ACI STRUCTURAL JOURNAL TECHNICAL PAPER

Title no. 94-S52

Effect of Web Reinforcement on High-Strength

Concrete Deep Beams

by Kang-Hai Tan, Fung-Kew Kong, Susanto Teng, and Li-Wei Weng

Results of an experimental investigation on the behavior and ultimate same can be said of the design provisions in the UK CIRIA
shear strength of 18 high strength concrete deep beams are summarized. Guide-2 and the Canadian Code. As high-strength concrete
The concrete cylinder compressive strength fc′ ranges from 55 to 86 MPa (HSC) is becoming more and more popular, it is timely to
(approximately 8000 to 12,500 psi). The test specimens are divided into
evaluate whether these design documents can still provide
three series based on the shear-span-to-overall-height ratio a/h. Each
series consists of six beams with different arrangements of horizontal and
safe design for HSC deep beams; HSC in this context refers
vertical web reinforcements, i.e., the main variables are the horizontal and to concrete with fc′ greater than 55 MPa (8000 psi). Previous
the vertical web steel ratios. Observations are made on mid-span deflec- work5 shows that the ACI design equations and the CIRIA
tions, crack widths, failure modes and ultimate strengths. The test results Guide-2 are applicable for deep beams with nominal web
show that for deep flexural members with a/h exceeding 1.00 (or shear- reinforcement and with fc′ exceeding 40 MPa. This paper
span-to-effective-depth ratio a/d ≥ 1.13), the vertical web reinforcement is further investigates the applicability of the codes for HSC
more effective than the horizontal web reinforcement. It is also shown that deep beams with a significant amount of web reinforcement.
orthogonal web reinforcement comprising both vertical and horizontal There has been relatively limited information on this aspect;
reinforcements is the most efficient in increasing the beam stiffness, most reported investigations are on HSC beams without web
restricting the diagonal crack width development and enhancing the ulti-
reinforcement.6-9 In the literature on HSC beams with web
mate shear strength.
reinforcement,10,11,12 the emphasis is on the shallow beam
The test results are then compared with the ultimate strength predictions
obtained using the current ACI Code, the Canadian Code, and the UK CIRIA
category (with a/d ≥ 2.50), with only limited test data in the
Guide. The deep-beam provisions in the ACI Code overestimate the contribu- short/deep beam category. The present investigation seeks to
tion of the horizontal web steel to shear strength. Based on the test results, a supplement such information since the effect of vertical and
revision to ACI Eq. (11-31) for web steel contribution is suggested. The horizontal web reinforcements on the behavior of low-
Canadian Code shows the most consistent and yet conservative predictions strength concrete deep beams has already been shown13-15 to
of the test beams with different web reinforcements, while the UK CIRIA be significant. Therefore, in the test program, the vertical and
Guide is unconservative for beams with horizontal web reinforcement. the horizontal web steel ratios ρv and ρh are considered as
two main variables. The test results are then compared with
Keywords: building codes; cracks; deep beams; deflections; diagonal the code predictions.
cracking; high-strength concrete; shear span; shear strength; ultimate
strength; web reinforcement.
RESEARCH SIGNIFICANCE
High-strength concrete is being used more and more
INTRODUCTION widely in the construction industry in recent years. Experi-
The design of reinforced concrete deep beam is a subject mental results described in this paper give further empirical
of considerable interest in structural engineering practice. It evidence on the behavior of HSC deep beams with fc′ greater
has various structural applications ranging from pile-caps than 55 MPa (8000 psi). In particular, the test program inves-
and wall foundations, to transfer girders in tall buildings. tigates the individual and combined effects of vertical and
Despite its wide structural applications, many national codes horizontal web reinforcements on HSC deep beams. The
do not include the design of deep beams. For instance, the study reveals that increasing the horizontal or vertical web
British Standards BS 81101 explicitly states that “for the steel ratio can considerably increase the ultimate shear
design of deep beams, reference should be made to specialist strength of the beams. In addition, when a/h exceeds 1.00
literature”. That design document may well refer to the (equivalent to a/d ≥ 1.13), the vertical web steel is more
CIRIA Guide-2,2 issued by the UK Construction Industry effective than the horizontal web steel. The current ACI
Research and Information Association. Besides the UK Code 318-89 and the UK CIRIA Guide-2 may yield uncon-
CIRIA Guide-2, the ACI Building Code,3 and the 1984
Canadian Code4 also provide guidance for the deep beam
design. ACI Structural Journal, V. 94, No. 5, September-October 1997.
Received October 24, 1995, and reviewed under Institute publication policies.
In the ACI Code, the empirical equations governing deep Copyright © 1997, American Concrete Institute. All rights reserved, including the
beam design are based on low-strength concrete specimens making of copies unless permission is obtained from the copyright proprietors.
Pertinent discussion will be published in the July-August 1998 ACI Structural
with fc′ in the range of 14 to 40 MPa (2000 to 6000 psi). The Journal if received by March 1, 1998.

