Experimental Behaviour of I Beam – SHS Column Steel Joints

with Hollo-bolts
1
Mesquita, A.B. , Simões da Silva, L.A.P.2, Jordão, S.3
1
Assistant, Civil Engineering & Architecture Department, University of Beira Interior, Covilhã, Portugal
2
Full Professor, Civil Engineering Department, University of Coimbra, Coimbra, Portugal
3
Assistant Professor, Civil Engineering Department, University de Coimbra, Coimbra, Portugal

ABSTRACT: The main objective of this paper is to present the results from an experimental analy-
sis aiming at the characterization of the behaviour of a new feature steel bolted joint, that is not yet
considered in Eurocode 3. The referred joint is I Beam – SHS Column using an innovative fasten-
ing solution with the commercial designation of hollo-bolt. This kind of device is a pre-assembled
unit that allows fixing and fastening when access is possible from one side only. This is particularly
useful for nodes in which the column is a hollow profile, which has recently become an increas-
ingly popular design solution. The test programme is currently being undertaken at the Civil Engi-
neering Department of the University of Coimbra and encompasses 17 tests, accounting for exter-
nal, corner and internal façade node configurations, flush and extended end-plate type, on full scale
prototypes. Some preliminary isolated hollo-bolt tests were also performed in order to fully charac-
terize its behaviour when subjected to tension and shear. The results are presented in terms of char-
acterization of initial stiffness and resistance for each configuration. A comparison is established,
for the referred parameters, between nodes with the same typology. Some preliminary conclusions
concerning the main parameters and phenomena influencing the behaviour of the new feature joints
are presented and discussed. Three main phenomena are in evidence: i) membrane behaviour re-
sponse of the tubular column wall; ii) influence of hollo-bolts on the column holes area yielding;
iii) three-dimensional effect when an orthogonal cantilever beam is added.

1 INTRODUCTION

In recent years, steel structures with hollow section elements have been an attractive choice for sev-
eral design solutions because of its structural and architectural potential. According to CORUS®
Group, the structural utilization of SHS profiles, comparatively to the open sections profiles (I, H,
U, L, T), presents several advantages: i) A very good resistance to the axial compression and
torque; ii) A good behaviour to the corrosion, due to the absence of alive edges; iii) A larger easi-
ness in the painting of the surfaces, taking into account the need of maintenance and consequent
economy. From the architectural point of view, the advantage is that hollow profiles can produce
fluid and slender surfaces with metallic texture.
Although the structural connections, that involve tubular profiles, can be majorly classified as par-
tial-strength and possess a simple geometry, the distribution of stresses and behaviour are more
complex comparatively to the profiles with rigid central web. The referred complexity requires the
definition of methods and calculation rules tailored for this type of joint, including formulation for
assessing resistance and rotational stiffness. In the particular case of the joint studied, it is also nec-
essary to develop formulation for the new component “hollo-bolt in combined bending, tension and
shear”. In fact, the inexistence in Part 1.8 of EC3 (2005) of calculation rules for tubular sections
compatible with column and beam sizes typical of current steel frames including the fixation system
problem, have been leading to the need of research on the behaviour of that type of cross-sections.
Some international institutions as International Institute of Welding (IIW) and Comité International
230
pour le Développement et l´Étude de la Construction Tubulaire (CIDECT), have been trying to
tackle the problem aiming at the uniformity of the calculation rules.
The main objective of the research work presented branches in this goal, since it aims at character-
ising the structural behaviour of I beam - SHS column steel joints with an innovative fastening solu-
tion, the hollo-bolts, as referred in Simões da Silva et al (2003).

2 CONFIGURATION AND TYPOLOGY OF PROTOTYPES

2.1 Joints configurations

Full-scale prototypes were used representing external (Figure 1a)) and corner (Figure 1b)) joint con-
figurations. Beams are welded to the end-plates, flush or extended, which are bolted to columns us-
ing hollo-bolts.

a) External joint - lateral and 3D view b) Corner joint - 3D view

Figure 1. Prototypes configurations

2.2 Connections typologies

It was decided to use typologies that are considered in Eurocode 3 in order to enable future com-
parisons with tests performed by other authors. Furthermore, it was intended to keep the number of
unknown variables to a minimum since the new joint encompasses a new uncertainty related with
the behaviour of the hollo-bolt. Figure 2a) shows the flush end-plate A configuration, and Figure
2b) exhibits the extended end-plate used for the tests with typology B, (A/B, detailed in Table 1).

