Strength Capacity of RC Beams Without Shear Reinforcement

Engineering Structures 317 (2024) 118634
Contents lists available at ScienceDirect
Engineering Structures
journal homepage: www.elsevier.com/locate/engstruct
Strength capacity of RC beams without shear reinforcement: Numerical

analysis and comparisons with code provisions
Nicola Caterino a, b, * , Vincenzo Di Cristo a , Francesca Ceroni a
a
University of Parthenope, Department of Engineering, Centro Direzionale di Napoli, Isola C 4, 80143 Naples, Italy
b
Construction Technologies Institute, National Research Council, Corso N. Protopisani n. 70, 80146 Naples, Italy
A R T I C L E I N F O A B S T R A C T
Keywords: The investigation of the behavior of reinforced concrete (RC) elements without shear reinforcement is a current
RC beams focal point, driven by the incomplete consolidation of predictive formulas for shear strength in RC elements.
Shear strength Currently, the literature provides, indeed, a limited number of experimental and numerical studies on this
Shear reinforcement
subject. This paper seeks to advance the comprehension of the behavior and collapse mechanisms of RC beams
Code provisions
FE numerical analyses
not provided of shear reinforcement, as commonly employed in RC slabs of bridges. The paper commences with a
critical review of several predictive formulas for the shear strength of RC elements lacking transverse rein-
forcement, as stipulated by various international codes. The objective is to identify the principal parameters
involved in the formulations and discuss their roles also by means of sensitivity analysis. Following this, the
results of nonlinear numerical analyses, based on a three-dimensional finite element (FE) model, are presented.
The FE model was initially calibrated using experimental results from a benchmark beam lacking shear rein-
forcement, retrieved from the literature, which exhibited a brittle shear failure. Subsequently, several specimens
were prepared for the numerical investigation, assuming multiple combinations of geometrical and mechanical
parameters. For specimens experiencing shear failure, the strength was compared with that provided by code
formulas, as well as using the strut-and-tie approach for the cases characterized by a reduced shear span-to-depth
ratio. In general, these analytical tools significantly underestimated the numerical strength, underscoring the
necessity for further insights based on experimental tests. The numerical outcomes have been, indeed, prodromal
to design an experimental campaign.
1. Introduction without any transversal steel reinforcement, surely requests investiga-

tion. The mechanical behavior of RC elements without shear reinforce-
The recent collapses of reinforced concrete (RC) bridges in Italy, such ment is, indeed, not well-established in both design practices and
as the Polcevera bridge in Genova (2018) and the Caprigliola bridge in standard codes. Additionally, the scenario of concentrated loads applied
Massa Carrara (2020), have raised concerns regarding the safety near to the supports, a frequent load condition for bridge lanes (i.e.,
assessment of these structures under both typical and extreme loading vehicle footprint close to the support beam), may lead to premature
conditions. The growing attention to the vulnerability of existing bridges brittle shear failure of slabs due to development of strut-and-tie mech-
in Italy is underscored also by the recent release of guidelines for risk anisms, thus, further complicating the analysis.
classification, management, safety evaluation, and monitoring [1]. This Few studies on the shear behavior of RC slabs without shear rein-
document outlines specific procedures for managing existing bridges to forcement, involving experimental tests and finite element (FE) models,
prevent and monitor damage through a multilevel risk approach. This are available in literature [2–11].
approach facilitates the determination of required safety checks and In the study by Bui et al. [5,6], both experimental tests and numerical
periodic monitoring protocols. analyses were conducted, varying different parameters. The experi-
Within the context of vulnerability assessments for existing bridges, mental tests revealed that as the compressive strength of the concrete
the shear strength of RC slabs, commonly present in deck bridges and the maximum size of the aggregate increase, the shear strength also
* Correspondence to: Associate Professor of Structural Engineering Department of Engineering University of Naples Parthenope Centro Direzionale, isola C480143
Naples, Italy.
E-mail address: [email protected] (N. Caterino).
https://doi.org/10.1016/j.engstruct.2024.118634
Received 8 February 2024; Received in revised form 25 June 2024; Accepted 14 July 2024
Available online 28 July 2024
0141-0296/© 2024 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
N. Caterino et al. Engineering Structures 317 (2024) 118634
increases. Conversely, it decreases as the shear span-to-depth ratio into the failure modes.
(av/d) increases, where av is the distance from the applied load to the Starting with a critical examination of various international code
closest support, and d is the effective depth of the slab cross-section. The formulations governing the evaluation of shear strength in RC elements
numerical analyses provided a good prediction of the experimental without transverse reinforcement, this paper primarily aims to elucidate
behavior. the role of the key parameters inherent in these formulations.
In Nana et al. [7] a comparison between experimental test results on Codes have been used in the scientific literature to compare experi-
thick and thin slabs and predictions from different standard codes in mental results regarding the strength of materials in different contexts,
terms of shear strength is presented. The comparisons highlighted an such as [7,11,16]. However, a comprehensive and comparative analysis
underestimation of the experimental shear strength by the code of the shear resistance for elements without transverse reinforcement,
formulations. covering a wide variety of Italian, European and international standards
Concerning the underestimation of shear strength by code formula- codes, is still lacking. Some interesting research outcomes are available
tions, it is noteworthy that when av/d assumes low values, strut-and-tie for shear collapse due to punching mechanism [17], while less contri-
based resistant mechanisms may occur. Eurocode 2 [12] defines the butions extensively discuss of one-way shear crisis. The parameters
extension of the length, also known as ’discontinuity regions’ or ’D-re- compared include longitudinal steel reinforcement, effective depth of
gions,’ where these mechanisms may occur as equal to the height of the the cross-section, concrete and steel strength, and the distance between
beam. Both Eurocode 2 [12] and several codes allow for an increase in the applied load and the supports. The first objective of the paper is to
shear strength when 0.5 < av/d < 2. However, determining the threshold pinpoint the parameters that exert the most significant influence on
values for av/d ranges where shear strength is related to ’beam’ behavior shear strength and the predominant failure mode (whether due to
or a strut-and-tie mechanism is still a subject worthy of investigation. bending or shear).
In Lantsoght et al. [3], the effect of concentrated loads on the shear A total of 243 beams are generated by systematically varying the
behavior of RC slabs was investigated. The tested slabs were considered most influencing parameters, providing an extensive parametric anal-
as part of bridges since the load footprint (200 mm × 200 mm) simulated ysis by means of the FE model. Prior to conducting the parametric an-
a vehicular load. All tested slabs exhibited brittle failures, either due to alyses, the FE model is rigorously validated through a comparison with
punching or one-way shear. Experimental shear strength increased with experimental results obtained from an RC beam lacking transverse
slab width until reaching a threshold value, beyond which shear reinforcement, as documented in the literature [18]. This benchmark
strength did not increase further. beam exhibited a clear brittle failure mode attributed to shear crisis.
Code formulations for RC elements without shear reinforcement are In the final stage, for the simulated beams that, according to the FE
generally adapted for RC beams by defining an effective width, beff, as model, experienced shear failure, the numerical shear strengths are
suggested in literature [4–7] and some codes [13–15]. compared with those predicted by the previously scrutinized code for-
Based on the literature review, the main parameters influencing mulations. The comparison reveals, in several instances, a notable un-
shear strength in RC beams without transversal reinforcement can be derestimation. Additionally, a simplistic model based on the strut-and-
summarized as: tie mechanism within the shear length is considered for cases falling
within the ’D-regions’.
• concrete strength, including both compressive and tensile strength: The authors are aware that the numerical outcomes have to be
most empirically assessed shear strength models assume a power further validated by experimental tests investigating the same crucial
function dependence (1/2 or 1/3) on the compressive strength of parameters also in order to solve the evidenced gap between the
concrete, fcm; analytical and numerical results. To this aim the numerical outcomes
• longitudinal steel reinforcement percentage, ρs (for beams in two have been, indeed, prodromal to design a specific experimental
orthogonal directions): this not only influences the shear strength campaign on this topic.
due to the ’dowel effect’ in closing shear cracks but also affects the
likelihood of shear failure. A higher ρs increases the probability of a 2. The shear strength of rc elements without transverse
shift from a ’ductile’ mode, if the flexural strength is attained, to a reinforcement: a critical review of code formulations
’brittle’ mode if shear strength is attained;
• maximum aggregate size: increasing the maximum aggregate size Formulations governing the shear strength in RC beams without
enhances the roughness of the concrete surface along cracks, leading transverse reinforcement, as stipulated by various codes, rely on diverse
to an increased interlocking effect and a beneficial impact on shear parameters that account for multiple resistance mechanisms. These
strength; mechanisms include the interlocking effect of the aggregate along
• loaded area footprint: an increase in the loaded area footprint results cracks, dowel action from longitudinal steel bars, shear stresses in un-
in higher shear strength due to the larger collaborating width of the cracked concrete (involving concrete compressive strength), behavior
slab and the higher load diffusion achieved in the slab; of struts constrained in un-cracked concrete, level of compression or pre-
• boundary conditions: experimental tests [7] have shown that beams compression in the element, and the arching effect. Based on experi-
constrained on four sides are less resistant and characterized by a mental evidence, key parameters in predicting equations for shear ca-
more diffuse cracking pattern in all directions; pacity in RC elements without shear reinforcement include concrete
• shear length-to-effective depth ratio, av/d: As the distance of the load compressive strength (fck), aggregate size (dmax), longitudinal rein-
from the support (av) decreases, i.e., falls within the range of 0.5- forcement ratios (ρl), effective depth (d) of the cross-section, and the
2.0d, the shear strength increases due to the establishment of an ratio of shear length to effective depth (av/d).
’arch mechanism’ instead of a ’beam mechanism’ to sustain the However, the complexity of resistance mechanisms and the limited
applied loads. number of experimental tests, particularly for beams, result in a lack of
consolidated knowledge and understanding of the phenomenon.
In general, given the lack of both experimental and theoretical Currently, there is no unified and comprehensive theory on shear
studies on the shear strength of reinforced concrete (RC) elements, behavior, and various theories and models exist in the literature.
encompassing both beams and slabs without transverse reinforcement, Commonly used models include those by Regan [19] and Bresler and
further validations through experimental and numerical means should MacGregor [20], derived from empirical relationships based on exper-
be warranted. In this context, finite element (FE) structural analysis imental results, Vecchio and Collins [21] based on the Modified
emerges as a valuable tool to meticulously observe the trends in internal Compression Field Theory (MCFT), and Muttoni [22] based on the
stress and crack propagation, track their evolution, and gain insights Critical Shear Crack Theory (CSCT).
2
International codes adopt different approaches to consider influen- load footprint.

