Content uploaded by Dan-Adrian Corfar
Author content
All content in this area was uploaded by Dan-Adrian Corfar on Jul 08, 2024
Content may be subject to copyright.
© 2024. This manuscript version is made available under the CC-BY-NC-ND 4.0
license https://creativecommons.org/licenses/by-nc-nd/4.0/
Cyclic tests on hybrid inter-module joints with high-
damping rubber
Konstantinos Daniel Tsavdaridis1[0000-0001-8349-3979] and Dan-Adrian Corfar1[0000-0002-
9843-2973]
1 City, University of London, London, UK
Konstantinos.Tsavdaridis@city.ac.uk
Abstract. Recently, a novel hybrid inter-module connection was proposed to
reduce the permanent damage in the volumetric module by employing a rub-
ber bearing. In this study, two cyclic tests were carried out to study the seis-
mic performance of the hybrid inter-module connection at joint level, while
also determining the effect of the beam-column connection detail. The stand-
ard FEMA/SAC loading sequence was employed on single-span, meso-scale
joint prototypes with bi-axial loading applied to the top post. The results
showed that the hybrid IMJs exhibited nonlinear, multi-stage hysteretic re-
sponses, governed by the bending resistance of the bolting assembly and the
stiffness of the intra-module connection. In terms of the aseismic perfor-
mance, the joints exhibited remarkably low residual drifts, below the repair-
ability limit of 0.5% up to 2% drift ratio, while displaying relatively high
equivalent viscous damping coefficients during the first stages of defor-
mation owing to the effectives of the high-damping rubber bearing. Overall,
the cyclic tests demonstrated the feasibility of the proposed connection,
which delayed the contribution of the members in the lateral response of the
joints, limiting the damage suffered by the volumetric module in the after-
math of an earthquake.
Keywords: Steel Modular Buildings, Inter-module Joints, Hybrid Connec-
tions, Intra-module Connections, High-Damping Rubber, Cyclic Load.
1 Introduction
Steel modular building systems (MBSs) have risen in popularity due to the widely
recognised advantages of modular construction [1], while the excellent strength-to-
weight ratio of structural steel has enabled the advancement of the technology to
new heights [2]. As previous research has showed, inter-module connections
(IMCs) play a crucial role in the lateral behaviour of self-standing steel modular
buildings [3–6]. Thus far, a lot of effort has been put into the development of full-
strength/rigid inter-module joints (IMJs) which demonstrate good seismic perfor-
mance based on capacity design which ensures that the seismic-input energy is dis-
sipated through the hysteresis of the steel beams [7–10]. However, the significant
2
© 2024. This manuscript version is made available under the CC-BY-NC-ND 4.0
license https://creativecommons.org/licenses/by-nc-nd/4.0/
permanent damage developed in the volumetric module hinders its demountability
and reusability, while the large residual deformations greatly affect the functionality
and repairability of a steel modular building in the aftermath of an earthquake.
Therefore, the subject of earthquake-resilient steel modular buildings with energy-
dissipating and/or self-centring components integrated within the inter-module
joints has captured the interest of researchers [11–14]. In this regard, the authors
have recently proposed a hybrid IMC employing a high-damping rubber bearing
and a bolting assembly and investigated its cyclic behaviour at connection level
through validated proof-of-concept finite element analysis (FEA) [15]. In this study,
two cyclic tests were carried out to study the seismic performance of the hybrid
IMC at joint level, while also determining the effect of the beam-column (intra-
module) connection.
2 The proposed hybrid inter-module connection
A schematic of the proposed connection was shown in Fig. 1a. 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 would be resisted by the bolting assembly.
Horizontal shear forces are transferred through a combined mechanism of friction
between the faying steel surfaces, shear resistance of the rubber bearing, and bend-
ing of the bolt rod, while the interlocking pins prevent accidental sliding. To im-
prove the energy dissipation capacity of the rubber bearing at the shear deformation
levels expected in the IMC (50%-100% shear strains), the rubber layers were made
of high-damping (filled) rubber instead of low-damping (unfilled) rubber, as the
addition of carbon black filler improves the hysteresis of rubber [16,17].
