Magazine of Concrete Research, 2003, 55, No.

1, February, 53–63

A direct method for deep beams with web

reinforcement
K. H. Tan
, C. Y. Tang
and K. Tong

Nanyang Technological University

This paper takes account of the effect of web reinforcement, be it in vertical, horizontal, inclined or orthogonal
configurations. This is because test results have indicated that the contribution of web reinforcement to shear
strength is indeed significant and should not be ignored. An original and rational strut-and-tie model for the tensile
contribution of both reinforcement and concrete is proposed, which takes account of their influence on the principal
tensile stress. The proposed model is verified against three case studies of a total of 116 beams. Generally, the
predictions are not only accurate and consistent in each case study, but also conservative.

la height of bottom node (Fig. 1)

Notation
lb width of support bearing plates (Fig. 1)
a shear span measured between concentrated load lc height of top node (Fig. 1)
and support point (Fig. 1) le effective span measured between centre-to-
Ac area of concrete section centre of supports
As area of longitudinal reinforcement k weighting factor taking account of positional
Astr cross-sectional area of diagonal strut influence of reinforcement
Aw area of web reinforcement S tensile force in principal direction (Fig. 2)
bw width of deep beam T tension force in the horizontal tie (Fig. 1)
d effective depth (Fig. 1) Tw force carried by vertical web reinforcement
dw distance from the beam top to the intersection v concrete utilisation factor
of web reinforcement with the line connecting Vexp experimental shear strength
the support centre and the load centre (Fig. 1) Vn predicted nominal shear strength
f1 principal tensile stress Łw angle between the web reinforcement and the
f2 principal compressive stress horizontal axis of beams at the intersection of
Fc compressive force in the diagonal strut (Fig. 1) the reinforcement and the diagonal strut (Fig. 1)
f 9c compressive strength of concrete Łs angle between the longitudinal tension
f ct tensile strength of concrete reinforcement and the diagonal strut (Fig. 1)
ft combined tensile strength of reinforcement and rh ratio of horizontal web reinforcement
concrete ri ratio of inclined web reinforcement
fy yield strength of longitudinal steel rs ratio of longitudinal reinforcement
reinforcement rv ratio of vertical web reinforcement
f yh yield strength of horizontal web reinforcement
The ACI code does not include the design of
f yv yield strength of vertical web reinforcement deep beams with inclined web reinforcement
f yw yield strength of web reinforcement
h overall height of deep beam (Fig. 1)
Introduction

School of Civil & Structural Engineering, Nanyang Technologicall
Recent years have seen a surge in interest in the
University, Singapore 639798.
applications of the strut-and-tie approach to the design
1–8
(MCR 1986) Paper received 6 November 2001; last revised 1 March of disturbed regions in concrete structures. In fact,
2002; accepted 21 May 2002 the approach has been codified into the CEB-FIP MC
53

0024-9831 # 2003 Thomas Telford Ltd

Downloaded by [ University of Ottawa Library System] on [12/09/16]. Copyright © ICE Publishing, all rights reserved.
 Tan et al.
a
Vn Vn resistance than horizontal web steel of the same area.
On the other hand, horizontal web steel is more effec-
14,16
lc tive for deep beams with a=d , 1. This also ex-
dw plains why inclined web steel, which provides the
θw greatest lever arm, is the most effective to enhance the
d h 16
ultimate shear resistance. Tan et al. recommend
the use of orthogonal web reinforcement, due to the
Fc
θs practical difficulty of providing inclined web reinforce-
T
la ment on site.
Against this background, the authors propose a sim-
Vn Vn
ple and direct method to account for the beneficial
lb effect of web reinforcement for the ultimate shear
strength of deep beams. A total of 116 experimental
Fig. 1. Strut-and-tie model for simply supported deep beams specimens are used to verify the proposed model. The
predictions are also compared to the ACI design equa-
17 9
tions, the FIP-CEB MC 90 and the Canadian CSA
10
Vn Building Code. The web reinforcement is represented
by a force to resist the diagonal splitting of the strut;
the further the web reinforcement from the beam soffit,
the less effective it becomes. In general, the proposed
θw model offers better agreement with the test results.
y = dw
Tw = fyAyw Proposed strut-and-tie model
Sw = Twsin(θs + θw) A strut-and-tie model for simply supported, partially
y=d
pre-stressed deep beams is given in Fig. 1. From the
θs equilibrium of forces
T
Vn
Fc ¼ (1)
y=h sin Łs
Linear Vn
stress T¼ (2)
block S = Tsinθs tan Łs
fmax = k(S/(A/sinθs))
where Fc is the compressive force in the diagonal strut;
Fig. 2. Determination of tensile stress f 1 at the bottom nodal
T is the tensile force in the horizontal tie; Vn is the
zone nominal shear strength; Łs is the inclined angle of the
diagonal strut.
Thus, the principal compressive stress f 2 in the diag-
9 10 onal strut can be computed by
90 and Canadian CSA Building Code, that clearly
stipulate that regions of members in which the plane Fc
f2 ¼ (Compressive stress) (3)
section assumption of flexural theory is not applicable A str
shall be proportioned for shear and torsion using the where the term Astr is the cross-sectional area of the
strut-and-tie model. However, the existing strut-and-tie strut
models often neglect the beneficial effect of web re-
inforcement. Tests in the 1970s by Kong et al.
11–13 A str ¼ b w (l a cos Łs þ l b sin Łs ) (4)
revealed that inclined web reinforcement greatly en- where bw is the web width; la is the depth of the
hanced the shear strength of deep beams. Smith and bottom nodal zone (Fig. 1); lb is the width of the
14
Vantsiotis also revealed that web reinforcement had support bearing plates (Fig. 1).
moderate effect on ultimate shear strength. From a The principal tensile stress f 1 across the diagonal
systematic experimental programme on deep beams by strut can be determined by
15–16
the first author, the tensile force in bottom main
kT sin Ł
steel is reduced due to the presence of web reinforce- f1 ¼ (5)
ment. The effectiveness of web steel is governed by its A c = sin Ł
8
lever arm to the top loading point; the greater the lever where Ac is the cross-sectional area of deep beam;
arm, the more effective is the web steel. Thus, for deep
T sin Łs
beams with a=d > 1:00, vertical web steel that has a
greater lever arm is more effective in restraining the A c =sin Łs
diagonal crack width and increasing the ultimate shear is the average tensile stress across the diagonal strut
54 Magazine of Concrete Research, 2003, 55, No. 1