572 ACI Structural Journal/September-October 1997

Kang-Hai Tan is a senior lecturer in the Division of Structures and Construction,
Nanyang Technological University (NTU), Singapore. He obtained his BSc and the
PhD degrees from the University of Manchester in England.His research interests
include concrete deep beams and thin-walled steel structures.

Fung-Kew Kong is a professor and Head of the Division of Structures and Construc-
tion at Nanyang Technological University, Singapore. Professor Kong is the Alternate
Chairman of Singapore's Structural Engineering Practice Technical Committee,
which oversees the drafting of all future structural standards and codes of practice in
Singapore. He has published several award-winning papers on structural concrete
and is a coeditor of the Handbook of Structural Concrete.

ACI member Susanto Teng is a lecturer in the Division of Structures and Construc-
tion, NTU, Singapore. He received his PhD from the University of Iowa. His research
interests include structural concrete deep beams, time-dependent effects and behavior
of concrete structures under dynamic loads.

Li-Wei Weng graduated from the Shanghai Jiao-Tong University, Department of

Engineering Mechanics, in 1991. Prior to joining NTU, he spent two years in the
Zhongnan Building Design Institute, working as a structural analyst. He obtaining his
MEng degree from NTU, after which he worked as a structural engineer in a civil/
structural consultancy firm in Singapore.

servative strength predictions for HSC deep beams with

significant amount of horizontal web steel. A revision to the
ACI Code Eq. (11-31) is suggested, based on the test results.
The Canadian Code gives the most consistent and safe esti-
mations of the ultimate shear strengths of these 18 beams, as
the Code does not take account of the contribution of web
reinforcement to shear strength.

EXPERIMENTAL PROGRAM
Beam notation
This paper addresses the effect of different web reinforce-
ments and a/h ratios on the shear behavior of HSC deep
beams. In the beam notation under Table 1 the series number
is given first; this is followed by the type of web reinforce-
ment and shear-span-to-overall-height ratio a/h. For
example, Beam II-5/1.00 refers to a specimen in Series II, Fig. 1—Different types of web reinforcements.
with Type 5 web reinforcement and an a/h ratio of 1.00.
Figure 1 shows the different types of web reinforcement the main longitudinal reinforcement, (ii) 10-mm (0.4-in.)
investigated. diameter high strength deformed bars with fy = 446.7 MPa
It should be noted that the web reinforcement details in the (65 ksi) as vertical or horizontal web bars, and (iii) 10 mm
two shear spans of Type 2 and Type 6 specimens are different (0.4 in.) diameter plain mild steel round bars with fy =
(Table 1). The shear span provided with lower strength plain 353.2 MPa (51 ksi) as vertical or horizontal web bars.
vertical web steel is designated as the “N” span while the Within each series, each beam had different web reinforce-
other shear span with high strength deformed vertical web ment details, as shown in Fig. 1.
steel is designated as the “S” span. For example, I-2N/0.75 Type 1: This specimen with no web reinforcement in either
refers to the “N” shear span of the specimen I-2/0.75, while I- shear span served as a control beam.
2S/0.75 refers to the “S” shear span of the same beam. Type 2: Vertical web steel consisting of lower strength
plain bars was placed in the “N” shear span whereas vertical
Beam details web steel of high strength deformed bars was placed in the
This program consisted of 18 rectangular beams of “S” shear span. The vertical web steel ratio in specimens
500 mm (19.50 in.) height and 110 mm (4.29 in.) width. Full II-2N/1.00, II-2S/1.00, III-2N/1.50 and III-2S/1.50 was kept
details are given in Table 1. Each beam had a longitudinal at 1.43 percent, while specimens I-2N/0.75 and I-2S/0.75
main steel ratio ρ of 2.58 percent, consisting of four 20-mm had ρv of 2.86 percent due to the fact that Series I specimens
(0.78-in.) diameter high strength deformed bars. At each had the shortest shear span. In this manner, both the effects
location of loading or support point, a built-in reinforcement of vertical web reinforcement ratio ρv and its yield strength
cage was placed to prevent premature bearing failure (Fig. 1). fyv could be investigated.
The effective span le varied from 1750 to 2500 mm (68 to 98 Type 3: Horizontal web reinforcement consisting of mild
in.). The shear span a varied from 375 to 750 mm (14.6 to 29 steel bars was provided, giving ρh = 1.59 percent. The
in.), resulting in three a/h ratios. Based on the a/h ratios, the vertical spacing between each layer of the bars was 90 mm
18 beams were divided into three series of six beams each: (3.50 in.).
Series I for a/h = 0.75 (a/d = 0.85), Series II for a/h = 1.00 Type 4: Horizontal web reinforcement was identical to
(a/d = 1.13) and Series III for a/h = 1.50 (a/d = 1.69). Type 3 specimen, except that high strength deformed bars
Three types of steel bars were used as reinforcements were used instead of lower strength plain bars.
(Table 1): (i) 20 mm (0.8 in.) diameter high strength Type 5: Horizontal web steel ratio using high strength
deformed bars with yield strength fy = 498.9 MPa (72 ksi) as deformed bars was doubled to ρh = 3.17 percent; the vertical