Fig. 1. Hybrid inter-module connection.
3
© 2024. 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 Experimental programme
3.1 Details of the test specimens
The test IMJ prototypes were made of top and bottom beam-column subassem-
blages, and the inter-module connection between them. The meso-scale configura-
tion (Fig. 1b) was adopted as a cost-effective solution that replicates the deformed
shape and anticipated points of inflection in planar modular frames subjected to in-
plane lateral loads by limiting the lengths of the posts and beams to half of those in
the full-size frame panel.
Table 1. Test specimens.
Joint prototype
Bolting assembly
Beam-column joint
JP01
M24, class 8.8 bolt
Unstiffened
JP02
M24, class 8.8 bolt
Stiffened
As summarised in Table 1, two specimens namely JP01 and JP02 were designed to
investigate the influence of the intra-module connection’s stiffness on the seismic
performance of the hybrid IMJs at joint level by considering two details of beam-
column joints (stiffened and unstiffened).
Fig. 2. Details of the beam-column subassemblages (shown for stiffened specimen).
4
© 2024. This manuscript version is made available under the CC-BY-NC-ND 4.0
license https://creativecommons.org/licenses/by-nc-nd/4.0/
The top and bottom beam-column subassemblages (Fig. 2) were S355J2H steel
cold-formed hollow members, joined to S355J2H steel-plated box corner fittings by
complete joint penetration (CJP) groove welds. The box corner fittings were fabri-
cated from 15-mm-thick steel plates. Two S355J2H steel stiffeners of 100 mm x
100 mm x 10 mm triangular plate were welded at the beam-column connection for
the stiffened specimen.
The rubber bearings (Fig. 3) were made of two outer S355 steel plates (150 mm
x 150 mm x 15 mm), four high-damping 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 improve the stability of the bearing and limit the bulging of the
rubber layers under the applied axial load.
Each bolting assemblies consisted of standard M24 x 150mm full-thread hexa-
gon head bolts for specimens JS-1 and JS-2 and an M27 x made of class 8.8 high-
strength steel (HSS).
The main properties obtained from material characterisation tests were summa-
rised in Table 2 for the S355 steel and Table 3 for the high-damping rubber com-
pound.
Fig. 3. Geometry of the rubber bearings.
Table 2. Material properties of the steel frame members.
Part
Steel grade
Ea (GPa)
fyb (N/mm2)
fuc (N/mm2)
Ad (%)
Top post
S355J2H
209
505
546
28.2
Bottom post
S355J2H
205
438
496
30.9
Floor beam
S355J2H
190
433
507
32.4
Ceiling beam
S355J2H
202
541
565
19.5
a Modulus of elasticity, b Yield strength, c Tensile strength, d Elongation percentage
Table 3. Material properties of high-damping rubber.
Hardness a
Shear modulus, G c
Effective damping ratio, ξ!"".$ c
86 IRHDb
0.61 MPa
18.46 %
a based on shear modulus at 5% shear strain, b International rubber hardness degree, c at
100% shear strain
5
© 2024. This manuscript version is made available under the CC-BY-NC-ND 4.0
license https://creativecommons.org/licenses/by-nc-nd/4.0/
3.2 Test setup
The test setup was depicted in Fig. 4. The loading system was designed to accom-
modate the bi-axial loading applied to the top column of the IMJ prototypes, while
realistically replicating the deformed shape of unbraced modular frames subjected
to a lateral load as recommended by Lacey et al. [18]. The joint prototypes were
assembled directly within the self-reacting test frame in a straightforward sequence
by stacking the lower sub-assemblage, the rubber bearing, and the upper sub-assem-
blage, followed by the insertion of the bolting assemblies. A snug-tight condition
was achieved by the full effort of manually tightening the bolts with a spanner
through the corner fitting access holes.
The test followed the prequalification and cyclic qualification testing provisions
as per ANSI/AISC 341-22 [19], applying the displacement-controlled standard
FEMA/SAC [20] loading sequence shown in Fig. 4, through the horizontal actuators
in a quasi-static manner at a rate of 10mm/min to limit the influence of dynamic
effects such as those arising from the inertial forces of the joint prototype members.
