You are on page 1of 13

Structures 23 (2020) 481–493

Contents lists available at ScienceDirect

Structures
journal homepage: www.elsevier.com/locate/structures

Maximum torsional reinforcement ratio of reinforced concrete beams T


a a a b,⁎
Hyunjin Ju , Deuckhang Lee , Jong R. Kim , Kang Su Kim
a
Department of Civil and Environmental Engineering, Nazarbayev University, 53 Kabanbay Batyr Ave., Nur-Sultan 010000, Kazakhstan
b
Department of Architectural Engineering, University of Seoul, 163 Seoulsiripdae-ro, Dongdaemun-gu, Seoul 02504, Republic of Korea

ARTICLE INFO ABSTRACT

Keywords: The amount of maximum torsional reinforcement is specified to prevent the failure of concrete compressive
Maximum torsional reinforcement struts before the reinforcement yielding in reinforced concrete members subjected to torsion. The maximum
Reinforced concrete beams torsional reinforcement presented in the design codes, however, differs by up to more than three times. This
Truss model study therefore aimed to theoretically derive the maximum torsional reinforcement ratio. The proposed ratio
Failure mode
was derived based on the equilibrium and compatibility conditions. The parametric study results of the torsional
behavior analysis were also used to reflect key variables, such as tensile resistance of concrete, the cross-sec-
tional area surrounded by shear flow, the concrete softening factor, and the average stress factor, in a simple
form. In addition, the maximum reinforcement was presented not only in transverse direction but also in
longitudinal direction to predict the failure modes of torsional members according to the yielding of re-
inforcements. The validation was performed by comparison with the experimental results of 98 beams under
pure torsion collected from the literature, and the maximum torsional reinforcement ratio proposed in this study
was found to reflect the failure modes of reinforced concrete beams subjected to pure torsion better than those
suggested by other researchers and the design codes.

1. Introduction transverse reinforcement bars (mm); fyt is the yield strength (MPa) of
the transverse reinforcement; and is the angle of the compression strut
Advances in construction materials and structural design technolo- with respect to the longitudinal axis. The design codes provide no ex-
gies have led to the construction of large-scale concrete structures plicit equations of the maximum torsional reinforcement, but the
composed of complicated and slender members. Very often, these maximum allowable stress of the concrete struts is specified to prevent
members require consideration of the torsion caused by various types of crushing failure of concrete. Therefore, the maximum reinforcement
loads, in addition to the general flexure and shear force [1,2]. In ad- ratios presented in Fig. 1 can be derived from the stress limit. Even
dition, the slender members generate very large stress in the cross- though each code has a different limit, the same value for angle was
section, and a relatively large amount of rebar should be placed to used for all cases in this study for comparison purposes. In addition, the
adequately resist the external force. If too much amount of reinforce- transverse torsional reinforcement ratio ( t ) is the form of the re-
ment, however, are placed in the reinforced concrete members that inforcement volume ratio per concrete volume, which is represented as
resist torsion, the members may fail because of the crushing of the At ph /(Ac s ) , where Ac is the total area of the member section. The
concrete before the rebars yield [3]. Therefore, the codes for concrete amount of maximum torsional reinforcement presented in ACI318-14
structures, including the ACI318 [4] standard, stipulate the maximum [4] is empirically calculated, whereas that in CSA [6] and EC2 [7] is
amount of rebar that can be placed in shear and torsion design to based on truss models [3,8,9], and the EC2 code uses 0.6(1 fc /250) as
prevent the crushing of concrete before the rebars yield and to ensure a coefficient ( ) for the softening phenomenon [10,11] in concrete.
the ductility of the members [4]. These equations were basically for shear design, but they were extended
Fig. 1 compares the maximum torsional reinforcement ratios ( t,max ) or modified to be used for both shear and torsional design. When the
presented in the ACI318 [5], CSA [6], and EC2 [7] codes, where fc is reinforced concrete member, however, is resistant to torsional loads, it
the compressive strength (MPa) of concrete; Aoh and ph are the area shows the strain distribution according to the depth within the effective
(mm2) and circumferential length (mm) surrounded by the centerline of thickness from the outermost surface of the member [12,13], unlike the
the transverse reinforcement, respectively; s is the spacing between the case of the shear. For this reason, not only two-dimensional membrane


Corresponding author.
E-mail addresses: hyunjin.ju@nu.edu.kz (H. Ju), deuckhang.lee@nu.edu.kz (D. Lee), jong.kim@nu.edu.kz (J.R. Kim), kangkim@uos.ac.kr (K.S. Kim).

https://doi.org/10.1016/j.istruc.2019.09.007
Received 26 July 2019; Received in revised form 10 September 2019; Accepted 12 September 2019
Available online 26 December 2019
2352-0124/ © 2019 Institution of Structural Engineers. Published by Elsevier Ltd. All rights reserved.
H. Ju, et al. Structures 23 (2020) 481–493