tial parameters, resulting in distinct calculation models. This paper
considers several codes, including Eurocode 2 [12], Italian Technical 2.2. American standard ACI 318 approach
Standards NTC [23], recent Italian guidelines for existing bridges [1],
two versions of the fib Model Code [14,15], American standard ACI 318 The American Standard ACI 318 [24] proposes the following semi-
[24], Swiss code SIA 262 [19], Canadian standard CSA A23.3.04 [26], –empirical equation for estimating the shear strength of RC elements
Japanese standard JSCE [27], British standard BS8110–1 [28], and without shear reinforcement:
Australian standard AS 3600 [29]. ( )
√̅̅̅̅̅ VEd • d √̅̅̅̅̅
The formulations provided by these codes are based on different VRd = 0, 16 • fck + 17 • ρl • • bw • d ≤ 0, 29 • fck • bw • d
MEd
theoretical or empirical approaches, as summarized in Table 1. Notably,
these codes are geographically associated with different continents, (2)
representing diverse approaches that can lead to significantly different
being VEd ed MEd the shear force and the bending moment acting in most
assessments of shear strength, as elucidated in the following.
stressed section, while the other parameters have the same meaning
previously introduced.
2.1. Eurocode 2 and NTC approach Eq. (2) is based on the experimental results of 194 beams tested by
Bresler and MacGregor [20] and for which the ratio VED⋅d/MED was < 1.
The Eurocode 2 [12] and the Italian Technical Standards NTC [23] It can be observed that Eq. (2) introduces, differently from Eq. (1), a
present identical equations for assessing the shear strength in RC ele- dependence not only on the ratio of the longitudinal reinforcement, ρl ,
ments without shear reinforcement: but also on the term VED⋅d/MED and, hence, it introduces a dependence
on the stress level in the longitudinal steel reinforcement. On the other
0.18 • k • (100 • ρl • fck )1/3 hand, the American code does not account for either the dimensional
VRd = max{[ + 0, 15 σcp ] • bw
γc effect or the size of the aggregate. The American code also offers a
• d; (vmin + 0, 15 σ cp ) • bw • d} simplified formulation that disregards the influence of the longitudinal
(1a) reinforcement, focusing solely on the dependence on the concrete
compressive strength and the dimensions of the section:
vmin = 0, 035 • k • fck 1/2 (1b) √̅̅̅̅̅
VRd = 0, 17 • fck • bw • d (3)
being: No indication is given by this code about the effective width in case
of slab.
• γc = 1.5, the partial safety factor of the concrete;
• σ cp = NEd /Ac (MPa) ≤ 0.2 fcd: the mean prestressing stress acting in 2.3. fib model code 2010 approach
the cross section Ac and due to the axial compressive load NEd, if
present; The previous version of the fib Model Code 2010 [14] introduces
• fck: the characteristic value of the cylindrical compressive strength of three levels of approximation (LoA) for assessing shear strength, pro-
the concrete (MPa); gressing from LoA I to LoA III with an increasing level of detail. LoA I and
• fcd = fck /γc: the design cylindrical compressive strength of the con- LoA III pertain to RC elements without transversal reinforcement, while
crete (MPa); LoA II addresses elements with dedicated steel reinforcement for shear.
• ρl = Abslw •d: the geometric ratio of the longitudinal steel reinforce- The general equation for evaluating the shear strength in RC ele-
ment, Asl; ments without shear reinforcement is as follows:
• d: the effective depth of the cross section (mm); √̅̅̅̅̅
• bw: the minimum width of the cross section (mm); fck
VRd = kv • • bw • d (4)
• k = 1 + (200/d)1/2 ≤ 2, a shape factor. γc
where d, bw, fck, and γc maintain the previously introduced meanings,

Both codes provide the option to account for the load distance from
while the definition of kv depends on the chosen approach (LoA I or LoA
the supports using the coefficient av/2d. Specifically, for av≤ 2d, the
II) and signifies the capacity of cracked concrete to convey tangential
effective design shear force can be reduced by av≤ 2d, resulting in an
stress through aggregate interlocking. LoA I is typically employed for the
enhancement of shear strength by 2d/av.
design of new RC elements when the compressive strength of concrete
However, neither code offers guidance on the effective width in the
does not surpass 70 MPa, the yielding stress of longitudinal reinforce-
case of beams. Conversely, the French Code [13] suggests employing an
ment (fyk) does not exceed 600 MPa, and the maximum aggregate size
effective width (beff), calculated as a ratio of bw by assuming a propa-
(dmax) exceeds 10 mm. Under these conditions, the parameter kv is:
gation of the shear stress along an inclination of 45◦ starting from the
180
kv = ≤ 0.15 (5a)
Table 1 (1000 + 1, 25 • d)
Methodologies employed by codes for evaluating shear strength in RC elements
The second level of approximation (LoA II) calculates the shear
lacking shear reinforcement.
strength based on a more accurate determination of kv, where, in addi-
MCFT- CSCT- Empirically
tion to the effective depth, a dependence on the maximum aggregate
based based derived
diameter, dmax, and on the mean longitudinal strain εx in the longitu-
EC2 [12] and NTC [23] √ dinal steel reinforcement is introduced, as follows:
fib MC2010 [14], fib MC2020 [15], √
AS 3600 [29] 0.40 1300
ACI 318 [24] kv = • Kdg
√
(1 + 1500 •εx ) (1000 + 0.7 • Kdg • d)
SIA 262 [25] √
MEd
DM 204 [1] √ 48 + VEd + 0.5 • NEd − Ap • fp0
A23.3.04 [26] = ≥ 1.15εx = d
√ 16 + dmax 2 • (Es • As + Ep • Ap )
JSCE [27] √
BS8110 − 1 [28] √ (5b)
3
with MEd, VEd, and NEd representing the bending moment, shear √̅̅̅̅̅
force, and axial force in the section of interest, respectively, and Ap, fp0, 0.3⋅ fck ⋅d • bw
VRd = (8)
and Ep representing the area, tensile stress, and Young’s modulus of the γ c ⋅(1 + 0.0022⋅d)
pre-stressed longitudinal reinforcement (if present), along with Es and As
representing the Young’s modulus and area of the ordinary longitudinal where all the parameters have the same meaning previously introduced.
reinforcement. When loads are applied within a distance variable in the Comparing Eq. (8) and Eq. (7c), it is evident that the term εv⋅kg from
range of d÷2d, i.e., d ≤ av ≤ 2d, the design shear force VEd can be SIA 262 [25], which is a function of steel strain and maximum aggregate
reduced by av/2d. This reduction can be interpreted as an increase in size, is simply replaced by 0.0022 in the DM 204 [1]. Fig. 1 illustrates the
shear strength by av/2d. variability of the parameter kd versus dmax for various values of εv ob-
In the case of beams subjected to concentrated loads, if the loading tained from Eq. (7b) and two values of d (500 and 200 mm, representing
area is less than the slab’s width, bw, the fib Model Code suggests using a beam and a slab, respectively). These curves are compared to the
an effective width beff. This effective width is a ratio of bw calculated by uniform values provided by DM 204 [1]. The kd curves from Eq. (7b)
assuming a propagation of the shear stress along an inclination of 60◦ assume fyd = 450 MPa, Es = 210 GPa, and three different bending
starting from the loading footprint. moment demand-to-capacity ratios MEd/MRd (0.25, 0.5, 0.75). As
MEd/MRd increases, εv also increases (from 0.5 ‰ to 1.6 ‰), and kd de-
creases. The values of kd from DM 204 are generally lower than those
2.4. fib model code 2020 and Australian code AS 3600 approaches calculated with Eq. (7b), indicating that the DM 204 [1] proposal is safer
for most cases compared to the SIA 262 [25] approach. However, for
The new fib Model Code 2020 [15] maintains the same approach as cases with the highest moment demand (MEd/MRd=0.75) and dmax less
given by Eq. (4) for evaluating shear strength. However, it modifies the than 20 mm, there is an exception to this trend. Notably, this code does
expression of kv for both levels of approximation, as follows: not provide guidance on the calculation of the effective width for beams.
LoA I:
0.40 1300 2.6. Canadian code A23.3.04 approach
kv = / • (6a)
(1 + 750 • fyd Es ) (1000 + 1.25 • d)
The Canadian code [26] provides the following equation for esti-
LoA II mating the shear strength in RC elements lacking transverse
0.40 1300 reinforcement:
kv = •
(1 + 1500 • εx ) (1000 + Kdg • d) √̅̅̅̅̅ 230 √̅̅̅̅̅
VRd =λ•β• fck • bw • dβ = fck ≤ 8 MPa (9)
( ) (1000 + d)
32 1 MEd 1
Kdg = ≥ 0.75εx = + VEd + NEd ≥0
16 + dmax 2 • Es • As d 2 where λ is a dimensionless factor designed to consider the concrete
(6b) density: it is 1 for normal density concrete, 0.85 for semi-low-density
concrete in which all the fine aggregate is natural sand, and 0.75 for
where fyd resents the design value of the yielding stress of steel (MPa), low-density concrete in which none of the fine aggregate is natural sand.
and the other parameters maintain the same meanings as introduced The other symbols have the same meanings as previously introduced.
previously. Additionally, the 2020 version considers the effect of the No guidance is provided regarding the calculation of the effective
load distance from the support through the factor av/2d, similar to the width in the case of a slab.
2010 version. The assumptions regarding the effective width beff, pro-
posed in the previous version [14], are also retained in [15]. The
Australian code AS 3600 [29] fully incorporates the formulation pro- 2.7. Japanese code JSCE approach
posed in Level of Approximation II of the fib Model Code 2020 [15].
The Japanese code JSCE [27] proposes the following equation for the
shear strength in RC elements without transversal reinforcement:
2.5. Swiss code SIA 262 and Italian guidelines DM 204 approach
VRd = βd • βp • fvcd • bw • d (10)
The Swiss code SIA 262 [25] propose the following equation for the
shear strength in RC elements without transversal reinforcement: where:
√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
Vrd = kd • τcd • d • bw (7) βd = 4 1000/d ≤ 1.5 (10a)
√ ̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
where the design shear stress τcd and the factor kd are: βp = 3
100 • ρl ≤ 1.5 (10b)
√̅̅̅̅̅
0.3 • fck √̅̅̅̅̅
τcd = (7a) fvcd = 0.2 • 3 fck ≤ 0.72 (10c)
γc
The symbols have the same meaning previously introduced. If the
1 f M 48 loads are applied at a distance from the supports variable in the range d-
kd = εv = yd • Ed kg = (7b)
1 + εv •d• kg Es MRd 16 + dmax 2d, i.e. d ≤av ≤2d, the acting design shear force VEd can be reduced by
av/2d, which can be interpreted as an increase of 2d/ av in terms of shear
being MRd the design flexural strength of the cross section and the
strength.
other parameters have the same meaning previously introduced.
No guidance is provided regarding the calculation of the effective
Substituting the expressions 7a e 7b in Eq. (7), the following equation is
width in the case of a slab.
obtained:
√̅̅̅̅̅
0.3⋅ fck ⋅d • bw
VRd = (7c) 2.8. British code BS8110-1 approach
γ c ⋅(1 + εv • d • kg )
A formulation similar to Eq. (7c), but with some simplifications, has The British code BS8110–1 [28] proposes the following equation for
been recently proposed by the Italian guidelines [1], as follows: the shear strength in RC elements without transversal reinforcement:
4
Fig. 1. Variability of kd (Eq. 7b) vs. dmax for different values of εv : (a) d= 500 mm, (b) d= 200 mm. Comparison of outcomes from SIA 262 [25] and DM 204 [1].
( )1 ( )1 ( )1 (Eq. 5b) and 2020 [15] (Eq. 6b), and the SIA 262 [25], i.e., Eq. (7).
As 3 400 4 fck 3
VRd = 0.79 100 • • • • bw • d (11) The longitudinal reinforcement ratio, ρl, is considered in several
bw • d d 25
codes: EC2 [12] and NTC [23] (Eq. 1a), JSCE [27], BS8110–1 [28] (Eq.
11), and Eq. (2) of ACI 318 [24].
with the following limitations:
The bending moment acting on the element is considered in ACI
( )1 ( )1 318–14 [24] (Eq. 3), in the LoA II of both fib Model Code versions [14,
As 3 fck 3
100 • ≤ 3and ≥1 (11b) 15] (Eq. 5b) and (Eq. 6b), in SIA 262 [25] (Eq. 7), and in ACI 318 [24]
bw • d 25
(Eq. 2). The effect of the shear force is considered in ACI 318 [24] (Eq.
The other symbols have the same meaning previously introduced. If 3), in the LoA II of both fib Model Code versions (Eq. 5b) and (Eq. 6b)
the loads are applied at a distance from the supports variable in the [14,15].
range d-2d, i.e. d ≤av ≤2d, the acting design shear force VEd can be In Fig. 2, the sensibility of the code formulations to the most sig-
reduced as already described in Section 2.7. No guidance is provided nificant parameters, i.e. fck, fy, As, d, is plotted. Along the x axis, the
regarding the calculation of the effective width in the case of a slab. variable parameters are normalized to the minimum value assumed in
the range, i.e. 10 MPa for fck, 300 MPa for fy, 200 mm2 for As, 100 mm
for d and, on the y axis, the shear strength given by each code is
2.9. Comparison among code formulations
normalized to the strength calculated assuming the minimum value of
the variable parameter. Each graph refers to the variation of only one
As previously noted, the codes propose different formulations to
parameter at time.
evaluate the shear strength of RC elements without shear reinforcement,
Fig. 2a shows that as the compressive strength of concrete increases,
and the considered parameters are not always the same or are not taken
the shear strength increases too; in particular, the British code BS8110–1
into account in the same way. In order to provide a synthesis, Table 2
[28], provides the lowest increasing and until fck < 25 MPa there is no
lists the parameters considered in all the examined codes.
effect of fck on the shear strength. The codes ACI 318 [24] (Eq. 3), DM
All the codes consider the influence of the effective depth, d, and the
204 [1], SIA 262 [25], LoA I and LoA III of fib MC2020 [15] and MC2010
width, bw, of the cross section, as well as the characteristic value of the
[14], A23.3–04 [26] gives the same trend and the highest effect of fck on
concrete compressive strength, fck.
the shear strength. ACI 318 [24] (Eq. 2), EC2 [12] and NTC [23] and
The maximum diameter of aggregate dmax is considered only in a few
JSCE [27] provide an intermediate trend.
approaches: the LoA II in both fib Model Code versions, i.e., 2010 [14]
Table 2
Shear strength calculation parameters in codes for unreinforced concrete elements.
Par. EC2 ACI 318-14 fib MC 2010 fib MC 2020 SIA 262 DM 204 A23-3 JSCE BS8110-1
NTC Eq. 2 Eq. 3 LoA I LoA II LoA I LoA II AS 3600
d √ √ √ √ √ √ √ √ √ √ √ √
bw √ √ √ √ √ √ √ √ √ √ √ √
ρl √ √ √ √
fck √ √ √ √ √ √ √ √ √ √ √ √
fyk √ √
Ved √ √ √
Med √ √ √ √
Ned √ √
Es √ √ √ √
As √ √
d,max √ √ √
av/d √ √ √ √ √ √ √ √ √
5
5 1.25
EC2 - JSCE2007
4.5 ACI 318-14 (Eq. 2)
ACI318-14 (Eq. 3) - MC10 LoA I - MC10 LoA III - MC20 1

