Transactions, SMiRT-22

San Francisco, California, USA - August 18-23, 2013

Division X (include assigned division number from I to X)

STRUCTURAL PERFORMANCE OF STEEL-CONCRETE-STEEL

SANDWICH COMPOSITE BEAMS WITH CHANNEL STEEL
CONNECTORS
Meng Chu1, Xiaobing Song2, Honghui Ge3
1
Senior Engineer, Shanghai nuclear engineering research and design institute, Shanghai, China
2
Associate Professor, School of Naval Architecture, Ocean & Civil Engineering, Shanghai Jiao Tong
University, Shanghai, China ([email protected])
3
Chief Engineer, Shanghai nuclear engineering research and design institute, Shanghai, China

ABSTRACT

This paper investigates the structural performance of steel-concrete-steel (SCS) sandwich

members comprising a concrete core sandwiched in between two steel plates which are interconnected by
channel steel connectors with large interval. Angle steel is adopted to serve as the connection between
channel steel and surface plates. The behavior of a series of 8 large-scale, simply supported tests was
investigated, covering a wide range of applied shear span to depth ratios from 1.5 to 5.5. Other parameters
include the interval of channel steel and axial tension force. Although critical inclined cracks appeared in
most of the beams (expect for the beam with shear span to depth ratio 5.5), the observed modes of failure
were all ductile failure characterized by yielding of the tension steel plate. Unlike the negligible dowel
action of longitudinal reinforcement in reinforced concrete members, the dowel action of steel plate in
SCS member seems to play a decisive role. Other shear performance differences between SCS members
and reinforced concrete (RC) members were also analyzed. Test results are compared with current design
recommendations.

INTRODUCTION

In recent 30 years, a new structural system, steel-concrete-steel (SCS) sandwich composite, has
been on the rise with good potential. This structure features good integrity, ductility, and the ability to
prevent leakage, impact and explosion. It has been widely used in infrastructure construction, such as that
for nuclear power plants, offshore structures, and high-rise buildings, etc.
There are a number of types of shear connectors have being used in practice such as welded shear
studs[1], friction-welded bar connectors in Bi-Steel SCS sections[2,3] and J-hook connectors[4,5], etc.
Welded studs terminate within concrete and their prime functions are to resist longitudinal shear and
plate/concrete separation. Bar connectors or J-hook connectors provide direct connection to the two face
plates and their main functions are to resist both longitudinal and transverse shear, and to prevent plate
buckling. The Bi-steel SCS system can only be manufactured in a factory environment and the core
thickness is limited to 200~700mm. When the sandwich depth is shallow, welding of straight bar
connectors at both ends to the steel face plates is not possible, and double J-hook connectors can be
adopted with minimum core thickness of 50 mm[4].
Many experimental research works have been conducted to investigate the load carrying capacity
of SCS composite beams subject to bending and shear [6,7].
T.M. Roberts and D. N. Edwards conducted test on eleven SCS sandwich beams (b=400 mm;
h=150mm). The observed modes of failure were yielding of the tension plate and slip yielding of the
tension plate connectors. Significant shear cracking occurred in several of the beams but was not the
primary cause of failure [6]. Eighteen beams having a range of span, depth, plate thickness and bar
spacing have been tested under static loading by M. Xie and N. Foundoukos [7], four elementary modes
 22nd Conference on Structural Mechanics in Reactor Technology
San Francisco, California, USA - August 18-23, 2013
Division X (include assigned division number from I to X)

of failure were observed: tension plate yielding, bar shear, bar tension and concrete shear. Results were
compared with Bi-Steel design and construction guide [8]. It was found that, for most cases,
experimental transverse shear resistance exceeds the failure load given by the guide.
When the thickness of SCS sandwich member is large, such as that used in nuclear power plant,
transverse shear becomes more significant. Information available on shear resistance of thick SCS
sandwich elements is so limited that it is imperative to generate more results.
This paper is concerned with the behavior of SCS sandwich beams with another kind of shear
connectors—channel steel connectors. This kind of connectors can extend the range of core thickness to a
value larger than 700mm. 8 large-scale simply supported beams under static loading were tested, covering
a wide range of applied shear span to depth ratios from 1.5 to 5.5. Other parameters include the interval of
channel steel and axial tension force.