overestimated. Chakraborty [16] used a three-dimensional truss model


to derive the maximum torsional reinforcement ratio by considering the
conditions for the equilibrium of forces, and proposed a semi-empirical
equation based on the regression analysis of the PCA test results [23]. In
addition, the maximum torsional reinforcement ratio in the long-
itudinal direction was estimated to be 1.3 times the maximum torsional
reinforcement ratio in the transverse direction, similar to the model
proposed by Hsu [23]. Later, Rahal and Collins [17] proposed a model
for predicting the torsional strengths of reinforced and prestressed
concrete members by considering the softening phenomenon of the
concrete compression strut, the spalling phenomenon, and the max-
imum compressive strain ( 2,max ) of the member surface, and also pre-
sented the maximum torsional reinforcement ratios of the longitudinal
and transverse directions. Although this approach has an advantage as a
simple form that does not require iterative calculations, it considered
only the compressive strength of concrete (fc ) and the yield strength of
reinforcement (fyt or f yl ) and did not reflect the contribution of concrete
Fig. 1. Maximum torsional reinforcement ratio in design codes.
to the tension and cross-sectional aspect ratio. As described above, the
maximum torsional reinforcement ratios proposed by the existing re-
elements but also three-dimensional member equilibrium and de- searchers are based on experimental results, but the softening phe-
formation compatibility should be considered. In addition, because the nomenon of compressive strength or the contribution of concrete to
member behaves differently from the member subjected to shear force, tension was not considered, even in the theoretical model. In addition,
as shown by the correlation between the longitudinal and transverse the strain gradients within the effective thickness, which occur in re-
reinforcement rebars [14–16] and the effects of the cover thickness inforced concrete members subjected to torsion, was not reflected ei-
[17–19], it may be inappropriate to apply the maximum reinforcement ther.
ratio determined from the shear behavior to the torsional member.
Further, each code presents the amount of maximum torsional re- 3. Proposed model
inforcement, which differs by up to three times, depending on the
compressive strength of concrete (fc ) , as shown in Fig. 1. These different 3.1. Derivation of the basic form
maximum torsional reinforcement ratios may not be adequate for pre-
venting the brittle failure mode of the torsional member. In particular, The thin-walled tube theory [24] suggests that reinforced concrete
when only the maximum value of the transverse reinforcement limit is members subjected to torsion (T ) resist the external force (T ) by the
presented, it is not possible to consider the failure mode of the member, shear flow (q ) within the effective thickness (td) from the outermost
which may differ depending on the yielding of the longitudinal and surface of the member, as shown in Fig. 2(a) [12,25]. The equilibrium
transverse reinforcements. of the forces in the transverse direction among the shear elements of the
Therefore, this study intended to present the maximum transverse thin-walled tubes with effective thickness td , as shown in Fig. 2(b) and
and longitudinal torsional reinforcement ratios based on clear theore- (c), can be represented by considering the average compressive stress
tical grounds and suggest them in a simple form to reflect the tensile ( 2 ) of the concrete compression strut and the stirrup stress and average
resistance of concrete and the compression softening, and to promote tensile stress ( 1 ) acting in a direction perpendicular thereto, as follows:
practical application. In addition, the proposed model was verified by
evaluating the failure modes of the members subjected to pure torsion. At fst = ( 2sin2 1cos
2 ) std (1)
where At and fst are the cross-sectional area and stress of the transverse
2. Existing models reinforcement, respectively. In addition, the equilibrium equation of
forces in the longitudinal direction can be represented, as follows:
The maximum torsional reinforcement ratio limits are intended to
Al fsl = ( 2cos2 1sin
2 ) po td (2)
prevent the compressive failure of concrete before the torsional re-
inforcement yields; the maximum torsional reinforcement ratio is where Al and fsl are the cross-sectional area and stress of the long-
therefore based on the balanced reinforcement ratio at which the itudinal reinforcement, and po and Ao are the perimeter length and area
crushing of the concrete compressive struts and the yielding of the re- surrounded by the centerline of the shear flow (q) in the cross-section of
bars occur at the same time [12,16,18,20–22]. The maximum torsional the torsional member, respectively, as shown in Fig. 2(a). The torsional
reinforcement ratio of reinforced concrete members have been sug- reinforcement ratios in the transverse and longitudinal directions are
gested by several researchers, starting from the study of Hsu [23] at the t = At poh /(A c s ) and , respectively, where poh is the perimeter length of
end of the 1960s. Based on the test results of the PCA (Portland Cement the stirrup, s is the stirrup spacing, and Ac is the concrete cross-sec-
Association), Hsu [23] presented the maximum torsional reinforcement tional area. Therefore, as in Eqs. (1) and (2), the transverse and long-
ratio as the sum of the transverse and longitudinal reinforcement ratios, itudinal reinforcement ratios ( t and l ) can be summarized, respec-
in which the longitudinal reinforcement-to-transverse reinforcement tively, as follows:
ratio was proposed as 1.2. Hsu [23] also observed that the maximum
( 2sin2 1cos
2 ) td poh
torsional reinforcement ratio was affected by the cross-sectional shape t =
of the member, but he did not reflect the effects of the member shape
fst Ac (3a)
and size in the proposed equation. ( cos2 2
2 1sin ) td po
Mitchell and Collins [17] first proposed a maximum torsional re- l =
fst Ac (3b)
inforcement ratio based on theoretical approaches with respect to
longitudinal and transverse rebars using the diagonal compression field According to Bredt’s thin-walled tube theory [11,25], the relation-
theory. As this model, however, does not take into account the soft- ship between the average shear strain ( ) of the thin-walled shear
ening effect of concrete, [10,11] a relatively high maximum re- element and the torsional angle ( ) per unit length can be expressed, as
inforcement ratio can be derived, because the concrete strength is follows:

482
H. Ju, et al. Structures 23 (2020) 481–493

Fig. 2. Equilibrium conditions.

po transverse reinforcement ( t ), and the strain of longitudinal reinforce-


=
2Ao (4) ment ( l ), as follows:

As shown in Fig. 3, the shear strain ( ) is again represented as a = 2( t 2 )tan (5a)


function of the average principal compressive strain ( 2 ), the strain of or

Fig. 3. Compatibility conditions.

483
H. Ju, et al. Structures 23 (2020) 481–493

Ao 2 2
l = ( 2 1tan )
Ac ( l 2) fsl (8b)

In the preceding Eq. (8a), poh / po can be simply taken as being 1.0.
The maximum torsional reinforcement ratio shown in Eq. (8) is a
function of the average compressive strain ( 2 ) of concrete in the thin-
walled tube, the strain in the rebars ( l, t ), and the inclination angle ( )
of the compression strut, which is related to the strain compatibility
condition. If the maximum torsional reinforcement ratio is calculated
from the conditions under which the compressive failure of the com-
pression struts occurs at the same time as the yielding of the re-
inforcement, the equations for the average compressive strain ( 2 ) in the
thin-walled tube and the strain of reinforcement ( t or l ) can be sim-
plified by substituting the ultimate strain of concrete and the yield
strain ( yt or yl ), respectively. It is not easy, however, to determine the
inclination angle ( ) of the compression strut, as in the shear problem.
Because the longitudinal reinforcement ratio is generally larger than
the transverse reinforcement ratio in a beam, it is often too conservative
to take the inclination angle ( ) of the compression zone as 45 degrees.
Therefore, EC2 [8] code allows to take 38.5 degrees as the inclination
angle ( ) of the compression strut. If the average compressive strain at
concrete crushing and yield strain of transverse reinforcement are fixed,
the strain of the longitudinal strain can be obtained from the compat-
ibility equation [i.e., tan2 = ( l + 2)/( t + 2) ] [8,27]. Since the com-
pressive strain ( 2,max ) in the outermost layer at crushing of concrete can
be adopted to be 0.0035 as presented in Hsu et al. [18], the average
compressive strain ( 2 = 2. max /2 ) can be taken as 0.00175, and the yield
strain of the grade 60 (420 MPa) torsional reinforcement is 0.0021.
Therefore, the strain in the longitudinal direction ( l ) is calculated as
0.0019 to satisfy the compatibility condition. Similarly, when de-
termining maximum torsional reinforcement in the longitudinal direc-
tion, the same logic can be applied as done for the maximum transverse
reinforcement ratio. The yield strain 0.0021 can be applied for l when
the longitudinal reinforcement yields prior to yielding of transverse
reinforcement, then the inclination angle ( ) is calculated to be 51.5
degrees. This means that 38. 5° and 51. 5° are used as the cracking angle
Fig. 4. Stress–strain relationships of concrete [26].
to derive the maximum torsional reinforcement in the transverse and
longitudinal directions, respectively. On this basis, by substituting
= 2( l 2 )cot (5b) 2 = 0.00175, t = 0.0021, = 38. 5° , fst = f yt in Eq. (8a), and
l = 0.0021, = 51. 5 , fsl = f yl in Eq. (8b), the maximum transverse and
°
In addition, a strain gradient occurs according to the depth of the
member within the effective thickness (td ) of the torsional member, as longitudinal torsional reinforcement ratios ( t,max , l,max ) can be ex-
pressed, as follows:
shown in Fig. 2(a), and assuming that it is linear, the curvature ( ) of
the strut in the effective thickness can be represented as a function of
Ao ( 2 1.58 1 )
the torsional angle ( ) per unit length and the angle ( ) of the com- t ,max = 0.45
Ac fyt (9a)
pression zone, as follows [12,18,26]:
2,max
= sin = Ao ( 1.58 1 )
td (6) = 0.45 2
l,max
Ac fyl (9b)
where 2,max is the maximum compressive strain in the outermost sur-
face of the member, which is twice the average principal compressive
where fyt and fyl are the yield strength of transverse and longitudinal
strain (2 2 ). In the combination of compatibility Eqs. (4)–(6), the ef-
reinforcement, respectively.
fective thickness (td ) of the torsional member can be summarized as
The constant ultimate compressive strain (i.e., 2,max = 0.0035)
follows:
taken in this study can provide un-conservative results of the maximum
2,max Ao torsional reinforcement ratio for high strength concrete members be-
td =
2sin2 po ( t 2) (7a) cause the concrete can fail before yielding of reinforcement.
Nevertheless, if 2,max is taken as a function of the compressive strength,
or it makes the derivation process too complex. In addition, as explained
2,max Ao by Jeng and Hsu [25], the ultimate compressive strain in a torsional
td =
2cos2 po ( l 2) (7b) member is expected to be larger than that failed in shear (i.e.,
2,max cu = 0.003) due to the out-of-plane bending, so-called the
If these are substituted into Eqs. (3a) and (3b), the transverse and warping effect within effective thickness. Thus, in this study, the as-
longitudinal torsional reinforcement ratios ( t and l ) can be derived, as sumption of 2,max = 0.0035 is kept for simplicity, and as will be dis-
follows: cussed later in the validation section, the proposed model is well cali-
Ao 2 2
poh brated to be conservative with the assumption.
t = ( 2 1cot )
Ac ( t 2) fst po (8a)