a) Typology A - Flush end-plate b) Typology B - Extended end-plate

Figure 2. Connections’ typologies: Front view
231
The values chosen for the distance between the middle lines of bolt holes, either vertical or horizon-
tal (e1, e2, p1 or p12 , referred on EC3), were obtained from a previous optimization calculation.
The Figure 3a) illustrate a 3D view of the corner joint and a cross section through the hollo-bolt
middle plane. Accordingly with the manufacturer information (Lindapter), hollo-bolts are a pre-
assembled five part fitting including a standard bolt, a cylindrical body with 4 legs, a collar, a
washer and a tronco-conical nut, as shown on Figure 3b).

M1
M2

a) 3D view of the corner joints b) Components of the hollo-bolt (in: British Steel) c) Hollo-bolt
Figure 3. Hollo-bolt : pre-assembled unit inserted in the column wall hole

Table 1 summarizes the data for each test in the experimental programme, describing the type of
node, details on the joint typology, number, diameter and tightening torque of the hollo-bolts and
loading strategy.

Table 1. Tests detail summary

Hollo- Torque Monotonic
Nº Column / beam / end- plate Configuration Typology
Bolts [Nm] loading
. E01_1 SHS 200 x 8 / IPE 330 / A External joint Flush M16 (4) 190 P1 (↓ )
. E01_2 SHS 200 x 8 / IPE 330 / A External joint Flush M20 (4) 480 P1 (↓ )
. E01_3 SHS 200 x 8 / IPE 330 / A External joint Flush M20 (4) 550 P1 (↓ )
. E03_1 SHS 200 x 12 / IPE 330 / B External joint Extended M20 (6) 550 P1 (↓ )
. E03_2 SHS 200 x 8 / IPE 330 / B External joint Extended M20 (6) 550 P1 (↓ )
. E04_1 SHS 200 x12 / IPE 330 (2) / A Corner joint Flush M20 (8) 500 P1 + P2 (↓ )
. E05_1 SHS 200 x 12 / IPE 330 (2) / B Corner joint Extended M20 (12) 500 P1 + P2 (↓ )
A – Flush: 370x170x20 ; B – Extended: 425x170x20

2.3 Material properties

The steel grade of the columns is S355 J2 H; for beams and end-plates the steel grade is S355 J0.
The bolts are 8.8 grade. Coupon tests are being prepared, to assess the full nonlinear material be-
haviour curve. Considering a previous set of push-out tests of hollo-bolts, contemplating simultane-
ously M16 and M20, it was obtained a tensile strength capacity of 92,5 kN and 188,5 kN respec-
tively for each bolt. These values were obtained for a 190 Nm (M16) and 500 Nm torque (M20).

3 EXPERIMENTAL SET UP

3.1 Layout
Figure 4 shows some views of the experimental layout for external and corner nodes. The layout
accounts for two orthogonal directions, needed for the corner nodes testing. The upper and lower
supports of columns are simulated by pinned joints, allowing that the rotations occur on the same
plane of the reaction wall, considering bolted plates on top and bottom connected do that pinned
joints, according to Figures 1 and 4.

232
Figure 4. Different perspectives of layout for external and corner joints

3.2 Measuring devices

The main purpose of the measuring devices is to obtain information on the behaviour of the node
and to obtain values for the calibration of the numerical model. Displacement transducers were used
to characterize the relevant node rotation components, so that the joint rotation could be precisely
assessed. They were also used to measure key displacements through out the structure that will
bring in useful information on local and global structural behaviour. Strain gauges were used to
characterize the stress field in key points through out the structure. The beam tip displacement, for
each test, was measured by a LVDT positioned at 1,0 m from column face under bottom beam’s
flange (Figure 1). Horizontal displacements were measured on heads bolts.
Figure 5a) illustrate a schematic disposition of displacement transducers adopted for tests with ex-
tended end-plates, while Figure 5b) shows a photo of the first corner joint test.
TML

17
L pilar sup

0,500
13
= 3,000 m

1 // 2 // 3 +P
TML
• TML
4 // 7 Célula 40 Tf (x2)
9
0,220
viga

TM L • TM L 5
pilar

TM L • TM L 6
L

10 TML Célula 10 Tf (x2)

8
14 15 16
L pilar inf

0,25 0,25 0,500

L viga = 1,0 m
TM L

18

SHS 200x 12
z
LVDT´s
TM L 1
TM L TM L
y 2
LVDT´s 9 // 10 TM L
3

IPE 330
LVDT´s 11 // 12 // 19 ?

