Beam CFDST 1
Beam CFDST 1
Beam CFDST 1
a r t i c l e i n f o a b s t r a c t
Article history: In the present study, concrete-filled double skin steel tubes (CFDSTs) with a joint, which are lighter and more
Received 9 July 2018 economical than concrete-filled steel tubes (CFSTs), were developed. The CFDSTs are made of two steel tubes
Received in revised form 3 December 2018 of different diameters with concrete filling the space between them. The length, the outer and inner diameters
Accepted 15 December 2018
of a CFDST are 10 m, 914.4 mm, and 514.4 mm, respectively. Two types of shear connectors consisting of M16
Available online 11 January 2019
and steel plate studs were used. A new type of connection in CFDSTs was developed to transfer the imposed
Keywords:
load effectively. Four different specimens were made to test all possible combinations of the two different
Concrete-Filled double Skin Tube shear connectors (M16 studs, steel plate studs) and joints. The bending performance of the CFDSTs was investi-
Composite tube gated through 4-point loading experiments and finite element analysis (FEA). Finally, a comparison of design
Finite element analysis predictions and experiment results was made for these tubes.
Joint © 2018 Elsevier Ltd. All rights reserved.
Shear connector
1. Introduction eccentric loads. The proposed model was validated by the experimental
results of DCFST slender columns reported in the literature.
In recent years, concrete-filled double skin steel tubes, which not Slips occur between steel and concrete during bending behaviors of
only have the advantages of concrete-filled steel tubes (CFSTs) but composite members. Numerous studies on shear connectors for
also supplement the weaknesses of CFSTs, have been developed. A preventing slips and inducing composite actions have been carried
CFDST is a composite member containing two steel tubes with concrete out. For example, Yoo SW et al. [13] studied the composite behavior of
filling the space between them. CFDSTs are lighter but have better seis- T-type girders with studs. Zhenyu H et al. [14] and Yan JB et al. [15] eval-
mic performance than CFSTs. Research studies related to CFDSTs have uated the composite behavior of steel-concrete-steel sandwich compos-
been extensively undertaken. For instance, Han and Zhao [1] conducted ite structures with studs, while Zhenyu H et al. [16] developed a shear
experimental research on beams, columns, and beams-columns made connector in the form of a J-shaped hook. Although shear connectors
of CFDSTs with various cross-sections. A series of tests on CFDSTs stub for composite actions of general composite members have been re-
columns and beam-columns were implemented by Zhong Tao et al. ported in a lot of studies, research related to shear connectors in CFDSTs
[2]. Han [3–5] carried out experiment and analysis studies on CFDSTs are rare. Shimizu et al. [17] conducted experimental research on the
subjected to cyclic and long-term sustained loadings. Wang [6] and ultimate strength of CFDSTs with shear connectors using a deformed re-
Huang [7] performed an experimental investigation of CFDSTs under inforcing bar, while Zhao et al. [18] investigated the behavior of large-
collision and torsion loads, respectively. Regarding analytical studies, scale hybrid FRP – CFDSTs beams with headed shear studs. In these
Pagoulatou [8] studied the behavior of CFDST stub columns subjected studies, shear connectors were installed partially in the steel tubes but
to concentric axial compression loads, and then he proposed a new for- not the entire steel tubes. The authors only studied the effect of one
mula to evaluate the strength of CFDSTs in accordance with EC4 [9]. type of shear connector on the ultimate strength of CFDSTs. Therefore,
Hassanein [10] investigated the compressive behavior of CFDSTs with the impact of various types of shear connectors placed in entire tubes
various cross-sections. By performing a finite element analysis (FEA), on the ultimate strength of CFDSTs should be investigated.
Huang [11] examined the influences of important parameters that de- In addition, sizes of CFDSTs members vary upon the purposes of use.
termine sectional capacities of CFDST stub columns. Recently, Liang Bulky CFDSTs are necessarily connected at the site because of transpor-
[12] suggested a mathematical model for the simulation of the perfor- tation constraints. Unlike existing CFSTs, which can be easily connected
mance of high strength circular double-skin concrete-filled steel tubular by welding outer steel tubes, connecting CFDSTs members through
slender columns (DCFST) with high-strength concrete subjected to welding is a significant challenge because inner steel tubes cannot be
welded. A new type of connection for CFDSTs, therefore, should be de-
veloped. However, no research has been done on the connection of
⁎ Corresponding author. CFDSTs.
