See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/281467748

Analytical Study of End-Plate Connection on Cruciform Column Section

Article  in  Jurnal Teknologi · May 2015

DOI: 10.11113/jt.v74.4602

CITATIONS READS

0 2,160

4 authors, including:

Poi Ngian Shek Mahmood Md. Tahir

Universiti Teknologi Malaysia 46 PUBLICATIONS   251 CITATIONS   
88 PUBLICATIONS   289 CITATIONS   
SEE PROFILE
SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Fire Resistance of Interlocking Brick Column under Normal Working Load View project

All content following this page was uploaded by Poi Ngian Shek on 22 January 2016.

The user has requested enhancement of the downloaded file.

 Jurnal Full paper
Teknologi

Analytical Study of End-Plate Connection on Cruciform Column Section

Boon Cheik Tan, Poi Ngian Shek*, Mahmood Md Tahir, Ker Shin Mu

UTM Construction Research Centre, Faculty of Civil Engineering, Universiti Teknologi Malaysia, 81310 UTM Johor Bahru, Malaysia

*Corresponding author: [email protected]

Article history Abstract

Received: 10 November 2014 This paper presents an analytical study on flush end-plate (FEP) and extended end-plate (EEP)
Received in revised form: connections connected to cruciform column section using component method. The objective of this study
23 January 2015 is to predict the moment resistance and initial stiffness of FEP and EEP connections on cruciform column
Accepted: 12 April 2015 section. A series of FEP and EEP connections are tested in laboratory. The connection tests consist of
four FEP and four EEP specimens with different configuration. Component method outlined in the
Graphical abstract publication of Steel Construction Institute and British Constructional Steelwork Association are based on
BS5950 and Eurocode 3 (EC3) are used to predict the moment resistance and initial stiffness of the tested
specimens. The experimental results are then used to validate the analytical predictions. As compare to
the experimental results, all moment resistance of the connections coincide well with analytical
predictions. Analytical prediction for initial stiffness using EC3 does not show good agreement with the
experimental results. This study shows that the component method can be used to predict the moment
resistance of FEP and EEP connections on cruciform column section. Further study need to be carried out
for initial stiffness to obtain accurate analytical representation.

Keywords: Flush end-plate; extended end-plate; component method; moment resistance; initial rotational
stiffness

© 2015 Penerbit UTM Press. All rights reserved.

1.0 INTRODUCTION empirical, analytical, mechanical and finite element models

[12].
The usage of build-up section is not new in engineering This study focused on the analytical prediction of steel
construction. Many researchers have studied on build-up connection calculated in accordance to the guideline published
section such as girders and columns. The fabrication and by Steel Construction Institute (SCI) and British Constructional
connection on build-up sections are always problematic and Steelwork Association (BCSA) using BS5950 and EC3 [13,
increase construction cost and time. 14]. The component method is used to study the moment
Cruciform column is made up of two universal beam resistance and initial stiffness of EEP and FEP connections. A
sections where one of the beams is cut into half and welded to series of FEP and EEP connections were tested experimentally
the other beam. By joining the two universal beams increase the to validate the analytical predictions. The arrangement of the
cross sectional area of the column and hence increase the test specimens are build-up hybrid beam sections connected to
compression resistance. The fabrication of cruciform column is cruciform column using either FEP or EEP connections as
simpler as compared to other build-up column such as laced shown in Figure 1.
column and battened column. Therefore the construction cost
and time may also be reduced [1-3].
One of the most important parts in designing steel 2.0 ANALYTICAL STUDY
structures is the connection between members. Connection is
the location where two or more elements meet. A steel structure 2.1 Connection Configurations
can perform its best to the desired function when the
connections between members are design adequately [4]. The Eight specimens consist of four FEP connections N1 to N4 and
fabrication cost of connections are high in steel structures. four EEP connections N5 to N8. All 8 specimens are using
Optimum connection design is often the main aim for build-up beam sections with four different sizes connected to
researchers to save cost and avoid wastage [5]. cruciform column section using end-plate connections. The
There are three important structural properties need to be configurations of the connections are shown in Table 1.
identified in connections which are the strength, stiffness and
ductility. By having the moment resistance and initial stiffness,
the moment rotation curve can be plotted. The moment rotation
curve represents the characteristics of connections which are
able to indicate the connection properties in terms of strength,
stiffness and ductility [6-11]. Experimental tests provide the
most accurate information to evaluate the reliability of