a) LVDT´s disposition for tests E03_1/2/E05 b) Test E04_1 general view

Figure 5. Exemplification of measuring devices

4 TEST PROCEDURES

A load, unload and reload history was in general considered for the performed tests. The load in
elastic range varies from 30 kN to 50 kN in some tests. For other tests, considering a more signifi-
cant plastic deformation, the value of 70 kN was selected. The control of equipment load was done
in terms of displacement considering velocities between 0,02 mm/s and 0,03 mm/s for the final re-
load until the end of external joints tests. For corner joints tests, where two load equipments were
involved the experimental procedure requires an incremental load applying in terms of displacement
control verifying in permanence the values of the beams tip displacements in order to adjust as
quickly as possible load P1 and load P2, once the electromechanical equipment controls are differ-
ent, respectively Servosis system and Dartec system.

233
5 EXPERIMENTAL RESULTS

5.1 Load-displacement diagrams - F-d

Figure 6a) shows the load-horizontal displacement diagram for external joint test E01_3 (flush end-
plate). On Figure 6b) it can be seen the correspondent hollo-bolts F-dx diagram. Figure 6c) shows
the load-vertical displacement diagram for the test E03_1 (extended end-plate).
80 180
80

70 160
70
140
60 60
120
50 50

Load (P1) [kN]

Load (P1) [kN]

Load [kN]
100
40 40
80
30 30
60
HB-1
20 F-dh _ 1 20
HB-2
F-dh_2 40
HB-3
10 F-dh_3 10
HB-4 20
F-dh _8
dz_LVDT 16
0 0
0
0 1 2 3 4 5 6 7 8 9 10 0 4 8 12 16 20 24 28
0 10 20 30 40 50 60 70 80
Horizontal displacement on SHS [mm] Horizontal displacement on hollo-bolts [mm]
Vertical displacement [mm]

a) Test E01_3 (F-dx on column) b) Test E01_3(F-dx on bolts) c) Test E03_1 (F-dz on beam)
Figure 6. F-d diagrams

These three diagrams are selected from the other cases in order to illustrate the general graphical
evolution of the load-displacement curves that represents the behaviour of tested SHS columns and
beams as well as hollo-bolts (upper and lower rows). The respective coherence with the moment-
rotation curves is observed in nest paragraph.

5.2 Moment-Rotation diagrams - M-Φ

5.2.1 External Joints (flush and extended end-plate)

The experimental moment-rotation characteristic curves obtained from tests corresponding to exter-
nal joints with flush end-plates are summarized on Figure 7.
80
80
70 70
60
60 50
Moment (Load P1) [kNm]

50 40
30
Moment [kNm]

40 20
30 10
0
20
-0 -0-10 0 0,02 0,04 0,06 0,08 0,1 0,12 0,14 0,16 0,18
10 -20
-30
0 -40
0 0,01 0,02 0,03 0,04 0,05 0,06 0,07 0,08 -50
Rotation [rad] Rotation [rad]

a) Test E01_1 b) Test E01_2

80
70
60
Moment [kNm]

50
40
30
20
10
0
0 0,01 0,02 0,03 0,04 0,05 0,06 0,07 0,08 0,09 0,1 0,11 0,12 0,13
Rotation [rad]

c) Test E01_3
Figure 7. M-Φ diagrams for external joints with connections by flush end-plate

234
With tests E01_1/2/3, it could be verified a deformability on SHS section wall and a rigid body ro-
tation of the flush end-plate when hollo-bolt legs failure happened, typical in all tests. It can be ob-
served on Figures 10a), d). A bigger deformability occurred on tests E01_2 and E01_3, considering
the higher hollo-bolt diameter in these last cases, according with Table 1, comparing with E01_1.
Figure 8 shows the experimental curves for extended end-plates on external joints.

