Load
direction
P1 (kN)
d1
(mm)
k1
(kN/mm)
P2 (kN)
d2
(mm)
k2
(kN/mm)
P3
(kN)
d3
(mm)
k3
(kN/mm)
(+)
10.2
20.6
0.49
50.9
81.9
0.66
63.1
110.0
0.44
(-)
-14.2
-20.7
0.68
-63.8
-82.4
0.80
-71.2
-109.7
0.27
At the end of the test, an evident loss of strength was noticed during the reverse loading in the 2nd cycle of
±110 mm lateral displacement (4% story drift angle) as the peak lateral load (Pmax = -61.6 kN) reached
approximately 85% of the peak lateral load (Pmax = -72.5 kN) previously recorded during the first cycle of the
same amplitude, and the onset of cracking of the fillet welds was observed at the inside corner of the ceiling
beam-column joint (Figure 8 (b)). This type of failure was in good agreement with that reported by Chen et al.
(2017) for similar one-sided IMJ frames with unstiffened beam-column joints subjected to quasi-static cyclic
loading, emphasising once more the critical role of the beam-column joints and the importance of weld quality
at these highly-stressed regions. Despite this initiation of strength degradation, the state of the specimen did
not hinder its stability under gravity load, such that it was possible to reach the end of the 2nd cycle of 4 %
storey drift, marking the end of the loading programme that could be accommodated by the test rig. After the
test, the plastic deformation of the bolt rod was limited such that it was possible to undo the nut with a regular
spanner to disassemble the joint assembly. According to acceptance criteria as defined in FEMA 350 (SAC
Joint Venture, 2000), the IMJ prototype met the minimum qualifying drift angle capacity for both strength
degradation (θSD = 0.02 rad) and connection failure (θu = 0.03 rad) for an ordinary moment frame, while also
meeting the qualifying drift angle capacity at strength degradation (θSD = 0.04 rad) for a special moment frame.
(a) (b)
crack
WCEE2024 Corfar & Tsavdaridis
10
4.2 Deformation patterns and strain evolution
Using the readings from transducers T01, T02, and T05, the deformation modes along the height of the test
frame have been constructed in Figure 9, based on data from the first cycle at each storey drift amplitude.
Because of operational issues with the linear potentiometers during the test, data was not available for the
negative loading direction corresponding to the 0.04 rad storey drift step. The gradients of the lateral
displacement along the height of the frame indicate a comparable contribution between the top and bottom
posts, while most of the inter-storey drift is concentrated in the IMC region. The stiffer behaviour in the negative
loading direction was exposed again, while the locking effect of the bolt assembly throughout the second stage
of the force-displacement curve was also highlighted by the constant relative displacement of the connected
box corners up to the level of 2% story drift angle.
The strain gauge data revealed that the steel members of the IMJ prototype worked within the elastic state of
deformation the whole time during the 0 to 0.03 rad storey drift angle load steps, while the onset of steel
yielding was reached in the top flange of the top beam (SG08), and inside flanges of the posts (SG02 and
SG09) only during the negative loading direction of the first 0.04 rad cycle. Yield strain was determined using
the yield strength values reported in Table 1 assuming a Young’s modulus of 210,000 MPa. These results
suggest that the hybrid IMC made of a high-damping rubber bearing and high-strength steel bolt is effective in
its passive damage control function by resisting most of the lateral load up to 0.03 rad storey drift angle, while
the stress in the main framing members is kept below yielding levels until θ = ±0.04 rad has been reached.
Figure 9. Deformation pattern of the IMJ prototype.
4.3 Energy dissipation capacity
The energy dissipation capacity was evaluated by the equivalent viscous damping coefficient (Chopra, 2020),
accounting for all energy dissipating mechanisms demonstrated during the cyclic loading. Based on data points
from the last hysteresis loop (2nd cycle of θ = ±0.04 rad), an equivalent damping coefficient of 0.12 was
obtained, which was lower than the values reported for similar tested specimens by Chen et al. (2017) at
comparable levels of applied lateral displacement. While a more ductile bolt may improve the low energy
dissipation capacity predicted also by the shape of the hysteresis loop, the limited permanent deformations
exhibited by frame members together with the pronounced re-centring effect remain as evidence for the
favourable effect of the hybrid IMC on the seismic resilience of the IMJ prototype.
