THE STRUCTURAL ASSESSMENT OF RIGID

CONNECTIONS BETWEEN REINFORCED CONCRETE

AND STEEL MEMBERS WITHIN FRAMED BUILDINGS

Karl Micallef

Faculty of Architecture and Civil Engineering, University of Malta, Malta

The use of composite construction utilising structural steel and reinforced concrete

beams and columns is a popular method implemented by engineers in order to produce

more efficient structures. Most of the literature available on the subject has relied

mainly on a form of construction in which steel beams run continuous through a

reinforced (or composite) column, forming a rigid connection between both members.

The scope of this study was to investigate alternative methods to the one cited above,

namely by the use of chemical and mechanical anchors. The latter were examined in

detail through the results of an experimental program carried out at the University of

Malta, where three hybrid specimens connected by means of cast-in-place anchors were

tested under monotonically increasing loading.

Stress-strain analysis was used to predict the specimens’ behaviour and predicted values

were compared with those measured. Lack of effective concrete containment was seen

to significantly imply a decrease in strength and stiffness, whereas it was noted that there

is little variation of strength and stiffness with concrete grade.

1. Introduction

The advantages of using mixed reinforced concrete and steel systems have long been

recognised. Much research and publications have been devised for composite structures

with reinforced concrete (or composite) columns and steel beams running continuous

though the joint to form a rigid “moment” connection.

Abstract

The aim of this study was to investigate the possibility of creating such a connection

using anchors. Therefore, an analytical and experimental program was carried out at the

University of Malta to investigate the ultimate capacity of such a connection method

between steel and reinforced concrete and the overall effect of the introduction of a joint

between the two materials.

2. Experimental test setup

2.1 Aim of research

The aim of the tests carried out at the University of Malta was to investigate a particular

method of creating a rigid connection between a reinforced concrete and a steel member,

namely, by means of threaded bars embedded in the concrete.

The study of the moment-carrying capacity of the connection was carried out by using a

hybrid simply supported beam consisting of two equal lengths of steel and reinforced

concrete. The test specimens were tested using a two-point loading arrangement such

that the central portion of the beam (which contained the actual connection) was

subjected to a constant moment (Figure 1).

Figure 1: Arrangement of applied loading

2.2 Variables and specimen sizing

It was originally intended to also vary the method of connection but, following

preliminary calculations, it was found that the moment-carrying capacity of a connection

approaching the plastic moment capacity of a steel beam could only be achieved by

means of anchored threaded bars. Thus, the variables under investigation were the grade

of concrete used (namely, fcu=25N/mm2, 30N/mm

2 and 35N/mm

2, where

fcu=characteristic concrete cube compressive strength) and the applied loading (a

monotonically increasing load).

The size of the test specimens was highly governed by the physical capacity of the

loading rig and jack at the University of Malta. Thus, the overall length of the beam, L,

was taken as L=4m. In order to prevent excessive deflections, a span/depth ratio of 20

was chosen and thus the depth of the steel beam was chosen to be at least 200mm. An

S275 IPE300 rolled I section was used.

The reinforced concrete member, c, was designed such that its stiffness was comparable

to that of the steel beam, s, thus achieving a uniform stiffness, EI, along the length of the

beam, where E=elasticity modulus and I=2nd

moment of area. Thus:

(EI)c=(EI)s (1)

For IPE300 section, Is=8356cm4

From B.S.5950-1:2000, clause 3.1.3, Es=205kN/mm2

Thus, (EI)s=17129.8kNm2

From B.S.8110-1:1997, figure 2.1, Ec=5.5(fcu/γm), where γm=material partial safety factor

Taking fcu as an average value of 30N/mm2 and γm=1.5, Ec=24.6kN/mm2

Substituting (EI)s and Ec in (1), Ic=696333333mm4

For a rectangular beam section of breadth, b, and depth, d, I=bd3/12

Taking I=Ic=696333333mm4 and b=200mm, then d=350mm

2.3 Specimen and connection design

The moment capacity of the steel beam was determined from B.S.5950-1:2000, clause

4.2.5.2, as Mp=pysxx, where Mp=plastic moment capacity, py=yield stress and sxx=section

plastic modulus. For the IPE300 section, Mp=172.7kNm.

Using the dimensions derived above, the reinforcement for the reinforced concrete beam

was designed to B.S.8110-1:1997 such that its moment-carrying capacity equalled the

plastic moment capacity of the steel beam. This required 2 No. T25 bars compression

steel and 4 No. T25 bars as tension steel.