E-mail address: [email protected] (S.-E. Kim).
https://doi.org/10.1016/j.jcsr.2018.12.012
0143-974X/© 2018 Elsevier Ltd. All rights reserved.
S.-S. Eom et al. / Journal of Constructional Steel Research 155 (2019) 260–272 261
10m
Inner steel
t = 6mm
Inner steel
t = 6mm
Considered Part It is noted that a large scale of composite tubes are necessarily con-
nected using joints at the site because of transportation constraints.
Each part of the composite tube filled with the concrete between the
shear connector inner and outer steel tubes is made in the factory. At the site, these
As
parts are groove-welded at the mid-joint (Fig. 4).
45° The design procedure for the joint based on AASHTO LRFD [19] is as
follows:
Ac
• Choose the dimensions and material properties of the joint.
Considered Part • Calculate the plastic moment of composite tubes without the joint,
Mp(tube) corresponding to Eurocode 4 [9] (EC4).
• The joint was designed to have a full plastic moment higher than com-
posite tubes. The plastic moment of the joint, Mp(joint) was given as
below:
Fig. 3. Partial details of the compositube considered for shear connector design. Mpðjo intÞ ¼ ZF y ð5Þ
when
• Compute the nominal shear resistance of a single shear connector, Qn
sffiffiffiffiffiffi
qffiffiffiffiffiffiffiffiffi D E
0 ≤0:45 ð6Þ
Q n ¼ 0:5Asc f c Ec ≤Asc F u ð2Þ t Fy
P D3 −D3
n¼ ð3Þ where, Z is the plastic section modulus of the splice tube, Z ¼ o 6 t ;
ϕsc Q n
Do and Dt represent the outer and inner diameters of the splice tube, re-
spectively; D and t are the diameters and thicknesses of the outer and
• Calculate the spacing of studs, dstud inner splice tubes, correspondingly; E and Fy stand for elastic modulus
and yield strength of the splice tubes, respectively.
Ltube
dstud ¼ ð4Þ This design procedure was applied for both composite tubes with
2n
M16 and plate studs.
where, As and Fy are the cross-sectional area and yield strength of outer 2.4. Material properties of composite tubes
tubes, respectively; fc’ and Ac represent the compressive strength and
area of the concrete in the considered part, correspondingly; Asc and Table 1 illustrates the material properties of the steel tubes and the
Fu are the cross-sectional area and minimum tensile strength of a stud concrete used in the current study. The material properties of SM490
shear connector, respectively; ϕsc stands for the resistance factor for steel tubes were determined by tensile coupon tests. The tensile cou-
shear connectors; Ltube is the length of the composite tubes. pons were extracted in the longitudinal direction of the steel tubes.
The number and spacing of studs obtained from the considered part The coupons were tested in accordance with the Korean Standard KS
can be used for the others part of the composite tube. D 3515:2014 [20] using a displacement controlled testing machine.
Data on the concrete material properties were obtained by means of
2.3. Design of joint concrete cylinder tests based on Korean Standard KS F 2405 [21].
These specimens were tested at 28 days after casting.
Since the inner steel tubes of the composite tubes cannot be welded,
joints were designed to splice two tubes as shown in Fig. 4. The devel- 2.5. Design prediction of the ultimate load for composite tubes
oped joint was made of 32 mm thick steel and was 600 mm in length.
In addition, eight steel plates of 20 mm thickness were welded to the The design prediction of the ultimate load for composite tubes is cal-
joint to effectively transfer loads from the composite tube to the joint. culated upon the nominal moment capacity of these tubes. The nominal
600
PL 200*176.2*20T
PL 200*176.2*20T PL 200*200*20T Steel plate
914.4
100
PL 850.8*850.8*20T
200
20
Table 1
Material properties of steel tubes and concrete.