Downloaded by [ University of Ottawa Library System] on [12/09/16]. Copyright © ICE Publishing, all rights reserved.
 A direct method for deep beams with web reinforcement

due to the component of T in the principal tensile duced to account for the positional influence of web
direction of the bottom nodal zone, T sin Łs (Fig. 2). reinforcement. It should be noted that the further the
The factor k in equation (5) is to account for the web reinforcement is from the beam soffit, the less
non-uniformity of stress distribution across the diagonal efficient it is to resist the splitting of diagonal strut, as
strut. Assumptions on stress distribution acting perpen- the corresponding lever arm is smaller. For web reinfor-
dicular to the inclined strut are necessary, as it is cement located in the top horizontal strut whose lever
difficult to determine the exact magnitude of f 1 and arm is 0, it experiences compression and therefore, is
the plane-section-remains-plane theory does not hold not effective to reduce the tensile stress across the
for deep beams. A linear stress distribution is found to diagonal strut. The third term represents the tensile
8,18
be in good agreement with test results. Thus, a value contribution from concrete. For conservatism, the term
of 2 is adopted in the current paper. f ct can be made equal to zero.
Conventionally, a utilisation factor v is applied to From equations 1, 2, 3, 5 and 6, an expression for
concrete strength f 9c , which takes account of the ultimate shear strength V n can be derived
strength reduction due to cracks in the diagonal 1
3,4,19,20 Vn ¼ (9)
strut. An original failure criterion that assumes a sin 2Łs 1
linear interaction relationship between f 1 and f 2, as þ
8 f t Ac f 9c A str sin Łs
modified from the Mohr–Columb theory, is proposed
f2 f1 In equation (9), the inclined angle of the diagonal
v¼ ¼1 <1 (6) strut Łs (Fig. 1) can be calculated from
f 9c ft
la lc
where f 1 and f 2 are the respective principal tensile and h 
tan Łs ¼ 2 2 (10)
compressive stresses; f 9c is the cylinder compressive a
strength, and it represents the maximum compressive
capacity; f t is the tensile strength contribution of both where lc is the depth of the top nodal zone and its
reinforcement and concrete, and it represents the maxi- value can be derived from the equilibrium of the top
mum tensile capacity. nodal zone
This failure criterion is adopted because it offers a V n =tan Łs
simple linear interactive relationship between two limit lc ¼ (11)
f 9c b w
failure modes, viz. concrete crushing (brittle failure)
The value of Vn can be determined by using equations
and yielding of steel reinforcement (ductile failure).
7, 9, 10 and 11 through iterations. In practice, however,
The terms f 2 and f 1 represent the actual applied stres-
the iterative procedure can be simplified by assuming
ses, which lie in between the two ultimate strengths f 9c
lc ¼ la . By doing so, it is found that the error intro-
and f t . The term f t in equation (6) is the combined
duced in Vn is generally less than 3%. This is because
tensile strength contribution of both reinforcement and
the width of horizontal strut lc in equation 10 is about
concrete, and is given by
ten times smaller than the overall height h. Therefore,
2A s f y sin Łs 2A w f yw sin(Łs þ Łw ) d w the inclined angle of the diagonal strut Łs can be still
ft ¼ þ  þ f ct
A c =sin Łs A c =sin Łs d accurately predicted even though a very rough value of
(7) lc is assumed. The steps for determining the shear
strength Vn directly are summarised as follows
where As and Aw are the respective total areas of long-
itudinal and web reinforcement; f y and f yw are the Step 1 Determine Łs from equation 10, assuming
respective yield strengths of longitudinal and web rein- lc ¼ la , where la ¼ 2(h  d).
forcement; f ct is the tensile strength of concrete Step 2 Determine f t from equation 7.
pffiffiffiffiffi
f ct ¼ 0:5 f 9c (SI) (8) Step 3 Determine Vn from equation 9.

The first term in equation 7 represents the tensile

capacity of longitudinal steel reinforcement and is de-
rived in a similar fashion as the term f 1 in equation 5, Comparison of test results with case
except that the full strength of longitudinal reinforce-
studies
ment As f y is used in place of T . The second term
represents the tensile capacity of inclined web reinfor- In the following, the proposed strut-and-tie model is
cement at an angle Łw to the horizontal axis in Fig. 2. verified using three case studies. Details of the deep
It takes account of different positions and arrangements beam specimens and predictions can be found in Tables
of web reinforcement, be it vertical, horizontal, inclined 1–6. The predictions are compared to those by the
17 9
or orthogonal. From the geometry of strut-and-tie mod- American ACI code, the CEB-FIP MC 90, and the
10
el, the tensile force contribution by web reinforcement Canadian CSA Building Code. It should be noted that
in the direction of f t is A w f yw sin(Łs þ Łw ). For vertical the material partial safety factors in these code provi-
web reinforcement, Łw ¼ 908. The factor d w =d is intro- sions have been set to unity for comparison purposes.
Magazine of Concrete Research, 2003, 55, No. 1 55

