A Review and Analysis of Testing and Modeling Practice of Extended Hollo-Bolt Blind Bolt Connections

Abstract

Steel Hollow Sections (SHS) offer many structural, economical and architectural advantages in multi-storey and high-rise construction. However, their use is not suitable for a wide range of applications due to the difficulties of site bolting as there is limited access to the inner part of the steel section for tightening of standard bolts. Blind bolts have been developed to overcome these difficulties in view of extending the application of SHS in construction. Special attention has been paid to blind bolts that could potentially be used in rigid or semi-rigid connections. This is the case of a modified blind bolt, termed the Extended Hollo-Bolt (EHB), which has shown to be able to achieve the required performance for its use in moment resisting connections. This paper critically reviews published work concerning the blind fastener, describes the loading procedures used for testing and failure modes produced, lists the assessed parameters with their respective applicability ranges, and summarises the analytical models developed for the EHB components. Additionally, a global sensitivity analysis is performed using information of two representative studies for the purpose of detecting key design parameters that influence the response of the connection in terms of strength and stiffness. The analysis shows that the concrete strength has the most influential effect on both the stiffness and strength of the column component as well as bolt component stiffness, while the bolt grade highly influences the bolt component strength.

Full text

A Review and Analysis of Testing and Modeling Practice of
Extended Hollo-Bolt Blind Bolt Connections
Manuela Cabreraa,∗
, Walid Tizania, Jelena Ninica
aDepartment of Civil Engineering, The University of Nottingham, Nottingham, UK
Abstract
Steel Hollow Sections (SHS) offer many structural, economical and architectural
advantages in multi-storey and high-rise construction. However, their use is not
suitable for a wide range of applications due to the difficulties of site bolting as there
is limited access to the inner part of the steel section for tightening of standard
bolts. Blind bolts have been developed to overcome these difficulties in view of
extending the application of SHS in construction. Special attention has been paid
to blind bolts that could potentially be used in rigid or semi-rigid connections. This
is the case of a modified blind bolt, termed the Extended Hollo-Bolt (EHB), which
has shown to be able to achieve the required performance for its use in moment
resisting connections. This paper critically reviews published work concerning the
blind fastener, describes the loading procedures used for testing and failure modes
produced, lists the assessed parameters with their respective applicability ranges, and
summarises the analytical models developed for the EHB components. Additionally,
a global sensitivity analysis is performed using information of two representative
studies for the purpose of detecting key design parameters that influence the response
of the connection in terms of strength and stiffness. The analysis shows that the
∗Corresponding author: M. Cabrera, Email: manuela.cabrera@nottingham.ac.uk
Preprint submitted to Journal of Constructional Steel Research May 8, 2021
Cabrera et al. (2021) A Review and Analysis of Testing and Modeling Practice of Extended
Hollo-Bolt Blind Bolt Connections. Journal of Constructional Steel Research. 183, 106763.
https://doi.org/10.1016/j.jcsr.2021.106763

concrete strength has the most influential effect on both the stiffness and strength
of the column component as well as bolt component stiffness, while the bolt grade
highly influences the bolt component strength.
Keywords: Extended Hollo-Bolt, Tubular Connection, Concrete-Filled Steel
Hollow Section, Experimental and Analytical Review, Sensitivity Analysis
1. Introduction
The use of Steel Hollow Sections (SHS) in multi-storey and high-rise construction
has grown over the years allowing the structural industry to explore new design
concepts. SHS members (e.g., rectangular, circular profiles) are desirable as columns
from architectural and structural points of view. They have superior axial load5
carrying capacity, higher strength-to-weight ratio, increased fire resistance and an
excellent torsional resistance compared to steel open section profiles (e.g.,I-shaped, T-
shaped profiles) [1]. Some alternatives to connect open beam-to-SHS involve welding
of fittings, threaded studs or diaphragms onto the face of the column to provide
access for bolting, or direct welding of the beam to the column [2]. However, welded10
components are prone to damage during transportation, could be impractical to
install [3], and have quality and inspection issues when done on-site [4]. Welded
(a) Shape before tightening.
(b) Shape after tightening.
Fig. 1. Extended Hollo-bolt (EHB).
2

connections have also exhibited brittle failure under seismic events [5–7]. Therefore,
bolting is broadly the preferred method, unless special circumstances dictate.
Blind bolts have been developed to overcome these limitations as they can be as-15
sembled and tightened from one-side only. There is an extensive range of commercial
blind bolts which, along with endplate connections, allow to connect open to closed
steel members. Each type of blind bolt has a particular geometry and installation
technique defined by the maker which allows on-site installation such as the Blind
Bolt [8], Huck BOM [9], Molabolt [10], Flowdrill system [11], Ajax-Oneside fastener20
[12], and Lindapter Hollo-Bolt (HB) [13].
Connections using the fasteners mentioned above provide sufficient shear and
tying resistance to satisfy structural integrity checks. However, such connections
tend to have low moment–rotation stiffness which is usually controlled by the inherent
flexibility of the SHS column face hindering the use of blind bolts in moment resisting25
connections. One of the effective ways to mitigate this problem is filling the SHS
with concrete [14]. The advantages of this technique have been highlighted by many
authors. For instance, the load carrying capacity, ductility and rotation capacity of
Concrete-Filled SHS (CFSHS) columns are enhanced by the confinement provided
by the column walls to the concrete which in turn limits concrete fracture [15]; the30
fire resistance of CFSHS members is higher than that of bare SHS since the infill
concrete absorbs part of the heat reducing the temperature increment rate of the
steel tube [16]; the bolt pull-out is limited by the anchoring effect produced by the
concrete around the bolt [17]; additionally, excessive localised deformation in the
column walls is prevented by the support provided by the concrete, specially in the35
compression zone of the connection [18–20].
Various authors have proposed modifications to the commercial blind bolts in
order to increase their tensile stiffness, and increase the bending stiffness of the face
3

of the hollow section. This is the case of the Extended Hollo-Bolt (EHB) developed
by Tizani and Ridley-Elis [21] as a modified version of the commercially available40
Lindapter Hollo-Bolt (HB). The modified fastener, Fig. 1, has an extended bolt shank
and an additional nut at the end of the bolt which creates an anchoring effect taking
advantage of the concrete around the bolt.
This paper presents a review of previous and ongoing research regarding the EHB
connection zones under different load types. Also, it describes the studied parameters45
and their impact on the connection response. A systematic literature survey has been
conducted to assess the effects of modifying the HB to the EHB, and to identify the
steps required to fully characterise the EHB. Three databases are chosen for the paper
retrieval, namely Scopus, Web of Science, and American Society of Civil Engineers
(ASCE) Library, among which ASCE is the core collection. Search results are then50
selectively reviewed based on refining topics to concentrate on the EHB connection.
For example, there are a total of 32 publications on this topic ranging from 2012 to
2021 (data accessed from Scopus on 28/10/2020) using the query strings in Scopus
as TITLE-ABS-KEY (”Extended Hollo Bolt” OR “Anchored Blind bolt”).
The purpose of this paper is to review all available information regarding the EHB55
connection in order to identify aspects that have not been addressed in the present
body of research and are required for deeper understanding of the EHB connection
such that design guidance can be produced. The development of a rigid bolted
connection system will allow for the use of open-section beams connected to hollow
sections as columns which represent a clear advantage for the building industry.60
The remaining of the paper is organized as follows: Section 2 presents the state
of the art of different blind bolts which are under current investigation; Section 3
presents a review of the available experimental and analytical information of the
EHB connection zones (i.e. tension, compression, and both); a sensitivity analysis
4

(a) Ajax-Oneside fastener [22](b) Thread-Fixed One-Side [23]
(c) T-shaped One-side Bolt [24](d) Lindapter Hollo-bolt [13]
Fig. 2. Different blind bolts under research.
is presented in Section 4; and finally, the conclusions of the work are presented in65
Section 5.
2. Blind Bolted connections background
Multiple studies have been conducted regarding blind bolted in steel and compos-
ite beam to column connections. Some studies have addressed the static behaviour of
CFSHS column connections with various blind fasteners, e.g: Loh et al. [25,26], Liu70
et al. [27], Ataei et al. [28,29], highlighting the benefits of using blind bolts in terms
of joint ductility, strength, and stiffness. Other researchers have investigated the
cyclic performance of blind bolted end plate connections to CFST columns, such as
5