4 LoA I - MC20 LoA II - SIA 262 - DM204 - A23.3
Vr,CODE/Vr, CODE, 10 MPa
BS 8110 - 1
Vr,CODE /Vr, CODE, 300 MPa

3.5
0.75
3
2.5 0.5
MC 2020 LoA I
(Eq. 4 e 6a)
2
SIA 262 (Eq. 7)
0.25
1.5
1 0
1 2 3 4 5 1 1.25 1.5 1.75 2
fck/10 MPa fy/300 MPa
a) b)
5 8
EC2 (Eq. 1)
EC2 - BS8110-1
ACI 318-14 (Eq. 2)
ACI 318-14 (Eq. 2) 7 ACI 318-14 (Eq.3)

MC 2010 LoA I (Eq. 4 e 5a)
Vr,CODE /Vr,CODE,200 mm2
4 MC10 LoA III - MC20 LoA II MC 2010 LoA III (Eq. 4 e 5b)
Vr, CODE/Vr,CODE,100 mm
6 MC 2020 LoA I (Eq. 4 e 6a)

JSCE2007 MC 2020 LoA II (Eq. 4 e 6b)
SIA 262 (Eq. 7)
5 DM 204 (Eq. 8)
3 A23.3 (Eq. 9)
JSCE 2007 (Eq. 10)
4 BS 8110- 1 (Eq. 11)
2
3
1 1
1 2 3 4 5 1 2 3 4 5
As/200 mm2 d/100
c) d)
Fig. 2. : Normalized strenght given by several codes vs. several normalized parameters: a) fck/10 MPa; b) fy/300 MPa; c) As/200 mm2; d) d/100 mm.
The presented data in Table 1 reveals that only SIA 262 [25] and LoA 3. Validation of the finite element model through comparative
I of fib MC2020 [15] account for the influence of fy on shear strength, as analysis with an experimental case study
illustrated in Fig. 2b. Both codes exhibit a decreasing trend with fy, with
fib MC2020 [15] displaying a more pronounced effect. In Fig. 2c, three A finite element (FE) model has been developed to analyze the
distinct increasing trends in shear strength with As are evident. The most nonlinear behavior of reinforced concrete (RC) elements lacking shear
significant trend is observed in LoA I and LoA III of fib MC2010 [14], reinforcement when subjected to concentrated vertical loads. The model
while the least prominent is found in Eq. 2 of ACI 318 [24]. EC2 [12], was initially validated through comparisons between numerical results
NTC [23], BS8110–1 [28], and JSCE [27] exhibit similar trends, slightly and the experimental behavior of a case study found in the literature,
surpassing that of ACI 318 [24]. specifically addressing an RC beam without shear reinforcement that
Finally, Fig. 2d illustrates distinct trends in shear strength concern- experienced shear failure.
ing the effective depth, d, of the cross-section within the considered code
formulations. Specifically, the most substantial increase is offered by
LoA II of fib MC2010 [14], while the smallest increase is observed in 3.1. The case study
BS8110–1 [28]. The pronounced dependence of shear strength on d in
LoA II of fib MC2010 [14] is mitigated in the 2020 version [15]. The RC beam, not reinforced against shear, selected as a benchmark
It’s worth noting that, for a few codes (LoA II of fib MC2010 [14], ACI (BN50D beam) was subjected to bending in a previous study [18],
318 [24], Eq. 2), shear strength increases with d more than propor- exhibiting a brittle failure attributed to a critical shear crack. The lon-
tionally. In contrast, for the other codes, the trend is less than propor- gitudinal steel reinforcement, in this case, remained within the elastic
tional, resulting in potentially significant variations in predictions. range until failure. The loading scheme and geometry of the BN50D
beam are illustrated in Fig. 3.
The mechanical properties provided by the authors were: charac-
teristic value of the cylindrical compressive strength of the concrete fck
6
150 mm 2700 mm 150 mm
(a) (b)
Fig. 3. : The benchmark beam BN50D [9]: a) loading scheme; b) cross section.
= 37 MPa, average yielding strength of the steel bars fy = 490 MPa. The assuming a perfect bond with the surrounding concrete.
steel reinforcement consisted into 2 bars with diameter 20 mm and one Boundary conditions are applied by constraining the nodes of the
bar with diameter 25 mm at the bottom side, while further 5 layers of first support solely along the vertical direction (Z-axis). In the symmetry
two bars with diameter 10 mm were placed at different height (Fig. 3b). section corresponding to the mid-span of the actual beam, nodes are
The longitudinal reinforcement ratio in tension (calculated with refer- constrained in both vertical and horizontal directions (Z and X axes) to
ence to the bottom layer and the first three layers of the 10 mm bars, replicate symmetry conditions.
which were actually loaded in tension, according to the numerical an- Determining the optimal mesh size involves finding a balance be-
alyses will be discussed later, Section 3.4) is ρl = 1.2 %. tween result accuracy and computational efficiency. Consequently, the
The beam was simply supported at both ends and was loaded by a concrete beam is discretized with a three-dimensional mesh (side length
concentrated load in the middle; thus, the shear span-to-depth ratio was 25 mm) composed of eight-node hexahedron elements, while the steel
av = 1350 mm, i.e. av/d = 3. For a shear force value of V= 162 kN, bars are discretized using two-node bar elements with a 10 mm mesh.
corresponding to half of the applied load at the mid-span, the beam
experienced a brittle failure caused by the abrupt opening of a critical 3.2.2. Materials
inclined crack near the left support. This crack propagated until reach- The concrete is simulated using a Smeared Crack Model, a total strain
ing the point of load application. Following the formation of this crack, crack model that offers various stress-strain relationships for both ten-
the beam exhibited no residual strength capacity. sion and compression behavior. The nonlinear compression behavior of
concrete is replicated using the model proposed by Thorenfeldt et al.
[31] (Fig. 5a), while the tension behavior is modeled according to
3.2. The FE model
Hordijk’s proposal [32] (Fig. 5b). Moreover, for the tensile behavior of
concrete, the softening behavior is assumed to be a function of the ratio
3.2.1. Geometry
G’f / h, where G’ is the fracture energy in tension, which depends on the
The finite element (FE) model of beam BN50D is created using
compressive strength of concrete fcm according to the Eq. (12) and h is
MIDAS FEA NX software [30]. Taking advantage of the symmetry in the
the total crack band width (suggested to be assumed as the mesh size of
beam and loading pattern, only half of the beam’s geometry is modeled
the concrete).
to reduce computational demands (see Fig. 4). To enhance the model’s
realism, stiff steel elements are employed to represent the steel plates Gf = 73 • fcm 0.18 (12)
used as supports at the ends and the plate where the vertical load was
For the reinforcement steel bars, a uniaxial elastic-plastic behavior is
applied during the test. The internal steel bars are individually modeled,
Fig. 4. : FEM model of BN50D beam: (a) 3D view, (b) detail of the cross section.
7
Fig. 5. : The stress-strain behavior assumed for concrete and steel.
adopted, both in tension and compression, utilizing the Von Mises [33] 168 kN, very close (+4 %) to the experimental value (162 kN). Gener-
yielding criteria. The yielding strength is assumed to be equal to the ally, the numerical curve closely follows the experimental trend,
experimental value, i.e., 490 MPa, and the Young’s modulus is set at particularly before the first cracking, although the numerical behavior is
200 GPa. stiffer than the experimental one at lower loading levels. The two curves
converge as concrete cracking intensifies. The increased stiffness of the
numerical curve may be attributed to the assumption of perfect bonding
3.3. Comparison of numerical and experimental results
between the steel bars and concrete. In reality, slips along the interfaces
contribute to the experimental behavior being more deformable. Addi-
The experimental test conducted on the BN50D beam is replicated
tionally, the beam may have been affected by shrinkage micro-cracks,
through a nonlinear analysis in displacement control. The vertical load
further enhancing its deformability and influencing the experimental
is applied to the nodes of the symmetry cross-section by applying in-
shear-deflection relation.
cremental displacements. The ’full Newton-Raphson’ iterative method is
The flexural capacity of the beam is calculated according to the usual
chosen for the implicit solution strategy, with a maximum displacement
assumptions for RC sections and assuming the average experimental
set at 8 mm over 100 steps.
values of the materials properties (i.e., fcm = 37 MPa, and fy = 490 MPa).
Fig. 6 illustrates the comparison between the experimental and nu-
It results MRd = 290 kNm and this value is used to calculate the shear,
merical results in terms of shear vs. mid-span displacement. The
Vpl, corresponding to the attainment of MRd at the mid-span section (i.e.,
experimental failure of the beam occurred at a shear force Vu,
Vpl = 2•MRd/L = 215 kN). Since such a value is greater than Vu,exp
exp= 162 kN. Up to this point, the beam did not exhibit any signs of
(162 kN), it is evidenced that the beam was designed to induce a pre-
imminent failure. After the development of the shear crack, it demon-
mature crisis due to shear, rather than a flexural collapse.
strated no residual capacity. The experimental shear force vs. mid-span
The comparison between experimental and numerical results is
displacement relation indicates a brittle failure in terms of global
presented also in terms of principal tension strain in concrete, crack
behavior, with a noticeable sudden drop in shear evident in Fig. 6 at a
pattern and status of crack opening for three values of the shear V in
displacement of approximately 6 mm.
Fig. 7 (77 %, 93 % and 100 % of the maximum experimental shear, Vu,
The numerical model depicts a linear elastic behavior up to a shear
exp).
force of about 52 kN, beyond which the first flexural cracks in the
The numerical distribution of the principal tension strain in the
concrete begin to appear. As the load increases, the flexural cracks
concrete (Figs. 7a, 7b, 7c) is consistent with the position of the experi-
become progressively more inclined until a critical shear crack forms,
mental cracks (see the left side of the beams plotted in Fig. 7) and
resulting in a significant reduction in strength at the same experimental
highlights the direction of tension and, thus, the direction of the opening
displacement of around 6 mm. This load drop occurs for a shear force of
cracks. The cyan represents a tensile strain level of 4 ‰, whereas the
light green correspond to 1 %. The tensile strain corresponding to
attaining the tensile strength of concrete ft, i.e. εt = 0.01 %, is also
indicated in the legend.
Similarly, it should be noted that the numerical crack pattern
(Figs. 7d, 7e, 7f), is in very good agreement with the experimental one.
When V = 125 kN, i.e. 77 % of Vu,exp (Fig. 7d), the cracks are the typical
flexural ones, i.e. sub-vertical. At V = 150 kN, i.e. 92 % of Vu,exp
(Fig. 7e), the most external cracks begin to become inclined, until at V
= Vu,exp (Fig. 7f), suddenly, an inclined crack forms from the support to
the load application point, confirming the brittle failure experimentally
observed.
Figs. 7d, 7e and 7f report also the opening level of cracks, provided
by the FE model by means of different colours. In particular, the blue
represents ‘partially open cracks in loading phase’, the magenta indicates
