Engineering Fracture Mechanics 74 (2007) 97–108

www.elsevier.com/locate/engfracmech

Size effect of compressed concrete in four point

bending RC beams
a,* b,1 a,2
Alessandro P. Fantilli , Ivo Iori , Paolo Vallini
a
Department of Structural and Geotechnical Engineering, Politecnico di Torino Corso Duca degli Abruzzi 24,
10129 Torino, Italy
b
Department of Civil and Environmental Engineering and Architecture, University of Parma,
Parco Area delle Scienze 181 A, 43100 Parma, Italy

Available online 3 March 2006

Abstract

A mechanical model is proposed to study the behaviour and ductility of reinforced concrete (RC) beams. The aim is to
reproduce numerically the experimental results of Weiss et al. (2001), where a size effect on the average strain of com-
pressed concrete has been measured by changing the distance of external loads in four point bending beams. Precisely,
the mechanical responses of RC members in bending are evaluated by modelling the progressive crushing of compressed
concrete. A good agreement between numerical and experimental results has been obtained for different types of beams.
The proposed approach also shows some differences from Eurocode 2, where size effect on the structural response of com-
pressed concrete is not taken into account.
Ó 2006 Elsevier Ltd. All rights reserved.

Keywords: Compressed concrete; Ductility; Four point bending; RC beams; Size effect

1. Introduction

The mechanical behaviour of compressed concrete has recently been the object of several investigations [1–
4]. The numerical and experimental models carried out in this field have mainly regarded the post peak soft-
ening branch of the stress–strain relationship. Generally, in this stage, progressive damage produces an
increase of compressive strains and a decrease of stresses. All these phenomena can affect the measure of duc-
tility, whose magnitude shows a size affect.
In the paper of Weiss et al. [5] for example, the flexural behaviour of reinforced concrete beams with
different constant-moment zone lengths is investigated. In particular, for each length, the mean value of

*
Corresponding author. Tel./fax: +39 011 5644889.
E-mail addresses: [email protected] (A.P. Fantilli), [email protected] (I. Iori), [email protected] (P. Vallini).
1
Tel.: +39 0521 90 5921; fax: +39 0521 90 5924.
2
Tel./fax: +39 011 5644849.

0013-7944/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.engfracmech.2006.01.013
98 A.P. Fantilli et al. / Engineering Fracture Mechanics 74 (2007) 97–108

Nomenclature

Ac area of concrete
As area of reinforcement
B width of a cross-section
c concrete cover
d effective depth of a cross-section
D width of a concrete specimen in compression
Ec tangent modulus of elasticity at concrete stress rc = 0
Ec1 secant modulus of elasticity at concrete stress rc = fc
Es tangent modulus of elasticity at reinforcement stress rs = 0
fc compressive strength of concrete
fc compressive strength of concrete in absence of confinement (r3 = 0)
fct tensile strength of concrete
fy yield strength of reinforcement
fu ultimate strength of reinforcement
h depth of a cross-section (h = d + c)
H, DH length of a concrete specimen in compression and its variation
Hi, DHi length of the ith concrete volume in compression (with ec > ec1) of a RC beam and its variation
K parameter of the proposed model
L length of the constant-moment zone in a four point bending beam
M bending moment
P applied load
s, Ds displacement on a sliding plane of a concrete specimen in compression and its variation
si, Dsi displacement on the sliding plane of the ith concrete volume in compression (with ec > ec1) of a
RC beam and its variation
w longitudinal inelastic shortening of a concrete specimen in compression
y vertical coordinate
yci distance between neutral axis and the ith concrete volume in compression (with ec > ec1) of a RC
beam
yc,max depth of neutral axis
y i width of the ith concrete volume in compression (with ec > ec1) of a RC beam
ys distance between neutral axis and reinforcement in tension
z horizontal coordinate
a angle between a sliding plane and horizontal axe
v scale factor
ec strain in compressed concrete
ec residual strain in a concrete specimen in compression during unloading
eci, Deci strain in the ith concrete volume in compression (with ec > ec1) of a RC beam and its variation
ec,max maximum compressive top fiber strain in cross-section
eme mean compressive top fiber strain in L
ec1 strain in compressed concrete at peak stress fc
ec1 compressive strain in concrete at peak stress fc in absence of confinement (r3 = 0)
es strain in reinforcement
u angle of internal friction
l cross-sectional curvature
x mechanical reinforcement ratio [x = Asfy/(B hfc)]
h plastic rotation
rc, Drc stress in a concrete specimen in compression and its variation
rci, Drci stress in the ith concrete volume in compression (with ec > ec1) of a RC beam and its variation
 A.P. Fantilli et al. / Engineering Fracture Mechanics 74 (2007) 97–108 99