Li et al. [30], Wang et al. [31–33], Waqas et al. [34], demonstrating that these blind
bolted connections perform satisfactorily in terms of yielding, maximum strength75
capacity, and ultimate displacement under seismic events.
These experimental studies have shown these connections to have a promising
prospect in practical engineering and to be an effective solution for modern struc-
tures. Four blind bolts under current investigation at different institutions are re-
viewed in this section for illustration, these are the Ajax-Oneside fastener, Thread-80
Fixed One-side Bolt, T-shaped One-side Bolt, and Lindapter Hollo-bolt.
2.1. The Ajax-Oneside fastener
The Ajax-Oneside fastener is a blind bolt comprised of a high strength bolt, a split
step-washer that expands once inside the hollow section, a solid step-washer, and a
structural nut, see Fig. 2a. This blind bolt can reach the full structural strength85
of high strength bolts under AS4291.1 specification, according to Ajax Fasteners
Innovations [35].
The Cogged Anchor Blind Bolt (CABB) is a modification of the Ajax-Oneside
fastener developed by Gardner and Goldsworthy [36]. This bolt has been studied
under cyclic tension by Gardner and Goldsworthy [37], and Yao et el. [20] studied90
groups of CABB to concrete-filled circular hollow sections. The studies showed that
the modified bolt has a higher connection failure load and initial stiffness compared
to the original bolt.
However, this modification was found impractical for manufacturing and instal-
lation, and therefore another modification was introduced by Yao et al. [38], the95
Headed Anchor Blind Bolt (HABB), which used the same method for anchorage as
the EHB. Yao et al. [38], Oktavianus et al. [39,40] performed a series of monotonic
tensile tests and parametric studies using FE analysis to assess the performance
6

of individual and groups of HABBs and compared them to the conventional Ajax-
Oneside bolts. The results indicated that the modified bolts could be suitable for100
moment-resisting connections with a high degree of strength and stiffness. Agheshlui
et al. [41] concluded that placing the HABBS close enough column corner prevents
full concrete cone failure and therefore the full tensile capacity of the bolt can be
used.
Further modification of the HABB, the Double-Headed Anchored Blind Bolt105
(DHABB), was introduced by Oktavianus et al. [4] who added a second embed-
ded head within the infill concrete. The individual and group behaviour of DHABB
under cyclic loading was studied by Oktavianus et al. [4,22], respectively. The
authors concluded that the DHABB exhibit higher secant stiffness if the extra em-
bedded head is installed in the appropriate location and the thickness of the T-stub110
flange has most influential effect on the secant stiffness of the connection.
The cyclic behavior of groups of DHABBs was experimentally and numerically
evaluated by Pokharel et al. [42]. The authors proposed the use of through bolt
along with the DHABBs and the test results show that stiffness of the connection is
increased while the cyclic deterioration is decreased. From parametric analysis, the115
variation of the flange thickness of the T-stub has shown to have the largest effect
on the tensile behavior of the DHABB connections.
2.2. The Thread-Fixed One-side Bolt (TFOB)
The Thread-Fixed One-side Bolt (TFOB) is similar to the Flowdrill system ex-
tended to thicker plates. In these bolts, a thread is created in the column wall holes120
and a bolt without nut can be installed and tightened, see Fig. 2b.
The TFOB has been studied by Iu et al. [43] under monotonic load to evaluate the
tension yield resistance of the connected T-stubs. Two failure modes were identified
7

from the experimental tests and a series of design methods were proposed. Zhu et al.
[44] found that using backing plates in combination with these kind of bolts improves125
the tension resistance of the connection. Using a validated FE model, Wulan et
al. [45] conducted parametric studies and concluded that threaded T-stubs provide
enough tensile capacity to fix the high strength bolt, preventing pre-mature thread
failure. Wang et al. [46] studied the TFOB on lap connections under shear load.
Finite Element Analysis (FEA) were also performed and conclusions show that the130
studied bolt and screwed shear plate could replace the traditional bolt and nut in
engineering applications.
The monotonic and cyclic loading response of the connection was investigated
by Wulan et al. [47]. It was observed that the cyclic loading caused the threads
on the wall become more vulnerable to failure compared to the monotonic case.135
Additionally, the authors concluded that available design methods for monotonic
load can be applied under cyclic loading as well.
Wang et al. [48] subjected beam to SHS connections using TFOBs to static bend-
ing moment. Strengthening methods were also used to improve the initial stiffness
and bending moment capacity. Test results showed that the yielding bending mo-140
ment, the ultimate bending moment and the ultimate rotation of a TFOB strength-
ened with backing plate were similar to those of traditional Nut-fixed bolts, and the
initial stiffness was enhanced. The yielding bending moment, the ultimate bend-
ing moment and the initial stiffness were also improved, but the ultimate rotation
decreased. Additionally, all tested specimens met the seismic ductility requirements.145
2.3. T-shaped One-side Bolt (TOB)
The T-shaped One-side Bolt (TOB), developed by Sun et al. [49], consists of a
bolt shank with T-head, a nut, and a washer, as illustrated in Fig. 2c. The authors
8

evaluated numerically the behaviour of the end plate connection of SHS to I-beam.
The numerical results showed that the bending moment capacity of the proposed150
TOB connection is higher than that of Standard High-strength Bolts.
Wang et al. [24] conducted tensile tests on TOB connections with vertical and
horizontal slotted bolt holes. It was concluded that compared to standard high
strength bolt connections, the initial stiffness of TOBs with vertical slotted bolt
holes was increased, while in the case of horizontal slotted bolt holes, it decreases.155
Theoretical models for calculating the bending yield strength were proposed.
2.4. Lindapter Hollo-bolt
The Lindapter Hollo-Bolt (HB) is comprised of a thread bolt, a collar, a sleeve,
a cone, and a rubber washer, as in Fig. 2d. Design guidance for simple joints using
the HB fastener is currently available in Eurocode 3 [50]. In order to extend the use160
of blind fasteners to moment resisting connections, the HB has been investigated in
combination with CFSHS.
Wang et al. [51] tested beam to concrete-filled column connections under sym-
metrical monotonic loading using HBs. According to the moment-rotation response,
the tested specimens were classified as semi-rigid and of partial strength according165
to the EC3 specification. A similar test programme was conducted by Wang et al.
[31] to evaluate the hysteretic performance of the connection. The authors concluded
that rotation capacities of this type of joint satisfied the ductility requirements for
earthquake resistance in most seismic regions.
Wang et al. [32] carried out experimental and analytical analysis of CFSHS to170
steel beam connections using HBs. Similar to [51], the specimens were classified
as semi-rigid and partial strength, and the rotation capacities satisfied the ductility
requirements suggested by FEMA-350 [52].
9

In spite of the advantages mentioned above, the use of HBs in combination with
CFSHS does not provide the required moment resistance and rotational stiffness to be175
classified as moment resistant. This is because the concrete filling only addresses the
flexibility of the tube face and the improvement is not sufficient to attain significant
moment resistance.
A modification of the HB, the Reverse Mechanism Hollo-bolt (RMH) [1], has an
inverted expanding sleeve that clamps directly to the underside of the joint. [53]180
tested the RMH a using back to back T-stubs test arrangement. Conclusions show
that the use of this fastener in moment resisting connections is feasible. However,
undesirable sudden failure occurs and the flexibility of the SHS may limit the moment
capacity of the connection. Tizani and Ridley-Elis [21] presented the results from
experimental tests carried out to RMH using SHS with and without infill concrete. It185
was concluded that the RMH connection without infill concrete has sufficient stiffness
to classify as moment-resisting but lower tensile strength than standard bolts. It was
also found that the insufficiency in strength can be improved by the use of concrete
infill.
Tizani and Ridley-Elis [21] proposed to add a nut at the end of an extended bolt190
shank in order to create an anchoring effect, take advantage of the infill concrete, and
improve the flexibility of the column face. Ellison and Tizani [54], and Tizani et al.
[14] compared the tensile behaviour of the modified blind bolt, termed the Extended
hollo-bolt (EHB), with standard bolts. The test results showed that both strength
and stiffness are enhanced by the modified EHB configuration. The use of infill195
concrete changes the failure mode from bolt pull-out to bolt shank tensile fracture
improving the strength of the connection. It also provides additional bending stiffness
to the face of the hollow section and the stiffness is enhanced by the embedded anchor
nut. This fastener has shown to have the potential to be used in moment-resisting
10