74:4 (2015) 7–11 | www.jurnalteknologi.utm.my | eISSN 2180–3722 |

8 Poi Ngian Shek et al. / Jurnal Teknologi (Sciences & Engineering) 74:4 (2015), 7–11

Figure 1 Flush end-plate and extended end-plate connection on cruciform column section [3]

Table 1 Configuration of FEP and EEP connections

Connection Column Bolt Row End Plate Bolt

Specimen Welded Beam Size
Type Size (T-B) Size Size
N1 FEP 400×140×41.13/12/5 200×12 20
1(2-2)
N2 FEP 500×180×63.59/16/5 200×12 20

N3 FEP 450×160×46.86/12/5 200×12 20

2(4-4)
N4 FEP 600×200×85.91/16/6 533x210x16 200×12 20

N5 EEP 400×140×41.13/12/5 CCUB 200×12 20

2(4-2)
N6 EEP 500×180×63.59/16/5 250×15 24

N7 EEP 450×160×46.86/12/5 200×12 20

3(6-4)
N8 EEP 600×200×85.91/16/8 250×15 24

2.2 Component Method Table 2 Component method design check on different zones [14]

The component method for determining the moment Zone Reference Item
resistance of the connection was based on the design
procedures outlined in the publication of SCI and BCSA
using BS5950 [13] and EC3 [14]. The component in a Bolt tension
b End-plate bending
connection is break down into its individual components
c Column flange bending
which consist of tension zone, compression zone, horizontal Tension d Beam web tension
shear zone and vertical shear zone as shown in Figure 2 and e Column web tension
Table 2. f Flange to end-plate weld
g Web to end-plate weld

Horizontal
h Column web panel shear
Shear
j Beam flange compression
Compression k Beam flange weld
l Column web
m Web to end-plate weld
Vertical
n Bolt shear
Shear
p Bolt bearing (plate or flange)

The component method starts with the calculation of the

resistance of bolt rows. Bolt resistance for each individual
bolt row in the tension zone is calculated. In calculating the
bolt resistance, the capacity of column flange bending,
column web in tension, end-plate bending and beam web in
Figure 2 Component method design check on different zones [14]
tension are calculated. Resistance of group bolt rows are then
considered using the same procedures as individual bolt row.
The least value from the calculation is considered as the
control value for the resistance of each bolt row.
9 Poi Ngian Shek et al. / Jurnal Teknologi (Sciences & Engineering) 74:4 (2015), 7–11

The bottom flange of the beam is checked under the Table 3 Component evaluation of specimen N4 using Eurocode 3
compression zone. The compression capacity is determined
from the least value of beam flange compression, column web
Specimen FRd Mj.Rd.EC3
crushing and column web buckling. Component
(N4) (kN) (kNm)
In shear zone, the shear is acting in two directions,
horizontal and vertical. However, horizontal shear is usually
End plate in bending 282.2
the most critical and is depending on whether it is one sided
a Column flange in bending 250.3
or two sided connection. For one sided connection, the Row 1
Beam web in tension 585.3
horizontal shear force is equal to the compression force at the Column web in tension 691.1
bottom flange of the beam. For vertical shear, the capacity is
calculated as the reduced capacity of bolt rows acting in the End plate in bending 282.2
tension zone while full shear values for bolt rows acting in the Column flange in bending 250.3
Row 2 a 221.6
shear zone. Beam web in tension 585.3
Theoretically, the forces in tension beam flange is Column web in tension 691.1
equivalent to the forces in compression beam flange, while
the horizontal shear force is equal to the compression force at Row 1 and
Column flange in bending 200.0
bottom of beam flange. Based on this assumption, the least row 2a
Column web in tension 760.8
value calculated in tension, compression and shear is taken as (in group)
the capacity of the connection. Then, mode of failure can be
predicted. The moment resistance of the connection is Column web in transverse compression 687.6
calculated from:
n
M j .Rd .EC 3   F
i 1,2
ri  hi  (1) Beam flange and web in compression 1492
a
No reduction on effective resistance of bolt row when sum of
where Fri is the resistance of bolt row in the tension zone, effective design tension resistance is less than compressive resistance of a
hi is distance of the joint.
ith bolt row from the centre of compression
n is the number of bolt rows.
3.2 Comparison between Experimental and Analytical
2.2 Initial Stiffness Results in Terms of Moment Resistance