484
H. Ju, et al. Structures 23 (2020) 481–493

Fig. 5. Strain and stress distribution of concrete struts and average stress block [25].

3.2. Development of a simple form used in this study are shown in Fig. 4, where is the compressive
softening factor of concrete, o is the compressive strain at which the
Reinforced concrete members subjected to torsion exhibit strain cylindrical concrete specimen exhibits compressive strength, fcr is the
gradients [19,28] in a thin-walled tube with the effective thickness, as tensile crack strength of concrete (= 0.652 fc ), cr is the tensile strain of
shown in Fig. 2(a). Therefore, in the previously derived Eq. (9a,b), the concrete at the cracking point, and Ec is the elastic modulus of concrete.
compressive and tensile stresses in concrete can be regarded as the That is, a model that reflects the deviation angle (β) and softening
average stresses considering these strain gradients [12,18,26,29]. That phenomenon of concrete was used for the compressive stress–strain
is, the average compressive and tensile stresses of concrete struts with relationship, and the model proposed by Jeng and Hsu [26] was used
the effective thickness (td ) can be calculated to analyze the 3D elements for the tensile stress–strain relationship.
of the thin-walled tube as 2D elements. The compressive and tensile Assuming that the compressive and tensile strains within the ef-
stress–strain relationships of the concrete section elements that were fective thickness have a linear distribution, the stress distributions and

485
H. Ju, et al. Structures 23 (2020) 481–493

Table 1 The factors (kc , kt , ) included in Eqs. (10) and (11) are hard to be
Ranges of variables for parametric study. determined as a constant value, because they are affected by the
Parameters Variable range compressive strength of concrete (fc ) and reinforcement ratio ( ). In
particular, it is very difficult to roughly estimate what the factors (kc , kt ,
Width (mm) 300 400 500 ) are when the compression strut reaches the ultimate strain and suf-
Height (mm) 300 400 500 fers compressive failure at the same time as the rebar yields. Therefore,
20 30 40 50 60 70 80 90 100
fc (MPa) in this study, a parametric study of the members with various variables
350 400 450 500
fyt (MPa) was performed using the softened membrane model for torsion (SMMT)
fyl (MPa) 350 400 450 500 [26], which is known to relatively accurately predict the torsional be-
t (%) 2.7%~6.9% (maximum reinforcement ratio of ACI318-11 ~ 1.5 times havior of reinforced concrete members. SMMT [26] was extended from
that of EC2)
the fixed-angle softened truss model that had been developed to analyze
(%) 2.7%~6.9% (maximum reinforcement ratio of ACI318-11 ~ 1.5 times
the shear behavior of reinforced concrete in order to adapt it for tor-
l
that of EC2)
sional behavior analysis, which can reflect Poisson’s effect and the
tensile resistance of concrete.
the equivalent stress blocks can be derived as shown in Fig. 5 [25,26], The ranges of variables for parametric study include cross-sectional
where the average compressive stress factor (kc ) and the average tensile sizes, the concrete compressive strength (fc ) , the yield strength of the
stress factor (kt ) are functions of the maximum compressive and tensile reinforcement ( fy ), and the torsional reinforcement ratio ( ), as sum-
strains ( 2,max , 1,max ) and the compressive softening factor ( ) of con- marized in Table 1; the total number of sections is 5184. Meanwhile,
crete. Thus, the average compressive stress ( 2 ) and tensile stress ( 1) of is the reinforcement ratio in one direction, and the reinforcement ratios
the concrete struts can be represented, respectively, as follows: in both directions are the same. The amount of torsional reinforcement
included in the parametric study ranges from as low as the maximum
2 = k c fc (10) torsional reinforcement ratio of ACI318 [5] to 1.5 times that of EC2 [7]
to cover all the ranges from under-reinforced to excessively over-re-
1 = kt fcr (11) inforced torsional members. With the use of SMMT, the distribution of

Fig. 6. Coefficients according to reinforcement ratios based on parametric study results.

486
H. Ju, et al. Structures 23 (2020) 481–493

Fig. 7. Coefficients according to compressive strengths of concrete based on parametric study results.

reinforcement ratio.
Fig. 6 shows that the rebar yields when the reinforcement ratio ( )
is less than a certain value, and the yielding point of the rebar is closely
related to the compressive strength of concrete (fc ) . As shown in
Fig. 6(a), the average compressive stress factor (kc ) at the yielding point
of the rebar (i.e., maximum reinforcement ratio) tends to increase with
the increasing compressive strength of concrete (fc ) . Fig. 6(c) shows
that the average compressive softening factor ( ) at the yielding point of
the rebar (i.e., maximum reinforcement ratio) tends to decrease with
the increasing compressive strength of concrete ( fc ). On the other hand,
there is no correlation in the average tensile stress factor (kt ) or Ao / Aoh .
Fig. 7 shows kc , kt , , and Ao / Aoh as related to the compressive strength
of concrete ( fc ). As can be observed in Fig. 6, the average compressive
stress factor (kc ) and the average compressive softening factor ( ) at the
yielding point of the rebar (i.e., maximum reinforcement ratio) tend to
Fig. 8. Effect of yield strengths of steel rebars. increase or decrease, respectively, with the increasing compressive
strength of concrete ( fc ). On the other hand, the tensile stress factor (kt )
and Ao / Aoh show constant values regardless of the changes in the
kc , kt , and when the maximum compressive strain ( 2,max ) of the
compressive strength of concrete (fc ) . Accordingly, it can be determined
member surface reaches −0.0035 was represented in terms of the re-
that kt = 0.285 and Ao = 0.96Aoh when compressive failure of the strut
inforcement ratio ( ) and concrete compressive strength ( fc ), respec-
and yielding of the rebar occur simultaneously in the thin-walled tube.
tively, as shown in Figs. 6 and 7. Each graph represents the reinforce-
The average compressive stress factor (kc ) and the compressive soft-
ment ratio according to the yielding of the rebar; the yielding point of
ening factor ( ), which show a distinct tendency according to the
the rebar means the coefficient value at the maximum torsional
concrete strength (fc ) , were calculated, respectively, by deriving the