175 175
150 150
125 125
100

Moment [kNm]
Moment [kNm]

100
75 75
50 50
25 25
0 0
0 0,01 0,02 0,03 0,04 0,05 0,06 0,07 0,08 0 0,01 0,02 0,03 0,04 0,05 0,06 0,07 0,08
Rotation [rad] Rotation [rad]

a) Test E03_1 b) Test E03_2

Figure 8. M-Φ diagrams for external joints with connections by extended end-plate

For the test E03_1, considering the major value of thickness wall of tubular profile (12mm) com-
paring with the first three tests (8 mm), it occurred also the failure of the hollo-bolt legs simultane-
ously with the bending of the extended end-plate and a very small deformability on the SHS col-
umn, according to Figure 10b). On the case of test E03_2, Figure 10c), even though this test
presents the same typology of test E03_1, the column has only 8 mm of thickness, lower than the
previous test. So, it could be verified a bigger deformability on the SHS wall and an almost rigid
body rotation of respective extended end-plate. The plastic deformations near of column upper
holes shows the tendency to the membrane behaviour of the load face column for sections with 8
mm of thickness.

5.2.2 Corner Joints

Figure 9 presents the moment-rotation diagrams for both corner tested joints.
120 250

100 200

80
150
Moment [kNm]
Moment [kNm]

beam 2 (Load 2 => M2)

60
100

40 Beam 2 (Load P2 => M2) joint-Rot 2 (beam)

50 joint Rot 1 (node)
beam 1 (Load 1 => M1)
joint Rot 1 (r/ node) joint_Rot 1 (node)
20 Beam 1 (Load P1 => M1)
joint Rot (r - beam)
joint Rot 2 (beam)
joint Rot 2 (r/ node)
joint Rot (r - beam)
0
0 0 0,01 0,02 0,03 0,04 0,05 0,06 0,07 0,08
0 0,01 0,01 0,02 0,02 0,03 0,03 0,04 0,04 0,05 0,05 0,06 0,06
Rotation [rad] Rotation [rad]

a) Test E04_1 b) Test E05_1

Figure 9. M-Φ diagrams for corner joints with connections by flush and extended end-plate

As in the other cases of configuration and typology, the rotation can be calculated according to the
expression,
θ jo int = θ node − θ web column (1)
where rotation of the node was obtained from difference between horizontal displacements on
heads bolts divided by the distance between their axis (LVDT 5 and 6 from Figure 5a)), and the ro-
tation on web column calculated from the difference between displacements measured on the oppo-
235
site loaded face (LVDT 9 and 10 from the Figure 5a)). Alternatively, rotation of the node can be ob-
tained taking into account LVDT 15 or 16 under flange beam as indicated on the same figure.

5.2.3 Deformabilities sources

Figure 10 shows the principal sources of deformability observed after laboratorial tests witch repre-
sents the general kinds of behaviour of the external joints also presented by corner joints.

a) Tests E01_1/2/3/04_1 b) Tests E03_1 and E05 c) Test E03_2 d) Failure legs of Hollo-bolts
Figure 10. Deformabilities verified during the execution of tests

5.3 Initial Rotational Stiffness and Plastic Moment for external and corner joints

Table 2 summarizes the values of the initial rotational stiffness, the plastic moment and the ultimate
moment for each tested joint, obtained from the experimental curves, by means of a graphical
method proposed by Jaspart (1991).

Table 2. Results of initial rotational stiffness, plastic moment and ultimate moment
HB Torque Sj,ini Mj,Rd Mu, exp
Nº Column / beam / end-plate
[Nm] [kNm/rad] [kNm] [kNm]
E01_1 SHS 200 x 8 / IPE 330 / A 190 (M 16) 2807,1 32,0 54,3
E01_2 SHS 200 x 8 / IPE 330 / A 480 (M 20) 4650,5 53,0 64,8
E01_3 SHS 200 x 8 / IPE 330 / A 500 (M 20) 3362,5 42,0 74,5
E03_1 SHS 200 x 12 / IPE 330 / B 550 (M 20) 12335,0 87,5 161,9
E03_2 SHS 200 x 8 / IPE 330 / B 550 (M 20) 11388,0 77,0 120,5
14276 - J1 79,0 101,7 (M1)
E04_1 SHS 200 x12 / IPE 330 (2) / A 500 (M 20)
17180 - J2 85,0 113,5 (M2)
15898 - J1 112 155,6 (M1)
E05_1 SHS 200 x 12 / IPE 330 (2) / B 500 (M 20)
27262,5- J2 141,5 204,9 (M2)

The referred method considers that the initial stiffness is obtained from the slope of the elastic por-
tion of the curve, and that the plastic moment is obtained from the intersection of two lines (“a” and
“b”), represented in Figure 11.
60
y = 494,32x + 26,514