No Composite tube types Item Yield strength Fy Ultimate strength Fy Compressive Strength f'c Young Modulus E
(Mpa) (Mpa) (MPa) (MPa)
1. Composite tubes with M16 studs Outer steel tube (8 mm) 486.5 533.9 – 205,000
Inner steel Tube (6 mm) 467.6 517.8 – 205,000
Joint (32 mm) 377.0 536.0 205,000
Concrete – – 48.9 –
2. Composite tubes with steel Outer steel tube (8 mm) 341.0 507.0 – 205,000
plate studs
Inner steel Tube (6 mm) 396.3 547.3 – 205,000
Joint (32 mm) 358.3 526.5 – 205,000
Concrete – – 47.3 –
moment capacity can be evaluated in line with such codes as ACI [22], Assuming two loads were applied on the composite tubes as shown
AISC [23], and EC4 [9]. In EC4 [9] and AISC [23], the strain compatibility in Fig. 6, the ultimate loads corresponding to the nominal moment ca-
method (SCM) and the plastic stress distribution method (PSDM) were pacity were calculated as 2340 kN and 1762 kN for the tube with M16
recommended. The SCM of EC4 [9] and ACI [22] methods were almost and steel plate studs, respectively.
similar, but the assumptions about the compressive strength of concrete
were different. Besides, the PSDM of EC4 [9] was also similar to that of
AISC [23], but the concrete stresses were not the same (1.0f'c in EC4 [9] 3. Experiment
and 0.95f'c in AISC [23]). In the design method, the PSDM of EC4 [9]
was employed to calculate the nominal moment capacity of the com- 3.1. Specimen preparation
posite tube.
The plastic moment of the composite tube was calculated based on In this experiment, four different specimens were fabricated. With
the plastic stress distribution, as shown in Fig. 5. The plastic neutral each type of shear connector (M16 and plate studs), two specimens
axis of the composite tube with the M16 and steel plate studs was were made with and without the joint. The specimens contained inner
246.4 mm and 225.88 mm from the top of the cross-section, respec- and outer steel tubes as well as concrete filling the space between
tively. As a result, the corresponding nominal moment capacities of tubes. The shear connectors were welded on these steel tubes as illus-
the composite tube with M16 and steel plate studs were 4972.96 kN- trated in Fig. 7. Details of the joint are presented in Fig. 8.
m and 3743.26 kN-m. The test specimens were labeled according to the type of studs and
specimens, as displayed in Table 2. For instance, in the label CM-J, the
Fy fc'
177.06
26.31
Concrete
97
Fy Fy Fy
P.N.A
276.8
445.17
N.A
P/2 P/2
10 m
Mn
4-point bending tests were conducted to evaluate the bending perfor- 3.3.2. Composite tubes with steel plate studs
mance of the composite tubes. The test arrangement of these tubes is The comparison between applied load and mid-span displacement
demonstrated in Fig. 9. The specimens were loaded by hydraulic jacks responses of the composite tubes with and without a joint is shown in
with a loading capacity of 5000 kN. The loads were applied at the position Fig.13, and the ultimate loads of these tubes are illustrated in Table 4.
of 750 mm away from the center of the composite tube on both sides. Table 4 witnesses a 14.4% increase in the ultimate load of specimen
Loading and boundary pads (25 mm thick) were placed to prevent stress CP-J compared with that of the specimen CP-WJ. The reason is that
concentration at the loading points and the supporting ends. Three the plate stud in the specimen CP-WJ failed; therefore, the ultimate
100 mm LVDTs were placed along the bottom of the specimens in a load was suddenly reduced. While for the specimen CP-J, no stud was
pure bending segment to measure the mid-span vertical displacement placed at the joint length. Moreover, the joint was designed to have a
of the tube. Hinge and roller conditions were applied to the bottom of much higher strength compared with the strength of the specimen
the boundary pads. The bending experiments were carried out with dis- with plate studs. As a result, the strength of the specimen without
placement control at the velocity of 2 mm per minute in the elastic region joint (CP-WJ) are less than that of the specimen with joint (CP-J).
and 4 mm per minute in the plastic region. Four different experiments Fig. 14 indicates the experimental load-stress of composite
were conducted for the four specimens shown in Table 2. tubes with steel plate studs. At the ultimate load state, the stresses
measured at the bottom middle point of specimens CP-WJ and CP-
3.3. Experimental results J were 342.1 MPa and 96.5 MPa, respectively. The stress measured
at 200 mm away from the joint edge of specimen CP-J was
Fig. 10 reveals the typical failure mode for the specimens. All the 343.5 MPa. It is worth noting that the stress at the joint was
specimens behaved in a ductile manner. Neither welding failure nor much lower than that at the composite tubes since the joint
S.-S. Eom et al. / Journal of Constructional Steel Research 155 (2019) 260–272 265
1500
200 200
Loading Pad
50 P/2 600 P/2 50
914.4
10m
Fig. 12. Experimental load-stress curves of composite tubes with M16 studs.