‘partially open cracks in unloading phase’, the red represents ‘fully open
cracks’, and the green represents ‘crack formed and then closed’ due to the
Fig. 6. : Comparison of numerical curve and experimental curve for the evolutionary redistribution of the stresses during the loading process. It
BN50D beam.
8
Fig. 7. Experimental vs. numerical results for three values of shear (125 kN, 150 kN, 162 kN): (a), (b) and (c) experimental crack pattern and numerical distribution
of main tensile strains in the concrete; (d), (e) and (f) experimental crack pattern and numerical prediction of crack opening.
can be noted that at V = Vu,exp (Fig. 7f), the shear crack appears fully the closest estimates to the experimental result, with an average un-
open. derestimation of about 13 %. Another group of formulations, including
Finally, Figs. 8a and 8b depict the stress and strain distribution in the LoA II of both fib MC2010 [14] and fib MC2020 [15], SIA 262 [25], DM
steel reinforcement at the point of beam failure. It can be observed that 204 [1], JSCE [27], BS8110–1 [28], provides similar predictions,
the maximum tensile stress in the bars is 393 MPa, which is lower than ranging from 116–124 kN, with an average underestimation of 26 %.
the average yielding stress (490 MPa). This confirms that the longitu- Lastly, predictions from LoA I of both fib MC2010 [14] and MC2020 [15]
dinal reinforcement remains in the elastic field when the shear failure of exhibit the highest underestimation, with values ranging from
the beam occurs. All the presented comparisons emphasize that the 95–96 kN, approximately 41 % lower than the experimental value. This
numerical approach implemented in the FE model is reliable and can conservative prediction aligns with the intended use of LoA I for pre-
serve as a valid numerical tool for investigating the shear behavior of RC liminary design of structural elements, where a safer calculation method
beams. is acceptable.
In conclusion, for RC beams with significant depth, such as the
benchmark beam analyzed in this study, certain code formulations offer
3.4. Comparisons of numerical and experimental shear strength with code
reliable predictions of experimental results. The last section of the paper
provisions for the benchmark beam
will address the lack of experimental information regarding the behavior
of shallow beams and beams, evaluating the reliability of code pre-
The experimental and numerical results related to the benchmark
dictions on these types of elements using the numerical analyses
beam are compared with the predictions given by the code formulations
described in Section 4 as benchmarks.
presented in Section 2. The values of the shear strength provided by all
the approaches, Vu,th, are listed in Table 3 together with the percentage
4. Parametric analysis of RC elements
difference with the experimental value, ΔV, and plotted in Fig. 9 in
comparison with the experimental value of the shear strength. All the
4.1. Database generation
values are calculated assuming the partial safety factor equal to 1 (γ c =
1).
The non–linear FE model introduced and validated in the previous
The formulations of ACI 318 [24] (Eq. 2) and of LoA II of fib Model
section is used to carry out an extensive campaign of numerical analyses
Code [14] (Eqs. 5b and 6b) depend on the applied bending moment,
that concerned 243 RC beams without shear reinforcement. The set of
which is calculated at the mid-span section. In the SIA 262 [25]
investigated beams is created by changing the mean concrete
formulation (Eq. 7), the flexural capacity previously calculated, MRd
compressive strength, fcm, the mean yielding stress of the steel bars, fy,
= 290 kNm, is introduced.
the beam height, H, the shear span-to-effective depth ratio, av/d, and the
The shear strength given by the codes EC2 [12], MC LoA II [14] and
diameter of the steel bars, ϕ, in the following ranges:
Eq. (2) of ACI 318 [24] are calculated assuming the reinforcement ratio
according to Section 3.1 (1.2 %), considering the number of bars in
- fcm: 20-25-28 MPa;
tension, actually below the neutral axis, as they result from the nu-
- fy: 350-450-550 MPa;
merical analysis (Fig. 8b).
- H: 150-200-250 mm;
All code formulations consistently underestimate the experimental
- av/d: 1.0-1.5-2.0;
value, with percentages ranging from 12 % to 41 %. Specifically, Eq. 1
- ϕː 10-14-18 mm.
from EC2 [12] and NTC [23] and Eqs. 2 and 3 from ACI 318 [24] yield
9
Fig. 8. : Loading conditions in the steel bars at the beam failure: (a) axial stress; (b) axial strain.
The 243 cases are obtained by varying one parameter at a time and
Table 3
keeping the others unchanged. Conversely, the following parameters
Shear strengths provided by code formulations for the benchmark beam.
and assumptions are considered fixed for all specimens:
Codes VRd ΔV the full length of the slab, L = 1.56 m, and the spacing between
[kN] [%]
supports, Ls = 1.16 m,
EC2 [12] and NTC[23], Eq. (1) 142 − 12.3
LoA I, fib MC2010 [14], Eq. (5a) 95 41.4
−
- the width of the slab: b = 400 mm,
LoA II, fib MC 2010 [14], Eq. (5b) 120 − 25.9
LoA I, fib MC 2020 [15], Eq. (6a) 96 − 40.7 - the concrete cover: c = 20 mm,
LoA II, fib MC 2020 [15], Eq. (6b) 120 − 25.9 - the Young’s modulus of steel, Es = 210 GPa,
ACI 318 [24], Eq. (3) 139 − 14.2 - the symmetrical reinforcement lay-out made of 4 steel bars at the
ACI 318 [24], Eq. (2) 140 − 13.6 bottom and 4 steel bars at the top,
SIA 262 [25], Eq. (7) 117 28.8
- absence of transversal reinforcement.
−
DM 204 [1], Eq. (8) 124 − 23.5
A23.3 − 04 [26], Eq. (9) 130 − 19.8
JSCE[27],Eq. (10) 116 − 28.4 The tensile strength and the Young’s modulus of the concrete are
BS8110 − 1 [28], Eq. (11) 124 − 23.5 obtained on the basis of the value of fcm by means of the correlations
provided in [12].
10
Fig. 9. : a) Mesh of concrete slab; b) Mesh of the steel bars.
4.2. Non–linear modelling with MIDAS FEA NX distinct cross sections considered for the beams are depicted in Fig. 10b.
The nonlinear analyses are conducted under displacement control by
The FE model for each of the 243 elements is constructed in a manner monitoring the point of force application on the upper surface of the
similar to the approach used for the benchmark beam using MIDAS FEA slab. The ’full Newton-Raphson’ iterative method is chosen to imple-
NX software [30]. Specifically, the concrete is represented by a ment the implicit solution strategy, imposing a maximum displacement
three-dimensional mesh (side length of 25 mm) composed of eight-node of 10 mm over 75 steps.
hexahedron elements, while the steel bars are discretized using The materials are modelled according to the same approaches dis-
two-node bar elements with a mesh resolution of 10 mm (see Fig. 9). cussed in Section 3.2.2. In Figs. 11a and 11b the compressive and the
The beams are subjected to a concentrated vertical force and are tension constitutive laws of the concrete are plotted considering the
supported at their ends by two one-sided restraints comprising steel variability of fcm.
plates (width 100 mm) positioned 100 mm from the beam ends (see Based on the previously described material properties and constitu-
Fig. 10a). These steel plates are connected to the beam through elastic tive laws, the analyses for each beam under the concentrated vertical
springs capable of undergoing compression only, possessing an axial load are conducted until one of the two following conditions is met:
stiffness of 100 MN/mm, sufficiently high to emulate an ’infinitely rigid’
constraint. These springs enable the consideration of the flexural • ’ductile’ failure: the flexural strength is reached in the most stressed
deformation effect in the slab. The vertical force is applied at a distance section, causing yielding of the inferior steel bars, and the maximum
av from the inner edge of the left support (Fig. 10a) and is uniformly strain in the compressed concrete reaches 7 ‰. It was verified that
distributed across the entire width b of the beam cross section. The three after this value, further advancements in the numerical analysis do
Fig. 10. : Geometry for the simulated beams (dimensions in mm): a) load position with av/d= 1, av/d= 1.5, av/d= 2; b) different dimensions of the cross section (H
= 150–200-250 mm, B = 400 mm).
11
Fig. 11. : a) Behavior of concrete in compression (Thorenfeldt et al. [31]); b) Behaviour of steel in tension (Hordijk [32]).
not influence shear-displacement curves significantly (the curves 1) Vpl > VFEM,max: the shear force Vpl corresponding to the theoretical
continue according to a constant trend associated with the adopted flexural strength of the section is not reached in the numerical
elastic-plastic behavior for the steel). simulation, meaning, thus, that, according to the FE model, a shear
• ’brittle’ failure: the shear strength is attained in the most stressed failure should occur; this should be confirmed by a brittle behaviour
section, usually with the inferior steel bars still in the elastic field. In after the peak load in the numerical shear-displacement curves;
this case, the limit condition of the analysis is represented by the 2) Vpl ≤ VFEM,max: the shear force Vpl corresponding to the theoretical
triggering of shear strength, resulting in a subsequent brittle reduc- flexural strength of the section is lower than the maximum shear
tion of strength. force provided by the FE model, meaning, thus, that a flexural crisis
occurs and that the flexural strength is overestimated by the FE
The global behavior of each beam is depicted in terms of shear- model; this should be confirmed by a ductile behaviour after the peak
displacement curves, where the shear is the sum of the reactions of load in the numerical shear-displacement curves.
the support (springs) closest to the loaded section, and the displace-
ments are measured at the midpoint of the bottom edge of the loaded The relative magnitude of Vpl and VFEM,max in the experience of the
section. authors actually lead to correctly predict the collapse mode in most
To determine if shear failure can precede flexural failure, the shear cases. Some exceptions however may occur when Vpl and VFEM,max as-
force corresponding to the theoretical flexural strength of the section, sume values very similar one each other. In this study, such cases con-
MRd, is calculated as Vpl = MRd/av for each slab. Here, MRd is calculated cerned only 6 % of the samples, where a brittle crisis has been observed
according to the usual assumptions for RC sections, assuming mean in the FEM model, even if Vpl was greater than VFEM,max. However, in this
values for material properties. The shear Vpl is then compared with the sub-set of samples, Vpl to VFEM,max ratio was always around one (spe-
maximum shear force resulting from the numerical analysis, VFEM,max, to cifically in the range 1.0–1-1).
establish the following scenarios: In Fig. 12 some examples of shear-displacement curves obtained by
MIDAS FEA NX are plotted with reference to three different types of
output and consistency checks.
300 300 300 VFEM, max
Vpl Vpl
250 250 250
VFEM, max
200 200 VFEM, max 200
Shear [kN]
Shear [kN]
Shear [kN]
150 150 150