With reference to Figure 1, the maximum load anticipated to be applied by the jack was

2Pu=3Mp/L=129.5kN. Thus, the reinforced concrete beam was similarly designed to

resist a maximum shear force of Pu. This required R6 closed links at a spacing of 50mm

centres, which was increased to 100mm from the point of contraflexure till mid-span to

avoid congestion of reinforcement.

The actual connection was designed as an extended end plate, with the geometric

configuration arranged such that the location threaded bars was limited to the zone of

confined concrete, i.e., within the links, and thus the bars were restricted to the zone

between flanges.

Using bolt spacing and end distance limitations as per B.S.5950-1:2000, the arrangement

was devised for 8 No. grade 8.8 M24 threaded bars (Figure 2a).

(a) Lever arm (b) Forces in bolts

Figure 2: Connection arrangement

The values for the lever arm of each bolt were thus determined: y1=245mm, y2=185mm,

y3=125mm, y4=65mm. The moment capacity of the bolt group was found by assuming a

linear force distribution with rotation occurring at the top of the plate and conservatively

ignoring the portion in compression under the top part of the plate (Figure 2b).

The nominal bolt tensile capacity for a grade 8.8 M24 bolt as per B.S.5950-1:2000,

clause 6.3.4.2, was found to be 158.14kN. Taking the maximum force F4 as this value,

the maximum moment capacity, M, was found:

M=2F Σy2/y4 (2)

Substituting F=158.14kN, y2=114100mm2 and y4=245mm in (2), M=147.29kNm. This

represented 85.3% of the hybrid beam’s moment capacity. Thus, it was expected that

the connection would just fail prior to the yielding of the steel beam and crushing of the

reinforced concrete beam.

The embedment length for the threaded bars was determined on the basis of the confined

concrete block assuming a 45º dispersion of stress from the concrete-plate interface.

This resulted in a block with a perimeter of 784mm. The bars were modelled as a

standard tensile pull-out arrangement, using the basic relation:

σ=F/A (3)

where σ=tensile strength of concrete, i.e., 0.12(fcu)0.7 from B.S.8007:1987

F=maximum bolt force defined earlier as 158.14kN

A=confined zone contact area, i.e., perimeter × embedment length

Taking fcu=25N/mm2, the required embedment length was found to be 177mm. Thus,

the total length of threaded bars used was taken as 333mm to cater for the ignored

(unconfined) concrete zone as well as the plate, washer and nut thickness.

The plate and welds were designed to B.S.5950-1:2000 as a normal column baseplate

subjected to a moment, resulting in a 15mm thick S275 plate welded to the IPE300

section with E35 6mm fillet welds.

2.4 Specimen fabrication

The steel components of the specimens were supplied and prepared by local steel

manufacturers. The reinforced concrete elements were cast in a local precast concrete

plant. 150×150×150mm cube samples taken from each batch to perform standard cube

compression tests to B.S.1881-116:1983 in order to verify the concrete strength after 7

and 28 days. All results were satisfactory.

2.5 Data measurement

Strain gauges were used to measure strains in the region of maximum moment in all

relevant materials in order to study stress-strain relationships.

Gauges were fixed to the tension and compression reinforcement in the reinforced

concrete beam, the compression zone of the concrete itself and both tension and

compression zones of the steel beam. Electrical foil-type strain gauges were used with a

gauge length of 60mm in the case of the concrete and 5mm for the other materials.

In addition, two linear variable displacement transducers (LVDTs) were mounted on

either side of each specimen near the connection prior to testing to measure the

maximum vertical displacement at mid-span.

2.6 Testing procedure

After casting, a minimum of 28 days was allowed to elapse prior to any form of testing

on the specimens in order for the concrete to develop its intentioned strength. The

specimens were lowered with the steel and reinforced concrete portions separately into

the loading rig by means of an overhead crane. The portions were then fixed in position

using washers and nuts (Plate 1), which were tightened firmly using a hand ratchet in a

staggered fashion.

Plate 1: Connection detail Plate 2: Assembled test setup

The applied loading was transferred in the form of two point loads by means of a

spreader beam (Plate 2). The steel spreader beam was placed on the specimen

symmetrically, supported on two steel round bars with neoprene pads between the

supports and the specimen. The load was applied by means of a 500kN hydraulic jack

fitted with a 500kN load cell. The latter, the LVDTs and the strain gauge wires were

connected to a data logging device via connector boards.

Load was applied gradually in small increments of approximately 5kN each in order to

allow proper recording of crack formation and also to accomplish as many load-

deflection readings as possible.

3. Experimental test results

All specimens were loaded to failure and a summary of the results is given in Table 1.