5 Concluding remarks
This study advanced the development of sustainable and seismic resilient steel modular buildings by
completing the meso-scale experimental test of an inter-module joint prototype designed to accommodate
large storey drift levels without sustaining excessive damage in the structural frame members. The IMJ
prototype consisted of a high-damping rubber bearing connected between the box corner fittings of top and
bottom subassemblages by means of a high-strength steel bolt and was subjected to the full FEMA-SAC cyclic
loading sequence in a quasi-static fashion to better understand the effect of the high-damping rubber bearing
on the seismic performance of the joint.
WCEE2024 Corfar & Tsavdaridis
11
The force-displacement hysteretic loop confirmed the presence of a good re-centring ability of the joint due to
the recoverable deformation of the HDR bearing. Based on a hybrid working mechanism defined by the HDR
shear straining and combined shear-bending deformation of the HSS bolt, the hybrid IMC fulfilled its passive
damage control role by delaying the full contribution of the IMJ frame members until large storey drift
amplitudes were reached.
While the onset of cracking was observed near the weld at the ceiling beam-column joint, the IMJ prototype
reached the end of the quasi-static cyclic loading protocol (2nd cycle of ±0.04 rad storey drift angle) with 85%
of its peak strength and limited damage suffered by the main framing members emphasising the resilience of
the connection system. Now, stiffened beam-column connections should be adopted to improve the lateral
load capacity and avoid weld cracking during the last loading step, while means of enhancing the joint’s energy
dissipation capacity without compromising on the re-centring effect and low-damage state of the frame
members should be investigated.
6 References
AISC (2022) Seismic Provisions for Structural Steel Buildings. Chicago: American Institute of Steel
Construction. Available at: https://www.aisc.org/publications/steel-standards/aisc-341/.
BSI (2015) Eurocode 3: Design of steel structures - Part 1-1: General rules and rules for buildings. London:
BSI.
BSI (2018) Anti-seismic devices. London: BSI.
Burtscher, S.L. and Dorfmann, A. (2004) ‘Compression and shear tests of anisotropic high damping rubber
bearings’, Engineering Structures, 26(13), pp. 1979–1991. Available at:
https://doi.org/10.1016/j.engstruct.2004.07.014.
Chen, Z. et al. (2017) ‘Experimental study of an innovative modular steel building connection’, Journal of
Constructional Steel Research, 139, pp. 69–82. Available at: https://doi.org/10.1016/j.jcsr.2017.09.008.
Chen, Z. et al. (2021) ‘Exploration of the multidirectional stability and response of prefabricated volumetric
modular steel structures’, Journal of Constructional Steel Research, 184. Available at:
https://doi.org/10.1016/j.jcsr.2021.106826.
Chopra, A.K. (2020) Dynamics of Structures: Theory and Applications to Earthquake Engineering. 5th edn.
Harlow: Pearson.
Corfar, D.-A. and Tsavdaridis, K.D. (2022) ‘A comprehensive review and classification of inter-module
connections for hot-rolled steel modular building systems’, Journal of Building Engineering, 50, p. 104006.
Available at: https://doi.org/10.1016/j.jobe.2022.104006.
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’, Engineering Structures, 289, p. 116281. Available at:
https://doi.org/10.1016/j.engstruct.2023.116281.
Deng, E.-F. et al. (2020) ‘Seismic performance of mid-to-high rise modular steel construction - A critical review’,
Thin-Walled Structures, 155, p. 106924. Available at: https://doi.org/10.1016/j.tws.2020.106924.
Farajian, M. et al. (2022) ‘Classification of inter-modular connections for stiffness and strength in sway corner-
supported steel modular frames’, Journal of Constructional Steel Research, 197, p. 107458. Available at:
https://doi.org/10.1016/j.jcsr.2022.107458.
Fujita, M. et al. (2023) ‘Japanese Efforts to Promote Steel Reuse in Building Construction’, Journal of Structural
Engineering, 149(1), p. 04022225. Available at: https://doi.org/10.1061/(ASCE)ST.1943-541X.0003473.
Gent, A.N. (2012) Engineering with Rubber: How to Design Rubber Components. 3rd edn. München: Hanser.
Grigorian, M., Sedighi, S. and Mohammadi, H. (2023) ‘Plastic design of sustainable steel earthquake resistant
structures’, Engineering Structures, 289, p. 116178. Available at:
https://doi.org/10.1016/j.engstruct.2023.116178.