ACI Structural Journal/September-October 1997 573

Table 1—Specimen details of Series I, II, and III
Effective Shear Effective fcu fc′ ft fy Age
span le, span a, a/h depth d a/d ρ, fyv, sv, ρv, fyh, sh , ρh, of
2
Beam* mm mm ratio mm ratio N/mm percent N/mm2 mm percent N/mm2 mm percent beams
I-1/0.75 99.4 56.3 4.2 0 — 0 0 — 0 146
I-2N/0.75 86.8 56.2 6.4 353.2 50 2.86 0 — 0 162
I-2S/0.75 86.8 56.2 6.4 446.7 50 2.86 0 — 0 162
I-3/0.75 78.8 59.2 6.6 0 — 0 353.2 90 1.59 160
1750 375 0.75 442.5 0.85 498.9 2.58
I-4/0.75 92.3 63.8 4.0 0 — 0 446.7 90 1.59 149
I-5/0.75 92.7 57.6 3.7 0 — 0 446.7 45 3.17 155
I-6N/0.75 80.9 59.7 5.4 353.2 50 2.86 446.7 90 1.59 168
I-6S/0.75 80.9 59.7 5.4 446.7 50 2.86 446.7 90 1.59 168
II-1/1.00 88.2 77.6 3.3 0 — 0 0 — 0 57
II-2N/1.00 88.2 77.6 3.3 353.2 100 1.43 0 — 0 56
II-2S/1.00 88.2 77.6 3.3 446.7 100 1.43 0 — 0 56
II-3/1.00 80.5 78.0 3.4 0 — 0 353.2 90 1.59 70
2000 500 1.00 442.5 1.13 498.9 2.58
II-4/1.00 89.4 86.3 3.3 0 — 0 446.7 90 1.59 63
II-5/1.00 89.4 86.3 3.3 0 — 0 446.7 45 3.17 63
II-6N/1.00 91.1 75.3 5.8 353.2 100 1.43 446.7 90 1.59 180
II-6S/1.00 91.1 75.3 5.8 446.7 100 1.43 446.7 90 1.59 180
III-1/1.50 88.2 77.6 3.3 0 — 0 0 — 0 56
III-2N/1.50 88.2 77.6 3.3 353.2 100 1.43 0 — 0 56
III-2S/1.50 88.2 77.6 3.3 446.7 100 1.43 0 — 0 56
III-3/1.50 80.5 78.0 3.4 0 — 0 353.2 90 1.59 70
2500 750 1.50 442.5 1.69 498.9 2.58
III-4/1.50 89.4 86.3 3.3 0 — 0 446.7 90 1.59 63
III-5/1.50 89.4 86.3 3.3 0 — 0 446.7 45 3.17 63
III-6N/1.50 91.1 78.9 5.6 353.2 100 1.43 446.7 90 1.59 182
III-6S/1.50 91.1 78.9 5.6 446.7 100 1.43 446.7 90 1.59 182
*Beam notation—The series number is given before the hyphen; this is followed by the different type of web reinforcement and then by the a/h ratio. In Type 2 and Type 6 beams,
the shear span provided with mild steel stirrups is designated as the “N” span while the other span with stirrups made of high strength deformed bars is designated as the “S” span.

Table 2—Summary of concrete mix design Concrete mix design

Characteristic cube strength 80 MPa The concrete for Series I, II, and III specimens were cast
Cement type ordinary portland cement from the same batch. The concrete mix design is given in
Aggregate type crushed granite and natural sand
Table 2. The design strength was specified in terms of the
characteristic concrete cube strength, i.e., fcu = 80 MPa
Slump for concrete 150 to 200 mm
(11600 psi). Ordinary portland cement was used, with an
Free water content 150 kg/m3 aggregate to cement ratio of 2.90 and a water-cement ratio of
Cement content 525 kg/m3 0.27. To avoid congestion in reinforcement, 10-mm (0.39
Coarse aggregate content 950 kg/m3
in.) chippings were used as aggregates. Two types of water-
reducing plasticizing admixture were added to maintain
Fine aggregate content 690 kg/m3 workability at a design slump of 150 to 200 mm. In Table 1,
Water-cement ratio (w/c) 0.27 the cube and cylinder compressive strengths fcu and fc′ were
Aggregate/cement ratio (a/c) 2.90 obtained from the averages of four 100 mm (3.90 in.) concrete
Admixture I 350 cc per 100 kg of cement cubes and four 150 mm (5.85 in.) diameter concrete cylinders
Admixture II 1700 cc per 100 kg of cement respectively. The fc′ values of Series I were lower than those of
Note: Admixture I = POZZOLITH 300N, a water-reducing plasticizing admixture for Series II and III, probably due to poor preparation of the
concrete; Admixture II = Rheobuild 1000. capping material. On the other hand, the fcu tests (which did
not require any capping) were consistent for all three series.
Two other cylinders were used to evaluate the splitting
spacing of the bars was reduced from 90 to 45 mm (3.50 tensile strength ft. Series II and III specimens were tested
to 1.75 in.). first, followed by Series I. Clearly, from Table 1, all the spec-
Type 6: Here, Type 2 vertical web steel and Type 4 hori- imens were made of HSC with fc′ > 55 MPa (8000 psi).
zontal web steel were combined to investigate the effect of
an orthogonal web reinforcement. Test procedure
Since the purpose in this program was to evaluate the effec- Vertical deflections were monitored by the LVDTs. At
tiveness of vertical and horizontal web steel (Av and Ah each load increment, the test data were captured by a data
respectively), the minimum requirements specified by the logger and automatically stored. All the beams were tested to
ACI Code for Av and Ah were waived. In addition, the width failure under two-point symmetric top loading. Specimens
of the bearing block was 150 mm (5.85 in.) at the two loading with Type 2 (only ρv) and Type 6 (combined ρv and ρh ) web
points and 100 mm (3.90 in.) at the two support points. Thus, reinforcements were tested twice. After the first shear span
the clear shear spans xe of Series I, II, and III beams were 250, had failed, the beam was externally clamped by a steel yoke.
375 and 625 mm respectively (9.75, 14.63, and 24.38 in.). The specimen was then re-loaded to failure.