Fig. 3. Joint components of an open section to CFSHS joint with an EHB flush end-plate connec-
tion. See Table 1 for component key.
Table 1. Key to Fig. 3. EHB joint components and evaluation rules availability.
Ref in Component EC3 Rotational stiff.
Fig. 2 availability contribution
Tension
a Bolt tension No Yes
b Endplate bending Yes Yes
c Column face bending No Yes
d Beam web tension Yes No
Compression
j Beam flange compression Yes No
connections and therefore, the following sections are focused on this type of blind200
bolt.
3. Review of studies on the EHB connection behaviour
Structural steel and composite joint systems are complex to characterize as a
whole due to their material and geometric non-linearities, residual stress conditions,
and complex geometrical configurations. Therefore, simplified mechanical models205
such as the component method in Eurocode 3 [50] have been developed to facilitate
11

the joint design procedure. In the component based approach, joints are decomposed
into a set of rigid and flexible components which contribute to the joint structural
properties and therefore constitute a powerful tool for the evaluation of the stiffness
and/or resistance properties of joints under different loading conditions [55]. The210
assembly of these individual basic components into a mechanical model can be used to
predict the response of any joint geometry as long as the behaviour of its components
(stiffness, resistance, and ductility) is fully characterized [19].
To extend the application of the component method to EHB blind-bolted connec-
tions between open and hollow sections, Pitrakkos et al. [55] reviewed the available215
data in terms of the relevant components of a single-sided joint between an open
section beam (I profile) and a CFSHS column connected using a flush endplate and
two rows of EHBs fasteners, one in tension and one in compression. The joint com-
ponents which contribute to the resistance and/or rotational stiffness of the EHB
joint and the availability of evaluation rules for each of them in Eurocode 3 [50] are220
presented in Fig. 3 and Table 1. The identification of these components is based on
the following assumptions:
The beam flange carries all compression and therefore, the beam web in com-
pression is not considered.
Due to the infill concrete stiffening action, the following components do no need225
to be taken into account: bolts shear, column face compression, side column
faces compression/tension, and punching shear failure around the bolt heads
in compression.
The weld components do not contribute to the rotational stiffness of the joint
(CEN 2005). However, their resistance must be checked against the existing230
rules available in Eurocode 3 Part 1-8.
12

Fig. 4. Summary of available literature regarding the EHB
Different authors have contributed to the existing gap in the knowledge regarding
the two components unavailable in Eurocode 3: bolts in tension and column face in
bending. Different studies addressing these components are presented below.
The study of the EHB connection behaviour have been made as a whole (beam235
and column connections) or dividing it into zones. Special attention has been paid
to the tension zone since the extension of the component method to the EHB con-
nection has been limited due to the lack of knowledge regarding the behaviour of
two components in this zone. A summary of the studied components and load types
applied for each zone is presented in Fig. 4.240
3.1. Tension zone
In the tension zone of the connection, three possible failure modes have been
identified: bolt failure in tension, Fig. 5a; column face failure in bending, Fig. 5b;
and combined failure mode (both bolt and column face can contribute to failure),
13

(a) Bolts in tension (b) Column face in bending (c) Combined failure
Fig. 5. Failure modes of the EHB connection [56].
Fig. 5c. The two extreme failure modes have also been identified and reported for245
other blind bolts, for instance, the TSOB with rigid T-Stub, displayed large column
face deformation when a thin column is used Fig. 6a, while bolt fracture is reported
for thick column face Fig. 6b [24]. Another example is the anchored Ajax-Oneside
fastener, for which the failure modes are bar fracture Fig. 7b, and the tube wall yield
and bar pullout Fig. 7a, depending on the bolt location (middle or side of the SHS),250
bolt diameter, SHS wall thickness, and compressive strength of the concrete infill
[38].
The first two EHB failure modes have been studied independently isolating the
component of interest. The tension zone of an end-plate connection between open
section members is modelled in Eurocode 3 [50] as a equivalent T-stub model, which255
represent the flange and web of the column, and the web and end plate of the beam
for open section steel members [1,57]. The component based approach and the T-
stub model have been adopted to study open beam-to-hollow column, as illustrated
in Fig. 8, in order to study the components in tension of the EHB connection. A
review of the available literature per component is presented next.260
14

(a) TSOB Tube wall yield and bar pullout
(b) TSOB Bar fracture
Fig. 6. TSOB failure modes [24].
(a) Ajax Bar fracture (b) Ajax tube wall yield and bar pullout
Fig. 7. Anchored Ajax-Oneside fastener failure modes [38].
15

3.1.1. Bolt component in tension
Extensive research has been carried out isolating the bolt component by means
of a rigid column face arrangement, the studied configurations include single-sided
and double-sided T-stub models under different loads.
Pitrakks [3] carried out 16 tests to evaluate the single EHB connection under a265
monotonic tensile pull-out test arrangement. The test set-up uses a reusable steel box
assembly comprised of four rigid flat plates limiting the bending of the top plate and
therefore isolating the bolt behaviour as illustrated in Fig. 9a. The authors identified
and assessed individually three components that contribute to the deformability of
the EHB component: 1) Internal bolt elongation, 2) Expanding sleeves, and 3) Bond270
and anchorage. Additional tests of EHBs without sleeves, and HBs with and without
infill concrete were performed in order to identify the contribution of each individual
component to the general behaviour of the connection.
It was observed that the EHB has better performance than the original version,
the HB, as the anchored nut distributes the applied force over the surrounding con-275
crete and therefore, concentration of stresses in the expanding sleeves is decreased,
limiting their failure and eliminating concrete breakout. Concrete strength was found
to have significant influence on the connection stiffness and negligible effect on its
strength and ductility. Higher bolt grade improves the stiffness, strength, and ductil-
ity. The study concluded that the EHB component can be compared to an standard280
bolt as the failure mode corresponds to bolt shank necking and fracture, showing
that it is able to develop the full tensile capacity of its internal bolt.
The mechanical properties of the bolts used in the testing programme were also
reported in [3] for seven bolt batches. Tensile tests were performed on machined and
full-size bolts in accordance with ISO 898-1:2009 [58]. Test results are summarised in285
16

Fig. 8. T-stub to steel hollow section model.
Table 2. Mechanical properties of different bolt batches.
Bolt Diameter Bolt fyb fy u E
batch (mm) grade (MPa) (MPa) (MPa)
A 16 8.8 907 1003 205
B 16 8.8 725 900 210
C 16 8.8 873 981 209
D 16 8.8 836 931 207
E 16 10.9 1086 1127 209
F 20 8.8 785 935 207
G 16 8.8 828 917 212
Table 2 where fyb ,fyu , and E are the yield, ultimate strength, and Young’s modulus
of elasticity, respectively. Variations were observed in the bolt properties for the
same bolt grade, which in turn caused some discrepancies in the yield and ultimate
states for the tensile results when different bolt batches were used for identical spec-
imens. The author highlighted the importance of considering the actual mechanical290
properties of the tested bolts as they influence the test results significantly.
Using the experimental results reported above, Pritrakkos et al. [55] developed
an analytical model based on a system of spring elements. This model takes into
17

account pre-load and deformation from the three components identified previously
for both the elastic and inelastic behaviour of the component. The proposed model295
showed to accurately predict the response of the component and contributed to the
development of a more detailed design method for the fastener. The analytical model
is presented in Section 3.3.
The performance of a group of EHBs under a monotonic tensile force was also
studied by Pitrakkos [3] using double-sided T-stub connections. Four bolts were300
used in each side of the CFSHS. The studied parameters included bolt grade, gauge
distance, pitch distance, and concrete grade. Apart from the benefits raised by
the use of concrete infill, high concrete strength, and bolt grade, it was found that
the bolt group action does not compromise the strength of the system as the total
connection strength is equal to the sum of the individual bolts.305
Tensile fatigue tests were conducted by Abd Rahman [17] using a single EHB.
The results indicate that the fatigue life and strength of an EHB were lower than
those of a standard bolt, but higher than those of a HB. The failure mode of the
connection was a fatigue fracture of the bolt shank which is comparable with that of
the standard bolt.310
Pascual et al. [59] evaluated the thermal behaviour of single unloaded HBs and
EHBs through experimental and FEA. Connections to SHS with and without infill
concrete were considered. It was concluded that the use of concrete has a noticeable
effect on the thermal behaviour of the connection and bolt temperature reduction.
On the other hand, the use of different section sizes and blind bolts (HB and EHB)315
has no effect on the thermal behaviour of the connection. Later, the same authors
[60] developed a loaded numerical model to predict the fire behaviour of blind bolts
in the tension zone of the connection. The failure occurred in the bolt shank near
the bolt head. This section is critical as high temperature at this location caused
18