487
Table 2
Database and failure modes estimated by test results and code provisions.
H. Ju, et al.

Ref. Specimen Name fc (MPa) f yt (MPa) f yl (MPa) b (mm) h (mm) x o (mm) yo (mm) s (mm) t (%) l (%) Yielding Failure Mode ACI318 CSA EC2

Transverse Longitudinal t,max Yielding t,max Yielding t,max Yielding

23 B1 27.57 341.2 313.7 254 381 215.9 342.9 152.4 0.540% 0.524% Yield Yield UR 0.680% Not 1.374% Not 1.579% Not
B2 28.61 319.9 316.4 254 381 215.9 342.9 181.0 0.809% 0.821% Yield Yield UR 0.675% Not 1.359% Not 1.573% Not
B3 28.06 319.9 327.5 254 381 215.9 342.9 127.0 1.152% 1.184% Yield Yield UR 0.649% Not 1.287% Not 1.518% Not
B4 30.54 323.3 319.9 254 381 215.9 342.9 92.1 1.589% 1.600% Not Yield POR 0.645% Not 1.312% Not 1.669% Not
B5 29.02 321.2 332.3 254 381 215.9 342.9 69.9 2.095% 2.108% Not Not OR 0.836% Not 2.104% Yield* 2.436% Yield*
B6 28.82 322.6 331.6 254 381 215.9 342.9 57.2 2.560% 2.666% Not Not OR 0.835% Not 2.118% Not 2.352% Not
B7 25.99 318.5 319.9 254 381 215.9 342.9 127.0 1.152% 0.524% Yield Yield UR 0.495% Not 0.762% Not 0.952% Not
B8 26.75 319.9 321.9 254 381 215.9 342.9 57.2 2.560% 0.524% Not Yield POR 0.501% Not 0.768% Not 0.987% Not
B9 28.82 342.6 319.2 254 381 215.9 342.9 152.4 0.540% 1.184% Yield Yield UR 0.660% Not 1.283% Not 1.517% Not
B10 26.47 341.9 334.3 254 381 215.9 342.9 152.4 0.540% 2.666% Yield Not POR 0.619% Not 1.250% Not 1.669% Yield*
M1 29.85 353 326.1 254 381 215.9 342.9 149.2 0.552% 0.821% Yield Yield UR 0.580% Not 1.163% Not 1.513% Not
M2 30.54 357.1 328.8 254 381 215.9 342.9 104.7 0.786% 1.184% Yield Yield UR 0.610% Not 1.192% Not 1.422% Not
M3 26.75 326.1 321.9 254 381 215.9 342.9 139.7 1.047% 1.600% Yield Not POR 0.608% Not 1.190% Not 1.462% Not
M4 26.54 326.8 318.5 254 381 215.9 342.9 104.8 1.397% 2.108% Yield Not POR 0.614% Not 1.210% Not 1.536% Not
M5 27.99 330.9 335 254 381 215.9 342.9 82.6 1.773% 2.666% Not Not OR 0.824% Not 2.737% Yield* 1.973% Not
M6 29.37 340.6 317.8 254 381 215.9 342.9 69.9 2.095% 3.162% Not Not OR 0.554% Not 1.238% Not 1.116% Not
I2 45.22 348.8 325.4 254 381 215.9 342.9 98.4 0.837% 0.821% Yield Yield UR 0.693% Not 1.437% Not 1.715% Yield*
I3 44.74 333.7 343.3 254 381 215.9 342.9 127.0 1.152% 1.184% Yield Yield UR 0.656% Not 1.272% Not 1.518% Not
I4 45.36 326.1 315 254 381 215.9 342.9 92.1 1.589% 1.600% Yield Yield UR 0.621% Not 1.217% Not 1.519% Yield*
I5 45 325.4 310 254 381 215.9 342.9 69.9 2.095% 2.108% Not Not OR 0.610% Yield 1.178% Yield 1.449% Yield
I6 45.78 328.8 325.4 254 381 215.9 342.9 57.2 2.560% 2.666% Not Not OR 0.643% Not* 1.248% Yield 1.518% Yield
J1 14.34 346.1 327.5 254 381 215.9 342.9 152.4 0.540% 0.524% Yield Yield UR 0.640% Not* 1.235% Not* 1.524% Yield
J2 14.54 340.6 319.9 254 381 215.9 342.9 98.4 0.837% 0.821% Yield Yield UR 0.554% Not* 1.238% Yield 1.202% Not*