To capture the effect of gravitational loading, an axial load equivalent to 5% of the
compressive yield capacity,
N!.#$
, (using nominal material properties as per Euro-
code 3, Part 1 [21]) was kept constant throughout the test. The vertical load exerted
a compressive stress of 4.7 MPa on the rubber bearing.
Due to the deflection of the rubber layers under the applied axial load, the bolting
assemblies experienced a partial loss of pre-tension, losing the preliminary snug-
tight condition. While this effect led to slightly lower shear stiffness in the connec-
tions during the early cycles, it allowed for a larger deformation capacity and pro-
vided a strength reserve by effectively delaying the full contribution of the studs
and bolts to the force-transfer mechanism of the connections.
Fig. 4. Test setup and loading protocol.
6
© 2024. This manuscript version is made available under the CC-BY-NC-ND 4.0
license https://creativecommons.org/licenses/by-nc-nd/4.0/
4 Results and discussion
4.1 Test phenomena and controlling failure modes
At the onset of cyclic loading, the inter-storey drift occurred mostly in the inter-
module connection, as expected due to the low shear stiffness of the rubber layers,
highlighting the rubber bearing as the major contributor to the joint’s response to
small-amplitude lateral loads. During these initial stages, it was observed that the
bolting assemblies were allowed to translate horizontally together with the upper
box corner fittings until the gap provided by the bolt hole tolerance was fully closed
and the bolting assemblies became locked in a bending deformation state. Then, as
the loading sequence progressed towards the larger drift ratio levels, the shear de-
formation of the rubber layers increased up to the point when the contribution of the
flexural stiffness of the beam-column sub-assemblages was evident as shown in Fig.
5 by the pronounced deflection of their framing members.
Fig. 5. Deflection of beam-column sub-assemblages at peak inter-storey drift ratio.
Taking a closer look at the connection level, the rubber bearing was subjected to
the combined effect of axial load and horizontal shear throughout the cyclic loading
protocol. Due to the quasi-static nature of the tests and the low thickness of the
rubber layers, heat build-up was not expected to be an issue, which was confirmed
with the aid of a thermal camera, registering a constant temperature throughout the
test. There was no gap opening between the box corner fitting end-plates and the
outer plates of the rubber bearing, as the top and bottom surfaces remained parallel
even at maximum shear deformation levels, as shown in Fig. 6.
During the test on specimen JP01, the M24 HSS bolt experienced significant
bending deformation as well as damage to the threads (Fig. 7). The onset of weld
cracking at the ceiling beam-column connection (Fig. 8a) was noticed after the peak
was reached in the positive loading direction during the second ±4% drift ratio cy-
cle, causing the peak lateral load in the reverse direction to fall to 85% of the previ-
ous maximum.
At the end of the load programme, the reduction was still well above the 20%
failure criterium defined in the FEMA/SAC guidance [20], yet the controlling
7
© 2024. This manuscript version is made available under the CC-BY-NC-ND 4.0
license https://creativecommons.org/licenses/by-nc-nd/4.0/
failure mode could be identified as the weakened beam-column connection. This
finding was in good agreement with results from similar tests [22–24], emphasising
once more the importance of the weld quality at these highly stressed regions.
Fig. 6. Shear deformation of the rubber layers.
Specimen JP02 suffered similar damage levels to the HSS bolt (Fig. 7), while the
test was completed without noticeable damage to the beam-column connection (Fig.
8b) and framing members due to the effect of the stiffener plates.
Even though no fracture occurred in the threads, the significant deformation in
the shanks of the HSS bolts may lead to the assumption that the bolts failed, and
therefore the ultimate limit state design of the hybrid joint would be controlled by
the bending resistance of the bolting assembly. Moreover, the results proved that it
is desirable to add stiffeners at the intra-module connections, to ensure that the con-
trolling failure mode is indeed governed by the bending of the bolt rod and by the
resistance of the weld at the beam-column joint.
Fig. 7. Bolt deformation and damage to threads.
Fig. 8. State of the intra-module connection in bottom module at the end of the test.