Downloaded by [ University of Ottawa Library System] on [12/09/16]. Copyright © ICE Publishing, all rights reserved.
 Tan et al.
15
Table 1. Case study 1 – details of 19 deep beams
Beam a: mm le : mm a=d le =d rs rv rh f cu : MPa f 9c : MPa
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
A-0·27-2·15 125 1000 0·27 2·15 0·0123 0·0048 0·0 67·80 58·84
A-0·27-3·23 125 1500 0·27 3·23 0·0123 0·0048 0·0 60·20 51·62
A-0·27-4·30 125 2000 0·27 4·30 0·0123 0·0048 0·0 62·55 53·85
A-0·27-5·38 125 2500 0·27 5·38 0·0123 0·0048 0·0 66·20 57·31
B-0·54-2·15 250 1000 0·54 2·15 0·0123 0·0048 0·0 64·80 55·98
B-0·54-3·23 250 1500 0·54 3·23 0·0123 0·0048 0·0 53·88 45·68
B-0·54-4·30 250 2000 0·54 4·30 0·0123 0·0048 0·0 62·55 53·85
B-0·54-5·38 250 2500 0·54 5·38 0·0123 0·0048 0·0 61·65 52·99
C-0·81-2·15 375 1000 0·81 2·15 0·0123 0·0048 0·0 59·70 51·15
C-0·81-3·23 375 1500 0·81 3·23 0·0123 0·0048 0·0 52·03 43·96
D-1·08-2·15 500 1000 1·08 2·15 0·0123 0·0048 0·0 56·57 48·20
D-1·08-3·23 500 1500 1·08 3·23 0·0123 0·0048 0·0 52·20 44·12
D-1·08-4·30 500 2000 1·08 4·30 0·0123 0·0048 0·0 55·08 46·81
D-1·08-5·38 500 2500 1·08 5·38 0·0123 0·0048 0·0 56·38 48·03
E-1·62-3·23 750 1500 1·62 3·23 0·0123 0·0048 0·0 59·07 50·56
E-1·62-4·30 750 2000 1·62 4·30 0·0123 0·0048 0·0 52·72 44·60
E-1·62-5·38 750 2500 1·62 5·38 0·0123 0·0048 0·0 53·50 45·33
F-2·16-4·30 1000 2000 2·16 4·30 0·0123 0·0048 0·0 48·91 41·06
G-2·70-5·38 1250 2500 2·70 5·38 0·0123 0·0048 0·0 50·79 42·80

15
Table 2. Case study 1 – summary of predictions for V exp
Beam Vexp : kN Vn : kN Author Vn =Vexp Author Vn =Vexp ACI Vn =Vexp FIP Vn =Vexp CSA
(1) (2) (3) (4) (5) (6) (7)
A-0·27-2·15 675 623 0·92 0·29 0·86 0·62
A-0·27-3·23 630 570 0·90 0·29 0·81 0·61
A-0·27-4·30 640 587 0·92 0·35 0·83 0·62
A-0·27-5·38 630 612 0·97 0·38 0·90 0·65
B-0·54-2·15 468 429 0·92 0·45 1·16 0·59
B-0·54-3·23 445 381 0·86 0·45 1·02 0·55
B-0·54-4·30 500 420 0·84 0·45 1·07 0·54
B-0·54-5·38 480 416 0·87 0·48 1·09 0·56
C-0·81-2·15 403 318 0·79 0·51 0·90 0·46
C-0·81-3·23 400 292 0·73 0·50 0·89 0·43
D-1·08-2·15 270 256 0·95 0·63 1·00 0·50
D-1·08-3·23 280 242 0·86 0·62 0·96 0·46
D-1·08-4·30 290 251 0·87 0·64 0·93 0·46
D-1·08-5·38 290 255 0·88 0·67 0·93 0·46
E-1·62-3·23 220 204 0·93 0·58 0·82 0·38
E-1·62-4·30 190 187 0·98 0·66 0·94 0·41
E-1·62-5·38 173 189 1·09 0·78 1·03 0·45
F-2·16-4·30 150 145 0·96 0·95 0·89 0·33
G-2·70-5·38 105 127 1·21 1·04 1·02 0·34
Mean 0·92 0·55 0·95 0·50
SD 0·104 0·203 0·097 0·097
COV 0·113 0·359 0·102 0·196

15
Case study: Tan et al. d ¼ 463 mm and a width b ¼ 110 mm. The beams
Nineteen high strength reinforced concrete deep were singly reinforced with the longitudinal steel per-
15
beams with effective span and shear span variations centage r s maintained at 1·23% and with the nominal
were tested under two-point top loading. All the beams percentage of shear reinforcement rv ¼ 0:48%. The
had an overall height h ¼ 500 mm, an effective depth yield strengths of longitudinal and web reinforcement
56 Magazine of Concrete Research, 2003, 55, No. 1