The initial stiffness of a connection can be calculated based The moment resistance and initial stiffness for all specimens
on the provision given in Eurocode 3: Part 1-8 Design of Steel are calculated based on the component method. The
Structures [4]. The initial stiffness of a joint is determined experimental results are plotted in moment rotation curves
from the flexibilities of its basic components represented by and the important values such as moment resistance and
an elastic stiffness coefficient ki. The formulation for the initial stiffness are extracted from the curves and summarized
initial stiffness of a joint is given by: in Table 4 and Table 5. The comparison between the
analytical and experimental results are shown in Table 4.
Ez 2
Sj  (2) From the comparison, it shows that the moment
1
 i resistance obtained from analytical model coincide well with
ki the test results where the moment resistance lies between the
where ki is the stiffness coefficient for basic joint ratios of 0.99 to 2.04 for BS5950 and 0.98 to 1.79 for EC3.
component i, The ratio for FEP is in the range of 1.22 to 2.04 for BS5950
z is the lever arm and 1.19 to 1.79 for EC3. For the extended end-plate lies in
μ is the stiffness ratio Sj,ini / Sj. the range of 0.99 to 1.61 for BS5950 and 0.98 to 1.41 for
EC3. As compared to EC3, moment resistance obtain from
The initial stiffness can be calculated from the BS5950 is more conservative. The highest percentage
expression by letting μ = 1.0. The stiffness coefficient ki is difference between EC3 and BS5950 is 12.76%. Both design
calculated for each component in the critical zones such as the guidelines results are closely match.
column web in shear, tension and compression zones (k1, k2 The failure modes of all tested specimens are either end-
and k3), column flange in tension zone (k4), bolts in tension plate yielding or column flange yielding followed by the bolt
zone (k5), and end-plate in tension zone (k6). slippage. According to analytical model from the component
method, the failure mode are end-plate yielding for specimens
with end-plate thickness of 12mm while specimens with end-
3.0 RESULTS AND DISCUSSION plate thickness of 15mm fail in column flange. In this study,
the failure modes are govern by the end-plate and column
3.1 Component Method flange. The failure will occur at column flange when the
thickness of end-plate is higher. Therefore it is suggested that
The moment resistance for all specimens are calculated using the thickness of column flange thickness should be higher
the component method. Table 3 is an example of calculated than the end-plate to avoid failure at column.
component method on specimen N4 using EC3. The lesser From the result, the analytical model by component
value in the component of a bolt row will be the bolt row method can be used to calculate the moment resistance and
resistance. Failure mode can be predicted from the analytical predict the failure mode of the proposed end-plate
calculation. From the table, specimen N4 is predicted to fail at connections.
column flange.
10 Poi Ngian Shek et al. / Jurnal Teknologi (Sciences & Engineering) 74:4 (2015), 7–11

3.3 Initial Stiffness are closer to the experimental value. This shows that the
assumption made is reasonable. Therefore, in predicting the
Along the uncut universal beam section is term as major axis initial stiffness for welded axis, the depth of the column web,
of cruciform column section. The other universal beam hc is taken as half of the depth of universal beam web, hb.
section is cut into half and welded to the uncut section is The ratios of experimental results to analytical initial
termed as welded axis. On the welded axis the section is not stiffness lie in the range of 0.49 to 1.39. The ratios are in the
continuous and jointed by fillet weld. In calculating the initial range of 0.64 to 1.07 for FEP connections and 0.49 to 1.39 for
stiffness, the depth of the column, hc is assumed half of the EEP connections. Analytical model based EC3 is not suitable
universal beam depth, hb for the welded axis. to predict the initial stiffness due to the limitation of the
In Table 5 shows the comparison of initial stiffness with connected column are not the standard size column. Further
hc is taken as hb and half of hb. The initial stiffness for the analysis are required to study the actual behaviour of the
welded axis, depth of cruciform column using half of the hb proposed connection.