rs stress in reinforcement
r3 confinement stress

compressive strains is evaluated. Since the damage is localised in different zones of constant-moment, the
beams show different ductility, despite the same cross-section.
The theoretical evaluation of ‘‘size effect’’ has generally been significant in structural analysis since the time
of Leonardo da Vinci or Galileo Galilei, when an increase of material strength with the decrease of structural
dimensions was observed [6]. Thus, in order to obtain a quantitative analysis of the mechanisms involved in
the response of reinforced concrete beams in bending, a theoretical model is proposed. We focus on the ulti-
mate stage before failure of reinforced concrete (RC) beams and, in particular, on the evaluation of structural
ductility and its size effect.
In this paper, the proposed model is validated by referring to the experimental analyses of Weiss et al. [5],
where four groups of beams, having different concrete strength (high HSC or normal NSC) and different per-
centages of reinforcement (high HR or low LR), have been tested. Each group contains four beams with the
same cross-section but different constant-moment zone length L (respectively 1, 2, 3, and 4 times the effective
depth d of the cross-section). Fig. 1(a) shows the tested beam, whereas the main mechanical properties of
materials and the amount of steel reinforcement are summarised in Table 1. Results indicate that damage

Fig. 1. The beams tested by Weiss et al. [5] (a) static conditions; (b) moment curvature responses.

Table 1
Mechanical and geometrical properties of the beams tested by Weiss et al. [5] (Fig. 1)
Beams fc (MPa) As (mm2) fy (MPa) fu (MPa) L/d
NSC–LR-1d 39.8 138.8 395 557 1
NSC–LR-2d 46.6 138.8 395 557 2
NSC–LR-3d 46.6 138.8 395 557 3
NSC–LR-4d 39.8 138.8 395 557 4
NSC–HR-1d 38.7 382.3 400 634 1
NSC–HR-2d 38.7 382.3 400 634 2
NSC–HR-3d 46.7 382.3 400 634 3
NSC–HR-4d 46.7 382.3 400 634 4
HSC–LR-1d 98.8 382.3 400 634 1
HSC–LR-2d 100.6 382.3 400 634 2
HSC–LR-3d 100.6 382.3 400 634 3
HSC–LR-4d 98.8 382.3 400 634 4
HSC–HR-1d 108.3 981.8 431 614 1
HSC–HR-2d 97.9 981.8 431 614 2
HSC–HR-3d 97.9 981.8 431 614 3
HSC–HR-4d 108.3 981.8 431 614 4
100 A.P. Fantilli et al. / Engineering Fracture Mechanics 74 (2007) 97–108

occurs in the compressed concrete of constant-moment zone, whose average moment curvature relationships
M–l (Fig. 1(b)) consequently show softening branches of different slopes. A decrease in ductility with the
increase of L is evident.
Therefore, it seems necessary to investigate the failure stage of reinforced concrete beams by modelling the
damage mechanisms localised in the compressed concrete.