Downloaded by [ University of Ottawa Library System] on [12/09/16]. Copyright © ICE Publishing, all rights reserved.
 A direct method for deep beams with web reinforcement
14
Table 3. Case study 2 – details of the 52 specimens
Beam a: mm le : mm a=d le =d rs rv rh f 9c : MPa
(1) (2) (3) (4) (5) (6) (7) (8) (9)
0A0-44 305 813 1·0 2·67 0·0194 0·0 0·0 20·48
0A0-48 305 813 1·0 2·67 0·0194 0·0 0·0 20·93
1A1-10 305 813 1·0 2·67 0·0194 0·0028 0·0023 18·69
1A3-11 305 813 1·0 2·67 0·0194 0·0028 0·0045 18·03
1A4-12 305 813 1·0 2·67 0·0194 0·0028 0·0068 16·07
1A4-51 305 813 1·0 2·67 0·0194 0·0028 0·0068 20·55
1A6-37 305 813 1·0 2·67 0·0194 0·0028 0·0091 21·06
2A1-38 305 813 1·0 2·67 0·0194 0·0063 0·0023 21·68
2A3-39 305 813 1·0 2·67 0·0194 0·0063 0·0045 19·75
2A4-40 305 813 1·0 2·67 0·0194 0·0063 0·0068 20·34
2A6-41 305 813 1·0 2·67 0·0194 0·0063 0·0091 19·13
3A1-42 305 813 1·0 2·67 0·0194 0·0125 0·0023 18·41
3A3-43 305 813 1·0 2·67 0·0194 0·0125 0·0045 19·24
3A4-45 305 813 1·0 2·67 0·0194 0·0125 0·0068 20·82
3A6-46 305 813 1·0 2·67 0·0194 0·0125 0·0091 19·93
0B0-49 368 940 1·21 3·08 0·0194 0·0 0·0 21·68
1B1-01 368 940 1·21 3·08 0·0194 0·0024 0·0023 22·06
1B3-29 368 940 1·21 3·08 0·0194 0·0024 0·0045 20·10
1B4-30 368 940 1·21 3·08 0·0194 0·0024 0·0068 20·82
1B6-31 368 940 1·21 3·08 0·0194 0·0024 0·0091 19·51
2B1-05 368 940 1·21 3·08 0·0194 0·0042 0·0023 19·17
2B3-06 368 940 1·21 3·08 0·0194 0·0042 0·0045 19·00
2B4-07 368 940 1·21 3·08 0·0194 0·0042 0·0068 17·48
2B4-52 368 940 1·21 3·08 0·0194 0·0042 0·0068 21·79
2B6-32 368 940 1·21 3·08 0·0194 0·0042 0·0091 19·75
3B1-08 368 940 1·21 3·08 0·0194 0·0063 0·0023 16·24
3B1-36 368 940 1·21 3·08 0·0194 0·0077 0·0023 20·41
3B3-33 368 940 1·21 3·08 0·0194 0·0077 0·0045 19·00
3B4-34 368 940 1·21 3·08 0·0194 0·0077 0·0068 19·24
3B6-35 368 940 1·21 3·08 0·0194 0·0077 0·0091 20·65
4B1-09 368 940 1·21 3·08 0·0194 0·0125 0·0023 17·10
0C0-50 457 1118 1·50 3·67 0·0194 0·0 0·0 20·69
1C1-14 457 1118 1·50 3·67 0·0194 0·0018 0·0023 19·24
1C3-02 457 1118 1·50 3·67 0·0194 0·0018 0·0045 21·89
1C4-15 457 1118 1·50 3·67 0·0194 0·0018 0·0068 22·68
1C6-16 457 1118 1·50 3·67 0·0194 0·0018 0·0091 21·79
2C1-17 457 1118 1·50 3·67 0·0194 0·0031 0·0023 19·86
2C3-03 457 1118 1·50 3·67 0·0194 0·0031 0·0045 19·24
2C3-27 457 1118 1·50 3·67 0·0194 0·0031 0·0045 19·31
2C4-18 457 1118 1·50 3·67 0·0194 0·0031 0·0068 20·44
2C6-19 457 1118 1·50 3·67 0·0194 0·0031 0·0091 20·75
3C1-20 457 1118 1·50 3·67 0·0194 0·0056 0·0023 21·03
3C3-21 457 1118 1·50 3·67 0·0194 0·0056 0·0045 16·55
3C4-22 457 1118 1·50 3·67 0·0194 0·0056 0·0068 18·27
3C6-23 457 1118 1·50 3·67 0·0194 0·0056 0·0091 19·00
4C1-24 457 1118 1·50 3·67 0·0194 0·0077 0·0023 19·58
4C3-04 457 1118 1·50 3·67 0·0194 0·0063 0·0045 18·55
4C3-28 457 1118 1·50 3·67 0·0194 0·0077 0·0045 19·24
4C4-25 457 1118 1·50 3·67 0·0194 0·0077 0·0068 18·51
4C6-26 457 1118 1·50 3·67 0·0194 0·0077 0·0091 21·24
0D0-47 635 1473 2·08 4·83 0·0194 0·0 0·0 19·51
4D1-13 635 1473 2·08 4·83 0·0194 0·0042 0·0023 16·07

were 504·8 and 375·2 MPa, respectively. The beams respectively, while predictions by the ACI code, the
were tested for seven shear span-depth a=d ratios, CEB-FIP MC 90, and the CSA code are also listed.
ranging from 0·27 to 2·70, and four effective span- Since the main objective is to compare the variations in
depth le =d ratios, ranging from 2·15 to 5·38. The de- predictions from different codes and the proposed
tails of the 19 beams are given in Table 1. Roller method with test results, Vexp is chosen as the common
bearings of lb ¼ 150 mm were used at the supports. denominator in all three case studies. It is also note-
The test results and predictions by the authors are worthy that beams E-1·62-5·38 and G-2·70-5·38 in
15
shown in the second and third columns of Table 2, Table 2 actually failed in flexure and are included in
Magazine of Concrete Research, 2003, 55, No. 1 57