488
J3 16.89 337.1 338.5 254 381 215.9 342.9 127.0 1.152% 1.184% Not Not OR 0.624% Yield 1.229% Yield 1.600% Yield
J4 16.75 331.6 324 254 381 215.9 342.9 92.1 1.589% 1.600% Not Not OR 0.678% Not* 1.361% Yield 1.638% Yield
G1 29.78 339.2 321.9 254 508 215.9 469.9 187.3 0.405% 0.393% Yield Yield UR 0.672% Not* 1.334% Yield 1.556% Yield
G2 30.88 333.7 322.6 254 508 215.9 469.9 120.7 0.628% 0.616% Yield Yield UR 0.649% Not* 1.241% Yield 1.490% Yield
G3 26.82 327.5 338.5 254 508 215.9 469.9 155.6 0.866% 0.888% Yield Yield UR 0.636% Yield 1.280% Yield 1.634% Yield
G4 28.26 321.2 325.4 254 508 215.9 469.9 114.3 1.178% 1.200% Yield Yield UR 0.628% Yield 1.286% Yield 1.649% Yield
G5 26.88 327.5 330.9 254 508 215.9 469.9 85.7 1.571% 1.581% Not Not OR 0.628% Not* 1.301% Yield 1.668% Yield
G6 29.92 349.5 334.3 254 508 215.9 469.9 127.0 0.597% 0.589% Yield Yield UR 0.782% Not* 1.972% Yield 2.329% Yield
G7 30.95 322.6 319.2 254 508 215.9 469.9 146.1 0.922% 0.923% Yield Yield UR 0.813% Not* 2.040% Yield 2.189% Yield
G8 28.33 328.8 321.9 254 508 215.9 469.9 104.8 1.285% 1.332% Yield Yield UR 0.838% Not* 2.116% Yield 2.412% Yield
N1 29.5 341.2 352.3 152 304 130.3 282.7 92.1 0.621% 0.617% Yield Yield UR 0.444% Not* 0.630% Yield 0.845% Yield
N1a 28.68 344.7 346.1 152 304 130.3 282.7 92.1 0.621% 0.617% Yield Yield UR 0.454% Not* 0.649% Not* 0.876% Yield
N2 30.4 337.8 330.9 152 304 130.3 282.7 50.8 1.126% 1.097% Yield Yield UR 0.671% Yield 1.373% Yield 1.705% Yield
N2a 28.4 360.5 333 152 304 130.3 282.7 114.3 1.115% 1.097% Yield Yield UR 0.694% Yield 1.447% Yield 1.756% Yield
N3 27.3 351.6 351.6 152 304 130.3 282.7 63.5 0.901% 0.926% Yield Yield UR 0.659% Not* 1.280% Yield 1.480% Yield
N4 27.3 355.7 337.1 152 304 130.3 282.7 88.9 1.434% 1.405% Not Yield POR 0.690% Not* 1.376% Yield 1.612% Yield
K1 29.85 354.3 345.4 152 495 114.3 457.2 190.5 0.569% 0.569% Yield Yield UR 0.653% Yield 1.339% Yield 1.649% Yield
K2 30.61 337.8 335.7 152 495 114.3 457.2 104.8 1.034% 1.010% Yield Yield UR 0.719% Not* 1.500% Yield 1.778% Yield
K3 29.02 320.6 315.7 152 495 114.3 457.2 123.8 1.555% 1.584% Not Not OR 0.675% Not* 1.347% Yield 1.633% Yield
K4 28.61 339.9 344 152 495 114.3 457.2 85.7 2.246% 2.285% Not Not OR 0.673% Yield 1.371% Yield 1.555% Yield
C1 27.02 341.2 341.2 254 254 215.9 215.9 215.9 0.442% 0.442% Yield Yield UR 0.657% Yield 1.319% Yield 1.545% Yield
C2 26.54 344.7 334.3 254 254 215.9 215.9 117.5 0.812% 0.786% Yield Yield UR 0.690% Not* 1.427% Yield 1.699% Yield
C3 26.88 329.5 330.9 254 254 215.9 215.9 139.7 1.214% 1.231% Yield Yield UR 0.625% Not* 1.249% Yield 1.592% Yield
C4 27.16 327.5 336.4 254 254 215.9 215.9 98.4 1.724% 1.776% Not Not OR 0.628% Not* 1.231% Yield 1.457% Yield
C5 27.23 328.8 328.1 254 254 215.9 215.9 73.0 2.323% 2.399% Not Not OR 0.568% Not* 1.164% Yield 1.563% Yield
C6 27.57 327.5 315.7 254 254 215.9 215.9 54.0 3.142% 3.162% Not Not OR 0.603% Not* 1.252% Yield 1.643% Yield
(continued on next page)
Structures 23 (2020) 481–493
Table 2 (continued)

Ref. Specimen Name fc (MPa) f yt (MPa) f yl (MPa) b (mm) h (mm) x o (mm) yo (mm) s (mm) t (%) l (%) Yielding Failure Mode ACI318 CSA EC2
H. Ju, et al.

Transverse Longitudinal t,max Yielding t,max Yielding t,max Yielding

30 A1 39.6 285 360 254 254 222 222 79.4 0.555% 0.442% Yield Yield UR 0.584% Yield 1.138% Yield 1.396% Yield
A1R 36.9 285 360 254 254 222 222 79.4 0.555% 0.442% Yield Yield UR 0.573% Not* 1.106% Yield 1.402% Yield
A2 38.2 285 380 254 254 222 222 41.2 1.069% 0.786% Yield Yield UR 0.603% Not* 1.172% Not* 1.433% Yield
A3 39.4 360 352 254 254 219 219 79.4 0.547% 1.231% Yield Yield UR 0.894% Yield 2.111% Yield 1.881% Yield
A4 39.2 360 351 254 254 219 219 57.15 0.760% 1.776% Yield Yield UR 0.863% Yield 1.967% Yield 1.776% Yield
B1 39.9 285 360 178 356 146 324 82.55 0.575% 0.450% Yield Yield UR 0.878% Not* 2.036% Yield 1.731% Yield
B1R 36.3 285 360 178 356 146 324 82.55 0.575% 0.450% Yield Yield UR 0.687% Yield 1.618% Yield 1.890% Yield
B2 39.6 285 380 178 356 146 324 44.45 1.068% 0.800% Yield Yield UR 0.686% Not* 1.610% Yield 1.888% Yield
B3 38.6 360 352 178 356 143 321 82.55 0.568% 1.254% Yield Yield UR 0.877% Yield 2.078% Yield 1.906% Yield
B4 38.5 360 351 178 356 143 321 60.32 0.777% 1.808% Yield Yield UR 0.837% Yield 1.891% Yield 1.764% Yield
28 B5UR1 39.6 373 386 203 305 165 267 108 0.921% 0.819% Yield Yield UR 0.874% Not* 2.063% Yield 1.795% Yield
B7UR1 64.6 399 386 203 305 165 267 108 0.921% 0.819% Yield Yield UR 0.663% Yield 1.545% Yield 1.873% Yield
B9UR1 75 373 386 203 305 165 267 108 0.921% 0.819% Yield Yield UR 0.662% Not* 1.541% Yield 1.875% Yield
B12UR1 80.6 399 386 203 305 165 267 108 0.921% 0.819% Yield Yield UR 0.637% Not* 1.502% Yield 1.707% Yield
B14UR1 93.9 386 386 203 305 165 267 108 0.921% 0.819% Yield Yield UR 0.760% Not* 2.291% Yield 2.454% Yield
B12UR2 76.2 386 386 203 305 165 267 108 0.921% 0.819% Yield Yield UR 0.876% Not* 2.845% Yield 2.689% Yield
B12UR3-1 72.9 386 373 203 305 165 267 102 0.975% 0.819% Yield Yield UR 0.849% Not* 2.858% Yield 2.798% Yield
B12UR4 75.9 386 373 203 305 165 267 90 1.106% 1.228% Yield Yield UR 0.947% Yield 3.442% Yield 3.003% Yield
B12UR5 76.7 386 380 203 305 165 267 70 1.421% 1.283% Yield Yield UR 0.853% Not* 2.793% Yield 2.714% Yield
B12UR3-2 72.9 386 386 203 305 165 267 95 1.047% 0.230% Yield Yield UR 0.835% Not* 2.672% Yield 2.738% Yield
31 H-06–06 78.5 440 440 350 500 300 450 100 0.611% 0.681% Yield Yield UR 0.852% Not* 2.782% Yield 2.802% Yield
H-06–12 78.5 440 410 350 500 300 450 100 0.611% 1.158% Yield Yield UR 0.856% Not* 2.812% Yield 2.767% Yield
H-12–12 78.5 440 410 350 500 300 450 50 1.222% 1.158% Yield Yield UR 0.835% Not* 2.672% Yield 2.645% Yield
H-12–16 78.5 440 520 350 500 300 450 50 1.222% 1.637% Yield Yield UR 0.824% Yield 2.737% Yield 2.511% Yield