Vpl
100 100 100
50 50 50
0 0 0
0 2.5 5 7. 0 2.5 5 7. 0 2.5 5 7.5
Displacement [mm] Displacement [mm] Displacement [mm]
a) b) c)
Fig. 12. : Typical numerical shear - displacement curves: a) case 1: Vpl > VFEM,max, shear failure; b) case 2: Vpl ≤ VFEM,max, flexural failure; c) case 3: Vpl ≤ VFEM,max,
shear failure.
12
Fig. 13. : Shear - displacement curves for beams with av/d= 1.0.
Fig. 13a shows an example of numerical curve where the brittle strength, indicating shear failure in the FE model. In this scenario, the
failure of the beam is clearly evident and, for this case, it is confirmed by maximum stress recorded in the tensioned steel bars is 435 MPa, close to
the analytical calculations too, since Vpl > VFEM,max, i.e. the shear value the yielding stress (450 MPa). This represents one of the simulated cases
(277 kN) corresponding to flexural crisis of the beam is not reached in where shear values corresponding to both flexural and shear crises of the
the FE model. The maximum numerical shear force is, indeed, 212 kN section are very similar, and hence, the difference in numerical and
and, at this stage, the maximum compressive strain in the concrete is analytical predictions may not be indicative of the failure mode occur-
only εc = 2.2 ‰, the steel reinforcement in tension is not yielded (about ring in the FE model. For the case presented in Fig. 12c, the maximum
201 MPa), and the displacement is 2.1 mm. Moreover, it can be noted compressive strain in the concrete at failure is only εc = 2.0 ‰, and the
that the numerical curve shows a clear change of slope due to the con- deflection at this stage is around 2.6 mm.
crete cracking.
Fig. 13b illustrates a numerical curve exemplifying ductile behavior. 4.3. Discussion of the numerical results
In this instance, the maximum shear force provided by the FE model
(177 kN) is slightly higher than the analytical value corresponding to In Figs. 13, 14, and 15, the numerical non-linear shear-displacement
flexural failure (Vpl = 162 kN), indicating potential yielding of the steel curves for all simulated beams are presented, organized based on the av/
bars in tension. This is supported by the recorded strains in the tensioned d ratio values, i.e., 1.0, 1.5, and 2.0. Each figure groups cases with the
steel bars and the plastic behavior exhibited by the beam post-peak. The same av/d ratio into columns and rows corresponding to the same height
analysis for this case is carried through until the maximum compressive (H = 150, 200, 250 mm) and yielding stress (fy = 350, 450, 550 MPa).
strain reaches 7 ‰, and the deflection at this stage is approximately Additionally, within each graph, curves of the same color (red, yel-
7.1 mm. low, green) represent beams with the same steel bar diameter (ϕ=10,
Lastly, Fig. 13c depicts the numerical curve of a beam where the 14, 18 mm). The three curves for each color correspond to beams with
maximum shear force from the FE model (303 kN) slightly exceeds the varying concrete compressive strength (fcm = 20, 25, 28 MPa).
analytical value corresponding to flexural failure (Vpl = 285 kN). Out of the 243 analyzed models, 75 beams exhibit brittle failure due
However, post-peak, the numerical curve displays a brittle reduction in to shear, with shear-displacement curves resembling the case depicted in
13
Fig. 12a. Conversely, the remaining beams experience flexural failure av/d = 1.5, the shear failure occurs when the height of the cross section
with ductile behavior post-peak, as evidenced by shear-displacement assumes the lowest value.
curves akin to the scenario illustrated in Fig. 12b. Finally, the compressive strength of the concrete plays a marginal
The Figs. 13, 14 and 15 evidence that the strength and the amount of role, since it is never able to change significantly the shear-displacement
the longitudinal steel reinforcement ρl play an important role, since the curves, with exception of some cases with av/d = 2.0, fy = 550 MPa, and
higher they are the lower is the possibility of attaining a flexural failure As = 4 ϕ18 (Fig. 14).
in the beams, due to the increasing of the flexural strength. It is worth Figs. 16 and 17 present detailed results for two beams manifesting
noting that for the beams reinforced with 4 ϕ10 bars, i.e. characterized distinct failure modes: a ductile flexural crisis for the first (Fig. 16) and a
by a reinforcement percentage ρl variable in the range 0.31–0.52 % brittle shear failure for the second one (Fig. 17). Specifically, for the load
according to the section height, the shear failure is never attained for step corresponding to the maximum shear value, the figures illustrate
whatever value of yielding stress and av/d. On the contrary, when the the distribution of principal tensile stresses in the concrete, the cracking
beams are more strongly reinforced, i.e. with 4 ϕ18 bars that correspond pattern indicating the opening condition of the cracks, the distribution
to have ρl variable in the range 1.02–1.70 %, the shear failure occurs in of vertical displacements, and the stress distribution in the steel bars.
all cases with exception of the cases with fy = 350 MPa and av/d = 1.5 or The results of Fig. 16 refer to a beam with H = 150 mm, av/d = 1,
2. fc= 20 MPa; fy= 350 MPa, As = 4ϕ10 mm. For this case, the principal
In case of av/d equal to 2 (Fig. 13), the beams have a ductile flexural tensile stresses are more concentrated at the section where the load is
failure for all the cases examined (with exception of the cases with fy = applied (Fig. 16a) and where an evident flexural crack forms followed by
550 MPa and As = 4 ϕ18), regardless of the bar diameter and height of a widespread cracking pattern around this section (Fig. 16b). Moreover,
the section. in this section and in the surrounding ones, the steel bars result yielded
In case of av/d = 1.0, the shear failure is attained only for the highest (Fig. 16c), evidencing, thus, a diffuse plasticization. The shape of the
values of the area or of the yielding stress of the steel reinforcement, vertical displacements is that common for flexural type collapse
while the height of the cross section is not influent. On the contrary, for (Fig. 16d).
14
Fig. 16. : Detailed results for the beam with H = 150 mm, av/d = 1, fcm = 20 MPa; fy = 350 MPa, ϕ = 10 mm at failure: (a) distribution of principal tensile stresses;
(b) cracking pattern status; (c) stress distribution in the steel bars; (d) vertical displacements distribution.
15
Fig. 17. : Detailed results for the beam with H = 250 mm, av/d = 1, fcm = 28 MPa; fy = 550 MPa, ϕ = 18 mm, at failure: (a) distribution of principal tensile stresses;
(b) cracking pattern status; (c) stress distribution in the steel bars; (d) vertical displacements distribution.
The beam in Fig. 17 is characterized by H = 250 mm, av/d = 1, fc parameters, while the curve with red dots refer to the beams with fy
= 28 MPa; fy = 550 MPa, As = 4ϕ18. In this scenario, the principal = 550 MPa and fy = 350 MPa (steel strength ratio 1.57).
tensile stresses at failure (Fig. 17a) are distributed along the diagonal In Fig. 18c, the curve with black dots represents the ratio of the
direction connecting the loading point to the nearest support (the left values of VFEM,Max, for the beams with d = 180 mm and d = 130 mm
one). This distribution corresponds to the formation and subsequent (effective depth ratio 1.33), being constant the other parameters, while
opening of a distinct diagonal crack (Fig. 17b). The longitudinal steel the curve with red dots refer to the beams with d = 180 mm and with
bars do not yield in any section of the beam (Fig. 17c), confirming that d = 130 mm (effective depth ratio 1.67).
the flexural strength of the cross section is not reached; instead, a shear In Fig. 18d only one curve is plotted and represents the ratio of the
failure occurs. This is further underscored by the shape of the vertical values of VFEM,Max for the beams reinforced with As = 4ϕ18 and As
displacement (Fig. 17c), which is characterized by a noticeable distor- = 4ϕ14 (steel area ratio 1.65), being constant the other parameters. It is
tion in the shear span length. worth noting that all the beams equipped with ϕ10 bars exhibit a flex-
ural failure and, thus, do not fall within the 75 cases examined.
5. Sensitivity of shear strength to geometrical and mechanical In Fig. 18e, the curve with black dots represents the ratio of the
parameters in the FE model values of VFEM,Max for the beams loaded under the condition av/d = 1
and av/d = 2.0 (shear span ratio 0.75), being constant the other pa-
Basing on the results of the numerical analyses, a summary of the rameters, while the curve with red dots refers to the beams with av/
influence of the different parameters assumed variable in the definition d = 1.5 and av/d = 2.0 (shear span ratio 0.50).
of the database (i.e., fcm, fy, H, As, av/d) on the maximum shear is herein Finally, Table 5 presents a summary of the results depicted in Fig. 18,
reported. The discussion refers to the 75 beams that exhibit a shear listing the minimum, maximum, and average percentage variation of
failure. Table 4 summarizes how many beams among the 75 cases are shear strength for each parameter change. It is important to note that, as
characterized by different couples of parameters. It can be observed that the comparisons of the maximum shear are conducted for pairs of beams
about 57 % of beams failed in shear are characterized by av/d = 1.0, characterized by the same values of geometrical and mechanical pa-
35 % by av/d = 1.5, and only 8 % by av/d = 2, evidencing, thus, that low rameters except one, the number of analyzed beams in the comparisons
values of parameter av/d (1.0–1.5) shift the failure versus a shear crisis. is lower than 75 and varies based on the parameter assumed to be var-
In particular, in Figure18a the effect of the compressive strength of iable in the comparisons.
concrete is considered: the curve with black dots represents the ratio of The trend reported in Fig. 18a and the summary in Table 5 show that
the numerical values of the maximum shear, VFEM,Max, obtained for the when fcm increases of 25 % (i.e., from 20 to 25 MPa), also VFEM,max tends
beams with fcm = 25 MPa and fcm = 20 MPa (concrete strength ratio to be higher with an average increase of 13 % (minimum 2 %, maximum
1.25), being constant the other parameters, while the curve with red 27 % depending on the values of the other parameters), which is, thus,
dots represents the same ratio calculated for the cases fcm = 28 MPa and about the half of the increase of fcm; similarly, while when fcm increases
fcm = 20 MPa (concrete strength ratio 1.40). of 40 % (i.e., from 20 MPa to 28 MPa), VFEM,max increases in average of
Analogously, in Fig. 18b, the curve with black dots represents the 19 % (minimum 4 %, maximum 28 %), which is again about the half of
ratio of the values of VFEM,Max for the beams with fy = 450 MPa and fy the variation of fcm.
= 350 MPa (steel strength ratio 1.29), being constant the other On the contrary, it can be observed in Fig. 18b that when the yielding
Table 4
Number of beams failed in shear distinguished for the associated parameters combination.
av/d Φ [mm] fc [MPa] fy [MPa] Tot
H [mm 1 1.5 2 10 14 18 20 25 28 350 450 550
150 15 11 4 0 9 21 11 9 10 5 10 15 30
200 15 9 1 0 6 19 9 8 8 6 9 10 25
250 13 6 1 0 4 16 7 7 6 3 7 10 20
16
Fig. 18. : Influence of the investigated parameters on VFEM,max: (a) fcm; (b) fy; (c) H; (d) As; (e) av/d.
17
Table 5 underestimations are significantly higher than those observed for the
Percentage variations of the shear strength. benchmark beam, varying from 50 % to 78 %. The British code [28],
Parameter parameter variation shear strength which considers the effect of av/d, provides the lowest underestimation,
while the simplified approach of ACI 318 [24], i.e., Eq. 3, gives the
min mean max
highest underestimation. Fig. 20 illustrates the distribution of the
fcm 20→25 MPa 25 % 2% 13 % 27 % average values of the ratio Vr,CODE/VFEM,max, and the corresponding
20→28 MPa 40 % 4% 19 % 28 %
fy 350→450 MPa 29 % − 16 % − 2% 8%
standard deviation provided for the 75 beams.
350→550 MPa 57 % − 6% 0% 8% In order to specifically investigate the impact of the beam height,
d 130→180 mm 38 % 0% 10 % 30 % Figs. 21a and 21b compare the theoretical shear strength values pro-
130→230 mm 77 % 9% 24 % 43 % vided by only EC2 [12] (Fig. 20a) and BS 8110–1 [28] (Fig. 21b) with
As 616→1018 mm2 65 % − 11 % 1% 7%
the numerical values. The data related to beams with the same height
av/d 2.0 → 1.5 − 25 % − 2% 3% 13 %
2.0 → 1.0 − 50 % 11 % 18 % 23 % (150, 200, and 250 mm) are grouped together with the same markers.
Both codes consider the effect of av/d in shear strength calculations
when av/d < 2.
stress fy increases by about 30 %, i.e. it changes from 350 to 450 MPa, in Fig. 21a illustrates that the underestimation by EC2 [12] tends to
average there is no increase of VFEM,max, even if a maximum increase of diminish with increased beam height. However, it is essential to
8 % and a maximum decrease of 16 % are registered. Again, when fy consider the effect of av/d within beams of the same height. Conversely,
increases of about 60 %, i.e. from 350 to 550 MPa, the average variation Fig. 21b demonstrates that, for BS 8110–1 [28], beam height does not
is zero, while the single variations vary from + 8 % to − 6 %. In exhibit a clear effect on underestimation. It is noteworthy that BS
particular, the beams where a reduction of VFEM,max is registered when fy 8110–1 [28] provides the lowest influence of H on shear strength among
increases have H= 150 mm, av/d= 1, fc= 20 MPa, variable fy, the examined formulations (refer to Fig. 2d).
ϕ= 18 mm, and ρl= 1.70 % and two of them fall within the few cases To understand the influence of av/d, in Fig. 22a-b, the ratio of the
(6 % of total) in which the consistency check on the values of Vpl and code provisions to the numerical ones for the shear strength are plotted
VFEM,max gives an uncertain outcome. In general, fy seems to have a little versus av/d (1.0, 1.5, 2.0) for the beams having As = 4 ϕ18. Again, only
effect on the shear capacity of the beams, as confirmed by most codes the formulations provided by EC2 [12] (Fig. 22a) and BS 8110–1 [28]
that do not consider such a parameter as significant. (Fig. 22b) are considered and the values referring to the beams having
The effective depth d of the cross section induces a greater variability the same height are indicated with the same markers. Moreover, the
in the shear strength; when d increases of about 40 %, i.e. from 130 to values of concrete compressive strength and of steel yielding strength
180 mm, the average increase of VFEM,max is 10 % (with a variation in the are indicated for the examined cases. The graphs show that, for both
range 0–30 %); when d increases of 75 %, i.e. from 130 mm to 230 mm, codes, the underestimation is lower when the parameter av/d reduces, i.
the average increase of VFEM,max is of 24 % (with a variation in the range e. for the cases when av/d = 1, evidencing, thus, that the shear strength
9–43 %). This means that the increasing of VFEM,max is about 1/3 of the increase suggested by the two codes is surely necessary, but is not suf-
variation of d. ficient to have a reliable prediction of the numerical shear strength.
As for the yielding stress, also the area of the steel reinforcement As Fig. 22a confirms that EC2 [12] (Fig. 22a) gives a lower underesti-
has a limited influence on the value of VFEM,max, in the cases charac- mation of the shear strength for the highest values of beam height, i.e.
terized by shear failure, since the maximum increase is 7 % and the H= 250 mm, while there is not clear trend for the code BS 8110–1 [28].
maximum decrease is 11 %. However, it is worth to underline that the Since the influence of the parameter av/d on the shear strength may
amount of longitudinal reinforcement and its yielding stress are central be relevant, in Fig. 23 the variability of the numerical values VFEM, max
in the definition of the failure mode: when the flexural failure does not provided by the FE model is plotted vs. av/d for the 75 beams that
occur, the shear strength is, indeed, low dependent on As. attained the shear failure. Fig. 23 shows that VFEM, max increases when
A significant effect of the shear span ratio av/d is observed mainly av/d reduces, probably because a strut and tie resistant mechanism de-
when av/d from 2.0 becomes 1.0, since an average increase of 18 % for velops in the shear length av between the support and the loading
VFEM,max is registered; in case of av/d = 1.5 only in few cases there is an application point.
increase for VFEM,max of about 10 %. It is worth to note that also such a EC2 [12] allows taking into account this mechanism by increasing
parameter has a more significant influence on the possibility of shifting the shear strength given by Eq. 1 of 2d/av, when 0.5 < av/d < 2, but it
the failure mode from bending to shear when it becomes lower, rather was previously observed that such an increase is not sufficient to achieve
than on the value of the shear strength, when the shear failure is shear strength values comparable with the numerical ones.
attained. Finally, to elucidate the impact of percentage variations in input
parameters (mechanical or geometric) on the shear resistance (both
6. Numerical shear strength vs analytical provisions from a FEM and code-based perspective), a sensitivity analysis was
conducted, and the results are summarized in Table 6. The table presents
6.1. Comparisons with code provisions the percentage variations of shear strength provided by the FE model
(VFEM, max) and the codes outlined in Section 2 (Vr,code) for the 75 beams.
For the 75 beams that experienced shear failure (see Table 4), the It is noteworthy that the average increase of VFEM, max with fcm (13 %)
maximum shear strength values provided by the FE model, VFEM, max, are aligns consistently with the variations provided by the codes (8–11 %).
compared in Fig. 19 with the shear strength, Vr,code, as given by the code As discussed earlier, for beams experiencing shear failure, the variation
formulations described in Section 2. of fy has no significant effect on VFEM, max, which is consistent with the
Since the beams are loaded with av/d varying from 1 to 2, and when consideration of this parameter in only two codes (LoA I of MC2020 [12]
explicitly required by the code, the effect of av/d on shear strength is and SIA 262 [25]).
considered. In Figs. 19a-19e, the data series with square markers pertain Concerning the parameter d, the FE model exhibits average varia-
to the standards that consider the effect of av/d, while the data series tions lower than those offered by the codes, except for BS 8110–1 [28],
with round markers are related to the standards that do not account for which demonstrates shear strength variations with d in excellent
that issue. agreement with the numerical outcomes. Section 2 highlighted the
All the codes offer safe provisions, as previously observed for the disparate influence of parameter d within the examined codes, with BS
benchmark beam examined. Less conservative predictions are obtained 8110–1 [28] presenting the least impact on shear strength.
when the effect of av/d is taken into account. However, the Furthermore, the influence of longitudinal steel reinforcement, as
18
Fig. 19. (a),(b),(c),(d) and (e) Comparisons between numerical and analytical values of shear strength in individual codes.
19
Fig. 20. Distribution of the average values of the ratio Vr,CODE/VFEM,max.
Fig. 21. : Comparisons between numerical and analytical values of shear strength considering the effect of beam height: (a) EC2 [12]; (b) BS8110 − 1 [28].
suggested by the codes, does not correspond to the numerical findings. A simple model based on a strut-and-tie mechanism within the shear
length can be assumed: the longitudinal steel reinforcement represents
6.2. Comparisons with strut-and-ties models the tie, while the strut is represented by the compressed concrete
element developing from the loading point of the vertical force to the
EC2 [12] (refer to section 5.6.4(1)), as previously mentioned, rec- support. This strut is characterized by an angle α, with tgα=av/d. Based
ommends utilizing the strut-and-tie approach for predicting shear on force equilibrium, the shear strength can be calculated as the mini-
strength within the designated ’D-regions’ (Discontinuity regions), mum vertical force leading to the yielding of the longitudinal steel
corresponding to approximately the height of the element. Therefore, reinforcement or failure in compression of the concrete, as follows:
when av/d ≤ 1, formulations such as Eq. 1a or Eq. 11 are expected to be VR,s = As • fy • tg α (13)
less effective, primarily designed for standard zones of the beam (i.e.,
the so-called ’B-regions’ or Bernoulli regions). In the subsequent section, β • d • b • fcm
shear strength for beams that collapsed numerically due to shear is VR,c = (14)
cosα
calculated using a strut-and-tie mechanism to assess whether improved
agreement can be achieved compared to the previous section. VR,strut-tie = min (VRd,s, VRd,c) (15)
The majority of beams analyzed in the FE model that failed due to
shear (92 % of 75 samples) have av/d values of 1.0 or 1.5. Stress con- where As and fy are the area and the yielding stress of steel longitudinal
tours in the FE model revealed that for these elements, a strut developed reinforcement in the ‘D region’ close to the support, d and b are effective
within the shear length. Cases with av/d equal to 1 or 1.5 fall into the depth and width of the element, fcm is the mean compressive strength of
category of elements where the load is applied within the designated ’D- the concrete, and β is a coefficient that defines the width of the strut in
region.’ Consequently, shear strength may not be accurately assessed by function of the effective depth d of the section, general assumed in the
models based on beam-like behavior commonly adopted in the previ- range 0.2–0.3.
ously described design formulations. This could explain the significant For the three values of av/d assumed in the FE analyses, the angle α is
underestimations observed in the numerical results provided by these as follow:
models.
20
by the FE model, VFEM, max, in terms of ratio VR,strut-tie/VFEM,max, under