Downloaded by [ University of Ottawa Library System] on [12/09/16]. Copyright © ICE Publishing, all rights reserved.
 Tan et al.
14
Table 4. Case study 2 – predictions for shear strength Vexp
Beam Vexp : kN Vn : kN Author Vn =Vexp Author Vn =Vexp ACI Vn =Vexp FIP Vn =Vexp CSA
(1) (2) (3) (4) (5) (6) (7)
0A0-44 140 124 0·89 0·50 0·91 0·91
0A0-48 136 125 0·92 0·52 0·95 0·95
1A1-10 161 119 0·74 0·57 0·72 0·72
1A3-11 149 118 0·79 0·61 0·75 0·75
1A4-12 141 109 0·77 0·60 0·71 0·70
1A4-51 171 130 0·76 0·56 0·74 0·74
1A6-37 184 134 0·73 0·53 0·71 0·71
2A1-38 175 136 0·78 0·56 0·77 0·77
2A3-39 171 129 0·75 0·55 0·72 0·72
2A4-40 172 132 0·77 0·56 0·73 0·73
2A6-41 162 128 0·79 0·57 0·73 0·73
3A1-42 161 125 0·78 0·56 0·71 0·70
3A3-43 173 130 0·75 0·54 0·69 0·69
3A4-45 179 139 0·78 0·54 0·72 0·72
3A6-46 168 136 0·81 0·56 0·73 0·73
0B0-49 149 114 0·76 0·48 0·90 0·68
1B1-01 148 119 0·80 0·69 0·93 0·69
1B3-29 144 112 0·78 0·68 0·87 0·65
1B4-30 141 116 0·82 0·71 0·92 0·69
1B6-31 154 112 0·73 0·63 0·79 0·59
2B1-05 129 109 0·84 0·74 0·92 0·69
2B3-06 131 109 0·83 0·73 0·90 0·68
2B4-07 126 103 0·82 0·73 0·86 0·65
2B4-52 150 122 0·81 0·68 0·90 0·68
2B6-32 145 115 0·79 0·67 0·85 0·63
3B1-08 131 97 0·74 0·67 0·77 0·58
3B1-36 159 117 0·74 0·62 0·79 0·60
3B3-33 159 112 0·71 0·60 0·74 0·56
3B4-34 155 114 0·73 0·62 0·77 0·58
3B6-35 166 121 0·73 0·60 0·77 0·58
4B1-09 154 106 0·69 0·59 0·69 0·52
0C0-50 116 94 0·81 0·52 1·11 0·56
1C1-14 119 92 0·77 0·74 1·00 0·50
1C3-02 124 102 0·83 0·87 1·10 0·55
1C4-15 131 106 0·81 0·83 1·07 0·54
1C6-16 123 104 0·84 0·87 1·10 0·56
2C1-17 124 96 0·77 0·77 0·99 0·50
2C3-03 104 94 0·91 0·97 1·15 0·58
2C3-27 116 94 0·82 0·87 1·03 0·52
2C4-18 125 99 0·79 0·83 1·01 0·51
2C6-19 124 101 0·82 0·84 1·04 0·52
3C1-20 141 103 0·73 0·74 0·92 0·46
3C3-21 125 85 0·68 0·74 0·82 0·41
3C4-22 128 93 0·73 0·76 0·89 0·45
3C6-23 137 97 0·70 0·73 0·86 0·43
4C1-24 147 99 0·67 0·69 0·83 0·42
4C3-04 129 94 0·73 0·77 0·89 0·45
4C3-28 153 98 0·64 0·66 0·78 0·39
4C4-25 153 95 0·62 0·65 0·75 0·38
4C6-26 160 106 0·67 0·66 0·82 0·42
0D0-47 74 71 0·96 0·48 1·28 0·40
4D1-13 88 66 0·75 0·89 1·14 0·27
Mean 0·77 0·67 0·87 0·60
SD 0·066 0·119 0·144 0·137
COV 0·086 0·179 0·165 0·229

the comparison study only for completeness. Table 2 the CEB-FIP MC 90, while the approach gives better
shows that the mean of predictions (equation (9)) to test predictions compared to the ACI code and CSA code.
results is 0·92 with a standard deviation (SD) of 0·104 14
and a coefficient of variation (COV) is 0·113. Thus, for Case study 2: Smith and Vantsiotis
deep beams with nominal web reinforcement, it seems Smith and Vantsiotis reported their test results of 52
that the proposed predictions are comparable to those by reinforced deep beams under two-point loading. Their
58 Magazine of Concrete Research, 2003, 55, No. 1

Downloaded by [ University of Ottawa Library System] on [12/09/16]. Copyright © ICE Publishing, all rights reserved.
 Magazine of Concrete Research, 2003, 55, No. 1

11, 12
Table 5. Case study 3 – details of the 45 specimens
Beam Web type h: mm d: mm a=d le =d rv rh r i

f yv : MPa f yh : MPa f yi : MPa f 9c : MPa
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13)
1-30 Type 1 762 724 0·35 1·05 0·0245 0·0 0·0 280 0·0 0·0 21·51
1-25 Type 1 635 597 0·43 1·28 0·0245 0·0 0·0 280 0·0 0·0 24·55
1-20 Type 1 508 470 0·54 1·62 0·0245 0·0 0·0 280 0·0 0·0 21·24
1-15 Type 1 381 343 0·74 2·22 0·0245 0·0 0·0 280 0·0 0·0 21·24
1-10 Type 1 254 216 1·18 3·53 0·0245 0·0 0·0 280 0·0 0·0 21·65