574 ACI Structural Journal/September-October 1997

Fig. 2—Mid-span deflection of Series I, I, and III specimens. Fig. 3—Diagonal crack width development: Series I, II, and
III beams).

Both surfaces of the beam were white-washed to aid observa-

tions of crack development during testing. Initial loading of orthogonal web reinforcement on stiffness is also more
was applied at an increment of 20 kN (4.5 kips) for each jack dominant at a/h ≥ 1.00.
until the first crack occurred. Subsequently, the load increment
was increased to 40 kN (9 kips) for each jack. Crack widths
In all the tests, the first crack in a specimen was a flexural
TEST RESULTS AND DISCUSSIONS crack of 0.02 to 0.04 mm (0.008 to 0.016 in.) wide initiated from
Deflections the beam soffit in the mid-span region. Compared to the
The mid-span deflection curves for Series I, II, and III diagonal cracks, the flexural cracks were narrower in width
beams are shown in Fig. 2(a), (b) and (c) respectively. It is although some of them exceeded the serviceability limit of 0.33
observed that in Series I (a/h = 0.75), beams with high mm (0.013 in.) (specified by the ACI Code) towards failure.
strength horizontal web reinforcement (Types 4 and 5) Figures 3(a) through (c) present the diagonal crack
appear to be stiffer than beams with lower strength hori- development for Series I, II, and III beams. Among the three
zontal web bars (Type 3) or with only vertical web steel series, the fastest development rate of the diagonal crack
(Type 2). However, this is not so in Series III (a/h = 1.50); occurred in Series III specimens with the highest a/h. Within
Beam III-2/1.50 is stiffer than the corresponding beams III- each series, different web reinforcements led to different
4/1.50 and III-5/1.50. Thus, it is deduced that the effect of rates of diagonal crack development. Diagonal cracks
vertical web reinforcement on beam stiffness is more propagated rapidly in Type 1 control beams with no web
significant when a/h exceeds 1.00 (a/d ≥ 1.13). The effect reinforcement. They also developed more quickly in Type 3

ACI Structural Journal/September-October 1997 575

Table 3—Test results of Series I, II, III specimens
TEST
2Vcr 2Vser 2V n V cr V ser
------------
- ------------
- 2Vs,
Beam* kN V nTEST V nTEST kN Failure mode
I-1/0.75 Diagonal-
160 240 1000 0.16 0.24 0 splitting
I-2N/0.75 Crushing of
240 1320 1520 0.16 0.87 520 strut
I-2S/0.75 280 520 — — — — —
I-3/0.75 Crushing of
240 560 1120 0.21 0.50 120 strut
I-4/0.75 Shear-
280 600 1160 0.24 0.52 160 compression
I-5/0.75 Crushing of
280 920 1550 0.18 0.59 550 strut
I-6N/0.75 240 1080 — — — — —
I-6S/0.75 Diagonal-
320 1320 1550 0.21 0.85 550 splitting
II-1/1.00 Diagonal-
Fig. 4—Serviceability strengths of Series I, II, and III
220 260 510 0.43 0.51 0 splitting specimens.
II-2N/1.00 Diagonal-
200 360 1040 0.19 0.35 530 splitting
II-2S/1.00 200 320 — — — — —
II-3/1.00 Diagonal-
200 400 780 0.26 0.51 270 splitting
II-4/1.00 Diagonal-
140 300 660 0.21 0.45 150 splitting
II-5/1.00 Diagonal-
260 540 940 0.28 0.57 430 splitting
II-6N/1.00 Crushing of
240 720 1340 0.18 0.54 830 strut
II-6S/1.00 240 760 — — — — —
III-1/1.50 Diagonal-
140 140 370 0.38 0.38 0 splitting
III-2N/1.50 Diagonal-
180 400 670 0.27 0.60 300 splitting
III-2S/1.50 Shear-
180 460 800 0.23 0.58 430 compression
III-3/1.50 Diagonal-
160 200 400 0.40 0.50 30 splitting
III-4/1.50 Diagonal-
200 280 380 0.53 0.74 10 splitting Fig. 5—Ultimate shear strengths of Series I, II, and III
III-5/1.50 Diagonal- specimens.
200 320 530 0.38 0.60 160 splitting
III-6N/1.50 Shear- total serviceability loads of all 18 specimens. On the whole,
200 480 920 0.22 0.52 550 compression
Vser is around 24 to 87 percent of the ultimate failure loads
III-6S/1.50 200 680 — — — — — VnTEST (Table 3). Beams with different web reinforcements
*Beam notation as in Table 1
attain different serviceability limits. The greatest Vser is
associated with Type 6 orthogonal web reinforcement,
(ρh = 1.59 percent) and Type 4 (ρh = 3.17 percent) specimens whereas the least is Vser with Type 1 control beams. Note
than in Type 2 (ρv = 1.43 percent or 2.86 percent) beams, that in Series II and III, the combined web reinforcement
suggesting that the vertical web steel was more effective ratios of Type 6 specimens (ρv and ρh) add up to 3.02
than horizontal web steel. In all the three series, Type 6 percent, which is less than ρh of 3.17 percent in Type 5
orthogonal web reinforcement was the most effective in beams in the two series. This confirms that orthogonal web
controlling the diagonal cracks. reinforcement offers the most effective restraint on diag-
Table 3 presents the diagonal cracking strengths (2Vcr), onal crack development of HSC deep beams. On the other
serviceability loads (2Vser) and ultimate strengths (2VnTEST) hand, the diagonal cracks in control beams without any web
of Series I, II, and III specimens. Generally, from Table 3, reinforcement can exceed the serviceability limit very
the diagonal cracking strength remains more or less constant quickly after the first cracking. Thus, to satisfy the service-
within the same series, relatively undisturbed by the type of ability requirement, it is prudent to provide nominal web
web reinforcement. It is also observed that Vcr decreases reinforcement in the shear span, even though it may not be
with increasing a/h. required by calculations.

