Escolar Documentos
Profissional Documentos
Cultura Documentos
www.elsevier.com/locate/engstruct
Received 8 September 2004; received in revised form 16 March 2005; accepted 25 April 2005
Available online 24 May 2005
Abstract
To better understand the response of high-strength concrete beams failing in shear with and without shear reinforcement, eighteen
reinforced concrete beams were tested as a part of extensive research on shear design of reinforced high-strength concrete beams. The
concrete compressive strength of the beams at the age of the tests ranged from 50 to 87 MPa. The primary design variables were the amount
of shear and longitudinal reinforcement. A minimum amount of shear reinforcement related to the concrete tensile strength was also proposed.
The details of the beam specimens, material properties, instrumentation and the testing procedure used are carefully described in this paper.
They will be useful for researchers to compare and analyze other design approaches. The test results are presented and discussed, and the
influence of each design parameter is studied separately. Furthermore, the minimum amount of web reinforcement proposed in this paper
is validated using experimental data from the literature. Test results are also compared with different shear design approaches and, finally,
conclusions are drawn.
© 2005 Elsevier Ltd. All rights reserved.
Keywords: High-strength concrete; Reinforced concrete; Beams; Tests; Shear failure; Building codes
1. Introduction
0141-0296/$ - see front matter © 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.engstruct.2005.04.010
1520 A. Cladera, A.R. Marí / Engineering Structures 27 (2005) 1519–1527
Table 1
Details of the beam specimens and summary of the experimental results
Moreover, shear failure in a beam without web reinforce- – To evaluate the efficiency of the amount of shear
ment is sudden and brittle. Therefore, it is necessary to reinforcement as a function of the concrete compressive
provide a minimum amount of shear reinforcement, which strength. Some authors believe that for high-strength
must prevent sudden shear failure on the formation of first concrete beams, stirrups are more efficient than for
diagonal tension cracking and, in addition, must adequately normal-strength beam specimens [4,5].
control the diagonal tension cracks at service load levels. – To evaluate the influence of the amount of longitudinal
Thus, the minimum area of web reinforcement is intended reinforcement on the shear strength. The majority of
to ensure that the capacity of the member after cracking current codes’ limitation of the amount of longitudinal
exceeds the load at which inclined cracking occurs. Due to reinforcement to 2% will be studied for high-strength
the higher tensile strength of high-strength concrete, a higher concrete beams.
cracking shear is expected and hence, would require a larger – To study the influence of web-distributed longitudinal re-
amount of minimum shear reinforcement [1]. inforcement for high-strength members without stirrups,
The experimental campaign described in this paper was as this variable has an important effect on the failure shear
part of extensive research that led to the proposal of a strength according to Ref. [6].
new shear design procedure for normal and high-strength
concrete beams [2,3].
3. Design of the test specimens
2. Objectives of the experimental campaign In order to achieve the prior objectives, eighteen beam
specimens were designed and tested. Table 1 and Fig. 2
The main objectives of the experimental campaign show the details of the 200 mm wide × 400 mm deep beam
carried out were: specimens that were tested with a shear span of 1080 mm.
The test program [7] consisted of four series of beams:
– To study the influence of the concrete compressive (1) the H50 series, designed to have a concrete strength of
strength on the shear strength in beams with and without 50 MPa; (2) the H60 series, designed to have a concrete
shear reinforcement. Some current procedures hold that strength of 60 MPa; (3) the H75 series, designed to have
the failure shear strength does not increase when concrete a concrete strength of 75 MPa with silica fume; and (4)
compressive strength is higher than 60 MPa for beams the H100 series, designed to have a concrete strength of
both with and without web reinforcement. 100 MPa. The actual concrete strength at the time of testing
– To propose and verify a minimum amount of web is presented in Table 1.
reinforcement for high-strength concrete beams in Beam specimen number one in each series (H50/1,
accordance with the increase in concrete tensile strength H60/1, H75/1, and H100/1) did not have shear reinforce-
for high-strength concretes. ment. The longitudinal reinforcement consisted of two
A. Cladera, A.R. Marí / Engineering Structures 27 (2005) 1519–1527 1521
32 mm diameter bars (ρl = 2.24%), with a characteristic The proposed amount of minimum shear reinforcement
yielding stress of 500 MPa. is given in Eq. (3) and compared in Fig. 3 with some other
Beam number two in each series (H50/2, H60/2, H75/2, code proposals.