2. Formulation of a theoretical model

The mechanical response of concrete specimens subjected to uniaxial compression (Fig. 2(a)) can be divided
into two stages (Fig. 2(b)). In the first stage, when the stress is lower than the concrete strength fc, the material
can be considered undamaged. As soon as the peak stress is reached, localised damage develops in the element
and the second stage begins. In this stage, the progressive sliding of two blocks of concrete (Fig. 2(a)) and the
softening branch (Fig. 2(b)) are evident. In Fig. 2(a), the angle a between the horizontal face of the specimen
and the sliding surface is assumed to be 17.5°, according to the experimental observations of Fujita et al. [7]. In
the cases of uniaxial compression, similar values of a are obtained through the Mohr–Coulomb failure crite-
rion, if the tensile strength of concrete is assumed to be 1/10 of that in compressive (fct = 1/10fc).
Both elastic shortening of the undamaged concrete and displacement between the two blocks along the slid-
ing surface (where crack is localised) rule the average post-peak compressive strain ec of the specimen (Fig. 2).
According to experimental results, the slope of softening branch of Fig. 2(b) increases in longer concrete spec-
imens because of the relationship w/H involved in the evaluation of ec. Therefore, longer specimens behave in
a more brittle manner, and sometimes snap-back can appear in the stress–strain relationship. In particular,
referring to the symbols depicted in Fig. 2, post peak stains and stresses can be respectively defined by the
following equations:
w Drc s  cos a
ec ¼ ec þ ¼ ec1 þ þ ð1Þ
H Ec H
Drc ¼ a  fc  w ¼ a  fc  s  cos a ð2Þ
where a = 0.1 is the reciprocal of the mean slope of the curve rc/fc  w [4]. Assuming K = a fc cos a, it is pos-
sible to obtain:
Drc ¼ K  s; rc ¼ fc  K  s ð3Þ
and substituting (3) into (1), and rearranging:
ec  ec1
s ¼ cos a K ð4Þ
H
 Ec
For the beam of Fig. 3(a), assuming plane sections compatibility, the strain profile ec(y) of a generic cross-
section (Fig. 3(b)) can be computed through the following equation:

Fig. 2. Damage localization in compressed concrete: the tests of Jansen and Shah [4].
 A.P. Fantilli et al. / Engineering Fracture Mechanics 74 (2007) 97–108 101

Fig. 3. Reinforced concrete element in bending: (a) cross-section; (b) longitudinal strain profile; (c) constant-moment zone.

e¼ly ð5Þ
The strains that exceed ec1, which rule the evolution of sliding planes, are split up into several rectangles of
base Deci (Fig. 3(b) and (c)). Due to the increase of ec,max, the increments Deci and the sliding planes progres-
sively involve wider zones of the beam. However, in case of constant bending moment, it is not possible to
define univocally both the evolution of the sliding planes and their slip increments Dsi. For the sake of sim-
plicity, in this paper, the number of planes originated by Deci is assumed to be as low as possible. This assump-
tion is founded on the hypothesis that new sliding planes are more probably originated by previous fracture
surfaces. Therefore, only one of the possible crack pattern is depicted in Fig. 3. In case of three point bending
beams (Fig. 4), new sliding planes start from the cross-sections with higher bending moment, so that the crack
pattern configuration is less ambiguous (Fig. 4(c)). In all cases, the strain increment Deci, related to the dis-
placement variation Dsi (Fig. 5), affects a volume of concrete in compression of width y i and length 2 Hi:
ec;max  eci
y i ¼ ð6Þ
l
y
Hi ¼ i ð7Þ
tan a
where a is the angle of the sliding planes with respect to the horizontal line (Fig. 5(c)).
For a given strain increment Deci (Fig. 5(b)), the displacement variation Dsi, affecting the width y i
(Fig. 5(c)), can be computed by substituting (6) and (7) into (4):

Fig. 4. Reinforced concrete element in bending: (a) cross-section; (b) longitudinal strain profile; (c) linear variation of the moment.
102 A.P. Fantilli et al. / Engineering Fracture Mechanics 74 (2007) 97–108

Fig. 5. Reinforced concrete element in bending: (a) cross-section; (b) longitudinal strain profile; (c) longitudinal sliding.