Table 4 Comparison between analytical and experimental results for moment resistance

Mj,Rd.EC3 /
Mj,Rd.EC3
Specimen Mj,max (kNm) Mj,Rd.BS5950 (kNm) Ratioa Ratiob Mj,Rd.BS5950
(kNm)
(%)
N1 149.5 73.3 2.04 83.6 1.79 12.3

N2 193.7 105.0 1.85 108.1 1.79 2.92

N3 205.4 135.2 1.52 154.9 1.33 12.76

N4 263.9 216.9 1.22 221.6 1.19 2.12

N5 222.3 138.5 1.61 158.1 1.41 12.41

N6 279.5 282.3 0.99 285.5 0.98 1.11

N7 304.9 209.0 1.46 233.2 1.31 10.38

N8 492.7 358.3 1.38 386.4 1.28 7.29

a
Ratio of experimental moment resistance, Mj,max to analytical moment resistance based on BS5950 Mj,Rd.BS5950.
b
Ratio of experimental moment resistance, Mj,max to analytical moment resistance based on EC3 Mj,Rd.EC3.

Table 5 Comparison between analytical and experimental results for initial stiffness

Specimen Sj,ini Sj,ini.EC3,h Sj,ini.EC3, h/2 Ratioa Ratiob Ratioc Axis

N1 17.2 18.7 28.8 0.92 0.60 0.92 Major

N2 55.2 32.9 51.6 1.68 1.07 1.07 Welded

N3 25.8 26.3 41.6 0.98 0.62 0.98 Major

N4 58.0 56.9 90.9 1.02 0.64 0.64 Welded

N5 23.5 32.8 50.5 0.71 0.46 0.71 Major

N6 125.6 56.4 90.3 2.23 1.39 1.39 Welded

N7 25.4 48.7 77.0 0.52 0.33 0.52 Major

N8 74.0 92.8 149.6 0.80 0.49 0.49 Welded

a
Ratio of experimental inital stiffness, Sj,ini to analytical initial stiffness calculated from hc = hb.
b
Ratio of experimental inital stiffness, Sj,ini to analytical initial stiffness calculated from hc = hb / 2.
c
Predicted analytical initial stiffness.

3.3 Performance of FEP and EEP N6, N3 with N7 and N4 with N8 are identical except for EEP
specimens have an extra bolt row above the beam flange.
Comparison between the performance of FEP and EEP The ratios of moment resistance between EEP and FEP
connections using EC3 are shown in Table 6. The comparison are from 1.51 to 2.64 while the ratios of initial rotational
is made to validate the performance between FEP specimen stiffness are 1.65 to 1.85. The extra bolt rows in extended
N1 to N4 and EEP specimen N5 to N8. The beam size, end-plate greatly increase the performance of the connections
column size and bolt configuration for N1 with N5, N2 with in terms of strength and stiffness.
 11 Poi Ngian Shek et al. / Jurnal Teknologi (Sciences & Engineering) 74:4 (2015), 7–11

Table 6 Comparison between flush end-plate and extended end-plate connections

Moment
Connection Resistance Initial Stiffness
Specimen Ratioa Ratiob
Type EC3 (kNm/mRad)
(kNm)

N1 (FEP) FEP 83.6 18.7

1.89 1.75
N5 (EEP) FEP 158.1 32.8

N2 (FEP) FEP 108.1 51.6

2.64 1.75
N6 (EEP) FEP 285.5 90.3

N3 (FEP) EEP 154.9 26.3

1.51 1.85
N7 (EEP) EEP 233.2 48.7

N4 (FEP) EEP 221.6 90.9

1.74 1.65
N8 (EEP) EEP 386.4 149.6
a
Ratio of moment resistance of EEP to FEP.
b
Ratio of initial rotational stiffness for EEP to FEP.