8
© 2024. This manuscript version is made available under the CC-BY-NC-ND 4.0
license https://creativecommons.org/licenses/by-nc-nd/4.0/
4.2 Hysteresis loops
The relationship between the applied lateral displacement and the corresponding
reaction force at the top of the IMJ prototypes was shown in Fig. 9.
Both specimens displayed similar reverse S-shaped curves with well-defined,
stable hysteresis and multiple stages indicating the successive activation of different
connection components.
During the first stage (0 to ±0.5% drift ratio), the rubber bearing was immediately
engaged by the horizontal shear developed between the box corner fittings of the
IMC due to the frictional resistance of the steel contact surfaces.
In the second stage (±0.5% to ±3%), the bolting assembly was completely en-
gaged in a bending deformation as the bolt hole clearances closed under the relative
horizontal displacement of the box corner fittings. This limited the shear defor-
mation in the rubber layers, initiating the partial involvement of the framing mem-
bers through their flexural stiffness.
The third and final stage (±3% and beyond) was identified by the comparison
between the stiffened and unstiffened beam-column connections, signalling the
prominent influence of the intra-module connection in the force resisting mecha-
nism developed at high drift ratios.
Overall, the IMJs displayed stiffer behaviour in the negative loading direction
(pulling direction for the actuators), which was attributed to the lack of symmetry
of the one-sided T-shaped specimens, with rigid intra-module beam-column con-
nections only on one side of the assemblage. This finding was in good agreement
with the results from other experimental work that employed one-sided T-shaped
joints with column loading [22,24].
Fig. 9. Hysteresis loops.
4.3 Residual drift
The residual deformation of the two specimens at each inter-storey drift level was
shown in Fig. 10, represented the percent ratio of the residual displacement corre-
sponding to zero lateral load during load reversal and the total specimen height.
9
© 2024. This manuscript version is made available under the CC-BY-NC-ND 4.0
license https://creativecommons.org/licenses/by-nc-nd/4.0/
Comparing the results to the residual drift limits defined by FEMA P-58-1 [25],
the specimens exhibited similar results up to 1% drift ratio, with residual drifts be-
low the realignment limit of 0.2% in the negative loading direction, while the resid-
ual deformation was generally below or slightly fluctuating near the 0.5% repaira-
bility limit until the 2% drift ratio was reached and still economically feasible to
repair up to 3% drift level. Above the 1% residual drift limit, the major realignment
required to restore the safety of the building may render the structure as unfeasible
to repair. At 4% drift ratio, the residual drift of the unstiffened specimen was larger
by 0.2% than that of the stiffened joint in the negative loading direction, emphasis-
ing the favourable influence of the intra-module connection on the re-centring abil-
ity when designed to perform elastically.
Fig. 10. Residual deformation levels in the hybrid IMJs.
4.4 Energy dissipation capacity
Fig. 11a shows the total energy dissipated in the first cycle, ED, at each drift ratio
level, represented by the area enclosed by the hysteresis loops. The reduced values
up to 2% drift ratio confirmed the limited inelastic deformation of the frame mem-
bers which performed elastically, while the sudden spike in the energy dissipated
during the 3%-4% drift ratios was mostly attributed to frame members hysteresis,
confirming their late activation during the third stage.
Fig. 11b illustrated the variation of the equivalent viscous damping coefficients,
xeq, at different drift ratios, showing significantly higher equivalent viscous damp-
ing coefficients during the early stages, followed by a progressive decrease and a
slight recovery after the 2% drift ratio level. This finding was quite unique when
compared to the variation of the equivalent viscous damping, provided by steel
yielding mechanisms, which usually starts close to zero followed by a steady in-
crease with the progression of the cyclic loading. The remarkable behaviour of the
hybrid joints was attributed to the activation of the rubber bearings during the early
stages of loading and to the mechanical properties of the high-damping rubber
10
© 2024. This manuscript version is made available under the CC-BY-NC-ND 4.0
license https://creativecommons.org/licenses/by-nc-nd/4.0/
compound used in the rubber layers, while the recovery after the 2% drift ratio was
correlated with the engagement of the steel frame elements during the third stage of
deformation observed on the hysteresis loops. The detail of the intra-module con-
nection did not have a significant influence on the energy dissipation capacity of
the joint up to the deformation levels achieved in the tests, while the differences
observed at 4% drift ratio resulted from the reduced displacement reached by the
actuators during the test on JP02 due to a malfunction of the control system.