(a) Side view of bolt test setup (b) Side view of column test setup
Fig. 9. Experimental configuration by: (a) Pitrakkos [3], and (b) Mahmood [56]
softening of the steel. Similar to the unloaded tests, no significant effect on the320
fire resistance was caused by changing the bolt type in concrete-filled specimens.
However, significant enhancement was observed from unfilled to concrete-filled SHS.
The group behaviour of the EHB connection was evaluated by Shamsudin [61]
using a test arrangement similar to the one used in [3]. A total of 36 tests with
one row of two EHBs were subjected to tensile loading. The effect of bolt gauge325
distance, concrete compressive strength, and embedment depth on the connection
strength and stiffness was investigated. The author concluded that small bolt gauge
distances lead to bolt interaction which results in low connection stiffness. The effect
of the concrete grade on the connection strength and failure mode was found to be
negligible, while the enhancement of the connection initial stiffness was significant330
up to 40MPa. The ductility of the connection was reduced with the use of small
embedment depth.
From the bolts in tension assessment, it is concluded that when a rigid column
wall is used, two failure modes are identified: bolt fracture and/or bolt pull-out. Dif-
ferent load types have been used for this component and a wide range of parameters335
19

studied.
3.1.2. Column face in bending
The column face component has only been assessed under tensile pull-out tests.
These studies isolate the column face component by using a simplified rigid replica
of the EHB usually denominated a dummy bolt. Dummy EHBs have a simplified340
geometry compared to the EHB and are fabricated with high strength steel. The
test arrangement is illustrated in Fig. 9b.
Mahmood [56] investigated the effect of the slenderness ratio (column face width
to its thickness ratio, µ=b/t), anchorage length, bolt gauge distance, concrete type
and strength on the bending behaviour of the connection using experimental and345
numerical methods.
In terms of slenderness ratio, it was concluded that increasing the column thick-
ness increases both the ultimate load carrying capacity and the stiffness of the con-
nection. However, the stiffness improvement is higher from thin to medium than
from medium to thick column thickness, indicating a possible optimum combination350
between concrete strength and column face thickness.
Regarding the anchorage length, it was found that increasing the anchorage length
significantly increases the component strength. For the bolt gauge distance, it was
observed improvement of both the ultimate strength and the initial stiffness of the
connection with the use of a larger bolt gauge distance. Besides, findings suggest355
that the use of small gauge distance leads to stress concentration in the concrete
between bolts limiting the anchorage effect.
From the concrete analysis, it was observed that the failure starts with anchorage
failure caused by concrete crushing in front of the anchor nut, followed by column
face bending and finally pull-out of the bolts. An increase in the concrete strength360
20

resulted in improvement of the component stiffness and significant enhancement in
the component strength. On the other hand, the use of self-compacting concrete
affected neither the strength nor the stiffness of the component while the use of light
weight concrete reduces both.
From the study summarised above, it can be seen that a wide range of parameters365
have been assessed under monotonic tensile load. However, other load types have
not been considered.
3.1.3. Combined failure
Cabrera et al. [62] developed and validated a Finite Element (FE) model com-
bining the results from research performed independently on the bolt [3] and column370
face components [56] in order to produce a combined failure. The effect of varying the
column face thickness on the connection behaviour was assessed showing that com-
ponents with small slenderness ratios (thick column walls) resist higher load before
concrete failure. It was concluded that the first failure signs are caused by concrete
crushing accompanied with SHS yielding. After this, the component strength is375
dependent mainly on the bolt properties in tension (bolt necking and rupture).
Debnath and chan [63] used the experimental results reported in [64] to validate
a numerical model and perform parametric studies to evaluate the influence of design
variables in the behaviour of the connection when using a single EHB under tensile
load. Investigated parameters include bolt embedment length, bolt grade, bolt di-380
ameter, concrete grade, and tube thickness. The authors concluded the connection
stiffness is influenced by slenderness ratio, concrete strength, bolt diameter, and em-
bedment depth, while strength is dependent on bolt diameter (when high stength
concrete is used), concrete grade, and embedment length.
From the tension zone assessment, it is observed that most studies have been385
21

carried out in the bolts in tension component, followed by the column face in bending
component. Up to date, only numerical analyses have been carried out to assess the
combined failure mode. Ultimately, this is the condition to which the connection
would be subjected to in construction so further studies are required to complement
the component method calculation for this kind of blind bolt.390
3.1.4. Bolts in combined tension and shear
It is generally assumed in plastic design of bolted connections that shear forces
are resisted mainly by bolts in the compression zone plus a small contribution (28%
of the shear resistance) of bolts in the tension zone [65], and therefore some bolts
are subjected to a combination of these forces. Pitrakkos et al. [66] studied the395
performance of a single EHB when subjected to various ratios of combined tension
and shear forces. A total of 13 tests were conducted, from pure tension to pure shear,
in order to propose an interaction curve for the studied blind bolt. The author found
that the EHB behaves better than the HB as the concrete infill reduces the effect of
bending in the bolt and prevents the pull-out failure. It was also observed that, at400
predominant tension angles, the load-capacity of the bolts has increased with respect
to predominant shear due to the fact that the shear stress area is increased by the
area of the sleeves
3.2. Beam and column connection
Tizani et al. [14] assessed the performance of the connection using connection405
stiffness classification methods from Eurocode 3 [50] and their suitability for use
as moment-resisting connections. The test arrangement consisted of a point load
applied to the beam 1m away from the column face producing a moment into the
connection. A total of eight specimens were tested with the samples designed to fail
by the EHB in tension either by its pull-out or bolt shank fracture.410
22

The authors used the beam-line method and Eurocode 3 [50] to classify the
connection in terms of stiffness and strength. The results showed that all the tested
connections are classified as semi-rigid and partial strength and none performed
as nominal pin demonstrating the capability of the fastener to provide semi-rigid
connections. Since the stiffness of the tested connections is relative to the attached415
beam, the normalised moment–rotation data was analysed varying the beam section
sizes. It was concluded the connection behaviour is mostly semi-rigid and that rigid
behaviour can be achieved in braced frames.
The seismic behaviour of CFSHS column joints with EHB blind bolts was studied
by Wang [67] and Tizani et al. [68]. The authors performed six full-scale connection420
tests under quasi-static cyclic loading in order to investigate the inelastic hysteretic
behaviour of the connection. The parameters investigated were amplitude of cyclic
loading procedure, bolt grade, tube wall thickness, and concrete grade. The authors
identified two failure modes. Mode I ”weak bolt – strong column face” was observed
in specimens with thick tube face and/or high strength concrete infill. Mode II425
”strong bolt – weak column face” had either thin column wall face or low concrete
strength.
23

Table 3. Design parameters and ranges assessed by different authors.
Ref. Analysis
type*
Benchmark Variables
N◦Bolt specimen** Column Name Range
Bolt component
[3] Exp
1 M16-8.8-NA-90-C40 200x10
Bolt diameter 16 & 20
Bolt Grade 8.8 & 10.9
Anchored length 85, 90 & 130
Concrete strength C40 & C60
4 200x10
Bolt grade 8.8 & 10.9
Double sided connection Gauge distance 90 & 120
M16-8.8-120-90-C40-P100 Pitch distance 100 & 140
Concrete strength C30, C40 & C50
[17] Exp 1 M16-8.8-NA-NR-C40 200x12.5 Load range (kN) 50, 60, 70 & 90
Frequency 0.2 to 5 Hz
[59]Exp &
Num 1 M16-8.8-NA-120-C30 220x10 Tube section 150x8, 250x150x10,
220x10 & 350x150x10
[61]Exp &
Num 2 M16-8.8-120-82-C20 240x180
x20
Gauge distance 120, 140 & 180
Anchored length 82, 92 & 102
Concrete strength C20, C40 & C80
Concrete type*** NW & LW
Column component
[56]Exp &
Num 2 M16-RIG-80-80-C40
200x6.3 Tube thickness 5, 6.3 & 8
Concrete strength C24, C36 & C90
200x8 Concrete type*** NW, NWSC, LW & LWSC
Gauge distance 80, 140 & 180
300x10 Anchored length 80, 103 & 112
Combined component
24