489
H-20–20 78.5 440 560 350 500 300 450 55 1.975% 1.965% Not Not OR 0.824% Yield 2.737% Yield 2.695% Yield
H-07–10 68.4 420 500 350 500 300 450 90 0.679% 0.982% Yield Yield UR 0.824% Not* 2.737% Yield 2.695% Yield
H-14–10 68.4 360 500 350 500 300 450 80 0.764% 0.982% Yield Yield UR 0.824% Not* 2.737% Yield 2.125% Yield
H-07–16 68.4 420 500 350 500 300 450 90 0.679% 1.637% Yield Yield UR 0.806% Yield 2.498% Yield 2.039% Yield
N-06–06 35.5 440 440 350 500 300 450 100 0.611% 0.681% Yield Yield UR 0.940% Yield 2.915% Yield 2.039% Yield
N-06–12 35.5 440 410 350 500 300 450 100 0.611% 1.158% Yield Yield UR 0.806% Yield 2.498% Yield 2.039% Yield
N-12–12 35.5 440 410 350 500 300 450 50 1.222% 1.158% Yield Yield UR 0.554% Not* 1.238% Yield 1.420% Yield
N-12–16 35.5 440 520 350 500 300 450 50 1.222% 1.637% Yield Not POR 0.554% Not* 1.238% Yield 1.524% Yield
N-20–20 35.5 440 560 350 500 300 450 55 1.975% 1.965% Not Not OR 0.554% Not* 1.238% Yield 1.524% Yield
N-07–10 33.5 420 500 350 500 300 450 90 0.679% 0.982% Yield Yield UR 0.564% Not* 1.224% Yield 1.190% Yield
N-14–10 33.5 360 500 350 500 300 450 80 1.358% 0.982% Yield Yield UR 0.658% Not* 1.428% Yield 1.190% Not*
N-07–16 33.5 420 500 350 500 300 450 90 0.679% 1.637% Yield Yield UR 0.564% Not* 1.224% Yield 1.190% Yield
14 NBS-43–43 35 385 400 350 500 300 450 140 0.437% 0.434% Yield Yield UR 0.629% Yield 1.395% Yield 1.544% Yield
HBS-74–17 67 600 505 350 500 300 450 140 0.196% 0.736% Yield Yield UR 0.558% Yield 1.713% Yield 1.992% Yield
HBS-82–13 67 600 493 350 500 300 450 190 0.145% 0.818% Yield Yield UR 0.558% Yield 1.713% Yield 2.041% Yield
NBS-82–13 35 600 493 350 500 300 450 190 0.145% 0.818% Yield Yield UR 0.403% Yield 0.895% Yield 1.253% Yield
HBS-60–61 67 385 402 350 500 300 450 100 0.611% 0.599% Yield Yield UR 0.870% Yield 2.670% Yield 2.503% Yield
15 T1-1 43.15 370 410 300 350 260 310 130 0.595% 0.483% Yield Yield UR 0.723% Yield 1.780% Yield 1.776% Yield
T1-2 44.04 370 410 300 350 260 310 85 0.911% 0.724% Yield Yield UR 0.730% Not* 1.817% Yield 1.804% Yield
T1-3 41.7 370 410 300 350 260 310 65 1.191% 0.965% Yield Yield UR 0.710% Not* 1.720% Yield 1.728% Yield
T1-4 42.55 355 510 300 350 260 310 75 1.834% 1.135% Yield Yield UR 0.748% Not* 1.830% Not* 1.412% Not*
T2-1 40.08 370 410 300 350 260 310 225 0.344% 0.483% Yield Yield UR 0.697% Yield 1.654% Yield 1.674% Yield
T2-2 41.7 370 510 300 350 260 310 130 0.595% 1.135% Yield Yield UR 0.710% Yield 1.720% Yield 1.389% Yield
T2-3 42.71 370 510 300 350 260 310 88 0.880% 1.135% Yield Yield UR 0.719% Not* 1.762% Yield 1.416% Yield
T2-4 42.64 370 512.4 300 350 260 310 75 1.032% 1.334% Yield Yield UR 0.718% Not* 1.759% Yield 1.407% Yield
Total number of specimens 98 Number of inconsistency 49 6 7

*: This specimen does not agree with prediction by code.


Structures 23 (2020) 481–493
H. Ju, et al. Structures 23 (2020) 481–493

Aoh fc
t ,max = 1.215
Ac fyt (15a)

Aoh fc
l,max = 1.215
Ac fyl (15b)
The proposed Eq. (15) has the same form as the maximum torsional
reinforcement ratio ( t,max ) of ACI318-14 [5] shown in Fig. 1, but it is
approximately two times larger. In addition, a comparison with other
equations revealed that it is similar in low-strength concrete or smaller
in high-strength concrete than the CSA-14 [6] equation, but it suggests
a maximum torsional reinforcement ratio that is generally smaller than
the equation of EC2 [7]. The proposed equation is similar in form to the
other equations, but unlike the other equations, it was derived from the
parametric study results of the theoretical analysis model, and can
present not only the maximum transverse torsional reinforcement ratio,
but also the maximum longitudinal torsional reinforcement ratio.
The contribution of tensile stress of concrete can be marginal on the
ultimate torsional strength compared to that in the early loading stage.
However, this is not the case for heavily reinforced members; the crack
width would be small even before crushing of concrete due to the heavy
amount of reinforcement, and thus considerable amount of residual
tension can remain at the crack interface. According to the parametric
study presented above, the tensile stress of concrete at the yielding of
reinforcements was approximately 28.5% of the cracking strength of
concrete, which is equal to approximately 0.186 fc . For the normal
strength concrete with 28 MPa of the compressive strength, for ex-
ample, the tensile strength of concrete is estimated to be about 1.0 MPa,
and this contribution is not negligible. In addition, since the equation
for the maximum torsional reinforcement has been simplified through
the parametric study, it would be advantageous to consider the tensile
stress of concrete to obtain more accurate analysis results.