Deci ðec;max  eci Þ  Ec  Deci

Dsi ¼ lsin a
¼ ð8Þ
ec;max eci
 EKc l  sin a  Ec  K  ðec;max  eci Þ

The derivation of si with respect to y can be also evaluated from the following equation:
dsi ðy c;max  y ci Þ  Ec  l
¼ ð9Þ
dy sin a  Ec  K  ðy c;max  y ci Þ
By means of Eq. (3), the concrete stress can also be obtained by a differential equation:
dr ds K  ðy c;max  y ci Þ  Ec  l
¼ K ¼ ð10Þ
dy dy sin a  Ec  K  ðy c;max  y ci Þ
and, according to Fantilli et al. [8,9], when ec1 < ec < ec,max, the stress of compressed concrete can be computed
with the equation:
"   #
E2c  l  sin a l  Ec  sin a  y c;max  K þ K  ec
rc ¼ fc   ln   þ Ec  ðec  ec1 Þ ð11Þ
K l  Ec  sin a  y c;max  K þ K  ec1

In this equation, the stress rc depends on the extension of the compressive zone yc,max, according to the size
effect model of Hillerborg [10]. However, in the proposed approach, the mechanical response of compressed
concrete is also a function of the cross-sectional curvature l.
It must be remarked that Eq. (11) is founded on the assumption that sliding planes are inclined of a = 17.5°
with respect to the load direction (Fig. 2(a)). According to the Mohr–Coulomb failure criterion, the uncer-
tainty of a depends on the ratio fct/fc. In ordinary concretes, the possible variation of this ratio (0.067 6
fct/fc 6 0.2) may give 11° 6 a 6 24°. By assuming all the values of a within this range, rc (ec = 0.0035) com-
puted with Eq. (11) can vary of about 4%, with respect to the values computed with a = 17.5°. Therefore,
Eq. (11) does not seem very sensitive to the possible uncertainties of a.
When ec < ec1 the following uniaxial stress–strain relationship of CEB–FIP model code [11] is adopted for
the ascending branch of concrete in compression:
 2
Ec ec ec
Ec1 ec1
 ec1
rc ¼    fc ð12Þ
1 þ EEc1c  2  eec1c

where Ec1 = fc/ec1.

When a confinement tension r3 is present (e.g., due to stirrups), the values of fc and ec1, which appear in the
Eqs. (11) and (12), increase. By assuming the linear envelope failure of Fig. 6(a), the relationship between the
concrete strength fc and r3 is easily evaluated as:
 A.P. Fantilli et al. / Engineering Fracture Mechanics 74 (2007) 97–108 103

Fig. 6. Confinement produced by stirrups [9]: (a) linear envelope for concrete strength; (b) increment of fc and ec1.

fc ¼ fc þ b  r3 ð13Þ
where:
1 þ sin u
b¼ ;
1  sin u
fc is the nominal strength without confinement stresses (r3 = 0) and u is the angle of internal friction. The
increase of fc is accompanied by an increment of the peak-strain ec1, whose growth is assumed to be directly
related to fc (Fig. 6(b)):
fc
ec1 ¼ ec1  ð14Þ
fc
where ec1 and ec1 are respectively the peak-strains for unconfined (r3 = 0) and for confined (r3 > 0) concrete.
In the constant-moment zone L of a four point bending beam, the unfailing scatter of concrete strength
makes the definition of crack pattern more ambiguous. Fig. 7(a), for example, shows a possible displacement
configuration when in Section 1 a reduced strength, with respect to Section 2, has been detected. The bearing