Table 3. Design parameters and ranges assessed by different authors.
Ref. Analysis
type*
Benchmark Variables
N◦Bolt specimen** Column Name Range
[62] Num 2 M16-8.8-80-80-C40 300x10 Tube thickness 5, 6.3 & 8
[63] Num 1 M20-8.8-NA-90-C40 250x8
Embedment depth 0,60, 72, 80 & 90
Bolt grade 8.8 & 10.9
Bolt diameter 12, 16 & 20
Concrete strength C40, C50, C60 & C70
SHS cross section 250, 275 & 300
Tube thickness 6, 10 & 12
Beam and column connections
[14]Exp &
Num 2 M16-8.8-120-NR-C40-P100 200x10
Tube thickness 5, 6.3 & 8
Concrete strength C40 & C60
Pitch distance 100 & 140
Endplate type Flush & Extended
[67]Exp &
Num 2 M16-8.8-120-NR-C50-P100 200x8
Tube thickness 5, 6.3 & 8
Bolt grade 8.8 & 10.9
Concrete strength C20 & C50
Beam section UB356 171 67
UB457 152 52
[69]
[70]
Exp &
Num 2 M16-8.8-120-90-C40-P100 250x5
Tube thickness 5 & 12
End plate thickness 12 & 24
Beam section HN350 175 7 11
HN300 150 6 9
Bolt grade 8.8, 10.9 & 12.9
Pitch distance 80, 100 & 120
Bolt diameter 16, 18 & 20
25

*Analysis type: Exp: experimental; Num: numerical.
**Specimen index: Bolt (1)-(2)-(3)-(4)-(5)-(6), where: (1) Bolt shank diameter;
(2) bolt grade, RIG: rigid bolt; (3) gauge distance, NA: not applicable; (4)430
anchored length, NR: not reported; (5) concrete grade; (6) pitch distance
(optional). Column (1)x(2), where: (1) SHS width; (2) SHS thickness. N◦: number
of bolts per sample. All dimensions in millimeters.
***Concrete types: (NW) Normal Weight; (LW) Light Weight; (NWSC) Normal
Weight Self Compacting; (LWSC) Light Weight Self Compacting.435
It was concluded that the EHB connection provides stable hysteretic behaviour
with appropriate level of strength and stiffness, and rigid behaviour can be achieved.
The connection behaviour was suitable for seismic applications as it offered adequate
energy dissipation capacity and ductility. This is particularly true for connections
that exhibited failure mode II (flexible column face) which have high ductility and440
relatively low strength degradation under cyclic loading. It is suggested to control
the connection failure mode in practice by designing for relatively thin tube face
and/or low strength concrete.
Even though the performance of the connection when using both column face
and bolt real mechanical properties was evaluated in [67], the failure modes reflect445
the extreme cases of either the bolt failure or the column face failure.
Wang et al. [69] tested six EHB flush endplate connections and developed a non-
linear FE model to assess the performance of the connection under quasi-static cyclic
loading. The test results showed the capability of the EHB connection to effectively
limit the deformation of the column face walls since the anchor nut transmitted the450
tensile force to the concrete. These results were closely examined in a FE model
which allowed to identify the transmission path as: beam - endplate - bolt - concrete
26

- column wall. The influence of bolt grade, endplate thickness, pitch distance, bolt
diameter, and pretension was assessed by means of FEA. The authors concluded that
all the studied configurations can be classified as semi-rigid connections.455
Wang et al. [70] conducted cyclic loading tests on seven extended-plate joints
between CFSHS columns and open section beams. The authors investigated the effect
of welding C-channels to locally strengthen the tube walls combined with the EHB
fastener. It was found that this combination allows the joint to fully utilize the bolt
strength and enhance its performance in terms of strength and strength degradation.460
Studied parameters included the end-plate thickness, steel tube wall thickness, beam
section size, local strengthening connection method, blind bolt anchorage method,
and the inclusion of stiffeners.
Table 3 summarises the parameter ranges considered in the studies mentioned in
this chapter and grouped according to the studied component.465
3.3. Analytical modelling review
Based on the results from experimental and FEA, different authors have proposed
equations to describe the global force-displacement response of the EHB connection
and its components. Table 4 summarises the proposed equations found in the liter-
ature.470
The numerical model developed by Pitrakkos [3] for a single EHB assumes three
sources of deformability for the bolts in tension component: elongation of the internal
bolt shank (kb), slippage of expanding sleeves (kHB ), and slippage of the mechanical
anchorage (kM). The massless spring model proposed in Fig. 10 is used for the assem-
bly of these individual components to estimate the EHB global force-displacement475
behaviour.
A regression analysis including a 95% prediction band was used to assess the reli-
27

Fig. 10. Spring component model for EHB [3].
ability of the proposed analytical model. It was concluded that at the chosen predic-
tion band level, the proposed component model predicts the experimental data with
a good level of accuracy when considering different bolt batches, concrete strength,480
bolt grade, bolt diameter, and embedded depth.
Shamsudin [61] further developed the model presented in [3] to extend it to groups
of EHBs using two regressions models: simple linear regression and multiple linear
regression. The proposed model is based on deformation calculations at four different
force intervals. Both models where validated against experimental and FE models485
displaying an error margin of 5%. It was concluded that the proposed equations show
good level of accuracy when predicting the group component behaviour withing the
ranges of validity of the analysis.
28

Table 4. EHB analytical models developed by different authors.
Ref Model Proposed equations Variable definition
Bolt component
FEH B =min(FHB +FM;Fb) Where:
[3]
Tetra-linear global
force-displacement
using Spring com-
ponent model
δEH B =min(δHB ;δM) + δbHB, M, b: sleeves,
mechanical anchorage,
and internal bolt
shank components
respective properties.
kEH B =


1
KHB +kM
+1
kb



−1
δ1=0.15Fu
cc,i,gki,g,80
,δ2=0.85Fu−0.15Fu
cc,i,gki,g,80
+δ1ki,g,80: stiffness for gauge
gand concrete C80.
cc,i,g: proposed coefficient
for gauge distance g.
Fu: bolt ultimate strength.
kex: bolt elastic stiffness.
Tetra-linear global
force-displacement
using simple linear
regression
δ3=0.90Fu−0.85Fu
cc,i,gki,g,80
+δ2
[61]
δ4=Fu−0.90Fu
0.02ke
x
+δ3
δ1=0.15Fu
−232.7+1.9fcu,i + 1.2Gi+ 2.2EDi
fcu,i: concrete strength
Gi: bolt gauge distance
EDi: embedment depth
kex: bolt elastic stiffness.
Tetra-linear global
force-displacement
using multiple linear
regression
δ2=0.85Fu−0.15Fu
−203.2+1.3fcu,i + 1.2Gi+ 2.2EDi
+δ1
δ3=0.90Fu−0.85F5
−219.4+0.5fcu,i + 0.8Gi+ 3.0EDi
+δ2
δ4=Fu−0.90Fu
0.02ke
x
+δ3
29

Table 4. EHB analytical models developed by different authors.
Ref Model Proposed equations Variable definition
Column component
Tetra-linear global
force-displacement
using yield line theory
and spring method
ki,single =Est3
eq
24γf(b−2t)2(1 −ν2)Es: SHS Young modulus.
teq:equivalent thickness.
γf: deflection coefficient.
b: SHS width, t: thickness,
&ν: Poison ratio.
ki,double =Est3
eq
12γf(b−2t)2(1 −ν2)
[56]
Fp,single = 2πMp


1 + Rs+r
Rs


+ 2Mp



2g−2r
Rs+r



Rs: yielded area radius
r: radius of bolt hole
g, p: gauge & pitch
Mp: plastic moment of
resistance for a unit
length of SHS plate.
Tetra-linear global
force-displacement
using yield line theory
and spring method
Fp,double = 4πMp


1 + Rs+r
Rs


+ 4Mp



2g−2r
Rs+r



Fp,comb = 2πMp


1 + Rs+r
Rs


+ 2Mp



3p+ 3g−4r
Rs+r



Fp: plastic strength.
Fd:lowest strength after
plastic load.
Fu: ultimate column face
strength.
δi: column face displ.
k1=0.75Fp
δ1
,k2=0.25Fp
δ2−δ1
,k3=Fd−Fp
δ3−δ2
,k4=Fu−Fd
δ4−δ3
Combined component
[62]
Tetra-linear global
force-displacement
using spring com-
ponent method
k1= 95t+ 263, k2=0.8Fp
δ2−δ1
,k3=Fd−Fp
δ3−δ2
,k4=Fu−Fp
δ4−δ3
t: column face thickness.
Fp: plastic load.
Fd:drop load.
Fu: ultimate load.
δ1=0.2Fp
k1
,δ2=Fp
0.28k1
,δ3= 8.7δ2,δ4=Fu
0.001k1
30