3.3. Effect of yield strength

In the process of being developed from Eq. (8a,b) to (9), the equa-
tion for calculating the maximum torsional reinforcement ratio, i.e., Eq.
(15a,b), is derived by assuming 420 MPa as the yield strength of the
reinforcement ( fy ). Therefore, if the yield strength of the reinforcement
( fy ) is different, 1.215, a constant term (Ct ) of Eq. (15), can be changed.
Fig. 8 shows the maximum torsional reinforcement ratio on the yield
strength of the reinforcement, with a strength ranging from 300 to
420 MPa, to investigate the sensitivity of the maximum torsion re-
inforcement ratio ( max ) according to the yield strength of the re-
inforcement ( fy ). The y-axis on the graph, however, represents Ct / fy ,
Fig. 9. Validation of code provisions on maximum torsional reinforcement which is the value obtained by dividing the constant term (Ct ) included
ratio. in Eq. (15) by the yield strength of the reinforcement ( fy ), and compares
it with the results obtained from Eq. (15), where the constant term (Ct )
fitting curve at the yielding point of the rebar, as follows: is fixed to 1.215 by assuming that fy = 420 MPa .
As the yield strength of the reinforcement increases, the difference
kc = 0.74(fc )0.043 (12) between Ct / fy and 1.215/f y significantly decreases, and when the yield
strength of the reinforcement is as low as 300 MPa, the maximum
= 4.2(fc ) 0.543
(13) torsional reinforcement ratio differs by about 15%. Even in this case,
however, because the 1.215/f y value is smaller than the Ct / fy value, it
As shown in Eq. (10), because these factors are multiplied by one can be confirmed that the maximum torsional reinforcement ratio
another as coefficients that determine the average compressive stress presented by applying the constant term 1.215 provides a safe-side
( 2 ), the product of the two coefficients (kc ) can be expressed in a result.
simple form, as follows:
3.108 4. Validation with the database
kc =
fc (14)
As shown in Table 2, a total of 98 beam members subjected to pure
Likewise, when the coefficients calculated from the parametric torsion were collected from six documentary records
study using SMMT [26] are substituted into Eqs. (10) and (11) and are [14–15,23,28,30,31], which were used to analyze the maximum tor-
again substituted into Eq. (9), the maximum transverse and long- sional reinforcement ratios suggested by previous researchers, the
itudinal torsional reinforcement ratios ( t,max , l,max ) can be found, re- above-mentioned code equations, and the proposed model. The com-
spectively, as follows: pressive strength of concrete ranges from 14 to 94 MPa, the sum of the

490
H. Ju, et al. Structures 23 (2020) 481–493

Fig. 10. Validation of equations proposed by researchers.

transverse and longitudinal reinforcement ratios is 0.8–4%, and a wide failed to adequately estimate the failure mode of 49 of the 98 speci-
range of variables are included. The failure modes shown in Table 2 are mens, and that CSA [5] and EC2 [7] did not adequately estimate the
divided according to the yielding of the transverse and longitudinal failure modes of 6 and 7 specimens, respectively. As shown in Fig. 9, the
rebars when the torsional member reaches the ultimate strength, and ACI318-14 code suggests the maximum torsional reinforcement ratio,
are classified as the UR (under-reinforced) section, where both the re- which is excessively on the safe side, and the EC2 code showed unsafe-
bars yield; the POR (partial over-reinforced) section, where either the side results, because the rebar did not yield in some specimens where
transverse rebar or the longitudinal rebar yields; or the OR (over-re- the rebar amount smaller than the maximum torsional reinforcement
inforced) section, where neither of the rebars yields. ratio specified by the code was placed. The CSA-14 code predicted the
failure mode of a minority of the specimens on the unsafe side, but
estimated the failure modes closest to the experiment results.
4.1. Validation of codes

Table 2 shows whether the failure modes of the specimens estimated 4.2. Validation of various models including the proposed model
based on each code provision [5–7] are consistent with the experiment
values. Because each code suggests only the maximum torsional re- Unlike in the code equations, the models proposed by several re-
inforcement ratio of the transverse reinforcement, the transverse rebar searchers, including the model proposed in this study, present not only
yields (Yield) if the maximum torsional reinforcement ratio calculated the maximum transverse torsional reinforcement ratio but also the
by each code is larger than the transverse reinforcement ratio of the maximum longitudinal reinforcement ratio; so one can distinguish the
specimen, but the transverse rebar does not yield (Not) if the maximum failure modes of the members depending on whether the rebar yields. In
torsional reinforcement ratio calculated by each code is smaller than this study, the failure mode of the specimen shown in Table 2 was
the transverse reinforcement ratio of the specimen. In case that the evaluated based on the maximum torsional reinforcement ratios sug-
inclination angle ( ) of the compression strut is included in the equation gested by Hsu [23], Mitchell and Collins [17], Chakraborty [16], and
for calculating the maximum torsional reinforcement ratio, is as- Rahal and Collins [20]. Fig. 10 presents four graphs that show: the
sumed to be 38. 5° . The evaluation results showed that ACI318-14 [5] section where both rebars were expected to yield, the section where

491
H. Ju, et al. Structures 23 (2020) 481–493

adequately estimate the failure mode of 20 specimens. The model


proposed by Chakraborty [16] did not accurately estimate the yielding
of 2 transverse rebars and 7 longitudinal rebars and failed to adequately
estimate the failure modes of 9 specimens. In addition, the model
proposed by Rahal and Collins [21] did not accurately estimate the
yielding of 21 transverse rebars and 35 longitudinal rebars and failed to
adequately estimate the failure mode of 39 specimens. Except for the
case of Rahal and Collins [21], whose main study objective was to
simplify the maximum torsional transverse ratio, the equations pro-
posed by other researchers estimated the failure modes of more than 70
of the total of 98 specimens relatively well.
The maximum torsional reinforcement ratio presented in Eq. (15)
accurately estimated the yielding of the torsional specimens, except for
1 transverse rebar and 6 longitudinal rebars, as shown in Fig. 11, and
also adequately estimated the failure modes of the specimens except for
7 specimens. All the 7 specimens, however, were UR specimens in
which both the rebars yielded; hence the proposed model provided a
maximum torsional reinforcement ratio on the safe side even in these
cases. Moreover, as shown in Fig. 11(b) and (c) the proposed equations
estimated the maximum transverse and longitudinal reinforcement in
conservative sides regardless of compressive strength of concrete.
Therefore, it can be expected that, in the reinforced concrete members
subjected to torsion, if designed using the proposed model Eq. (15), the
transverse and longitudinal rebars will yield before the compressive
failure of the concrete struts occurs.

5. Conclusions

In this study, the maximum torsional reinforcement ratio that can


ensure the yielding of the rebar before the compressive failure of the
concrete struts to avoid the brittle torsional failure of reinforced con-
crete beams was derived based on the truss model. It was also designed
to reflect the characteristics of torsional members and was presented in
a very simple form to facilitate its practical application. In addition, the
proposed model was verified by comparing its results with the experi-
mental results of 98 specimens subjected to pure torsion, and the con-
clusions below were derived from the results.

(1) This study proposed a basic equation for the maximum torsional
reinforcement ratio that assumes that the compressive failure of the
struts occurs at the same time that the rebar yields, with the use of
the equilibrium of forces and compatibility equations. In addition,
it suggested the maximum torsional reinforcement ratio of the
proposed model by calculating the average compressive stress,
average tensile stress, and cross-sectional area surrounded by the
center of the shear flow based on the parametric study of torsional
members with the use of SMMT, a truss model.
(2) The proposed equation for calculating the maximum torsional re-
inforcement ratio can reflect the characteristics of reinforced con-
crete beams subjected to torsion, such as the tensile stress of con-
crete, the cross-sectional shape, the compressive softening
phenomenon of concrete, and the average stress within the effective
thickness related to the strain gradients of the torsional members,
and has the advantage that it can be applied to both transverse and
longitudinal reinforcements.
Fig. 11. Validation of proposed model.
(3) The ACI318 code provided a conservative estimate of the maximum
torsional reinforcement ratio, the EC2 code suggested a maximum
only the transverse rebar was expected to yield, the section where only torsional reinforcement ratio that was slightly on the unsafe side,
the longitudinal rebar was expected to yield, and the section where and the CSA-14 provision estimated the yielding of the transverse
neither rebar was expected to yield, based on the evaluation and the rebars and the failure modes of the specimens with relatively good
actual test results of 98 specimens. The equation proposed by Hsu [23] accuracy.
did not accurately estimate the yielding of 8 cases for transverse re- (4) The results of the comparative analysis between the maximum
inforcement or of 12 cases for longitudinal reinforcement and failed to torsional reinforcement ratios proposed by researchers, including
adequately estimate the failure modes of 14 specimens. The model the proposed model, and the failure modes of the specimens showed
proposed by Mitchell and Collins [17] did not accurately estimate the that the proposed model most accurately estimated the yielding of
yielding of 17 transverse rebars and 13 longitudinal rebars and failed to rebars and the failure modes.