Fig. 7. Reinforced concrete element in bending: (a) cross-section; (b) longitudinal strain profile; (c) damage localization in the constant-
moment zone.
104 A.P. Fantilli et al. / Engineering Fracture Mechanics 74 (2007) 97–108

capacity of the beam (i.e., the maximum moment Mu1) must be evaluated in Section 1. In the zones closer to
this section, the increase of curvature after the peak produces both a reduction of the bending moment M and
an extension of damaged zone in compressed concrete. In other zones (Section 2), where the maximum
moment Mu2 has not been reached, the decrease of M produces an elastic decrease of curvature l. Therefore,
as for the compressed concrete specimens of Fig. 2, the failure mechanisms is localised in a part of the beam. In
particular, the softening branch of the M–l diagram, evaluated in the constant-moment zone, shows an
increasing slope with the increase of L. Such a failure is similar the failure of compressed concrete cylinders.
It must be also observed a reduced ductility of four point bending beams with longer L, because localised
damage is smeared on wider zones. In other words, the post peak behaviour is ruled by the softening branch of
the moment curvature M–l relationship of the weakest section (Section 1 in Fig. 7(c)), whereas, in the other
zones elastic unloading should be considered (Section 2 in Fig. 7(c)).
Although the position of the weakest section in the constant-moment zone is always unknown, its existence
cannot be neglected. In fact, due to series-connection between all the sections in these zones, only the weakest
one can follow the quasi-horizontal branch that precedes the peak of the moment curvature diagrams
(Fig. 7(c)). In this way, except for the zone with the lower strength (where the damage is localised), the other
sections of the beam cannot release all their potential ductility. Starting from these assumptions, it is possible
to explain the dependence of four point bending beams ductility on the constant-moment zone length, as
observed by Weiss et al. [5].

3. Comparison between the experimental results of Weiss et al. [5] and the numerical results of the
proposed model

The higher ductility of the beams with a shorter constant-moment zone length cannot be theoretically
reproduced through Eqs. (11) and (12). In these equations, both damage localization and confinement pro-
duced by external load P are not taken into account. Due to P, the compressed concrete is confined and its
strength shows a local increase according to Eqs. (13) and (14) (Fig. 8). To reproduce numerically this phe-
nomenon, it is necessary to define the geometry of confined zones. As is shown in Fig. 9, the confined zones
are bounded on the top of the beam by the sliding planes of compressed concrete, which start from the points
where the loads are applied. Since they extend to the border of compressed zone where the concrete strength is
reached (incipient sliding), the whole constant-moment zone can be confined (Fig. 9(a)). This is true in the
cases of short distance between the applied forces or deep compressed zones (high reinforcement ratio).
For the sake of simplicity, the mechanical response is obtained by analysing two representative sections:
Section AA and Section BB of Fig. 9. As is explained before, since a series connection between cross-sections
is established in the constant-moment zone, damage starts in the cross-section with the lowest concrete
strength. In particular, in proximity of loads P, sliding planes grow from unconfined concrete to confined
zones (with a higher resistance). Schematically this behaviour is reproduced by the Section BB of Fig. 9(a)
and (b), where an average confinement r3 ¼ P =ð2By i = tan aÞ is assumed for y i > y 0 . In the zones closer to Sec-
tion BB, concrete sliding displacements increase, thus their values are doubled with respect to the ones
obtained from Eq. (9). Section AA is located in the middle of the constant-moment zone, where confinement
stresses r3 are lower compared to Section BB (Fig. 9(a)), or null as in the case of Fig. 9(b).

Fig. 8. Confinement produced by the applied forces P.

 A.P. Fantilli et al. / Engineering Fracture Mechanics 74 (2007) 97–108 105

Fig. 9. Confinement in the constant-moment zone: (a) confinement in the whole length L; (b) confinement in a part of L.