Fig. 13. Experimental load-displacement curves of composite tubes with steel plate studs.
thickness was larger than that of the composite tubes. In addition, In this section, the commercial software ABAQUS [24] was used to
it is clear that the measured stress of specimen CP-WJ was higher analyze 3D finite element models of CFDSTs. The element types, the ma-
than the yield strength (341 MPa); therefore, it can be concluded terial models, the loading and boundary conditions were described.
that specimen CP-WJ has failed due to the yielding of the steel. Different element types, including solid, truss, and beam elements,
Opposingly, since the stress of specimen CP-J did not exceed the were employed for composite tube components. An 8-node solid ele-
yield strength (358.3 MPa), the joint was observed to have the ment (C3D8R) was used for the steel pads, the steel and concrete of
Table 4
Table 3 Ultimate load of composite tubes with steel plate studs.
Ultimate load of composite tubes with M16 studs.
No Specimen Ultimate load Pui/Pu2
No Specimen Ultimate load, Pu (kN) Pui/Pu1 Pu (kN)
Fig. 15. Components and FE model of the composite tubes with joint.
268 S.-S. Eom et al. / Journal of Constructional Steel Research 155 (2019) 260–272
Fig. 18. Load-displacement curves of composite tubes with M16 studs. Fig. 19. Load-stress curves of composite tubes with M16 studs (FEA).
S.-S. Eom et al. / Journal of Constructional Steel Research 155 (2019) 260–272 269
for the specimen with the joint, the stress focused around the locations 5. Comparison of experiment and analysis
adjacent to the joint and failure modes occurred at these locations but
not at the joint position. The conclusion to be drawn here is that the In this section, the results of the experiment and the FEA are com-
joint has sufficient strength to connect the two composite tubes. pared for the composite tubes with M16 and steel plate studs.
4.2.2. Composite tubes with steel plate studs 5.1. Composite tubes with M16 studs
Based on the FEA results, the load-displacement curves of specimens
CP-WJ and CP-J are indicated in Fig. 21. The ultimate loads of specimens Fig. 24 shows the load-displacement relationship of experiment, and
CP-WJ and CP-J were 1788 kN and 2129 kN, respectively. Therefore, FEA of two specimens, CM-J and CM-WJ, whereas Table 5 presents the
there was a 16% increase in the ultimate load of the specimen with comparison of ultimate loads obtained by the experiment, and analysis.
the joint compared with that of the specimen without the joint. It can be seen in Fig. 24 that the load-displacement curves obtained from
Fig. 22 shows the load-stresses, and Fig. 23 reveals the stress distri- the analyses were close to those obtained from the experiments for
butions of these two composite tubes. Based on these figures, the these specimens. Moreover, as shown in Table 5, good agreement be-
same conclusion on the composite tubes with M16 studs could be tween the analysis and the experimental results was found with the
drawn for those with steel plate studs. maximum difference of ultimate loads of only about 6%. For this reason,
Fig. 21. Load-displacement curves of composite tubes with steel plate studs. Fig. 22. Load-stress curves of composite tubes with steel plate studs (FEA).
270 S.-S. Eom et al. / Journal of Constructional Steel Research 155 (2019) 260–272
Fig. 23. Stress distributions of composite tubes with steel plate studs.
Fig. 24. Comparison of load-displacement curves of composite tubes with M16 studs obtained from FEA and test results.
it can be concluded that the FE model provides accurate predictions for and analysis is shown in Table 6. It is clear from Fig. 25 that the
the response of composite tubes without the joint. load-displacement curves obtained from the FEA well agreed with the
experimental results. In addition, as shown in Table 6, the difference
in ultimate strength between the FEA and experimental results was
5.2. Composite tubes with steel plate studs
minor. In other words, the FE model developed above is reliable, provid-
ing accurate predictions for the behavior of composite tubes having
For the composite tubes with steel plate studs, the load-
steel plate studs.
displacement of the experiment and the FEA is illustrated in Fig. 25,
while the comparison of ultimate loads obtained by the experiment
6. Comparison of test strengths with design strengths
Fig. 25. Comparison of load-displacement curves of composite tubes with steel plate studs obtained from FEA and test results.