Table 1: Summary of load and deflection results

The deflection reading was found by taking the mean of the two measurements recorded

from the two LVDTs. The load-deflection chart for all specimens is shown in Chart 1.

0

10

20

30

40

50

60

70

80

0 2 4 6 8 10 12

Deflection (mm)

Loa

d (

kN

)

C25

C30

C35

Chart 1: Superimposed load-deflection graph for all specimens

4. Discussion on experimental test results

All specimens failed at an ultimate load which was considerably lower than the one

expected. No damage was visible in the steel beam and connection itself. Failure

occurred in the reinforced concrete elements, with cover spalling and crushing in the

compression region and severe cracking in the tension zone.

Cracking in the compression zone was more pronounced as the concrete grade decreased

while that in the tension zone was equally rigorous in all specimens.

In all cases, the cracks were initially vertical but then generally followed a slightly

inclined pattern as they propagated upwards.

Specimen At formation of first crack(s) At failure

Load Deflection Load Deflection

kN mm kN mm

C25 58.2 8.43 64.0 10.61

C30 45.6 6.81 58.4 11.15

C35 68.6 8.94 80.6 10.92

From a physical examination of the specimens

after loading, it was noticed that the crack

pattern in the reinforced concrete elements

followed closely the position of the

reinforcement (Plate 3). In all specimens, the

reinforcement bars were not manufactured as

specified (Figure 3). This discrepancy created a

mass of unconfined concrete at the point of

maximum moment in the concrete-plate

interface.

Plate 3: C35 specimen after loading

(a) Reinforcement as fabricated (b) Reinforcement as specified

Figure 3: Specimen reinforcement comparison

However, the threaded bars remained embedded in the concrete throughout their whole

lengths and did not suffer any local bending; these were verified by extracting the bars

from the concrete.

4.1 Embedment length evaluation

To verify whether premature failure occurred due to insufficient embedment length, the

required length for the maximum load applied in each case was computed and compared

with the actual embedment length provided. For each maximum recorded load, 2Pu, the

moment generated, M, was computed.

Using equation 2, the forces in each bolt were found and thus the resultant force, F, was

computed.

From equation 3, the embedment length required for each case was computed and

compared to that provided as estimated from photographs of the specimens prior to

casting (Table 2).

Table 2: Computation and comparison of embedment lengths

Specimen 2Pu M F Required

length

Specified

length

Specimen

length

kN kNm kN mm mm mm

C25 64.0 42.67 115.89 173.8 259.0 159.0

C30 58.4 38.93 105.78 139.5 259.0 134.0

C35 80.6 53.73 146.03 172.4 259.0 184.0

From these results, it is evident that slip failure possibly occurred in the C25 and C30

specimens. However, the C30 specimen failed at the lowest load and this is explained

by the fact that this specimen had the least provided embedment length.

4.2 Load-deflection relationships

As seen from Chart 1, the three load-deflection charts are of the same shape. In the

initial stage of loading, deflection increased linearly with load since the stresses and

associated strains of each material in the specimens were small and all materials were in

the elastic portion of their respective responses.

Until a load of circa 35kN, all curves have approximately the same gradient. Beyond

this, a point of tri-furcation occurs: the C25 and C30 exhibit a slight drop in gradient

while the C35 specimen progresses with the same gradient. All specimens proceed

linearly until failure, after which linearity ceases. This is explained by the fact that, as

the yield stress is approached and attained, elastic behaviour ceases and plastic

deformation commences. Non-linear behaviour is associated with the plastic zones in a

load-deflection curve.

It was noted that in the region where the first cracks were observed, all specimens

showed a drop in gradient, thus a drop in stiffness. This is due to the fact that, as a

cracked section forms, both the effective 2nd moment of area and also the effective

modulus of elasticity drop, thus explaining the recorded drops in stiffness.

Due to the constant stiffness, EI, along the hybrid beam’s length, then at low load levels,

deflection was assumed to be directly proportional to the applied loading and inversely

proportional to the stiffness.

For various load values, P, the estimated value of deflection, δ, was computed and

compared with that measured, ∆, using the relation δ=23PL3/684EI (Table 3).

Table 3: Computation and comparison of deflections

The almost constant ratio between measured to estimated deflection suggests that there is

a correlation in the behaviour between the measured and estimated stiffness, albeit

biased towards the measured data. Thus, it can be concluded that, even in the elastic

region, the presence of the joint resulted in a decrease in stiffness of circa 60%.

4.3 Stress-strain relationships

From the data recorded by the strain gauges, stress-strain relationships were studied for

the different materials.