50 Line “a”
y = 2808,9x + 0,1153
Moment (Load P1) [kNm]

40

Point “P” Line “b”

30

y = 2805,7x - 84,022
20

Sj,ini
10
Sj Sj = Sj,ini /2

0
-0,01 0 0,01 0,02 0,03 0,04 0,05 0,06 0,07 0,08

Rotation [rad]

Figure 11. Graphical determination of structural experimental properties: general consideration

236
Line “a” intersects the axes origin and its slope is the initial rotational stiffness (elastic stiffness).
Line “b” is obtained by linear regression adjusting to a portion of the curve, in the after yielding
range, starting on point P, represented in Figure 11. Point P is defined by the intersection of the ex-
perimental curve with a line with a slope of Sj,ini/η (η =2, bolted connection – EC3). Taking into ac-
count the moment-rotation diagrams for the first tests as well as the results listed in Table 2, it can
be concluded that the values of Sj,ini and Mj,Rd increase with the hollo-bolt diameter and respective
torque applied. On tests E03_1 and E03_2 the values of initial stiffness and plastic moment are con-
siderably superior to those obtained with the series E01. This fact occurs due to the use of the ex-
tended end-plate. However, in the second of these last tests, the lower thickness of SHS wall origi-
nated a decrease of structural properties. These same properties are higher on test E04_1, relatively
to the previous tests. Analogous conclusion can be verified between this one and test E05_1. For the
corner joints connections, we can verify different results for initial rotational stiffness and plastic
moment between beam 1 and beam 2 for each test.
It is interesting to state that the different behaviour of the two beams and the correspondent or-
thogonal hollo-bolts results also of the different static loads applied in several steps of the incre-
mental process, inspite the intention to minimize the displacements difference during the execution
of tests. When the load 1 is applied on beam 1, the second beam moves due to the instantaneous un-
balance between orthogonal loaded SHS faces. Considering the relation Mj,rd 1/Mj,rd 2 we obtain 93%
for test E04_1 and 79% for test E05_1.

6 CONCLUSIONS

With the performed group tests it is possible to verify the failure of the hollo-bolt legs under tension
and shear introduced by the bending on beam as well as the increase of Sj,ini and Mj,Rd values not
only with the hollo-bolt diameter and respective torque applied but also with an extended end-plate.
On the other side, the membrane behaviour effect, typical of tubular columns under orthogonal
forces on loaded faces (Neves, 2004) will be more visible on the SHS sections with smaller thick-
ness, which will be taken into account in the second part of experimental programme. However,
such tendency can be observed on prototypes with columns of 8 mm of thickness and M20 hollo-
bolts. Also the influence of a second orthogonal cantilever beam can produce a modification in the
structural joint behaviour, as the example between tests E03_1 and E05_1: equal sections, equal
hollo-bolts, similar torque, but higher values for the last test in both directions, in what concerns to
strength and stiffness.

7 REFERENCES

British Steel, Tubes & Pipes. 2005. London. British Steel plc.
CEN (2005) Eurocode 3: Design of Steel Structures, Part 1-1: General rules and rules for buildings, EN-
1993-1-1, European Committee for Standardization, Brussels.
CEN (2005) Eurocode 3: Design of Steel Structures, Part 1.8: Design of Joints, EN-1993-1-8, European
Committee for Standardization, Brussels.
CORUS Tubes library publication.
ECCS, Technical Committee TC 8 – Structural Stability, Technical Working Group 8.1: Components, Test-
ing Procedures, First edition, nº. 36, 1984.
Jaspart , J.P., Étude de la semi-rigidité des noeuds poutre-colonne et son influence sur la résistence et la sta-
bilité des ossatures en acier, Ph.D. Thèse, 1991, Fac. des Sciences Appl., Université de Liège, (in french).
Lindapter International, cavity fixings: Hollo-bolt. Bradford.
Mesquita, A. (2008), “Hollo-bolt Push-out Tests”, Internal Report, DEC, University of Coimbra.
Neves, L.F.C. (2004), “Comportamento de ligações de eixo fraco e entre pilares tubulares e vigas em secção
I”, Ph.D. Thesis, Universidade de Coimbra, Coimbra, (in portuguese).
Simões da Silva, L. e Santiago, A. (2003), Manual de Ligações Metálicas. Coimbra. CMM ,(in portuguese).

237