Mahmood [56] used the yield line theory to derive equations for the the column
face plastic load (Fp) and initial stiffness. It was assumed that Fpis equal to the re-490
sistance provided by the SHS plate and the anchorage action. The overall behaviour
of the component is divided into four stages: initial, secondary, drop and membrane
action. The proposed analytical models showed the ability to represent the compo-
nent behaviour with acceptable level of accuracy when compared to experimental
and numerical data.495
Cabrera et al. [62] combined the analytical models proposed by Mahmood [56]
for the column face in bending and Pitrakkos [3] for the bolts in tension in order to
represent the global behaviour of the combined component. The proposed equations
were validated using numerical results from FEA obtaining reasonable agreement
within an error band of 15%.500
4. Sensitivity Analysis
Sensitivity Analysis (SA) allows to study how the output of a model is affected
by the input variation or uncertainty. In this way, SA has been used in different
engineering models to determine which parameters are key in a model and rank
them according to their importance. Different applications of SA can be found in505
[71].
As summarised in the previous section, different authors have studied the in-
fluence varying design parameters on the EHB connection response under different
loading cases. In this section, the influence of varying the studied parameters is
assessed by means of SA.510
Two representative studies have been chosen in the present work to perform a SA:
Mahmood [56] for the column component and Shamsudin [61] for the bolt component.
31

4.1. Scatter Plots
Scatterplots allow for the investigation of the behaviour of the models by visual
inspection when the number of important components is low. Fig. 11 and Fig. 12515
show the scatterplots obtained after performing data standardization to the connec-
tion variables and response for the column and bolt components, respectively. Eq. 1
was used to standardize the data.
Zi=xi−µ
σ(1)
Where Ziis the standardized value, xiis the observed value, µis the mean, and
σis the standard deviation of the sample.520
The scatterplots for the bolt component in Fig. 11a show that the connection
strength is only influenced by the bolt grade. This expected as the failure corresponds
to bolt fracture and therefore the strength properties of the bolt define the connection
strength, this is also in agreement with Pitrakkos [3]. On the other hand, Fig. 11b
shows that all studied parameters have a positive correlation with the connection525
stiffness such that the parameter influence can be rank as: concrete strength >
gauge distance >anchored length >bolt grade.
In the case of the column component, a linear relationship between component
strength and concrete grade can be observed in Fig. 12a with this parameter being
the most influential. In the case of gauge distance and anchored length, parabolic530
correlations are observed. On the other hand, the slenderness ratio has a negative
correlation with the strength of the connection as the wall thicknesses is inversely
proportional to the slenderness ratio, this parameter has the smallest influence on
the component strength.
Fig. 12b shows gauge distance to have a bi-linear tendency which is in agreement535
32

(a) Strength sensitivity analysis. (b) Stiffness sensitivity analysis.
Fig. 11. Scatterplots of connection response versus design parameters from bolt component studies
by Shamsudin [61].
(a) Strength sensitivity analysis. (b) Stiffness sensitivity analysis.
Fig. 12. Scatterplots of connection response versus design parameters from column component
studies by Mahmood [56].
33

Table 5. Parameter sensitivity measures calculated using EE method.
Parameter Strength Stiffness
µ∗µ σ µ∗µ σ
Bolt Component
Concrete strength 0.006 -0.006 0.025 0.786 0.786 0.762
Bolt grade 0.737 0.737 0.087 0.063 0.050 0.049
Gauge distance 0.014 0.014 0.039 0.680 0.680 0.415
Anchored length 0.012 -0.012 0.041 0.359 0.359 0.265
Column Component
Concrete strength 0.797 0.797 0.456 1.147 0.071 1.634
Slenderness ratio 0.352 -0.352 0.112 0.545 -0.441 0.342
Gauge distance 0.471 0.471 0.268 0.596 0.596 0.531
Anchored length 0.468 0.468 0.339 0.523 0.523 0.086
with the literature which states that the initial stiffness is improved by an insignificant
amount when large bolt gauges are used. Similar trends are observed for concrete
strength and anchored length. The parameter influence on the component stiffness
is classified as: concrete strength >gauge distance >slenderness ratio >anchored
length.540
4.2. Elementary Effects Method
Different SA measures have been developed to provide the information provided
by scatterplots in a condensed format. The Elementary Effect (EE) is a SA method
introduced by Morris in 1991 [72] and used to identify the most important model
parameters when a relatively small number of sample points is available.545
This method uses two sensitivity measures to identify the input factors to have
more effects on the output of the system: the mean µand the standard deviation
σof a finite distribution Fi. Consider a model Ywith knormalized independent
inputs Xi, i = 1, ..., k, hence varying in a k-dimensional unit cube across pselected
levels. Therefore, the input spaced is discretized into a p-level grid Ω. For a given550
34

point X in this grid, the elementary effect of the ith input factor is given as:
EEi=Y(X1, X2, ..., Xi−1, Xi+∆ , ...Xk)−Y(X1, X2, ..., Xk)
∆(2)
Where ∆ is a value in {1/(p−1), ..., 1−1/(p−1)},X= (X1, X2, ...Xk) is any
selected value in Ω such that the transformed point (X+ei∆) is still in Ω for each
index i= 1, ..., k, and eiis a vector of zeros but with a unit as its ith component.
The distribution of elementary effects associated with the ith input value is ob-555
tained by randomly sampling different Xfrom Ω, denoted by Fi, i.e. EEi∼Fi.
The mean µestimates the overall influence of the input factor to the system re-
sponse, while the standard deviation σassesses the interaction effects with the other
parameters as well as the nonlinear relation between the input [71].
The sign of the elementary effect might vary between different evaluation points,560
and therefore the value of the mean can lead to erroneous conclusions. To overcome
this limitation, Campolongo et al. [73] proposed using µ∗which is the mean of
the distribution of the absolute values of the elementary effects, denoted as Gi, i.e.
EEi∼Gi. For the purpose of completeness, all sensitivity measures are calculated
in this study.565
The sensitivity indices for the studied standardized parameters are given in Ta-
ble 5. The mean of the elementary effect absolute value µ∗allows to rank the param-
eters according to their influence in the strength and stiffness response of the system.
For the bolt component, the most influential parameter in the component strength
is the bolt grade. This result is expected as the failure mode of these components570
is bolt fracture, which is determined by the bolt ultimate strength. The following
parameters are gauge distance and anchored length, which have similar sensitivity
measures, and finally the concrete strength. The µ∗value for the bolt grade is sig-
35

nificantly larger than the other three studied parameters concluding that the latest
have low to insignificant influence in the system response. In the case of the stiffness575
response, the concrete strength and gauge distance are the most influential with sim-
ilar values of µ∗, followed by the anchored length and the least influential parameter
is the gauge distance.
In the case of the column component, the parameters are ranked as: concrete
strength >gauge distance >anchored length >slenderness ratio for the component580
strength, and concrete strength >gauge distance >slenderness ratio >anchored
length for the component stiffness.
The EE method also identifies the nonlinear relationship between the studied
parameters and the connection response. Large σvalues, like the one obtained
between concrete strength and connection stiffness for the column component, reflect585
the bi-linear behaviour observed in the scatterplots discussed in the previous section.
The classification obtained with scatterplots and EE method shows similar results
increasing the reliability of the study.
5. Conclusions
A modified blind bolt, termed the Extended Hollo-Bolt (EHB), provides a conve-590
nient and reliable means of connecting to steel hollow sections. The EHB has shown
to have superior performance in terms of moment and strength resistance, and initial
stiffness when compared to the commercially available Hollo-Bolt (HB), showing po-
tential to be used in moment-resisting connections. Studies available in the literature
regarding this type of fastener have been reviewed here. It is found that there are595
areas which have not been addressed yet and therefore there is insufficient knowledge
at present for the safe design of moment-resisting connections using the EHB. Other
findings and recommendations from this research include:
36