Serviceability limit Ultimate strength

In this context, the serviceability load Vser is defined as Figure 5 compares the percentage ratios of the ultimate
the load corresponding to an observed crack width of 0.33 mm strengths (VnTEST) of different beams to that of Type 1 beam
(0.013 in.) as specified in the ACI Code. Figure 4 shows the (Vn(Type1)TEST) within each series; for the three control

576 ACI Structural Journal/September-October 1997

specimens, VnTEST/Vn(Type1)TEST = 100 percent. This
comparison is valid as within each series there is no signif-
icant variation in the concrete strengths (Table 1). The term
Vn(Type1)TEST represents the pure concrete contribution to
shear strength Vc. To obtain the web steel contribution to
shear strength Vs, Vn(Type1)TEST is subtracted from VnTEST of
each beam in a series. The values of Vs for different types of
web reinforcement are shown in the seventh column of
Table 3. Clearly, the web steel contribution associated with
Type 6 orthogonal web reinforcement is the highest,
followed on by the Type 2 vertical web steel and the Type 5
heavy horizontal web steel. Thus, the effective contribu-
tions of different types of web reinforcements can be ranked
in a descending order as follows: Type 6 > Type 2 > Type 5
> Type 3 ≈ Type 4. The failure loads of specimens with high
strength horizontal web reinforcement (Type 4) were about
the same as those of beams with smooth horizontal bars
(Type 3). But beams with double the horizontal web rein-
forcement ratio (Type 5) achieved significantly greater load
as compared with either Type 3 or Type 4 specimens.
However, when a/h ≥ 1.00, Type 5 specimens (ρh = 2.58
percent) with high strength deformed horizontal bars, had
smaller Vs as compared with Type 2 beams (ρv = 1.43
percent) with low strength smooth bars, showing that the
vertical web reinforcement was more effective. Thus, it is
confirmed that the effect of horizontal web reinforcement
on shear strength diminishes when a/h exceeds 1.00.
There was a total of six specimens with two different web
reinforcements in both shear spans (Type 2 and Type 6
specimens from the three series in Table 1). The initial
objective was to compare the shear contributions from high
strength deformed bars and plain mild steel bars. In the six
tests, after the first shear span had failed, the beams were
externally clamped with steel yokes and re-loaded to failure.
Unfortunately, only Beam III-2/1.50 was successfully retested;
two values of VnTEST were obtained for both shear spans
(Table 3). Re-testing was not successful for the other five
beams as the concrete strut linking the loading and support
points in the failed shear span was severely damaged.
Apart from Beam I-6/0.75, the other five specimens
failed first in the “N” spans which were reinforced with
lower strength vertical web steel, implying that the ultimate
strengths of the “S” spans with high strength deformed bars
were higher. This indicates that the shear strength contribu-
tions of high strength deformed bars are greater than that of
mild steel bars.