DETAILS OF TEST SPECIMENS

Fabrication details of the test beams are presented in Table 1 and Figure 1. All the beams had
concrete depth h=734 mm, width b=762 mm, nominal tension and compression plate thickness ts=14 mm,
C14b channel steels with interval Sc=1220 or 610mm. The stud connectors were 19 mm diameter by 150
mm long. All studs in a beam were equally spaced longitudinally and transversely with spacing s=152
mm. Stud connectors were welded to the plates using a stud welding gun (Figure 2). Stud could be tapped
to form a 90 degree bending to confirm the integrity of the weld (Figure 3).

Figure.1 Schematic arrangement of the SCS sandwich structure with channel steel connectors

Table 1: Details of test specimens and modes of failure.

Axial Sc Load Model of
Beam L [mm] λ
tension [mm] condition failure
WS5.5 10480 5.5 - 1220 A Bend
.WS3.0 7815 3.0 - 1220 A Bend+Shear
WS2.0 4148 2.0 - 1220 B Shear
WS1.5 3386 1.5 - 1220 B Shear
WS'1.5 3386 1.5 - 610 B Shear
WS'2.0 4148 2.0 - 610 B Shear
WS3.0T 9586 3.0 0.5AsFy 1220 C Shear
WS1.5T 7400 1.5 0.5AsFy 1220 C Shear
Load condition: A—Two point loads; B—One point load; C—Two point loads with tension
 22nd Conference on Structural Mechanics in Reactor Technology
San Francisco, California, USA - August 18-23, 2013
Division X (include assigned division number from I to X)

Figure.2 Welding of studs Figure.3 Tapped stud

The beams were filled with ready-mix concrete of Grade C35, which would nominally give a
cube strength fcu=35N/mm2. Before concreting, sections of steel plate were cut from the ends of the plates,
from which tensile test specimens were made and tested. Tensile tests were also conducted on C14b
channel steels. Material test results for concrete, plates and channel steels are given in Table 2. After
concreting the beams were cured under wet hessian, for approximately 28 days, prior to testing.

Table2: Material Properties of Test Specimens- Mean Values [N/mm2].

Steel plate C14b L100*80*10 Stud

Beam fcu
fy fu fy fu fy fu fy fu
WS5.5 43.4 330 485 293 401 255 430 395 418
WS3.0 34.5 295 450 270 390 255 430 395 418
WS2.0 33.7 290 440 270 390 255 430 395 418
WS1.5 38.8 305 445 270 390 255 430 395 418
WS'1.5 37.7 300 400 270 390 255 430 395 418
WS'2.0 45.2 280 430 270 390 255 430 395 418
WS3.0T 36.4 315 440 270 390 255 430 395 418
WS1.5T 27.5 285 425 270 390 255 430 395 418

SUPPORT AND LOADING CONDITIONS

The test beams were simply supported and subjected to either one or two point loads, as shown in
Figure.4. Loads were applied to the beams under force control before the load capacity was reached, and
then, transferred to deflection control to obtain the descending segment of load vs. deflection relationship.
Specimen WS1.5T and WS3.0T were subjected to proportionally increasing axial tension and
out-of-plane shear loading. The ratio of the axial tension (T) to out-of-plane force (P) was equal to 3:1.
The maximum axial loading was limited to 50 percent of the yield strength (AsFy, steel area multiplied by
nominal yield stress) of the SCS section. The axial loading was maintained constant after reaching this
value, and the out-of-plane force was increased either to failure or to the loading capacity of the test setup.
 22nd Conference on Structural Mechanics in Reactor Technology
San Francisco, California, USA - August 18-23, 2013
Division X (include assigned division number from I to X)