From the EHB joint tension zone review, it is found that the bolts in tension
and column face in bending components are not fully characterized yet. These600
components are required in order to extend the component method from EC3
for this type of blind bolted connection. From the range of studies found in
the literature assessing the joint zones independently, it is found that special
attention has been paid to the bolts in tension component. A wide range of
design parameters such as: bolt diameter and grade, anchored length, concrete605
grade and type, and gauge and pitch distance have been assessed. Additionally,
the connection has been subjected to different loading procedures: monotonic
tensile pull-out, quasi-static cyclic and thermal. On the other hand, for the
column face in bending component, studies addressing the behaviour of the
connection when varying the tube thickness, anchored length, gauge distance,610
concrete strength and grade are found only under monotonic tensile pull-out.
In the case of combined failure mode, only numerical and analytical models are
presented with no experimental tests performed. It is concluded that further
studies are required in the combined failure mode in order to fully characterize
the connection behaviour.615
The whole connection (beam and column) has been experimentally and nu-
merically studied under quasi-static cyclic loading for different tube thickness,
concrete strength, pitch distance, endplate type and thickness, bold grade and
diameter, and beam section. The moment-rotation behaviour of the connec-
tion shows semi-rigid and rigid behaviour as well as adequate performance for620
seismic applications. Additionally, the moment-rotation behaviour of the con-
nections has been addressed when a rigid column is used. However, the whole
connection behaviour has not been fully characterized when all components
37

can deform and therefore further studies are required..
A review of the available analytical studies performed by different authors625
shows that the spring component method is widely adopted for the compo-
nents in the tension zone of the EHB connection. All models describing the
global force-displacement behaviour of the connection adopt tetra-linear mod-
els and present equations for the stiffness and/or displacement for the four
linear sections of the graph. Analytical models for whole connections are not630
found in the literature.
Sensibility Analysis (SA) was performed using two representative studies of
the column and bolt components of the tension zone of the EHB connection.
Scatterplots and the Elementary Effect (EE) method were used to rank the
importance of the model parameters with respect to their effect on the con-635
nection response to tensile loading. Both methods yielded similar results. The
concrete grade shows to be the most influential parameter in terms of stiffness
of the bolt component, and both strength and stiffness of the column compo-
nent. In the case of the bolt component, the bolt grade has shown to have the
highest effect on the component strength. All parameters considered in the SA640
have shown to influence in the connection response, either in terms of strength
and/or stiffness, and therefore, it is recommended to continue considering them
in future studies.
The considered parameters produce different effects in each independent com-
ponent, i.e., bolts in tension and column face in bending. Therefore, further pa-645
rameter studies are recommended to be performed for combined failure mode,
and beam and column connections in order the identify the most influential
38

parameters for the whole joint.
Funding
This research received funding from the University of Nottingham and materials650
from Lindapter International.
Acknowledgments
The authors wish to acknowledge TATA Steel, Lindapter International, and the
University of Nottingham HPC, for supporting this research.
References655
[1] T. C. Barnett, The Behaviour of a Blind Bolt for Moment Resisting Connections
in Hollow Steel Sections, Ph.D. thesis, University of Nottingham, 2001.
[2] Z.-Y. Wang, Q.-Y. Wang, Yield and ultimate strengths determination of a
blind bolted endplate connection to square hollow section column, Engineering
Structures 111 (2016) 345 – 369.660
[3] T. Pitrakkos, The Tensile Stiffness of a Novel Anchored Blind-bolt Component
for Moment-resisting Connections to Concrete-filled Hollow Sections, Ph.D. the-
sis, University of Nottingham, 2012.
[4] Y. Oktavianus, H. Chang, H. Goldsworthy, E. Gad, Component model for
pull-out behaviour of headed anchored blind bolt within concrete filled circular665
hollow section, Engineering Structures 148 (2017) 210 – 224.
[5] S. A. Mahin, Lessons from damage to steel buildings during the northridge
earthquake, Engineering Structures 20 (1998) 261 – 270.
39

[6] M. Nakashima, K. Inoue, M. Tada, Classification of damage to steel buildings
observed in the 1995 hyogoken-nanbu earthquake, Engineering Structures 20670
(1998) 271 – 281. Innovations in Stability Concepts and Methods for Seismic
Design in Structural Steel.
[7] Federal Emergency Management Agency (FEMA-351), Recommended Seis-
mic Evaluation and Upgrade Criteria for Existing Welded Steel Moment-frame
Buildings, volume Chapter 2, SAC Joint Venture, 2000.675
[8] Blind Bolt Company Ltd, Blind bolt uk, https://www.blindbolt.co.uk/,
2012. Accessed: 13th of February 2020.
[9] Arconic Fastening Systems and Rings, Huck bom brochure, https://www.
arconic.com/, 2017. Accessed: 14th of February 2020.
[10] Advanced Bolting Solutions, Design resistance of molabolt peg anchors, http:680
//molabolt.co.uk/, 2016. Accessed: 14th of February 2020.
[11] Flowdrill Ltd, Flowdrill brochure, https://www.flowdrill.com/, 2020. Ac-
cessed: 14th of February 2020.
[12] Ajax Engineered Fasteners, Oneside brochure b-n012 data sheet, 2002.
[13] Lindapter, Hollo-bolt product brochure, uk, 2019.685
[14] W. Tizani, A. Al-Mughairi, J. Owen, T. Pitrakkos, Rotational stiffness of a
blind-bolted connection to concrete-filled tubes using modified hollo-bolt, Jour-
nal of Constructional Steel Research 80 (2013) 317 – 331.
[15] Y. Kurobane, J. Packer, J. Wardenier, N. Yeomans, Design Guide for Structural
Hollow Section Column Connections, volume 9 of Construction with hollow steel690
40

sections, Comite international pour le developpement et letude de la construc-
tion tubulaire, 2004.
[16] L.-H. Han, W. Li, R. Bjorhovde, Developments and advanced applications of
concrete-filled steel tubular (cfst) structures: Members, Journal of Construc-
tional Steel Research 100 (2014) 211 – 228.695
[17] N. Abd Rahman, Fatigue behaviour and reliability of Extended Hollobolt to
concrete filled hollow section., Ph.D. thesis, University of Nottingham, 2012.
[18] J. France, J. B. Davison, P. Kirby, Moment-capacity and iffness of endplate con-
nections to concrete-filled tubular columns with flowdrilled connectors, Journal
of Constructional Steel Research 50 (1999) 35 – 48.700
[19] L. C. Neves, L. S. da Silva, P. C. da S. Vellasco, Experimental behaviour of
end plate i-beam to concrete-filled rectangular hollow section column joints, in:
Advances in Steel Structures (ICASS ’02), pp. 253 – 260.
[20] H. Yao, H. Goldsworthy, E. Gad, Experimental and numerical investigation of
the tensile behavior of blind-bolted t-stub connections to concrete-filled circular705
columns, Journal of Structural Engineering 134 (2008) 198–208.
[21] W. Tizani, D. J. Ridley-Elis, The performance of a new blind-bolt for moment-
resisting connections., in: Tubular structures X: proceedings of the 10th inter-
national symposium on tubular structures, pp. 395–400.
[22] Y. Oktavianus, H. M. Goldsworthy, E. Gad, Group behavior of double-headed710
anchored blind bolts within concrete-filled circular hollow sections under cyclic
loading, Journal of Structural Engineering 143 (2017) 04017140.
41

[23] Y. Zhang, M. Liu, Q. Ma, Z. Liu, P. Wang, C. Ma, L. Sun, Yield line patterns
of t-stubs connected by thread-fixed one-side bolts under tension, Journal of
Constructional Steel Research 166 (2020) 105932 1–17.715
[24] P. Wang, L. Sun, B. Zhang, X. Yang, F. Liu, Z. Han, Experimental studies on
t-stub to hollow section column connection bolted by t-head square-neck one-
side bolts under tension, Journal of Constructional Steel Research 178 (2021)
106493.
[25] H. Loh, B. Uy, M. Bradford, The effects of partial shear connection in composite720
flush end plate joints part i — experimental study, Journal of Constructional
Steel Research 62 (2006) 378–390.
[26] H. Loh, B. Uy, M. Bradford, The effects of partial shear connection in composite
flush end plate joints part ii—analytical study and design appraisal, Journal of
Constructional Steel Research 62 (2006) 391–412.725
[27] Y. Liu, C. M´alaga-Chuquitaype, A. Elghazouli, Response and component char-
acterisation of semi-rigid connections to tubular columns under axial loads, En-
gineering Structures 41 (2012) 510–532.
[28] A. Ataei, M. A. Bradford, H. R. Valipour, Experimental study of flush end
plate beam-to-cfst column composite joints with deconstructable bolted shear730
connectors, Engineering Structures 99 (2015) 616–630.
[29] A. Ataei, M. A. Bradford, Numerical study of deconstructable flush end plate
composite joints to concrete-filled steel tubular columns, Structures 8 (2016)
130–143.
42