different assumptions for the β value (namely 0.20, 0.25 and 0.30).
For all the beams with H = 150 mm, the minimum shear is related to
the failure in compression of the strut. For the beams with av/d = 1, Eq.
15 gives good agreement for β = 0.3 (overestimation of the numerical
strength of 3 %) or 0.25 (underestimation of the numerical strength of
15 %),), while for av/d= 1.5 and 2, the best agreement is attained for
β = 0.2 (overestimation of the numerical strength of 4–5 %), since for
β = 0.25 or 0.30, there is an overestimation ranging in 30–50 %.
For the beams with H = 200 mm, in most cases the minimum shear is
related to the failure in compression of the strut, while the yielding of
the reinforcement occurred only in 8 cases, mainly when β = 0.30. For
the beams with av/d = 1, Eq. 15 gives good agreement for β = 0.25
(overestimation of the numerical strength of 9 %), while for av/d = 1.5
and 2 for β = 0.20 the overestimation of the numerical strength ranges
in 16–24 % and for β = 0.25 or 0.30 is even higher.
For the beams with H= 250 mm, in most cases the minimum shear is
related to the failure in compression of the strut, while the yielding of
the reinforcement occurred only in 11 cases, mainly when β = 0.30 and
0.25. For the beams with av/d = 1, Eq. 15 gives good agreement for
β = 0.20 (underestimation of the numerical strength of 13 %), while for
av/d = 1.5 and 2.0 the overestimation of the numerical strength is larger
than 30 % for whatever value of β.
These comparisons evidence that:
• yielding of steel reinforcement occurs more frequently when the