A direct method for deep beams with web reinforcement

2-30 Type 2 762 724 0·35 1·05 0·0086 0·0 0·0 303 0·0 0·0 19·2
2-25 Type 2 635 597 0·43 1·28 0·0086 0·0 0·0 303 0·0 0·0 18·62
2-20 Type 2 508 470 0·54 1·62 0·0086 0·0 0·0 303 0·0 0·0 19·86
2-15 Type 2 381 343 0·74 2·22 0·0086 0·0 0·0 303 0·0 0·0 22·75
2-10 Type 2 254 216 1·18 3·53 0·0086 0·0 0·0 303 0·0 0·0 20·13
3-30 Type 3 762 724 0·35 1·05 0·0 0·0245 0·0 0.0 280 0·0 22·55
3-25 Type 3 635 597 0·43 1·28 0·0 0·0245 0·0 0.0 280 0·0 20·96
3-20 Type 3 508 470 0·54 1·62 0·0 0·0245 0·0 0.0 280 0·0 19·24
3-15 Type 3 381 343 0·74 2·22 0·0 0·0245 0·0 0.0 280 0·0 21·93
3-10 Type 3 254 216 1·18 3·53 0·0 0·0245 0·0 0.0 280 0·0 22·62
4-30 Type 4 762 724 0·35 1·05 0·0 0·0086 0·0 0.0 303 0·0 22·00
4-25 Type 4 635 597 0·43 1·28 0·0 0·0086 0·0 0.0 303 0·0 20·96
4-20 Type 4 508 470 0·54 1·62 0·0 0·0086 0·0 0.0 303 0·0 20·13
4-15 Type 4 381 343 0·74 2·22 0·0 0·0086 0·0 0.0 303 0·0 22·00
4-10 Type 4 254 216 1·18 3·53 0·0 0·0086 0·0 0.0 303 0·0 22·62
5-30 Type 5 762 724 0·35 1·05 0·0061 0·0061 0·0 280 280 0·0 18·55
5-25 Type 5 635 597 0·43 1·28 0·0061 0·0061 0·0 280 280 0·0 19·24
5-20 Type 5 508 470 0·54 1·62 0·0061 0·0061 0·0 280 280 0·0 20·13
5-15 Type 5 381 343 0·74 2·22 0·0061 0·0061 0·0 280 280 0·0 21·93
5-10 Type 5 254 216 1·18 3·53 0·0061 0·0061 0·0 280 280 0·0 22·55
59

(continued overleaf )

Downloaded by [ University of Ottawa Library System] on [12/09/16]. Copyright © ICE Publishing, all rights reserved.
 60

Tan et al.
Table 5. (continued)
Beam Web type h: mm d: mm a=d le =d rv rh r i

f yv : MPa f yh : MPa f yi : MPa f 9c : MPa
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13)
6-30 Type 6 762 724 0·35 1·05 0·0 0·0051
0·0 0·0 303 0·0 26·08
6-25 Type 6 635 597 0·43 1·28 0·0 0·0061
0·0 0·0 303 0·0 25·10
6-20 Type 6 508 470 0·54 1·62 0·0 0·0077
0·0 0·0 303 0·0 26·08
6-15 Type 6 381 343 0·74 2·22 0·0 0·0102
0·0 0·0 303 0·0 26·08
6-10 Type 6 254 216 1·18 3·53 0·0 0·0153
0·0 0·0 303 0·0 25·1
7-30A Type 7 762 724 0·35 1·05 0·0 0·0 0·0 0·0 303 0·0 25·10
7-30B Type 7 762 724 0·35 1·05 0·0 0·0017
0·0 0·0 303 0·0 26·08
7-30C Type 7 762 724 0·35 1·05 0·0 0·0034
0·0 0·0 303 0·0 25·10
7-30D Type 7 762 724 0·35 1·05 0·0 0·0068
0·0 0·0 303 0·0 21·27
7-30E Type 7 762 724 0·35 1·05 0·0 0·0085
0·0 0·0 303 0·0 21·27
#S-30 Type 8 762 724 0·35 1·05 0·0 0·0 0·0076 0·0 0·0 338 22·13
#S-25 Type 8 635 597 0·43 1·28 0·0 0·0 0·0066 0·0 0·0 338 21·24
#S-20 Type 8 508 470 0·54 1·62 0·0 0·0 0·0050 0·0 0·0 338 21·79
#S-15 Type 8 381 343 0·74 2·22 0·0 0·0 0·0063 0·0 0·0 338 27·65
#S-10 Type 8 254 216 1·18 3·53 0·0 0·0 0·0060 0·0 0·0 338 23·31
#D-30 Type 8 762 724 0·35 1·05 0·0 0·0 0·0252 0·0 0·0 296 23·17
#D-25 Type 8 635 597 0·43 1·28 0·0 0·0 0·0252 0·0 0·0 296 23·79
#D-20 Type 8 508 470 0·54 1·62 0·0 0·0 0·0235 0·0 0·0 296 24·75
#D-15 Type 8 381 343 0·74 2·22 0·0 0·0 0·0231 0·0 0·0 296 27·65
#D-10 Type 8 254 216 1·18 3·53 0·0 0·0 0·0216 0·0 0·0 296 24·20

r ¼ (total area of web reinforcement)/(cross section of beam).
s


r ¼ (total volume of web steel)/(total volume of concrete).
i
Magazine of Concrete Research, 2003, 55, No. 1