Fig. 11. Energy dissipation capacity of the IMJs.
5 Conclusions
In this study, two inter-module joint (IMJ) prototypes equipped with a novel hybrid
inter-module connection were tested under cyclic loading to evaluate their aseismic
performance and determine the effect of the intra-module connection detail.
The hybrid IMJs exhibited nonlinear hysteretic responses characterised by the
sequential activation of different connection components at three different stages.
The first stage was governed by the low shear stiffness of the rubber bearing, the
second stage reflected the key influence of the bolting assembly, now fully engaged
by the shear deformation in the rubber bearing, while the third stage highlighted the
role of the intra-module connection stiffness.
The intra-module connections should be designed to perform elastically by add-
ing stiffeners to the beam-column joint in order to ensure that the controlling failure
mode of the joints is governed by the bending of the HSS bolts.
Both specimens exhibited remarkably low residual drifts, below the repairability
limit of 0.5% up to 2% drift ratio, while the higher equivalent viscous damping
coefficients registered during the first stages of deformation emphasised the effec-
tiveness of the high-damping rubber compound.
Overall, the cyclic tests demonstrated the feasibility of the proposed connection,
which effectively delayed the full participation of the members in the lateral
11
© 2024. This manuscript version is made available under the CC-BY-NC-ND 4.0
license https://creativecommons.org/licenses/by-nc-nd/4.0/
response of the joints, limiting the damage suffered by the volumetric module in the
aftermath of an earthquake. Now, design recommendations should be developed to
allow the necessary deformation capacity in the bolt rod at the required lateral load
capacity of the hybrid joint, while further effort should be put into maximising the
energy dissipation capacity of the hybrid IMC without affecting its re-centring ca-
pabilities or increasing the permanent damage in the volumetrics.
References
1. Ferdous W, Manalo A, Sharda A, Bai Y, Ngo TD, Mendis P (2022) Construction Indus-
try Transformation Through Modular Methods. In: Ghaffar SH, Mullett P, Pei E, Rob-
erts J, editors. Innovation in Construction, Cham: Springer, p. 259–76.
https://doi.org/10.1007/978-3-030-95798-8_11.
2. Thai H-T, Ngo T, Uy B (2020) A review on modular construction for high-rise buildings.
Structures 28:1265–90. https://doi.org/10.1016/j.istruc.2020.09.070.
3. Corfar D-A, Tsavdaridis KD (2022) A comprehensive review and classification of inter-
module connections for hot-rolled steel modular building systems. Journal of Building
Engineering 50:104006. https://doi.org/10.1016/j.jobe.2022.104006.
4. Farajian M, Sharafi P, Eslamnia H, Kildashti K, Bai Y (2022) Classification of inter-
modular connections for stiffness and strength in sway corner-supported steel modular
frames. Journal of Constructional Steel Research 197:107458
https://doi.org/10.1016/j.jcsr.2022.107458.
5. Wang Z, Tsavdaridis KD (2022) Optimality criteria-based minimum-weight design
method for modular building systems subjected to generalised stiffness constraints: A
comparative study. Engineering Structures 251:113472. https://doi.org/10.1016/j.eng-
struct.2021.113472.
6. Wang Z, Rajana K, Corfar D-A, Tsavdaridis KD (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
286:116121. https://doi.org/10.1016/j.engstruct.2023.116121.
7. Dai X-M, Zong L, Ding Y, Li Z-X (2019) Experimental study on seismic behavior of a
novel plug-in self-lock joint for modular steel construction. Engineering Structures
181:143–64. https://doi.org/10.1016/j.engstruct.2018.11.075.
8. Chen Z, Wang J, Liu J, Khan K (2021) Seismic behavior and moment transfer capacity
of an innovative self-locking inter-module connection for modular steel building. Engi-
neering Structures 245:112978. https://doi.org/10.1016/j.engstruct.2021.112978.