height of the beam is higher or the width of the strut is assumed
larger (β = 0.30) because of the increased strength of the strut;
• Eq. (15) tends to overestimate more significantly the numerical shear
strength VFEM, max when av/d or the height of the section increase.
Thus, it is confirmed that the ‘D-region’ where ‘strut and tie’-based

models can be considered reliable in predicting the shear strength cor-
responds to the condition av/d = 1 and a very good prediction of the
numerical values can be attained assuming β = 0.25 or 0.20 (for greater
Fig. 22. Code-to-numerical provisions ratio vs. av/d: (a) EC2 [12]; (b)
BS8110 − 1 [28].
height). On the contrary, for av/d > 1, the strut and tie’-based models
provide a significant overestimation of the numerical strength, espe-
cially for greater heights, confirming that the ‘beam-like behaviour’-
based models may be more realistic. In the previous sections, it was,
indeed, evidenced that the code formulations have a better agreement
with the numerical results when av/d and the height increase.
In Fig. 24 the values of shear strength given by Eq. (15) are compared
with the ones provided by the FE model, VFEM, max, for the 75 beams that
attained the shear failure, only for the case β = 0.25. Fig. 24a, b and c
refer to the beams with H = 150 mm, 200 mm and 250 mm,
respectively.
Fig. 24 confirm that the shear strength provided by the strut-and-tie
based model is reliable for estimating the numerical values, VFEM, max,
for av/d = 1 and assuming β = 0.25, especially for the beams with H
= 150 mm and 200 mm.
7. Conclusions
The recent structural collapses of reinforced concrete bridges in Italy

have highlighted the high vulnerability of such infrastructures even
under the influence of gravitational and traffic loads alone.
One of the less-explored aspects of the structural behavior of RC
bridges is the shear behavior of the deck. Therefore, this paper focuses
Fig. 23. VFEM, vs. av/d.
max
on the numerical simulation of the shear behavior of RC beams without
specific transverse steel reinforcement with a dual objective: 1) inves-
• av/d= 1.0→α = 45◦ tigating the impact of various geometric and mechanical parameters on
• av/d= 1.5→α = 56◦ shear strength, and 2) assessing the reliability of several code recom-
• av/d= 2.0→α = 63◦ mendations predicting the shear strength of RC elements without
transverse reinforcement.
In Table 7, for the 75 beams that attained the shear failure, the values Utilizing available experimental data from the literature, a numerical
of shear strength given by Eq. (15) are compared with the ones provided model developed in the MIDAS FEA software [30] was trained and
21
Table 6
Numerical vs. code variations of the shear strength due to the main parameters varied in the FE analyses.
Variation of parameters FEM EC2 and ACI 318 fib MC2010 fib MC2020 SIA DM A23- JSCE BS8110-
NTC 262 204 3.04 1
Eq. 2 Eq. 3 LoA I LoA LoA I LoA II AS
II 3600
fcm 20→25 MPa 25 % 13 % 8% 11 % 11 % 11 % 11 % 11 % 11 % 11 % 11 % 11 % 8% 8%