and H100/2) contained the proposed minimum amount of
shear reinforcement. The longitudinal reinforcement was f ct,m bw s
Aw,min = MPa (3)
equal to that provided in series 1. 7.5 fy
It is proposed that the minimum amount of shear
reinforcement be proportional to the tensile strength of the where Aw,min is the amount of minimum shear reinforce-
concrete, fct,m . In the ‘Design guidance for high-strength ment, bw is the web width, s the stirrup spacing and f y is the
concrete’ [8] it is suggested that traditional tensile strength yield strength of the reinforcing steel.
can be unconservative for high-strength concrete, and it The above formulation was derived taking into account
proposes the following equation: that the minimum shear reinforcement is necessary to avoid
a sudden failure after the diagonal cracking and, therefore, it
should be proportional to the tensile strength of the concrete.
f ct,m = 0.30 3 f c2 MPa if f c ≤ 60 MPa (1)
To totally derive the proposed equation, the minimum
f ct,m = 0.58 2 f c MPa if fc > 60 MPa. (2) amount was fixed to around 0.35 MPa for a 25 MPa concrete
1522 A. Cladera, A.R. Marí / Engineering Structures 27 (2005) 1519–1527
Table 2
Properties of web reinforcing bars
Size-series Area fy fu
Fig. 3. Comparison of the minimum amount of web reinforcement of (mm2 ) (MPa) (MPa)
different specifications and proposed equation proportional to the concrete
tensile strength. φ6-H60 and H75 28.27 530 680
φ8-H60 and H75 50.27 530 685
φ6-H50 and H100 28.27 530 680
(Fig. 3), as the ACI 99 Code proposed for conventional φ8-H50 and H100 50.27 540 672
concretes.
The third specimen in each series (H50/3, H60/3, H75/3,
and H100/3) had the same amount of web reinforcement for To monitor the behaviour of the tested beams, the
all beams, 8 mm diameter stirrups spaced by 210 mm. It was applied loads, strains at the reinforcement and at the
designed to have the highest amount of web reinforcement concrete surface, and displacements were measured using
in order to produce a shear failure with the provided different instruments such as load cells, strain gauges and
amount of longitudinal reinforcement. Additionally, the magnetostrictive transducers (LVDTs). All the variables
flexural tension reinforcement for all the beams consisted of were monitored continuously by the data acquisition system.
2φ32 bars. Photography and video equipment were also utilised.
The fourth beam in each series (H50/4, H60/4, H75/4, The web strain, υx y , was measured by means of a rosette
and H100/4) had the same shear reinforcement as the of two LVDTs mounted in one side of the beams of series
third series but the flexural tension reinforcement consisted H50 and H100 (see Fig. 4). Using the Mohr circle it is easy
of 2φ32 bars plus a 1φ25 bar. Hence, the amount of to deduce how the web strain can be obtained from the
longitudinal reinforcement was equal to 2.99%. reading of the transducers [7]. However, the direction of the
The fifth beam specimen in each series (H50/5 and strains in the web may not be found with this set-up.
H100/5) did not have stirrups but instead contained four The load was applied at midspan of the beam specimen
small longitudinal bars (8 mm diameter) distributed along by a 150 mm wide and 28 mm thick neoprene pad under
the web plus two bars in the compression zone, resulting
a distance between them of 110 mm. As was mentioned
earlier, Collins et al. [6] postulated that the size effect
is not only a function of the beam depth, but also of
the distance between distributed longitudinal reinforcement
(sz ). According to the AASHTO Specifications, each layer
of this crack control reinforcement must have an area of at
least 0.004bwsz . The area provided in specimens H50/5 and
H100/5 verified the previous expression.
Fig. 6. Cracking prior to failure and at failure in a beam with web reinforcement. Concrete spalling near the diagonal crack.
Table 3
Minimum amount of web reinforcement, observed failure, yielding and cracking shear for each specimen provided with the proposed minimum amount
Fig. 7. Shear deformation in beams without shear reinforcement and with Fig. 8. Beam specimens with web reinforcement. Failure shear strength
the proposed minimum. versus concrete compressive strength for series 3 and 4.
Fig. 9. Beam specimens with distributed longitudinal reinforcement. Shear Fig. 10. Validation of the proposed amount of minimum reinforcement with
strain in beams H50/5 and H100/5. experimental data of the literature.
one shear crack and failure shear strength was 30.5% higher
reinforcement. For members with stirrups, the procedure
in beam H50/5 and 18.9% in beam specimen H100/5 than
is more conservative, especially for high-strength concrete
in other, similar beams without web reinforcement (Fig. 9).
specimens.