The mean deformability of the constant-moment zone can be obtained by considering the effect of damage
localization in the Section (AA or BB) with the lower peak moment in the M–l relationship. Consequently,
the descending post-peak branch is ruled by the softening behaviour of the damaged cross-section and by the
linear elastic unloading of the undamaged ones. Both these contributes are assumed as directly proportional to
referring zone length.
According to the tests, a strict correlation between the ductility and constant-moment zone length L is
shown by the numerical results of the proposed model. This is evident in Figs. 10 and 11, where the mean

Fig. 10. Average compressive top fiber strain eme in the constant-moment zone L: comparison between the numerical results of the
proposed model and those experimentally measured in [5] for the beams of normal strength concrete (NSC) with high (HR) and low (LR)
reinforcement ratio.
106 A.P. Fantilli et al. / Engineering Fracture Mechanics 74 (2007) 97–108

Fig. 11. Average compressive top fiber strain eme in the constant-moment zone L: comparison between the numerical results of the
proposed model and those experimentally measured in [5] for the beams of high strength concrete (HSC) with high (HR) and low (LR)
reinforcement ratio.

compressive top fiber strains eme experimentally evaluated are compared with the numerical ones. A little dis-
crepancy is revealed for beams HSC–LR, made of high strength concrete and low reinforcement ratio
(Fig. 11), for which the maximum bending moment and ductility are overestimated. In the case of lower rein-
forced concrete beams the magnitude of damaged zones, approximately equivalent to depth of compressed
concrete, can drastically reduce the ductility contribution of the undamaged cross-sections. Therefore, the
peak moment is reached when the inelastic strains are lower than the ones of undamaged concrete.

4. Evaluation of plastic rotation with the proposed model

From a practical point of view, it can be also interesting to evaluate the ductility of RC beams by measuring
the plastic rotation h. In this paper, the value of h is conventionally computed by referring to a portion of
beam having a length equal to the depth h of its cross-section:
h ¼ lp  h ð15Þ
where lp is the cross-sectional curvature at peak of the moment curvature diagram. Since the rotation h is
evaluated at the peak of the M–l curve, it also includes the effects of the inelastic behaviour of the beam.
In the serviceability stage, cracks are limited in width and the M–l diagrams are remarkably affected by
the tension-stiffening phenomenon [12]. On the contrary, during the ultimate stage, the contribution of con-
crete in tension can be neglected. In fact, due to a heavy crack pattern, bond between steel and concrete is
completely damaged. Therefore, in the evaluation of lp of RC beams, it seems appropriate to neglect the ten-
sion stiffening effect.
 A.P. Fantilli et al. / Engineering Fracture Mechanics 74 (2007) 97–108 107

Generally, Code rules [11,13] define the ultimate stage when the conventional limit strains in concrete or
steel are reached. However, to avoid a connection between the numerical results obtained with the proposed
model and the conventional limit strains, in the present work compressed concrete is modelled without any
restriction for the strain ec. In particular, the approach described in the previous paragraphs is here applied
to different beams, obtained from the cross-sections depicted in Fig. 12(a) by scaling its sizes by a factor v.
In the numerical simulations, the mechanical properties of the adopted steel bars are summarized in
Fig. 12(b). The maximum bending moments and the relative curvature lp are obtained with the proposed
approach and with Eurocode 2 [13] requirements for the same cross-sections. In the latter case the compressed
concrete is simply modelled through the parabola–rectangle r–e diagram. Thus in a generic cross-section the
maximum moment and lp are reached when one of the materials attains its limit strain. Both for the proposed
model and from Eurocode 2 approach, the analyses are developed in a deterministic way, without any partial
factor for the materials.
To highlight the effect of beam size, the plastic rotation h is normalized with respect to h(v = 1), and the
diagrams of h/h(v = 1) versus v are shown in Fig. 13. Size effects are particularly evident in beams with low
mechanical reinforcement percentage x = Asfy/(B h fc) (Fig. 13(a)). They cannot be reproduced in the frame-
work of the classical approach proposed by Eurocode 2 [13], where compressed concrete is modelled through
the parabola–rectangle r–e diagram.