20→28 MPa 40 % 19 % 11 % 18 % 18 % 18 % 18 % 18 % 18 % 18 % 18 % 18 % 11 % 12 %
fy 350→450 MPa 29 % − 2% χ χ χ χ χ − 13 % χ − 11 % χ χ χ χ
350→550 MPa 57 % 1% χ χ χ χ χ − 23 % χ − 16 % χ χ χ χ
d 130→180 mm 38 % 10 % 26 % 29 % 38 % 34 % 21 % 34 % 21 % 26 % 28 % 21 % 26 % 14 %
130→230 mm 75 % 24 % 44 % 35 % 77 % 65 % 45 % 65 % 45 % 50 % 52 % 45 % 44 % 27 %
As 616→1018 mm2 65 % 1% 18 % 10 % χ χ 38 % χ 39 % χ χ χ 18 % 18 %
Table 7
Variation of ratio VR,strut-tie/VFEM, max for different values of av/d (1.0, 1.5, 2.0) and three values of β (0.20, 0.25 and 0.30).
Vstrut_and tie/VFEM,max
β av/d= 1.0 av/d= 1.5 av/d= 2.0
150 mm 200 mm 250 mm 150 mm 200 mm 250 mm 150 mm 200 mm 250 mm
0.30 1.03 1.22 1.30 1.56 1.87 2.01 1.52 1.74 1.58
0.25 0.85 1.09 1.20 1.30 1.56 1.67 1.32 1.45 2.15
0.20 0.69 1.18 0.87 1.04 1.24 1.33 1.05 1.16 1.72
700 700
min (Eq. 12; 13) with β=0,25 min (Eq. 12; 13) with β=0,25
600 600
500 500
V strut-tie [kN]
Vstrut-tie [kN]
400 400
300 300
h=150 av/d=1
h=200 av/d=1
200 200
h=150 av/d=1.5
h=200 av/d=1.5
100 100
h=150 av/d=2 h=200 av/d=2
0 0
0 100 200 300 400 500 600 700 0 100 200 300 400 500 600 700
V FEM, max [kN] V FEM, max [kN]
a) b)
700
min (Eq. 12; 13) with β=0,25
600
V strut-tie [kN]
500
400
300 h=250 av/d=1
200 h=250 av/d=1.5
100
h=250 av/d=2
0
0 100 200 300 400 500 600 700
V FEM, max [kN]
c)
Fig. 24. Numerical vs. analytical values of shear strength provided by the strut-and-tie model for β = 0.25: a) beams with H = 150 mm, b) beams with H = 200 mm,
c) beams with H = 250 mm.
22
validated. Subsequently, an extensive campaign of nonlinear numerical CRediT authorship contribution statement
analyses (approximately 240 specimens) was conducted on RC beams
without transverse reinforcement, considering multiple combinations of Nicola Caterino: Writing – review & editing, Writing – original
geometric and mechanical parameters to assess the influence of each draft, Validation, Supervision, Methodology, Investigation, Funding
parameter. The results were analyzed in terms of shear-displacement acquisition, Conceptualization. Francesca Ceroni: Writing – review &
curves, principal stress distributions, evolution of the cracking pattern, editing, Writing – original draft, Validation, Supervision, Methodology,
maximum tensile stresses in the steel bars, and maximum compressive Investigation, Funding acquisition, Conceptualization. Vincenzo Di
strains in the concrete. Cristo: Writing – review & editing, Writing – original draft, Visualiza-
Moreover, for the 75 beams that failed in shear, the numerical tion, Investigation, Formal analysis, Data curation.
strengths were compared with the specifications of several codes. The
main findings from the numerical analysis and comparisons with code Declaration of Competing Interest
provisions include:
The authors declare that they have no known competing financial
• In the FE analysis, the collapse was predominantly ductile for most interests or personal relationships that could have appeared to influence
samples (168 out of 243), while the remaining 75 exhibited brittle the work reported in this paper.
shear failure. The occurrence of shear failure in the beams is mainly
influenced by the steel percentage in tension, its yielding stress, and Data availability
the shear span-to-depth ratio; lower values of these parameters in-
crease the likelihood of shear failure; Data will be made available on request.
• for the 75 beams experiencing shear failure, the compressive
strength of concrete and the effective depth of the cross-section were Acknowledgements
found to be the most influential parameters. The increase in shear
strength is approximately half of the increase in fcm and about one- The research activities presented in the paper have been carried out
third of the variation in effective depth; in the framework of the research project ‘Reluis Bridges’, Work Package
• the influence of the main parameters observed in the numerical an- 4 – Task 6 - Testing on Structural and/or Special Components." (CUP
alyses does not always align with that provided by code formula- I65F21000730005) with the additional financial support of the ‘Sus-
tions. Good agreement between numerical and analytical variations tainable Mobility Center’ (Centro Nazionale per la Mobilità Sostenibile)
is only observed for concrete compressive strength, while variations funded under the Italian National Recovery and Resilience Plan, Mission
for other parameters are typically lower in the FE model than in the 4 Component 2 Investment 1.4 NextGenerationEU (CUP
codes; I63C22000340001).
• all codes significantly underestimate the strengths provided by the
FE model for the simulated beams, with underestimation rates References
ranging from 50 % to 78 %, even when considering the effect of av/d;
• some code provisions are more reliable, resulting in lower underes- [1] DM 2042022, Decreto Ministeriale n. 204 del 1.7.2022, Linee Guida per la
classificazione e gestione del rischio, la valutazione della sicurezza ed il
timation of the numerical shear strength, particularly when applied monitoraggio dei ponti esistenti. (in Italian).
to RC beams with large heights, as in the benchmark case used for [2] Lantsoght EO, van der Veen C, Walraven JC. Shear tests of reinforced concrete
checking the reliability of the FE model; slabs with concentrated loads near to supports. In: Proceedings of the eighth fib
PhD symposium in Kgs. Lyngby, Denmark 2010:81–6.
• the comparison between numerical results and normative formulas [3] Lantsoght EO, van der Veen C, Walraven JC. Shear capacity of slabs and slab strips
confirms the existence of ’discontinuity regions’ or ’D regions’ in the loaded close to the support, ACI SP-287. A recent development in reinforced
case of concentrated loads near the supports, where ’strut-and-tie’ concrete slab analysis, design and serviceability 2012. 5–1.
[4] Lantsoght E, der Veen CV. Shear in one-way slabs under concentrated load close to
mechanisms better correspond with the results of FE models. Despite support. Acids Struct J 2014;110(2):275–84.
the predicted increase in shear strength by many codes for 0.5 ≤ av/ [5] Bui TT, Abouri S, Limam A, Nana WSA, Teodoldi B, Rougre T. Experimental
d ≤ 2, expressed as a function of 2d/av, it is insufficient to accurately investigation of shear strength of full-scale concrete slabs subjected to concentrated
loads in nuclear buildings. Eng Struct 2017;131:405–20.
predict the numerical shear strength;
[6] Bui TT, Limam A, Nana WSA, Ferrier E, Bost M, Bui QB. Evaluation of one-way
• in the ’D region’ (av/d=1), the model based on a strut-and-tie shear behavior of reinforced concrete slabs: experimental and numerical analysis.
mechanism shows a good correspondence with the numerical Eur J Environ Civ Eng 2017;24(2):1–27.
values, especially for the shallowest beams, assuming a strut width [7] Nana W, Bui TT, Bost M, Limam A. Shear bearing capacity of RC slabs without
shear reinforcement: design codes comparison. KSCE journal of civil engineering,
equal to 0.25 times the effective depth of the slab; 23. Springer; 2019. p. 321–34. https://doi.org/10.1007/s12205-018-0612-7.
• for av/d values of 1.5 and 2.0, code formulations significantly un- [8] Taylor R, Hayes B. Some tests on the effect of edge restraint on punching shear in
derestimate the shear strength provided by the numerical models, reinforced concrete slabs. Mag Concr Res N 17 1965;Vol. 50:39–44.
[9] Papanikolaou KV, Tegos IA, Kappos AJ. Punching shear testing of reinforced
while formulas based on the strut-and-tie mechanism overestimate concrete slabs, and design implications. Mag Concr Res 2005:167–77.
it. These results indicate that in the case of av/d equal to 1.5 or 2.0, [10] Grimaldi, A., Meda, A., Rinaldi, Z., Devito Franceschi, A.2011, Solette da ponte in
the resistance mechanism may be intermediate between an arch calcestruzzo fibrorinforzato, proceedings of the workshop “Progettare e realizzare
elementi strutturali in calcestruzzo fibrorinforzato a cinque anni dal Documento
mechanism and a beam mechanism and, thus, code formulas simu- CNR DT – 204″, Roma, Italy. (in Italian).
lating the latter behaviour need better calibration. [11] Natário F, Fernández Ruiz M, Muttoni A. Shear strength of RC slabs under
concentrated loads near clamped linear support. Eng Struct 2014;76:10–23.
[12] Eurocode 2, EC22005, Design of concrete structures - part 1–1: general rules and
All these issues require detailed investigation through experimental rules for buildings, Brussels, EN, CEN.
tests on RC beams too. To this aim, the numerical studies presented [13] FD P 18–7172013. Eurocode 2 - Calcul des structures en béton - Guide
herein served as a precursor to the design of an extensive experimental d′application des normes NF EN. (in French).
[14] fib MC20102013. Model Code for concrete structures 2010, Vol 2 Bull. 65,
campaign on RC beams without transverse reinforcement, aimed at
Fédération Internationale du Béton, Lausanne Switzerland.
studying the impact of geometrical and mechanical parameters on shear [15] fib MC20202023. Model Code for concrete structures 2020, Fédération
strength evidenced by the numerical outcomes. Internationale du Béton, Lausanne, Switzerland, ISBN: 978–2-88394–175-5.
Based on the forthcoming experimental results, proposals for modi- [16] Ewees MH, Gabr ASA, Farrag MR. Effect of tension and compression flexural
reinforcement on punching shear strength of reinforced concrete flat slab. Alex Eng
fications to code formulations could also be developed to enhance their J 2024;99:282–302.
reliability. In addition, further analysis is imperative to assess the impact
of reinforced concrete deterioration on the shear strength of the beams.
23
[17] Pani L, Stochino F. Punching of reinforced concrete slab without shear [25] SIA 262, Construction en béton; Société Suisse des Ingénieurs et Architectes;
reinforcement: standard models and new proposal. Front Struct Civ Eng 2020;14: révision partielle de 2012 Zürich; Suisse. (in French).
1196–214. [26] CSA A23.3–042004 CSA Standard A23.3–04, Design of Concrete Structures,
[18] Podgorniak-Stanik BA. The influence of concrete strength, distribution of Canadian Standard Association, Toronto (Ontario), Canada.
longitudinal reinforcement, amount of transverse reinforcement and member size [27] JSCE. Standard specification for concrete structures. Jpn Soc Civ Eng 2007.
on shear strength of reinforced concrete members. [MSc thesis], 1998. University [28] BS8110–11997, Structural use of concrete, British Standards Institution BSI.
of Toronto,; 1998. [29] AS 3600, Standards Australia2018, Concrete structures. Sydney: SAI Global
[19] Regan, P.E.1987, Shear resistance of members without shear reinforcement; Limited.
proposal for CEB model code MC90. London, UK: Polytechnic of Central London. [30] Midas FEA NX v. 1.12022. Manual: analysis and algorithm MIDAS FEA, advanced
pp. 1–28. nonlinear and detail analysis program, CSP SEA.
[20] Bresler B, MacGregor JG. Review of concrete beams failing in shear. J Struct Div [31] Thorenfeldt, E., Tomaszewicz, A., Jensen, J.J.1987, Mechanical properties of high-
1967;93(1):343–72. strength concrete and applications in design, In: Proceedings of the symposium
[21] Vecchio FJ, Collins MP. The modified compression-field theory for reinforced utilization of high-strength concrete, Trondheim, Norway, Tapir.
concrete elements subjected to shear. J Am Concr Inst 1986;83(2):219–31. [32] Hordijk, D.A.1967, Local approach to fatigue of concrete [Ph.D. thesis]. Delft
[22] Muttoni, A.2003, Eléments sans armature d′effort trenchant, Documentation SIA. University of Technology.
D0182 Introduction à la norme SIA 262. pp. 47–55. (in French). [33] Von Mises, R.1913, Mechanik der festen Körper im plastisch-deformablen Zustand,
[23] NTC, Norme Tecniche per le Costruzioni. Decreto Ministeriale del 17 gennaio Nachrichten von der Gesellschaft der Wissenschaften zu Göttingen, Mathematisch-
2018, Roma. (in Italian). Physikalische Klasse, (1). pp. 582–92. (in German).
[24] ACI 3182014, American Concrete Institute, Building Code Requirements for
Structural Concrete.
24

Strength Capacity of RC Beams Without Shear Reinforcement

Uploaded by

Copyright:

Available Formats

Strength Capacity of RC Beams Without Shear Reinforcement

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Strength Capacity of RC Beams Without Shear Reinforcement

Uploaded by

Copyright:

Available Formats

Engineering Structures 317 (2024) 118634

Contents lists available at ScienceDirect

Strength capacity of RC beams without shear reinforcement: Numerical

1. Introduction without any transversal steel reinforcement, surely requests investiga-

International codes adopt different approaches to consider influen- load footprint.

where d, bw, fck, and γc maintain the previously introduced meanings,

NTC Eq. 2 Eq. 3 LoA I LoA II LoA I LoA II AS 3600

4.5 ACI 318-14 (Eq. 2)

ACI318-14 (Eq. 3) - MC10 LoA I - MC10 LoA III - MC20 1

Vr,CODE /Vr, CODE, 300 MPa

ACI 318-14 (Eq. 2) 7 ACI 318-14 (Eq.3)

6 MC 2020 LoA I (Eq. 4 e 6a)

150 mm 2700 mm 150 mm

Fig. 5. : The stress-strain behavior assumed for concrete and steel.

Fig. 9. : a) Mesh of concrete slab; b) Mesh of the steel bars.

300 300 300 VFEM, max

150 150 150

100 100 100

H [mm 1 1.5 2 10 14 18 20 25 28 350 450 550

Fig. 20. Distribution of the average values of the ratio Vr,CODE/VFEM,max.

by the FE model, VFEM, max, in terms of ratio VR,strut-tie/VFEM,max, under

• yielding of steel reinforcement occurs more frequently when the

Thus, it is confirmed that the ‘D-region’ where ‘strut and tie’-based

The recent structural collapses of reinforced concrete bridges in Italy

fcm 20→25 MPa 25 % 13 % 8% 11 % 11 % 11 % 11 % 11 % 11 % 11 % 11 % 11 % 8% 8%

β av/d= 1.0 av/d= 1.5 av/d= 2.0

150 mm 200 mm 250 mm 150 mm 200 mm 250 mm 150 mm 200 mm 250 mm

300 h=250 av/d=1

200 h=250 av/d=1.5

V FEM, max [kN]

You might also like