Modes of failure
The failure modes of the 18 specimens are indicated in the
last column of Table 3. Three failure modes are identified,
i.e. crushing of strut failure, diagonal-splitting failure, and
shear-compression failure. In the crushing of strut failure,
there normally exist more than one inclined cracks. The
concrete portion between the inclined cracks is in compres-
sion, forming a concrete compression strut, which crushes Fig. 6(a)—Crack patterns of Series I specimens.
under high compression. This mode of failure is brittle and
sudden. An equally brittle failure mode is the shear-
compression mode in which after the appearance of the edge of the bearing block at loading point and the inside edge of
inclined crack, the concrete portion above the upper end of the bearing block at support point. No explosive sound was
this crack experiences high compression. When the inclined heard for this mode of failure, which was akin to the splitting
crack further propagates upward, the concrete above the failure of a concrete cylinder in a tensile splitting test.
crack fails by crushing, accompanied by a loud noise. The The crack patterns at failure of Series I, II and III specimens
diagonal splitting mode is a less brittle mode in comparison, are shown in Fig. 6(a), (b) and (c) respectively. For Series I
characterized by a critical diagonal crack joining the outside specimens, Beams I-1/0.75 and I-6/0.75 failed in the diagonal-

ACI Structural Journal/September-October 1997 577

 Fig. 6(c)—Crack patterns of Series III specimens.

Fig. 6(b)—Crack patterns of Series II specimens.

specimens. The results are given in Table 4 and Fig. 7.
Only the relevant deep beam design equations are included
splitting mode. The failure was not as explosive as the in this paper. The meanings of the notations used are
crushing of strut mode experienced by Beams I-2/0.75, I-3/ explained under the list of symbols.
0.75, and I-5/0.75 in which loud noise could be heard. An
equally abrupt failure was that of Beam I-4/0.75, which ACI Building Code (ACI 318-89)
failed in shear-compression mode. The failure modes of These design equations are applicable to beams with ln/d
Series II and III specimens are indicated in Table 3. less than 5. However, for Series III beams in which ln/d
exceeds 5, the same two equations are used for comparison.
COMPARISON OF TEST RESULTS WITH SHEAR This is because the shear span-depth ratio a/d of Series III
DESIGN EQUATIONS beams is 1.69 (Table 1), much less than 2.5 which is
Three design methods, namely, the ACI Code 318-89,3 generally16 taken to be the transition point between a deep
the Canadian CSA Code,4 and the UK CIRIA Guide-22 are beam and a shallow beam. The shear design at the critical
used to estimate the ultimate shear strengths of the 18 section (at distance a/2 from the support) is based on:

578 ACI Structural Journal/September-October 1997

 Table 4—Summary of predictions: Series I, II,
V c = ⎛ 3.5 – 2.5 --------u-⎞ ⎛ 1.9 f c′ + 2500ρ w --------
M V u d⎞
⎝
- bw d (ACI Eq. [11-30]) (1)
V u d⎠ ⎝ Mu ⎠ and III
Vn Vn Vn Vn
-,
------------ -,
------------ -,
------------ -,
------------
A 1 + l n ⁄ d⎞ A vh ⎛ 11 – l n ⁄ d⎞ V nTEST V nTEST V nTEST V nTEST
V s = -----v ⎛ -------------------
- + ------- ---------------------- f y d (ACI Eq. [11-31]) (2)
s ⎝ 12 ⎠ s 2 ⎝ 12 ⎠ ACI Canadian CIRIA Modified
Beam* Code Code Guide-2 method
I-1/0.75 0.36 0.79 0.78 0.36
The usual restrictions on Vc, Vs , and Vn are imposed. It I-2N/0.75 0.36 0.52 0.48 0.37
should be noted that in ACI Eq. (11-31), the coefficients in
I-2S/0.75 — — — —
parenthesis are weighting factors16 for the relative effective-
I-3/0.75 0.51 0.73 0.62 0.43
ness of the vertical and horizontal web steel. At ln /d = 5 (the
limit of deep beams according to the ACI Code), vertical and I-4/0.75 0.51 0.75 0.65 0.45
horizontal web steel are taken to be equally effective, i.e. the I-5/0.75 0.36 0.52 0.49 0.36
vertical web steel contribution Vsv is equal to the horizontal I-6N/0.75 — — — —
web steel contribution Vsh provided Av/s is equal to Avh/s2 . I-6S/0.75 0.37 0.53 0.46 0.37
As ln/d decreases below 5, the horizontal web steel becomes II-1/1.00 0.78 1.36 1.45 0.78
more effective. II-2N/1.00 0.59 0.66 0.71 0.59
II-2S/1.00 — — — —
CIRIA Guide-2 “Supplementary Rules” II-3/1.00 0.87 0.89 0.90 0.61
The CIRIA Guide-2 method was based on Kong’s II-4/1.00 1.08 1.13 1.13 0.78
work13,14 in the seventies. Basically, it is applicable for the II-5/1.00 0.76 0.79 0.79 0.65
range of 0 ≤ xe /h ≤ 0.7. In this comparison, it is applied to
II-6N/1.00 0.50 0.51 0.56 0.50
specimens with 0.5 ≤ xe /h ≤ 1.25.
II-6S/1.00 — — — —
III-1/1.50 0.61 0.90 1.39 0.61
x
V c = λ 1 ⎛ 1 – 0.35 -----e ⎞ f cu bh a (CIRIA Guide-2: Cl 3.4.2) (3) III-2N/1.50 0.73 0.50 0.87 0.73
⎝ h a⎠
III-2S/1.50 0.70 0.42 0.84 0.70
III-3/1.50 1.20 0.84 1.29 0.63
⎛ n 100A r y r sin 2αr⎞ III-4/1.50 1.47 0.95 1.45 0.71
V s = λ 2 ⎜ ∑ ----------------------------------
-⎟ bh a (CIRIA Guide-2: Cl.3.4.2) (4)
⎝i = 1 bh a
2
⎠ III-5/1.50 1.46 0.68 1.13 0.57
III-6N/1.50 0.80 0.37 0.68 0.57
III-6S/1.50 — — — —
where λ1 = 0.44 for normal weight concrete, and λ2 = 0.85
Mean 0.74 0.73 0.88 0.57
MPa for plain round bars and 1.95 MPa for deformed bars.
The meanings of the various notations are explained in the SD 0.35 0.25 0.33 0.14
list of symbols. COV 0.47 0.35 0.38 0.25
Note: Vn refers to the predicted ultimate strength by various methods whereas VnTEST
is the experimental failure load of the beam.
Canadian CSA Code (CAN3-A23.3-M84) *Beam notation as in Table 1.
The 1984 Canadian Code recommended a strut-and-tie
design approach for deep beams. In a single-span deep beam
with two-point top-loading, concrete compression “struts”
web steel is overestimated by the current ACI Code. The
would lie in between the upper nodal zone and the lower
reason for the discrepancy is that in ACI 318-89 Eq. (11-31)
nodal zone. Each strut is inclined at an angle αs to a hori-
for web steel contribution, the threshold for ln /d at which
zontal tension tie, which represents the main longitudinal
both horizontal and vertical web steel are equally effective,
reinforcement. The compressive force of the inclined
is probably unrealistic. By modifying the threshold ln /d
concrete strut is balanced at the lower nodal zone by the
from 5 to 2.50 (thereby reducing the weighting factor from
support reaction and the tension tie force, and at the upper
11 to 6 for Vsh), a more consistent result is obtained, as
nodal zone by the external load and the horizontal thrust. The
shown in Fig. 7(b). The proposed revision for Eq. (11-31) is
Canadian Code stipulates that concrete compressive stresses
as follows
in the nodal zones may not exceed certain stress limits.4 The
compressive stress f2 in the inclined concrete strut should not
A 1 + l n ⁄ d ⎞ A vh ⎛ 6 – l n ⁄ d ⎞
V s = -----v ⎛ -------------------
exceed f2max given by:
- + -------- -------------------- f d (6)
s ⎝ 12 ⎠ s 2 ⎝ 12 ⎠ y
λφ c f c ′
- ≤ λφ c f c′
f 2max = --------------------------- (CAN3-A23.3-M84: Eq.11-19) (5)
0.8 + 170ε 1
The revision of the threshold of web steel effectiveness
is in line with the experimental observation for the 18 spec-
where f2max is the diagonal crushing strength of concrete, λ imens. It has been shown that the horizontal web steel
= 1.0 for normal weight concrete, ε1 is the principal tensile contribution is most effective in Series I beams (equivalent
strain crossing the inclined concrete strut, and φc is set to to a/d = 0.85); this effect diminishes for Series III beams
unity. (equivalent to a/d = 1.69). From linear interpolation, the
Figure 7(a) shows that the ACI predictions are conservative threshold works out to be at a/d = 1.27 or thereabouts. If we
for a/h = 0.75; the conservatism disappears for higher a/h assume the clear span ln to be twice the shear span a (as in
and for beams with Type 4 and Type 5 horizontal web steel. the case of central point load), then the value of the
Thus, it is deduced that the beneficial effect of horizontal threshold ln/d is about 2.5. From the last column of Table 4, it

ACI Structural Journal/September-October 1997 579

Fig. 7—Ultimate shear strength predictions by different design equations.