However, failure was also sudden.
The EC-2 equations are also excessively conservative
for members with shear reinforcement, in particular for
7. Comparison of test results with different design members with the minimum amount of web reinforcement,
approaches while for beam specimens without stirrups it is slightly
unconservative.
Table 4 summarises the shear procedures included in the
ACI 318-02 Code [9], AASHTO LRFD 2001 [10], the 2002
Final Draft of Eurocode 2 [11] and the method proposed 8. Validation of the proposed minimum amount of web
by Cladera and Marí [2,3]. To calculate the predictions reinforcement
(Table 5) all the safety factors were taken equal to 1.00.
The amount of web reinforcement was obtained by consid- It has already been seen that, for the beam specimens
ering the stress state of the stirrups during the test just before described in this paper, the minimum amount of web
reaching the failure load (Table 1). For the beams with web reinforcement proportioned adequate reserve shear strength.
reinforcement, the EC predicted failure loads depend on the However, as significant amount of data exist in the current
angle of the compression struts, θ . It has been taken, to ob- literature, a broader analysis was carried out using test
tain the highest prediction, as cotg θ equal to 2.5 (Table 5). results by [1,7,12–17].
For the ACI procedure the angle is fixed to 45◦ (cotg θ = 1). Fig. 10 represents the amount of web reinforcement
For the AASHTO LRFD shear procedure and the method provided in tested beams over the minimum amount
proposed by Cladera and Marí the angle of the compression proposed versus the reserve shear strength index (Vfail /Vcr ).
strut is obtained by compatibility (Table 5). Both methods Ozcebe et al. [12] suggested that to insure an adequate
are based on the Modified Compression Field Theory. margin of safety the value of the index should be greater than
For the eighteen beam specimens, the average Vtest / 1.30. This condition is satisfied (Fig. 10) by the minimum
Vpredicted ratio is 1.34 for the ACI 318-99 formulation, 1.26 amount of web reinforcement proposed by Eq. (3).
for the EC-2, 1.18 for the AASHTO LRFD, and 1.14 for the
Refs. [2,3] predictions. The coefficient of variation (standard
deviation over the average) is 11.21% for the ACI Code, 9. Conclusions
6.66% for the AASHTO LRFD, 20.2% for the EC-2, and
6.93% for the Cladera and Marí shear procedure. Based on the test results of the eighteen beam specimens,
Both the AASHTO LRFD and Cladera and Marí predic- the following conclusions can be drawn:
tions prove very satisfactory when compared with the EC-2
predictions. – Beams without web reinforcement presented a very
The ACI procedure correlates better with members brittle behaviour. The higher their concrete compressive
without web reinforcement, than with members with shear strength, the brisker their failure.
1526 A. Cladera, A.R. Marí / Engineering Structures 27 (2005) 1519–1527
Table 4
Summary of different shear procedures
Equations Comments
Table 5
Summary of the predictions made by the ACI Code 318-99, the 2001 AASHTO LRFD Specifications, April 2002 Final Draft of the Eurocode 2 procedures
and method proposed in Refs. [2,3]
– For beams without web reinforcement, the failure shear Vy shear strength at yielding of the stirrups;
strength generally increased as the concrete compressive φ diameter of the rebar;
strength increased, except for beam H75/3. ρl amount of longitudinal reinforcement (%);
– High-strength concrete beams with stirrups presented a ρw f y amount of shear reinforcement (MPa).
less fragile response than similar beams without web
reinforcement.
References
– The minimum amount of web reinforcement proposed in
this paper was sufficient in terms of the demand of reserve [1] Yoon YS, Cook WD, Mitchell D. Minimum shear reinforcement in
of strength after shear cracking. normal, medium and high-strength concrete beams. ACI Struct J 1996;
– For beams with the same geometric amount of transverse 93(5):576–84.
reinforcement, the higher their concrete compressive [2] Cladera A, Marí AR. Shear design procedure for reinforced concrete
beams using artificial neural networks. Part I: Beams without stirrups.
strength, the more effective stirrups are. Eng Struct 2004;26(7):917–26.
– For high-strength concrete beams with stirrups, the [3] Cladera A, Marí AR. Shear design procedure for reinforced concrete
limitation of the amount of longitudinal reinforcement to beams using artificial neural networks. Part II: Beams with stirrups.