4.0 CONCLUSIONS References

The main conclusions that can be drawn from this study are: [1] Tahir, M. M. and P. N. Shek. 2005. Performance of Cruciform
1. The component method adopted in this study shows Column Using Universal Beam Sections under Axial Compression
Load. Jurnal Teknologi. 43(B): 51–66.
reasonable agreement with the experimental results in [2] Tahir, M. M., P. N. Shek, A. Sulaiman and C. S. Tan. 2009.
terms of moment resistance for FEP and EEP. The ratio Experimental Investigation of Short Cruciform Columns Using
between analytical values and experimental values are in Universal Beam Sections. Construction and Building Materials. 23:
the range of 0.99 to 2.04 for BS5950 and 0.98 to 1.79 for 1354–1364.
EC3. [3] Shek, P. N., M. M. Tahir, C. S. Tan and A. B. H. Kueh. 2011.
Experimental Investigation of End-Plate Connection with Cruciform
2. Both BS5950 and EC3 design guidelines results are Column Section. Advanced Materials Research. 250–253: 3730–
closely match and can be used to predict the moment 3733.
resistance of the proposed cruciform column end-plate [4] British Standard Institution. 2005. Eurocode 3: Design of Steel
connections. The highest percentage difference between Structures, Part 1-8: Design of Joints. London: BSI.
EC3 and BS5950 is 12.76%. [5] Tahir, M. M., P. N. Shek amd C. S. Tan. 2009. Push-Off Tests on
Pin-Connected Shear Studs with Composite Steel-Concrete Beams.
3. The initial stiffness calculated using Eurocode 3: Part 1.8 Construction and Building Materials. 23(9): 3024–3033.
Design of Steel Structures shows the ratio between [6] Yee, Y. L. and R. E. Melchers. 1986. Moment Rotation Curves for
analytical values and experimental values are in the range Bolted Connections. Journal of Structural Engineering. 112(3):
of 0.49 to 1.39. 615–635.
4. EC 3: Part 1.8 can be used to predict initial stiffness of [7] Azizinamini and J. B. Radziminski. 1989. Static and Cyclic
Performance of Semi-Rigid Steel Beam-to-Column Connections.
cruciform column end-plate connections with the Journal of Structural Engineering. 115(12): 2979–2999.
assumption of the depth of cruciform column for welded [8] Bjorhovde, R., A. Colson and J. Brozzetti. 1990. Classification
axis is taken as half of the depth of universal beam. System for Beam-to-Column Connections. Journal of Structural
5. The performance of EEP connection is higher than the Engineering. 116(11): 3059–3076.
FEP connection in terms of strength and stiffness with the [9] Abolmaali, A., J. H. Matthys, M. Farooqi and Y. Choi. 2005.
Development of Moment-Rotation Model Equations for Flush End-
same beam depth and connection configurations. The Plate Connections. Journal of Structural Steel Research. 61: 1595–
ratios of moment resistance between EEP and FEP are 1612.
from 1.51 to 2.64 while the ratios of initial rotational [10] Shi, Y., G. Shi and Y. Q. Wang. 2007. Experimental and Theoretical
stiffness are 1.65 to 1.85. Analysis of the Moment Rotation Behaviour of Stiffened Extended
End-Plate Connections. Journal of Constructional Steel Research.
63: 1279–1293.
[11] Lee Y. H., C. S. Tan, Y. L. Lee, M. M. Tahir, S. Mohammad and P.
Acknowledgement N. Shek. 2013. Numerical Modelling of Stiffness and Strength
Behaviour of Top-Seat Flange-Cleat Connection for Cold-Formed
The authors would like to thank MOHE (FRGS vot 4F050) Double Channel Section. Applied Mechanics and Materials. 284–
for funding this research, Universiti Teknologi Malaysia 287: 1426–1430.
[12] Faella, C., V. Piluso and G. Rizzano. 2000. Structural Steel
(UTM) and UTM Construction Research Centre for providing Semirigid Connections: Theory, Design and Software. CRC Press
assistance and laboratory facilities. Special thanks to Lim LLC.
Chia Wei and Philip Heng Jin We who contributed in this [13] Steel Construction Institute and British Constructional Steelwork
project. Association. 1996. Joints in Steel Construction–Volume 1: Moment
Connections. Berkshire: SCI.
[14] Steel Construction Institute and British Constructional Steelwork
Association. 2013. Joints in Steel Construction–Moment-Resisting
Joints to Eurocode3. Berkshire: SCI.

View publication stats