Table 6 • While the M16 studs did not fail, the steel plate studs failed under ul-
Comparison of the experiment and analysis for composite tubes with steel plate studs. timate load. Therefore, the M16 studs are more suitable than the steel
No Specimen Experiment, PuEXP (kN) FEA, PuFEA (kN) P uFEA plate studs.
P uEXP
• For the composite tubes with the joint, FEA and experiment results in-
1. CP-WJ 1863 1788 0.96
dicated that the joint has sufficient strength to connect these compos-
2. CP-J 2164 2129 0.984
ite tubes.
• Based on this study, it is concluded that the developed CFDST mem-
bers have sufficient flexural strength, and they can be used for
Table 7 construction.
Comparison of ultimate loads obtained from design predictions with test and FEA results.
CM-J 2359 2328 0.99 2340 0.99 2156 0.91 2376 1.01 This work was supported by the National Research Foundation of
CP-J 2164 1753 0.81 1762 0.81 1742 0.80 1886 0.87 Korea (NRF) grant funded by the Korea government(MSIT No.
CM-WJ 2444 2328 0.95 2340 0.96 2156 0.88 2376 0.97 2018R1A2A2A05018524).
CP-WJ 1863 1753 0.94 1762 0.95 1742 0.93 1886 1.01
[14] Z. Huang, J.R. Liew, Nonlinear finite element modelling and parametric study of [21] Standard test method for compressive strength of concrete. KS F 2405:2010. , Ko-
curved steel–concrete–steel double skin composite panels infilled with ultra- rean Standards Association, 2010.
lightweight cement composite, Constr. Build. Mater. 95 (2015) 922–938. [22] American Concrete Institute, "ACI 318," Building Code Requirements for Structural
[15] J.-B. Yan, J.R. Liew, X. Qian, et al., Ultimate strength behavior of curved steel– Concrete and Commentary, Farmington Hills, 2014.
concrete–steel sandwich composite beams, J. Constr. Steel Res. 115 (2015) [23] Specification for Structural Steel Buildings. , American Institute of Steel Construction,
316–328. 2010 ANSI/AISC 360-10.
[16] Z. Huang, J.R. Liew, Structural behaviour of steel–concrete–steel sandwich compos- [24] ABAQUS Analysis User's Manual version 6.14, Dassault Syst. (2014).
ite wall subjected to compression and end moment, Thin-Walled Struct. 98 (2016) [25] W. T. Tsai, “Uniaxial compressional stress-strain relation of concrete,” J. Struct. Eng.
592–606. ASCE, Vol. 114, 9, pp. e, 1988.
[17] M. Shimizu, F. Tatsumi, T. Ishikawa, et al., Experimental Study on Ultimate Strength [26] J. Lubliner, J. Oliver, S. Oller, et al., A plastic-damage model for concrete, Int. J. Solids
of Concrete Filled double Tubular Steel with Shear Connector, Int. J. Steel Struct. 13 Struct. 25 (3) (1988) 28.
(1) (2013) 6. [27] J. Lee, G.L. Fenves, Plastic-damage model for cyclic loading of concrete structures, J.
[18] J.L. Zhao, J.G. Teng, T. Yu, et al., Behavior of Large-Scale Hybrid FRP–Concrete–Steel Eng. Mech. ASCE 124 (8) (1998) 9.
Double-Skin Tubular Beams with Shear Connectors, J. Compos. Construct. ASCE [28] K. Uenaka, H. Kitoh, K. Sonoda, Concrete filled double skin tubular members sub-
(2016) 11. jected to bending, Steel Compos. Struct. 8 (4) (2008) 297–312.
[19] AASHTO LRFD, Design Specifications, 2012 1661. [29] L.H. Han, H. Huang, X.L. Zhao, Analytical behaviour of concrete-filled double skin
[20] Rolled steels for welded structures. KS D 3515:2014, Korean Standards steel tubular (CFDST) beam-columns under cyclic loading, Thin-Walled Struct. 47
Association2014. (2009) 13.