(a) Two point loads (b) One point load

(c) Two point loads with tension

Figure.4 Schematic test setup

INSTRUMENTATION

Strains in the tension and compression plates were measured at several locations using resistance
strain gauges. Concrete strains at the centers of the beams were measured using a 100 mm gauge. Slip
between steel and concrete was indicated by grid lines along the length of the bottom and top plate, and
measured using a displacement transducer at four points on the bottom plate. Vertical deflections at the
mid span and the supports were measured by displacement transducers. Magnitudes of applied load, strain,
deflection and slip were recorded by data acquisition equipment. Digital camera was fixed at a position to
record successive visible changes, principally concrete cracking process at each load steps.

TEST RESULTS AND COMPARISONS

Failure loads and modes

The modes of failure for the tested 8 beams are presented in Table 1. All the beams failed (rapidly
increasing central deflection or reduction in load carrying capacity) due to either tension plate failed at
mid span (bending failure) or tension plate failed at shear span followed with critical inclined cracks
opening (shear failure). The load/deflection relationships of the specimens with different shear span to
depth ratio are illustrated in Figure 5. Load capacity increased with the decreasing of shear span/depth
ratio.
 22nd Conference on Structural Mechanics in Reactor Technology
San Francisco, California, USA - August 18-23, 2013
Division X (include assigned division number from I to X)

Figure.5 load deflection relationship of specimens with different shear span to depth ratio

Tension plate failure at mid span

The failure mode of specimen WS5.5 is characterized by the tension plate failure at mid span.
The failure behavior is almost the same as the bending failure of doubly reinforced concrete beams with
symmetrically placed reinforcements. When the load increased to 50kN, the initial flexural crack
appeared at the bottom of the beam in pure bending zone, and penetrated rapidly towards the top of the
section, showed the first turning point on the load/deflection curve. The second turning point appeared at
about 700kN due to the yielding of tension steel plate at mid span, and after which, the load ceased to
increase with the increasing of the deflection. The failure of the beam is characterized by the large
deflection and pronounced curvature at the failure section in mid span (figure 8).

Figure.6 Crack pattern of specimen WS5.5

Figure.7 Load condition of specimen WS5.5 and strain gages arrangement on bottom plate
 22nd Conference on Structural Mechanics in Reactor Technology
San Francisco, California, USA - August 18-23, 2013
Division X (include assigned division number from I to X)

80kN 345kN 500kN 700kN 800kN

12000

10000
Strain (με)

8000

6000
`
4000

2000

0
G-3/4 F-3/4 E-3/4 D-3/4 C-3/4 B-3/4 A-3/4
Strain gages
Figure.8 Bottom steel plate strain distribution at different load level of WS5.5

Tension plate failure at shear span

Expect for the specimen WS5.5, all the other 7 specimens are characterized by the tension plate
failure at shear span. Figure 9 shows the crack pattern of specimen WS1.5. Figure 11 shows the strain
distribution of bottom steel plate at different load level. At failure stage, maximum strain appears at A3/4
in shear span. There exist distinct differences for the diagonal failure performance between RC beams and
SCS Beams.
For SCS beams, after the appearing of critical inclined crack, there is a sudden increasing of the
tensile stress in the bottom steel plate within shear span, due to which the diagonal crack widens and
extends directly to the load point. The tension force in bottom steel plate at mid span is balanced by the
compression force in symmetrically placed top steel plate. Nearly no concrete contributes to the
compression in top zone. This forms the essential difference with RC beams, in which a certain quantity
of compression concrete is needed.