Gunawardena, T. and Mendis, P. (2022) ‘Prefabricated Building Systems—Design and Construction’,
Encyclopedia, 2(1), pp. 70–95. Available at: https://doi.org/10.3390/encyclopedia2010006.
WCEE2024 Corfar & Tsavdaridis
12
Hao, H. et al. (2023) ‘Towards next generation design of sustainable, durable, multi-hazard resistant, resilient,
and smart civil engineering structures’, Engineering Structures, 277, p. 115477. Available at:
https://doi.org/10.1016/j.engstruct.2022.115477.
Jing, J. et al. (2021) ‘Seismic protection of modular buildings with galvanised steel wall tracks and bonded
rubber units: Experimental and numerical study’, Thin-Walled Structures, 162. Available at:
https://doi.org/10.1016/j.tws.2021.107563.
Lacey, A.W. et al. (2020) ‘Effect of inter-module connection stiffness on structural response of a modular steel
building subjected to wind and earthquake load’, Engineering Structures, 213. Available at:
https://doi.org/10.1016/j.engstruct.2020.110628.
Lindley, P.B. and Gough, J. (2015) Engineering Design with Natural Rubber. 6th edn. Bertford: Tun Abdul
Razak Research Centre.
McConnell, J.R. and Fahnestock, L.A. (2015) ‘Innovations in Steel Design: Research Needs for Global
Sustainability’, Journal of Structural Engineering, 141(2), p. 02514001. Available at:
https://doi.org/10.1061/(ASCE)ST.1943-541X.0001185.
McCormick, J. et al. (2008) ‘Permissible Residual Deformation Levels For Building Structures Considering
Both Safety And Human Elements’, in. The 14th World Conference on Earthquake Engineering.
Mullins, L. (1969) ‘Softening of Rubber by Deformation’, Rubber Chemistry and Technology, 42(1), pp. 339–
362. Available at: https://doi.org/10.5254/1.3539210.
Nguyen, T.D.H.N. et al. (2023) ‘An innovative approach to temporary educational facilities: A case study of
relocatable modular school in South Korea’, Journal of Building Engineering, 76, p. 107097. Available at:
https://doi.org/10.1016/j.jobe.2023.107097.
Payne, A.R. (1962) ‘The dynamic properties of carbon black-loaded natural rubber vulcanizates. Part I’, Journal
of Applied Polymer Science, 6(19), pp. 57–63. Available at: https://doi.org/10.1002/app.1962.070061906.
SAC Joint Venture (2000) Recommended Seismic Design Criteria for New Steel Moment-Frame Buildings.
Federal Emergency Management Agency.
Sendanayake, S.V. et al. (2021) ‘Enhancing the lateral performance of modular buildings through innovative
inter-modular connections’, Structures, 29, pp. 167–184. Available at:
https://doi.org/10.1016/j.istruc.2020.10.047.
Sultana, P. and Youssef, M.A. (2018) ‘Seismic Performance of Modular Steel-Braced Frames Utilizing
Superelastic Shape Memory Alloy Bolts in the Vertical Module Connections’, Journal of Earthquake
Engineering, 24(4), pp. 628–652. Available at: https://doi.org/10.1080/13632469.2018.1453394.
United Nations, Department of Economic and Social Affairs (2015) Transforming our World: The 2030 Agenda
for Sustainable Development. A/RES/70/1. United Nations. Available at: https://sdgs.un.org/2030agenda
(Accessed: 8 June 2023).
Wang, Z. et al. (2023) ‘Automated minimum-weight sizing design framework for tall self-standing modular
buildings subjected to multiple performance constraints under static and dynamic wind loads’, Engineering
Structures, 286, p. 116121. Available at: https://doi.org/10.1016/j.engstruct.2023.116121.
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’, Engineering
Structures, 251, p. 113472. Available at: https://doi.org/10.1016/j.engstruct.2021.113472.
Wu, C.X. et al. (2019) ‘Research on seismic behaviour analysis of shock absorbing structure and connecting
joints of container assembly structures’, Steel Construction, 34(4), pp. 1-8+73.
Zhang, G., Xu, L.-H. and Li, Z.-X. (2022) ‘Experimental evaluation on seismic performance of a novel plug-in
modular steel structure connection system’, Engineering Structures, 273, p. 115099. Available at:
https://doi.org/10.1016/j.engstruct.2022.115099.