# Test specimens were obtained from Reference 12, while the remaining specimens from Reference 11.

Downloaded by [ University of Ottawa Library System] on [12/09/16]. Copyright © ICE Publishing, all rights reserved.
 A direct method for deep beams with web reinforcement
11, 13
Table 6. Case study 3 – predictions for shear strength V exp
Beam Vexp (kN) Vn (kN) Author Vn =Vexp Author Vn =Vexp ACI
Vn =Vexp FIP Vn =Vexp CSA
(1) (2) (3) (4) (5) (6) (7)
1-30 239 174 0·72 0·66 0·31 0·53
1-25 224 174 0·77 0·62 0·38 0·65
1-20 190 139 0·73 0·57 0·39 0·65
1-15 164 114 0·70 0·49 0·45 0·60
1-10 90 83 0·92 0·62 0·47 0·60
2-30 249 155 0·62 0·53 0·27 0·46
2-25 224 138 0·61 0·48 0·29 0·49
2-20 216 125 0·58 0·41 0·32 0·53
2-15 140 109 0·78 0·50 0·45 0·71
2-10 100 70 0·70 0·44 0·41 0·50
3-30 276 216 0·78 0·63 0·28 0·48
3-25 226 186 0·82 0·61 0·32 0·55
3-20 208 151 0·72 0·50 0·32 0·54
3-15 159 128 0·80 0·51 0·32 0·62
3-10 87 81 0·94 0·66 0·36 0·65
4-30 242 191 0·79 0·71 0·32 0·54
4-25 201 166 0·82 0·69 0·36 0·62
4-20 181 137 0·76 0·59 0·39 0·65
4-15 110 112 1·03 0·74 0·47 0·90
4-10 96 73 0·77 0·55 0·32 0·59
5-30 240 163 0·68 0·66 0·27 0·46
5-25 208 152 0·73 0·64 0·32 0·55
5-20 173 135 0·78 0·62 0·41 0·68
5-15 127 112 0·88 0·64 0·46 0·78
5-10 78 76 0·98 0·73 0·48 0·72
6-30 308 217 0·70 0·61 0·30 0·50
6-25 266 191 0·72 0·57 0·33 0·56
6-20 245 168 0·69 0·50 0·30 0·57
6-15 173 131 0·76 0·51 0·30 0·57
6-10 99 82 0·83 0·61 0·32 0·59
7-30A 253 185 0·73 0·54 0·35 0·59
7-30B 300 201 0·67 0·62 0·30 0·51
7-30C 260 204 0·79 0·71 0·34 0·57
7-30D 264 190 0·72 0·64 0·28 0·48
7-30E 297 194 0·65 0·57 0·25 0·42
S-30 288 218 0·76 / 0·27 0·52
S-25 282 196 0·70 / 0·26 0·48
S-20 239 177 0·74 / 0·32 0·51
S-15 208 165 0·79 / 0·36 0·52
S-10 110 96 0·87 / 0·39 0·43
D-30 279 233 0·83 / 0·29 0·56
D-25 270 222 0·82 / 0·31 0·55
D-20 278 203 0·73 / 0·31 0·48
D-15 237 173 0·73 / 0·32 0·46
D-10 119 102 0·86 / 0·37 0·41
Mean 0·77 0·59 0·34 0·56
SD 0·092 0·082 0·062 0·096
COV 0·120 0·140 0·181 0·171

The ACI code does not include the design of deep beams with inclined web reinforcement.

experimental objectives were to study the effect of formation of inclined cracks but had moderate effect
vertical and horizontal web reinforcement and shear- on ultimate shear strength. The addition of vertical web
span-to-effective-depth ratios on inclined cracking reinforcement improved the ultimate shear strength but
strength, ultimate shear strength, mid-span deflection, the addition of horizontal web reinforcement did not
tension reinforcement strain and crack width. They have any noticeable effect. A considerable increase in
found that web reinforcement had no effect on the load-carrying capacity was observed with increasing
Magazine of Concrete Research, 2003, 55, No. 1 61

Downloaded by [ University of Ottawa Library System] on [12/09/16]. Copyright © ICE Publishing, all rights reserved.
 Tan et al.

concrete strength and decreasing shear span=depth

ratio.
The details of the 52 beams are given in Table 3. All
the beams had an overall height h ¼ 356 mm, an effec-
tive depth d ¼ 305 mm and a width b ¼ 102 mm. The
52 beams contained three #5 bars as longitudinal re-
inforcement; As ¼ 600 mm2 , f y ¼ 431 MPa. For both
the vertical and horizontal web reinforcement, the Type 1 web reinforcement Type 2 web reinforcement
measured yield strengths were f yv ¼ f yh ¼ 483:5 MPa.
Bearing plates of 102 3 102 3 25 mm were used at the
supports and at the two loading points. The ultimate
shear strength predictions of 52 beams are given in
Table 4, which also shows that the proposed method
gives good predictions with a mean value of 0·77, the
lowest SD and COV of 0·066 and 0·086, respectively.
11–13
Case study 3: Kong et al. Type 3 web reinforcement Type 4 web reinforcement

Kong et al. performed a systematic study on deep

beams with different web reinforcement. They investi-
gated a total of 35 simply supported normal-weight
concrete deep beams with span=depth ratio ranging
from 1·05 to 3·53 and with seven different types of web
reinforcement. Two years later, they reported a second
batch of test results that consisted of 38 simply sup-
12
ported lightweight concrete deep beams. From the
Type 5 web reinforcement Type 6 web reinforcement
test results, they found that inclined web reinforcement
was very effective in increasing the ultimate strength.
Consequently, they tested yet another batch of 10 sim-
ply supported normal weight deep beams with inclined
13
web reinforcement. The tests re-confirmed the bene-
ficial effect of inclined web reinforcement on enhan-
cing shear strength.
The details of the two batches of normal weight deep
11,13
beams (35 plus 10 beams) are given in Table 5. All
Type 7 web reinforcement Type 8 web reinforcement
the beams had an effective span le ¼ 762 mm, a shear
span a ¼ 254 mm and a width b ¼ 76 mm. Each beam Fig. 3. Types of web reinforcement
contained one #6 bar as the longitudinal reinforce-
ment; As ¼ 287 mm2 , f y ¼ 287 MPa. The eight types
11
of web reinforcement (seven types of beams were that the proposed strut-and-tie approach is conservative
13
tested in 1970 and the last type was tested in 1972) even for deep beams with different types of web rein-
are shown in Fig. 3. While clearly some of the web forcement. It is noteworthy that the FIP predictions in
reinforcement investigated may not be practical (such this case are ultra conservative. This is due to the fact
as inclined web reinforcement), this case study never- that lower strength of concrete and higher percentage
theless provides a comprehensive and thorough means of web reinforcement were used in the current case
to validate the accuracy of the proposed strut-and-tie study, which indicates that web steel plays a much
model which takes account of web reinforcement. 9
more dominant role. Here, the CEB-FIP MC 90 sug-
The test results and predictions are given in Table 6. gested that shear capacity of deep beams is governed
It should be noted that the ACI code does not include by two criteria
the design of deep beams with inclined web reinforce-
ment, e.g. Beam S-30, S-15, etc. Table 6 shows that the Vn < f y As tan Łs þ 0:5Tw (12)
proposed method gives good predictions for deep
(yielding of main reinforcement)
beams with eight different types of web reinforcement.
The mean of predictions to test results is 0·77 with a and
low SD of 0·092 and the lowest COV of 0·120. It is also
noticed that all the predictions by the new approach are Vn < vf 9c lb bw
generally lower than the test values, with the ratio (crushing of concrete in the diagonal strut) (13)
V n =V exp ranging from 0·58 (Beam 2-20 in Table 6) to
1·03 (Beam 4-15 in Table 6). This shows conclusively where Tw is the force carried by the vertical web
62 Magazine of Concrete Research, 2003, 55, No. 1