Fig. 12. The cross-sections considered in the evaluation of plastic rotation h: (a) geometry and concrete properties; (b) stress–strain
diagram of reinforcing steel.

Fig. 13. Normalized plastic rotation h/h (v = 1) versus size factor v diagram: (a) results of the proposed model and Eurocode 2[13] for RC
beams having x = 0.1; (b) results of the proposed model and Eurocode 2 [13] for RC beams having x = 0.3.
108 A.P. Fantilli et al. / Engineering Fracture Mechanics 74 (2007) 97–108

5. Conclusions

To analyse the ultimate stage of reinforced concrete beams, a mechanical model for compressed concrete
has been proposed by Fantilli et al. [8]. Afterwards, this approach has been extended to confined concrete sub-
jected to compressive loads in order to simulate the effectiveness of stirrups confinement [9]. In the present
paper, the model has been further refined with the aim of reproducing different boundary conditions of tests.
The good agreement between the numerical results and the experimental measures of Weiss et al. [5] demon-
strate the effectiveness of the model in computing the mechanisms that affect reinforced concrete beams during
the ultimate stage. In particular, it is possible to define the zones where the maximum strains are localised, and
consequently to evaluate the flexural response of reinforced concrete beams and its size dependence. The post
peak branch of the r–e diagram remarkably affects the cross-sectional strength of a RC beam and the corre-
sponding plastic rotation h. The decrease of h, observed when dimensions of beam increase, can be explained
as a reduced capability of bigger structures to bear plastic deformations. Since size effect is not currently com-
puted by Eurocode 2, it is desirable that, in future, Code requirements will be able to consider explicitly this
phenomenon.

References

[1] Bazant ZP, Xiang Y. Size effect in compression fracture: splitting crack band propagation. J Engng Mech ASCE 1997;123(2):162–72.
[2] Markeset G, Hillerborg A. Softening of concrete in compression—Localization and size effects. Cem Concr Res 1995;25(4):702–8.
[3] Wang EZ, Shrive NG. Brittle fracture in compression: mechanisms, models and criteria. Engng Fract Mech 1995;52(6):1107–26.
[4] Jansen DC, Shah SP. Effect of length on compressive strain softening of concrete. J Engng Mech ASCE 1997;23(1):25–35.
[5] Weiss WJ, Güler K, Shah SP. Localization and size-dependent response of reinforced concrete beams. ACI Struct J 2001;98(5):
686–95.
[6] Benvenuto E. An Introduction to the History of Structural Mechanics. Part I: Statics and Resistance of Solids. New York: Springer-
Verlag; 1991.
[7] Fujita Y, Ishimaru R, Hanai S, Suenaga Y. Study on internal friction angle and tensile strength of plain concrete. In: Mihashi H,
Rokugo K, editors. Fracture Mechanics of Concrete Structures FRAMCOS 3. Freiburg Aedificatio Publishers; 1998. p. 325–34.
[8] Fantilli AP, Ferretti D, Iori I, Vallini P. A mechanical model for the failure of compressed concrete in R/C beams. J Struct Engng
ASCE 2002;128(5):637–45.
[9] Fantilli AP, Iori I, Vallini P. Mechanical model for the confined compressed concrete of RC beams. 1st fib congress: Concrete
structures in the 21st century 2002:43–52.
[10] Hillerborg A. Fracture mechanics concepts applied to moment capacity and rotational capacity of reinforced concrete beams. Engng
Fract Mech 1990;35(1/2/3):233–40.
[11] CEB 1991. CEB-FIP Model Code 1990. Bulletin d’information 203–205. Lausanne, 1991.
[12] Fantilli AP, Ferretti D, Iori I, Vallini P. Flexural deformability of reinforced concrete beams. J Struct Engng ASCE 1998;124(9):
1041–9.
[13] European Committee for Standardization. Eurocode 2: Design of concrete structures—Part 1: general rules and rules for buildings.
Brussels, 2001.