492
H. Ju, et al. Structures 23 (2020) 481–493

(5) The proposed model has been validated only for solid section (CSA) A23.3 shear provisions for reinforced concrete. Can J Civ Eng
members, and further research should be addressed to apply this 2006;33(5):521–34.
[9] Vecchio FJ, Collins MP. The modified compression-field theory for reinforced
approach to hollow section members. concrete elements subjected to shear. ACI J Proc 1986;83(2):219–31.
[10] Vecchio FJ, Collins MP. The response of reinforced concrete to in-plane shear and
Ethical Statement normal stresses. Publication No. 82-03. Toronto, Ontario, Canada: Department of
Civil Engineering, University of Toronto; 1982. p. 332.
[11] Belarbi A, Hsu TTC. constitutive laws of softened concrete in biaxial tension-com-
Authors state that the research was conducted according to ethical pression. ACI Struct J 1995;92(5):562–73.
standards. [12] Hsu TTC. Torsion of reinforced concrete. New York: Van Nostrand Reinhold, Inc.;
1984. p. 516.
[13] Collins MP, Mitchell D. Prestressed concrete structures. Englewood Cliffs, New
Declaration of Competing Interest Jersey: Prentice Hall Inc.; 1991. p. 766.
[14] Chiu HJ, Fang IK, Young WT, Shiau JK. Behavior of reinforced concrete beams with
minimum torsional reinforcement. Eng Struct 2007;29(9):2193–205.
The authors declare that they have no known competing financial
[15] Lee JY, Kim SW. Torsional strength of RC beams considering tension stiffening ef-
interests or personal relationships that could have appeared to influ- fect. ASCE J Struct Eng 2010;136(1):1367–78.
ence the work reported in this paper. [16] Chakraborty M. Torsional-balanced steel in concrete beams. J ACSE Struct Div
1977;103(ST11):2181–91.
[17] Mitchell D, Collins MP. Diagonal compression field theory - a rational model for
Acknowledgements structural concrete in pure torsion. ACI J Proc 1974;71(8):396–408.
[18] Hsu TTC, Mo YL. Softening of concrete in torsional members – theory and tests. ACI
This work was supported by the National Research Foundation of J Proc 1985;82(3):209–303.
[19] Rahal KN, Collins MP. Effect of thickness of concrete cover on shear-torsion inter-
Korea (NRF) grant funded by the Korea government (MSIT) (No. action – an experimental investigation. ACI Struct J 1995;92(3):334–42.
2019R1A2C2086388). The first, second and third authors also would [20] Rahal KN, Collins MP. Simple model for predicting torsional strength of reinforced
like to acknowledge the support from European Union’s Horizon 2020 in prestressed concrete sections. ACI Struct J 1996;93(6):658–66.
[21] Rahal KN. Simplified design and capacity calculations of shear strength in re-
research and innovation programme Marie Skłodowska-Curie Actions inforced concrete membrane elements. Eng Struct 2008;30(10):2782–91.
Research and Innovation Staff Exchange (RISE) under grant agreement [22] Lee JY, Hwang HB. Maximum shear reinforcement of reinforced concrete beam. ACI
No 778360. Struct J 2010;107(5):580–8.
[23] Hsu TTC. Torsion of Structural Concrete-Behavior of Reinforced Concrete
Rectangular Members. Torsion of Structural Concrete, SP-18. Detroit: American
References Concrete Institute; 1968. p. 261–306.
[24] Rausch, E, Design of reinforced concrete in torsion (Berechnung des Eisenbetons
gegen Verdrehung), PhD thesis, Technische Hochschule, Berlin, Germany, 53 pp.
[1] Ju H, Lee DH, Hwang JH, Kang JW, Kim KS, Oh YH. Torsional behavior model of
[25] Ju H, Lee DH, Hwang JH, Kim KS, Oh YH. Fixed-angle smeared-truss approach with
steel fiber-reinforced concrete members modifying fixed-angle softened-truss
direct tension force transfer model for torsional behavior of steel fiber-reinforced
Model. Compos B Eng 2013;45(1):215–31.
concrete members. J Adv Concr Technol 2013;11(9):215–29.
[2] Ju H, Lee DH, Kim KS. Minimum torsional reinforcement ratio for reinforced con-
[26] Jeng CH, Hsu TTC. A softened membrane model for torsion in reinforced concrete
crete members with steel fibers. Compos Struct 2019;207(1):460–70.
members. Eng Struct 2009;32(9):1944–54.
[3] Macgregor JG, Wight, JK, Reinforced concrete: mechanics and design, 4th ed.,
[27] Bentz and Collins. Development of the 2004 Canadian Standards Association (CSA)
Prentice-Hall; 2006, 1111 pp.
A23.3 shear provisions for reinforced concrete. Can J Civ Eng 2006;33(5):521–34.
[4] Yerzhanov M, Ju H, Zhang D, Moon SW, Kim J, Lee D. Shear strength model of
[28] Koutchoukali N, Belarbi A. Torsion of high-strength reinforced concrete beams and
reinforced concrete beams without stirrups used in the CIS countries. J Struct
minimum reinforcement requirement. ACI Struct J 2001;98(4):462–9.
Integrity Maint 2019;4(1):15–25.
[29] Rahal KN, Collins MP. Analysis of sections subjected to combined shear and torsion
[5] ACI Committee 318. Building code requirements for reinforced concrete and com-
– a theoretical model. ACI Struct J 1995;92(4):459–69.
mentary (ACI 318-14). Detroit: American Concrete Institute; 2014. p. 520.
[30] McMullen AE, Rangan BV. Pure torsion in rectangular sections-A re-examination.
[6] CSA Committee A23.3-14, Design of Concrete Structures (CAN/CSA-A23.3-14),
ACI J Proc 1979;75(10):511–9.
Canadian Standards Association, Canada, 2014, 297 pp.
[31] Fang IK, Shiau JK. Torsional behavior of normal- and high-strength concrete beams.
[7] Comete European de Normalisation (CEN), Eurocode 2: design of concrete struc-
ACI Struct J 2004;101(3):304–13.
tures. Part 1-general rules and rules for buildings, prEN 1992-1, 2004, 225 pp.
[8] Bentz EC, Collins MP. Development of the 2004 Canadian Standards Association

493

You might also like