Downloaded by [ University of Ottawa Library System] on [12/09/16]. Copyright © ICE Publishing, all rights reserved.
 A direct method for deep beams with web reinforcement

reinforcement, T w < f yv Av ; v is concrete utilisation 5. SCHLAICH J. and SCHAFER K. Design and detailing of structural
9 concrete using strut-and-tie models. The Structural Engineer,
factor, taken as 0·6 in the CEB-FIP MC 90.
1991, 69, No. 6, 113–125.
Thus, for beams with lower strength of concrete
6. RAMIREZ J. A. Strut-tie design of pretensioned concrete members.
and/or high percentage of web reinforcement, the ulti- ACI Structure Journal, 1994, 91, 572–578.
mate shear strength Vn is only governed by equation 7. FIP RECOMMENDATIONS 1996 Practical Design of Structural
(13). Clearly, equation (13) does not take account of Concrete, SETO, London, 1999.
the beneficial effect of web reinforcement on the capa- 8. TAN K. H., TONG K. and TANG C. Y. A direct struct-and-tie
city of the diagonal strut, as a fixed value of v ¼ 0:6 is model for prestressed deep beams. Journal of Structural Engi-
neering, ASCE, 2001, 127, 1076–1084.
adopted. In fact, since the provision of adequate web 9. CEB-FIP MC 90 Design of Concrete Structures, CEB-FIP-
steel (horizontal or vertical) effectively controls diag- Model-Code 1990, Thomas Telford, London, 1993.
onal cracks and reduces the principal tensile stress 10. CANADIAN CSA BUILDING CODE. Design of Concrete Structures:
across the diagonal strut, a greater value may be used Structures (Design)—A National Standard of Canada (CAN-
for the concrete utilisation factor v. A23.3-94), Clause 11.1.2, Canadian Standards Association,
Toronto, 1994.
11. KONG F. K., ROBINS P. J. and COLE D. F. Web reinforcement
effects on deep beams. ACI Journal, 1970, 67–73, 1010–1017.
Conclusion 12. KONG F. K. and ROBINS P. J. Web reinforcement effects on light-
A simple and direct method is presented which takes weight concrete deep beams. ACI Journal, 1972, 68–47,
514–520.
into account the contribution of web reinforcement, be 13. KONG F. K., ROBINS P. J., KIRBY D. P. and SHORT D. R. Deep
it vertical, horizontal, inclined or orthogonal. Compari- beams with inclined web reinforcement. ACI Journal, 1972,
son with three case studies shows that the approach 69–76, 172–176.
yields safe and consistent predictions for deep beams 14. SMITH K. N. and VANTSIOTIS A. S. Shear strength of deep beams.
with. ACI Journal, 1982, 79–22, 201–213.
15. TAN K. H., KONG F. K., TENG S. and GUAN L. High-strength
(a) different geometrical ratios (case study 1). concrete deep beams with effective span and shear span varia-
tions. ACI Structure Journal, 1995, 92, S37, 395–405.
(b) different ratios of uniform vertical and horizontal
16. TAN K. H., KONG F. K., TENG S. and WENG L. W. Effect of web
web reinforcement (case study 2). reinforcement on high strength concrete deep beams. ACI Struc-
(c) different arrangements of non-uniform web reinfor- tural Journal, 1997, 94, 572–582.
cement including inclined reinforcement (case 17. AMERICAN CONCRETE INSTITUTE. Building Code Requirements
study 3). for Reinforced Concrete (ACI 318-1999) and Commentary—ACI
318R-1999, American Concrete Institute, Detroit, 1999.
18. TONG K. Strut-and-Tie Approach to Shear Strength Prediction of
References Deep Beams. MEng thesis, Res. Proj. RG 64/94, Nanyang Tech-
nological University, Singapore, 1997.
1. MARTI P. Basic tools of reinforced concrete beam design. ACI 19. ALSHEGEIR A. and RAMIREZ J. A. Strut-tie approach in preten-
Structure Journal, 1985, 82–84, 46–56. sioned deep beams. ACI Structure Journal, 1992, 89, S29, 296–
2. ROGOWSKY D. M. and MACGREGOR J. G. Design of reinforced 304.
concrete deep beams. Concrete International: Design & Con- 20. MACGREGOR J. G. Reinforced Concrete – Mechanics and Design,
struction, 1986, 8, No. 8, 49–58. 3rd edn, Prentice Hall, New Jersey, 1997.
3. COLLINS M. P. and MITCHELL D. A rational approach to shear
design – the 1984 Canadian Code provisions. ACI Structural
Journal, 1986, 83, 925–933.
4. SCHLAICH J., SCHAFER K. and JENNEWEIN M. Towards a consis-
tent design for structural concrete. PCI Journal, 1987, 32, No. 3, Discussion contributions on this paper should reach the editor by
75–150. 1 August 2003

Magazine of Concrete Research, 2003, 55, No. 1 63

Downloaded by [ University of Ottawa Library System] on [12/09/16]. Copyright © ICE Publishing, all rights reserved.