2% is not experimentally justified. Eng Struct 2004;26(7):927–36.
[4] Elzanaty AH, Nilson AH, Slate FO. Shear capacity of reinforced
– Beam specimens with longitudinally distributed web re-
concrete beams using high-strength concrete. ACI J Proc 1986;83(2):
inforcement along the web showed a better behaviour 290–6.
than similar beams without any kind of shear reinforce- [5] Duthinh D, Carino NJ. Shear design of high-strength concrete
ment. Although their failure was also fragile, several beams: a review of the state-of-the-art. Building and Fire Research
shear cracks, rather than a single one, were reported, and Laboratory, National Institute of Standards and Technology; 1996.
p. 198.
the failure shear strength increased about 25%. [6] Collins MP, Kuchma D. How safe are our large, lightly reinforced
– Methods based on the modified compression field theory concrete beams, slabs and footings? ACI Struct J 1999;96(4):482–90.
showed a close correlation to the empirical results in [7] Cladera A. Shear design of reinforced high-strength concrete
comparison to the other codes’ correlations. beams. Pd.D. thesis published by ACHE (Spanish Concrete
Association). 2003. Also available on the word wide web:
http://www.tdx.cesca.es/TDCat-0328103-122019/.
[8] Concrete Society Technical Report 49. Design guidance for high
Acknowledgments strength concrete. United Kingdom; 1998. p. 168.
[9] American Concrete Institute. ACI building code requirements for
reinforced concrete, ACI 318-02; 2002.
The research described in this paper comprises part of [10] AASHTO LRFD bridge design specifications and commentary. 2nd
the Spanish Ministry of Science and Technology’s projects ed. (1998) and 2000 update. Washington DC: American Association
TRA99/0974 and MAT2002-00615. The beam specimens of State Highway Transportation Official; 1998, 2000.
were cast at the Alvisa precast concrete plant located in [11] European Committee for Standardization. Eurocode 2: Design of
concrete structures, Part 1: General rules and rules for buildings. Final
Selgua (Huesca, Spain), the authors extend their sincere
Draft. July 2002, p. 226.
thanks to them. The beams were tested at the Structural [12] Ozcebe G, Ersoy U, Tankut T. Evaluation of minimum shear
Technology Laboratory of the Construction Engineering reinforcement requirements for higher strength concrete. ACI J 1999;
Department, with the assistance of the former undergraduate 96(3):361–8.
students Josep Capell and Jorge Suárez. [13] Etxeberria M. Experimental study on microstructure and structural
behaviour of recycled aggregate concrete. Ph.D. thesis, Universidad
Politécnica de Cataluña; 2004, p. 230.
[14] González-Fonteboa B. Hormigones con áridos reciclados procedentes
de demoliciones: dosificaciones, propiedades mecánicas y compor-
Appendix. Notation
tamiento estructural a cortante. Ph.D. thesis, Universidad de la
Coruña; 2002.
Aw,min amount of minimum shear reinforcement; [15] Teoh BK, Mansur MA, Wee TH. Behaviour of high-strength concrete
a/d shear span to depth ratio; I-beams with low shear reinforcement. ACI Struct J 2002;99(3):
299–307.
bw web width; [16] Adebar P, Collins MP. Shear strength of members without transverse
d effective depth; reinforcement. Canad J Civil Eng 1996;23:30–41.
fc concrete compressive strength; [17] Johnson MK, Ramirez JA. Minimum shear reinforcement in beams
f ct,m average concrete tensile strength; with higher strength concrete. ACI Struct J 1989;86(4):376–82.
f sp average concrete splitting strength;
fu ultimate tensile strength of the reinforcing steel; Dr. A. Cladera is an assistant professor at the Department of Physics
at the University of Balearic Islands. He obtained his Ph.D. in 2003 at
fy yield strength of the reinforcing steel; the Technical University of Catalonia researching on shear strength of
s stirrup spacing; reinforced high-strength concrete beams.
sz vertical distance between distributed longitudinal
reinforcement; Dr. A.R. Marí is a Professor of Construction Engineering at the Technical
University of Catalonia, Barcelona, Spain. He is a member of the fib task
Vcr cracking shear strength; group 4.1 “Serviceability models” and general reporter of the Committee in
Vfail failure shear strength; charge of updating the Spanish Concrete Code.