Figure.9 Crack pattern of specimen WS1.5

Figure.10 Load condition of specimen WS1.5 and strain gages arrangement on bottom plate
 22nd Conference on Structural Mechanics in Reactor Technology
San Francisco, California, USA - August 18-23, 2013
Division X (include assigned division number from I to X)

400kN 1200kN 2400kN 3000kN 3440kN 2700kN(descending segment)

3500

3000

2500
Strain(με)

2000

1500

1000

500

0
C-3/4 B-3/4 A-3/4
Strain gages
Figure.11 Bottom steel plate strain distribution at different load level of WS1.5

For RC beams, the direct cause of diagonal section failure is the compression failure or tension
failure of the top compression zone concrete. For SCS composite beams, top steel plate is strong enough
to resist the compressive force, top concrete compression zone is usually small, and no obvious failure is
observed in this zone. The diagonal section failure of a SCS composite beam is caused mainly by yielding
of tension steel plate in shear span followed with crushing of concrete in the web zone (Figure 12)
Unlike the unreliable dowel action of longitudinal reinforcement in RC beams, which may
develop tearing tensile stress in the concrete between the rods, the studs between steel plate and concrete
endow the connection with reliable integrity. The dowel action of steel plate in SCS sandwich beams can
take an amazing portion of the total shear.

Figure.12 Typical crack pattern of SCS beams subject to combination of shear and moment

Influence of amount of shear connectors

Compared with the specimen WS1.5, which has no channel steel within shear span, there are two
channel steels arranged in the shear span of specimen WS1.5’. Test results show the shear capacity
difference caused by these two channel steels is 580kN, the contribution for one channel steel is about
290kN.
 22nd Conference on Structural Mechanics in Reactor Technology
San Francisco, California, USA - August 18-23, 2013
Division X (include assigned division number from I to X)

Compared with the specimen WS2.0, which has one channel steel within shear span, there are
two channel steels arranged in the shear span of specimen WS2.0’. Test results show the contribution of
the additional one channel steel is about 250kN.
The cross area of C14b channel is 2131mm2, and the yield stress is 270MPa. The tension capacity
of one C14b channel is 575.4KN. Under the experimental condition, the contribution of transverse
channel steel to shear capacity is about 50% of the tension strength.

Influence of tension force to shear capacity

In general, tension force will reduce the shear capacity of RC beams. However, test results show
that significant axial tension does not seem to have an influence on the shear strength of SCS beams
(Figure 13). The fundamental cause of this phenomenal may lies in the fact that, after the appearing of
critical inclined cracks, axial tension force is resisted mainly by steel plates, As the inclined cracks
develop throughout the concrete section, axial tension force has little effect on the stress state of concrete
beside inclined cracks, and the ultimate shear capacity is equal to the vertical component of the
compressive force in concrete strut.
For specimens WS3.0 and WS3.0T, the initial stiffness is almost the same. At load level of
800kN, yielding of bottom steel plate causes stiffness reduction. As the load conditions for WS1.5 and
WS1.5T are different, one point loading for WS1.5 and two points loading for WS1.5T, it is hardly to
compare the stiffness difference between them.

2000
WS3.0
WS3.0T
WS1.5
1500 WS1.5T
Shear force (kN)

1000

500

0
0 10 20 30 40 50 60 70 80 90
Deflection (mm)
Figure.13 Load deflection relationship of specimens WS 1.5, WS1.5T, WS3.0, WS3.0T

The test load and code recommendation

In order to demonstrate whether the shear strength of SCS beams can be estimated using ACI
code provisions for shear strength of RC beams, the test results of ultimate shear load are compared with
the recommendation values of ACI349-01 [9] in Figure 14. The code provisions for shear strength in
ACI349-01 are summarized using the equations below. In these equations, Vc is the shear strength
contribution of the concrete and Vs is the shear strength contribution from transverse shear connectors.
V = V +V ACI349, Equation 11-2
V =2 f b d ACI349, Equation 11-3
V =A f ds ACI349, Equation 11-15
V = 2(1 + ) f b d ACI349, Equation 11-8
 22nd Conference on Structural Mechanics in Reactor Technology
San Francisco, California, USA - August 18-23, 2013
Division X (include assigned division number from I to X)

Where fc’ is the compressive strength of concrete, b w is concrete section width, d is concrete
section height, Nu is axial force (Negative for tension), Ag is the gross cross-sectional area, A v is the cross
of transverse shear steel, fyv is yield strength of transverse shear steel.
As shown in Figure 14, shear strength is estimated conservatively by the ACI code provisions, the
underestimated percentage ranged from 25% to 77%, and the least accurate results appeared for
specimens WS1.5, WS1.5T and WS3.0T.