9. Zhai S-Y, Lyu Y-F, Cao K, Li G-Q, Wang W-Y, Chen C (2023) Seismic behavior of an
innovative bolted connection with dual-slot hole for modular steel buildings. Engineer-
ing Structures 279:115619. https://doi.org/10.1016/j.engstruct.2023.115619.
10. Yang C, Chen H, Wen H, Wang Q, Zhang B, Ou J (2024) Experimental study on seismic
performance of internal cruciform joints of grouting sleeve connection for modular in-
tegrated construction. Engineering Structures 301:117325.
https://doi.org/10.1016/j.engstruct.2023.117325.
11. Sultana P, Youssef MA (2016) Seismic performance of steel moment resisting frames
utilizing superelastic shape memory alloys. Journal of Constructional Steel Research
125:239–51. https://doi.org/10.1016/j.jcsr.2016.06.019.
12
© 2024. This manuscript version is made available under the CC-BY-NC-ND 4.0
license https://creativecommons.org/licenses/by-nc-nd/4.0/
12. Sendanayake SV, Thambiratnam DP, Perera NJ, Chan THT, Aghdamy S (2021) En-
hancing the lateral performance of modular buildings through innovative inter-modular
connections. Structures 29:167–84. https://doi.org/10.1016/j.istruc.2020.10.047.
13. Zhang G, Xu L, Li Z (2023) Development and Experimental Verification of Self-Cen-
tering Haunched Plug-In Modular Connections. J Struct Eng 149:04023047.
https://doi.org/10.1061/JSENDH.STENG-12029.
14. He J, Zhou X, Xu F, Shi Y, Okazaki T (2024) Seismic performance of self-centring
connections with two energy dissipation stages for reusable modular steel buildings.
Thin-Walled Structures 196:111442. https://doi.org/10.1016/j.tws.2023.111442.
15. Corfar D-A, Tsavdaridis KD (2023) A hybrid inter-module connection for steel modular
building systems with SMA and high-damping rubber components. Engineering Struc-
tures 289:116281. https://doi.org/10.1016/j.engstruct.2023.116281.
16. Ahmadi H, Kingston J, Muhr AH (2008) Dynamic Properties of Filled Rubber - Part I:
Simple Model, Experimental Data and Simulated Results. Rubber Chemistry and Tech-
nology 81:1–18. https://doi.org/10.5254/1.3548196.
17. Lindley PB, Gough J (2015) Engineering Design with Natural Rubber. 6th edn. Tun
Abdul Razak Research Centre, Bertford.
18. Lacey AW, Chen W, Hao H (2022) Experimental methods for inter-module joints in
modular building structures – A state-of-the-art review. Journal of Building Engineering
46:103792. https://doi.org/10.1016/j.jobe.2021.103792.
19. AISC (2022) Seismic Provisions for Structural Steel Buildings. American Institute of
Steel Construction, Chicago.
20. SAC Joint Venture (2000) Recommended Seismic Design Criteria for New Steel Mo-
ment-Frame Buildings. Federal Emergency Management Agency, Washington, D.C.
21. BSI (2015) Eurocode 3: Design of steel structures - Part 1-1: General rules and rules for
buildings. BSI, London.
22. Chen Z, Liu J, Yu Y, Zhou C, Yan R (2017) Experimental study of an innovative mod-
ular steel building connection. Journal of Constructional Steel Research 139:69–82.
https://doi.org/10.1016/j.jcsr.2017.09.008.
23. Sanches R, Mercan O, Roberts B (2018) Experimental investigations of vertical post-
tensioned connection for modular steel structures. Engineering Structures 175:776–89.
https://doi.org/10.1016/j.engstruct.2018.08.049.
24. Zhai S-Y, Lyu Y-F, Cao K, Li G-Q, Wang W-Y, Chen C (2022) Experimental study on
bolted-cover plate corner connections for column-supported modular steel buildings.
Journal of Constructional Steel Research 189:107060.
https://doi.org/10.1016/j.jcsr.2021.107060.
25. FEMA (2018) Seismic Performance Assessment of Buildings, Volume 1 – Methodol-
ogy, 2nd edn. Federal Emergency Management Agency, Washington, D.C.