[30] X. Li, Y. Xiao, Y. Wu, Seismic behavior of exterior connections with steel beams735
bolted to cft columns, Journal of Constructional Steel Research 65 (2009) 1438–
1446.
[31] J.-F. Wang, L.-H. Han, B. Uy, Hysteretic behaviour of flush end plate joints to
concrete-filled steel tubular columns, Journal of Constructional Steel Research
65 (2009) 1644 – 1663.740
[32] J. Wang, L. Zhang, B. Spencer, Seismic response of extended end plate joints to
concrete-filled steel tubular columns, Engineering Structures 49 (2013) 876–892.
[33] J. Wang, J. Wang, H. Wang, Seismic behavior of blind bolted cfst frames with
semi-rigid connections, Structures 9 (2017) 91–104. Advances in Steel-Concrete
Composite Structures.745
[34] R. Waqas, B. Uy, H.-T. Thai, Experimental and numerical behaviour of blind
bolted flush endplate composite connections, Journal of Constructional Steel
Research 153 (2019) 179–195.
[35] Ajax Fasteners Innovations, Joint design usin oneside structural fastener, 2005.
[36] A. Gardner, H. Goldsworthy, Moment-resisting connections for composite750
frames, in: Mechanics of structures and materials conference. Balkema, Rotter-
dam, pp. 309–314.
[37] A. Gardner, H. Goldsworthy, Experimental investigation of the stiffness of
critical components in a moment-resisting composite connection, Journal of
Constructional Steel Research 61 (2005) 709 – 726.755
43

[38] H. Yao, H. Goldsworthy, E. Gad, S. Fernando, Experimental study on modified
blind bolts anchored in concrete-filled steel tubular columns, in: Australian
Earthquake Engineering Society Conference, Barossa Valley, Australia, pp. 1–9.
[39] Y. Oktavianus, H. Goldsworthy, E. Gad, Behaviour of headed anchor blind
bolts embedded in concrete filled circular hollow section column, in: Australian760
Earthquake Engineering Society Conference, Lorne, Vic, Australia, pp. 1–9.
[40] Y. Oktavianus, H. Yao, H. Goldsworthy, E. Gad, Pull-out behaviour of blind
bolts from concrete-filled tubes, Structures & Buildings 168 (2015) 747 – 759.
[41] H. Agheshlui, H. Goldsworthy, E. Gad, S. Fernando, Tensile behaviour of an-
chored blind bolts in concrete filled square hollow sections, Materials and Struc-765
tures 49 (2016) 1511–1525.
[42] T. Pokharel, H. M. Goldsworthy, E. F. Gad, Tensile behavior of groups of
double-headed anchored blind bolts within concrete-filled square hollow sections
under cyclic loading, Journal of Structural Engineering 147 (2021) 04020349.
[43] M. Liu, X. Zhu, P. Wang, W. Tuoya, S. Hu, Tension strength and design method770
for thread-fixed one-side bolted t-stub, Engineering Structures 150 (2017) 918
– 933.
[44] X. Zhu, P. Wang, M. Liu, W. Tuoya, S. Hu, Behaviors of one-side bolted t-stub
through thread holes under tension strengthened with backing plate, Journal of
Constructional Steel Research 134 (2017) 53 – 65.775
[45] T. Wulan, P. Wang, Y. Li, Y. You, F. Tang, Numerical investigation on strength
and failure modes of thread-fixed one-side bolted t-stubs under tension, Engi-
neering Structures 169 (2018) 15 – 36.
44

[46] P. Wang, T. Wulan, M. Liu, H. Qu, Y. You, Shear behavior of lap connection
using one-side bolts, Engineering Structures 186 (2019) 64 – 85.780
[47] T. Wulan, Q. Ma, Z. Liu, M. Liu, J. Song, J. Cai, P. Wang, Experimental study
on t-stubs connected by thread-fixed one-side bolts under cyclic load, Journal
of Constructional Steel Research 169 (2020) 106050.
[48] P. Wang, L. Sun, M. Liu, B. Zhang, X. Hu, J. Yu, Experimental studies on
thread-fixed one-side bolted connection of beam to hollow square steel tube785
under static bending moment, Engineering Structures 214 (2020) 110655.
[49] L. Sun, M. Liu, Y. Liu, P. Wang, H. Zhao, J. Sun, Y. Shang, Studies on t-shaped
one-side bolted connection to hollow section column under bending, Journal of
Constructional Steel Research 175 (2020) 106359.
[50] European Committee for Standardisation (CEN), Design of steel structures,790
Part 1-8: Design of joints, Eurocode 3, 2005. EN 1993-1-8.
[51] J.-F. Wang, L.-H. Han, B. Uy, Behaviour of flush end plate joints to concrete-
filled steel tubular columns, Journal of Constructional Steel Research 65 (2009)
925 – 939.
[52] Federal Emergency Management Agency (FEMA-350), Recommended seismic795
design moment-frame buildings, SAC Joint Venture, 2000.
[53] T. Barnett, W. Tizani, D. Nethercot, The practice of blind bolting connections
to structural hollow sections: A review, Steel and Composite Structures 1 (2001)
1–16.
[54] S. Ellison, W. Tizani, Behaviour of blind bolted connections to concrete filled800
hollow sections, Structural Engineering 82 (2004) 16–17.
45

[55] T. Pitrakkos, W. Tizani, Z. Wang, Pull-out behaviour of anchored blind-bolt:
a component based approach, Proceedings of the International Conference on
Computing in Civil and Building Engineering (ICCCBE), pp. 509 1–7.
[56] M. Mahmood, Column Face Bending of Anchored Blind Bolted Connections to805
Concrete Filled Tubular Sections, Ph.D. thesis, University of Nottingham, 2015.
[57] J. Ribeiro, A. Santiago, C. Rigueiro, L. S. da Silva, Analytical model for the
response of t-stub joint component under impact loading, Journal of Construc-
tional Steel Research 106 (2015) 23 – 34.
[58] I. O. for Standardization, Mechanical properties of fasteners made of carbon810
steel and alloy steel — Part 1: Bolts, screws and studs with specified property
classes — Coarse thread and fine pitch thread, Standard, ISO, 2009.
[59] A. M. Pascual, M. L. Romero, W. Tizani, Thermal behaviour of blind-bolted
connections to hollow and concrete-filled steel tubular columns, Journal of Con-
structional Steel Research 107 (2015) 137 – 149.815
[60] A. M. Pascual, M. L. Romero, W. Tizani, Fire performance of blind-bolted
connections to concrete filled tubular columns in tension, Engineering Structures
96 (2015) 111 – 125.
[61] M. F. Shamsudin, Group Behaviour of Extended HolloBolts (EHBs) in Tension,
Ph.D. thesis, University of Nottingham, 2019.820
[62] M. Cabrera, W. Tizani, M. Mahmood, M. F. Shamsudin, Analysis of extended
hollo-bolt connections: Combined failure in tension, Journal of Constructional
Steel Research 165 (2020) 105766 1–14.
46

[63] P. P. Debnath, T.-M. Chan, Tensile behaviour of headed anchored hollo-bolts in
concrete filled hollow steel tube connections, Engineering Structures 234 (2021)825
111982.
[64] T. Pitrakkos, W. Tizani, Experimental behaviour of a novel anchored blind-bolt
in tension, Engineering Structures 49 (2013) 905 – 919.
[65] Steel Construction Institute (SCI) and British Constructional Steelwork Associ-
ation (BCSA), Joints in steel construction: Moment-resisting joints to Eurocode830
3, Steel Construction Institute, 2013.
[66] T. Pitrakkos, W. Tizani, M. Cabrera, N. Faqe Salh, Blind bolts with headed
anchors under combined tension and shear, Journal of Constructional Steel
Research 179 (2021) 106546.
[67] Z. Wang, Hysteretic response of an innovative blind bolted endplate connection835
to concrete filled tubular columns, Ph.D. thesis, University of Nottingham, 2012.
[68] W. Tizani, Z. Y. Wang, I. Hajirasouliha, Hysteretic performance of a new
blind bolted connection to concrete filled columns under cyclic loading: An
experimental investigation, Engineering Structures 46 (2013) 535 – 546.
[69] Y. Wang, Z. Wang, J. Pan, P. Wang, Nonlinear finite element analysis of an-840
chored blind-bolted joints to concrete-filled steel tubular columns, International
Journal of Performability Engineering 15 (2019) 676–687.
[70] Y. Wang, Z. Wang, J. Pan, P. Wang, J. Qin, S. Chen, Cyclic behavior of an-
chored blind-bolted extended end-plate joints to cfst columns, Applied Sciences
10 (2020).845
47

[71] A. Saltelli, M. Ratto, T. Andres, F. Campolongo, J. Cariboni, D. G. M. Saisana,
S. Tarantola, Global sensitivity analysis, The primer. John Wiley & Sons, 2009.
[72] M. Morris, Factorial sampling plans for preliminary computational experiments,
Technometrics 33 (1991) 161–174.
[73] F. Campolongo, J. Cariboni, A. Saltelli, An effective screening design for sensi-850
tivity analysis of large models, Environmental Modelling & Software 22 (2007)
1509 – 1518.
48