2500
WS3.0

2000 WS2.0
Experimental Value (kN)

WS'2.0
1500 WS1.5

WS'1.5
1000
WS1.5T

WS3.0T
500

0
0 500 1000 1500 2000 2500
ACI349 Recommended Results (kN)

Figure.14 Tested shear resistances and ACI349 recommended results

CONCLUSIONS

Test results show that the bending failure mechanism of SCS composite member is almost the same
as that of RC beams with symmetrically reinforced section, for transverse shear, the way in which SCS
beams fail are rather different in their nature from the ways in which RC beams do. The main differences
are:
1. While the direct cause of diagonal section failure of a RC beams is the compression failure
or tension failure of the compression zone concrete, the diagonal section failure of a SCS
composite beam is caused mainly by yielding of tension steel plate in shear span followed
with crushing of concrete in the web zone.
2. Dowel action of steel plate contributes significantly to shear resistance and should not be
neglected.
3. The contribution of transverse channel connecters crossing the diagonal section is about 50%
of the tension yielding strength of channel connecters.
4. Axial tension promotes the yielding of steel plates, thereby reduces the bending stiffness, but
significant axial tension does not seem to have an influence on the shear strength of SCS
beams.
Although the calculation results are conservative, ACI349-01 code provisions for shear strength
of RC beams cannot provide satisfactory assessment for SCS beams. In consideration of the behavior
differences between SCS beams and RC beams under combined shear and bending, it is necessary to
grope for a new shear capacity model which can correspond with the unique shear mechanism of SCS
members. This work is in progress in Shanghai Nuclear Engineering Research and Design Institute and
Shanghai Jiaotong University, and the study is to be published later on.
 22 nd Conference on Structural Mechanics in Reactor Technology
San Francisco, California, USA - August 18-23, 2013
Division X (include assigned division number from I to X)

References
B. A. Burgan and F. J. Naji. Steel-concrete-steel Sandwich Construction. Journal of Constructional Steel
Research. 1998, Vol. 46. Nos. l-3: 219-344.
Bowerman HG, Gough MS, king CM. Bi-steel design and construction guide. British Steel Ltd, 1999.
Min Xie, J.C. Chapman. Developments in sandwich construction. 2006, Journal of Constructional Steel
Research, 62: 1123–1133.
Liew, J.Y.R. and K.M.A. Sohel. Lightweight steel–concrete–steel sandwich system with J-hook
connectors. Engineering structures, 2009. 31(5): p. 1166-1178.
Liew, J.Y.R., Sohel, K.M.A.. Structural performance of steel-concrete-steel sandwich composite
structures. Advances in Structural Engineering, 2010, 13(3): 453-470.
T. M. Roberts, D. N. Edwards & R. Narayanan. Testing and Analysis of Steel-Concrete-Steel Sandwich
Beams. J. Construct. Steel Res. Vol. 38, No. 3, pp. 257-279, 1996.
M. Xie, N. Foundoukos, J.C. Chapman. Static tests on steel–concrete–steel sandwich beams. Journal of
Constructional Steel Research, 63 (2007) 735–750.
European Committee for Standardization. Design of Concrete Structures. Eurocode 2, Brussels; 1992.
ACI349, Code Requirements for Nuclear Safety-Related Concrete Structures and Commentary, American
Concrete Institute, January 2001.