Specimen ∆ δ ∆/δ

mm mm

C25 1.5 0.94 1.596

C30 2.0 1.26 1.587

C35 2.5 1.57 1.592

For the concrete, the strain measured at low levels of loading, ε, were converted to

stresses using the Desayi-Krishnan relationship, i.e., σ=Ecε/({1+(ε/ε0)2}, where

ε0=0.00024(fcu/γm) as given in B.S.8110-1:1997, figure 2.1.

These stresses were compared to those which were expected to be generated by flexure

using the simple equation of bending σ’=My/I. The results are given in Table 4.

Table 4: Computation and comparison of concrete stress-strain relationships

In all specimens, it is evident that the concrete was highly stressed in the compression

zone, even in early loading stages. It is apparent that, as the beam rotated, further local

stresses in addition to those due to bending were induced, causing progressive internal

cracking and leading to surface cracking and thus compressive failure.

A similar procedure was computed for the steel beam using strains measured for each

specimen. In this case, stresses were obtained from measured strains using the stress-

strain relation from elastic theory σ=Esε/(1-ν)2, where ν=Poisson’s ratio, taken as 0.3 as

per B.S.5950-1:2000, clause 3.1.3. The strains were similarly compared to those

induced by bending using the relation σ’=My/I. The results are given in Table 5.

Table 5: Computation and comparison of steel stress-strain relationships

As in the case of the concrete stresses, the data measured indicates that, in the region of

the joint, high levels of stresses were present, even at low load levels. It was noted that

the strains in the compression zone were marginally larger than those in the tension

zone, possibly due to the fact that, as the beam rotated, addition local stresses were

generated near the point of rotation, i.e., in the compression zone.

4.4 Possible causes of failure

The unconfined concrete zone by the deficiency in reinforcement bar length decreased

the anchors’ effective embedment length, affecting the pull-out resistance and thus

implying transfer of tensile stresses generated by flexure to be transferred from the steel

beam to the anchors through the concrete.

Specimen fcu ε ε0 σ σ’

N/mm2 µm/m µm/m N/mm2 N/mm2

C25 25 0.7177 0.9798 10.49 2.84

C30 30 0.5981 1.0733 11.23 2.39

C35 35 0.4187 1.1593 9.84 3.67

Specimen ε σ σ’

µm/m N/mm2 N/mm

2

C25 0.7075 159.38 27.05

C30 0.2359 53.14 23.22

C35 0.5896 132.82 19.63

Although the C35 specimen had a sufficiently large embedment length, the specimen

failed at a lesser load than predicted, indicating that the unconfined zone proved to be a

weak link for transfer of tension.

The nature of confining steel also has an affect on the ductility of a reinforced concrete

member, particularly when stress levels approach the concrete uni-axial strength and

thus progressive internal cracking starts to occur. At this point, the concrete bears on the

transverse steel; this provides the concrete with a confining reaction. The provision of

widely spaced mild steel closed links near the zone of maximum moment further

jeopardised the confinement of concrete.

5. Conclusions and recommendations for future work

Based on results described in this paper, the following conclusions can be drawn:

� The capacity of a moment connection using cast-in-place bars is not directly related

to the concrete compressive strength but rather to the provision of adequate

confinement to the concrete.

� The use of high concrete grades and steel with high yield values are favoured since

materials were highly stressed in the zone of maximum moment.

� Adequate confinement should be provided by transverse steel, which should take the

form of high yield steel, closely spaced links.

� The introduction of the connection resulted in a decrease in stiffness of circa 60%.

Thus, adequate margins should be introduced in design to cater for flexural

performance and also to satisfy serviceability limit state requirements to prevent

excessive rotations and/or deflections.

The following are recommendations for further study related to the subject:

� The effect of proper confining steel to assess the ultimate capacity of anchors, using

different variables for transverse steel, including rectangular and circular hoops and

welded fabric reinforcement as transverse steel.

� The effect of cycling loading on connections achieved by means of anchors.

� Alternative connection methods using bolt couplers to develop full strength joints

between reinforcement bars and standard metric bolts.

6. References

1. B.S. 5950-1:2000: Structural use of steelwork in building. Part 1. Code of practice

for design-Rolled and welded sections, British Standards Institution, London, 2000.

2. B.S. 8110-1:1997: Structural use of concrete. Part 1. Code of practice for design

and construction, British Standards Institution, London, 1997.

3. B.S. 8007:1987. Code of practice for design of concrete structures for retaining

aqueous liquids, British Standards Institution, London, 1987.