is shown that the modified ACI Eq. (11-31) (Eq. [6]) yields resistance of HSC deep beams than the horizontal web steel
the least standard deviation among the three methods of the same steel ratio.
considered, and its conservatism is consistent for different 3. It is also confirmed that the web steel contribution of
types of web reinforcements. high strength deformed bars is significantly greater than that
The CIRIA Guide-2 predictions (Fig. 7[c]) also overestimate of lower strength plain mild steel bars.
the benefit associated with ρh; the conservatism reduces with The ACI Code overestimates the contribution of horizontal
increasing a/h. Besides, the Guide seems to overestimate the web reinforcement for Series III beams (a/h = 1.50). With a
ultimate shear capacity of Type 1 control beams without any suitable revision to the threshold of web steel effectiveness
web reinforcement, suggesting that the CIRIA Guide-2 may (Eq. [6]), the conservatism of the predictions is maintained.
probably overestimate Vc for HSC deep beams. As for the 4. The UK CIRIA Guide-2 also gives unconservative
1984 Canadian Code, which is essentially a strut-and-tie predictions for specimens with high percentage of horizontal
approach, the shear strength estimations for these 18 specimens web bars. The Guide also overestimates the concrete contri-
are generally conservative and consistent. This is expected bution from high strength concrete.
as the Canadian Code does not take the contribution of web The Canadian Code gives conservative predictions for
steel into account. most of the 18 specimens as the method does not take the
web steel contribution into account.
CONCLUSIONS ACKNOWLEDGMENTS
From the study of 18 HSC deep beams with six different web The funding for this research was provided by the Nanyang Technological
reinforcement details, the following conclusions are made: University, Singapore, through the Applied Research Project No. 9/91.
1. It is evident that web reinforcement can play an important
role for HSC deep beams. The most favorable pattern is the CONVERSION FACTORS
1 mm = 0.039 in
Type 6 orthogonal web reinforcement; it is the most effective in 1 mm2 = 0.00152 in2
increasing the beam stiffness, restricting the diagonal crack 1 kN = 0.2248 kips
width development (thereby increasing the serviceability 1 MPa = 145 psi
load) and in increasing the ultimate shear resistance. At a/h
= 1.50, the effectiveness of orthogonal web reinforcement is NOTATION
greater than the combined individual contributions of the a = shear span measured from center of support to center of loading
point
horizontal and vertical web reinforcements (Column 7 in Avh = area of shear reinforcement parallel to flexural tension
Table 3: 2Vs = 550 kN (Type 6) c.f. 2Vsv = 300 kN (Type 2) reinforcement
and 2Vsh = 10 kN (Type 4)). Ar = area of reinforcing bar
As = area of main longitudinal reinforcement
2. For deep beams with a/h ≥ 1.00 (equivalent to a/d ≥
Av = area of shear reinforcement perpendicular to flexural tension
1.13), the vertical web steel has greater effect on restraining reinforcement
the diagonal crack width and increasing the ultimate shear b, bw = beam thickness

580 ACI Structural Journal/September-October 1997

d = effective depth 3. ACI Committee 318, “Building Code Requirements for Reinforced
h = overall height of beam Concrete (ACI 318-89) and Commentary (318R-89),” American Concrete
ha = active beam height (the lesser of h or le) Institute, Detroit, 1989.
fcu = concrete cube compressive strength 4. Canadian Standards Association, “Design of Concrete Structures for
f c′ = concrete cylinder compressive strength Buildings (CAN3-A23.3-M84),” CSA, Rexdale, Ontario, Dec. 1984, 281 pp.
ft = concrete cylinder tensile splitting strength 5. Tan, K.H.; Kong, F. K., Teng, S.; and Guan, L., “High Strength
fy = yield strength of reinforcement Concrete Deep Beams with Effective Span and Shear Span Variations,”
fyv = yield strength of vertical reinforcement ACI JOURNAL, V. 92, No. 4, July-Aug. 1995, pp. 395-405.
fyh = yield strength of horizontal web reinforcement 6. Mphonde, A. G., and Frantz, G. C., “Shear Tests of High- and Low-
le = effective span as measured from center to center of support Strength Concrete Beams Without Stirrups,” ACI JOURNAL, Proceedings
points V. 81, No. 4 , July-Aug. 1984, pp. 350-357.
ln = clear span measured face-to-face of supports 7. Elzanaty, A. H.; Nilson, A. H.; and Slate, F. O., “Shear Capacity of
s2 = spacing of horizontal web reinforcement Reinforced Concrete Beams Using High-strength Concrete,” ACI JOURNAL,
s = spacing of vertical web reinforcement Proceedings V. 83, No. 2, Mar.-Apr. 1986, pp. 290-296.
V = shear force 8. Ahmad, S. H.; Khaloo, A. R.; and Poveda, A., “Shear Capacity of
Vc = calculated nominal shear strength provided by concrete Reinforced High-Strength Concrete Beams,” ACI JOURNAL, Proceedings
Vcr = measured diagonal cracking strength V. 83, No. 2, Mar.-Apr. 1986, pp. 297-305.
Vn = nominal shear strength (= Vc + Vs) 9. Ahmad, S. H., and Lue, D. M., “Flexure-Shear Interaction of Reinforced
VnTEST = measured ultimate shear strength High-Strength Concrete Beams,” ACI Structural Journal, V. 84, No. 4,
Vs = calculated nominal shear strength provided by shear July-Aug. 1987, pp. 330-341.
reinforcement 10. Roller, J. J., and Russell, H. G., “Shear Strength of High-strength
Vser = measured serviceability load (ACI 318-89: Clause 10.6.4) Concrete Beams with Web Reinforcement,” ACI Structural Journal, V. 87,
Vsh = calculated nominal shear strength provided by horizontal web No. 2, Mar.-Apr. 1990, pp. 191-198.
steel 11. Sarsam, K. F., and Al-Musawi, J. M. S., “Shear Design of High- and
Vsv = calculated nominal shear strength provided by vertical web steel Normal Strength Concrete Beams with Web Reinforcement,” ACI Structural
ρ = longitudinal main steel ratio (= As /bd) Journal, V. 89, No. 6, Nov.-Dec. 1992, pp. 658-664.
ρv = vertical shear reinforcement ratio (= Av /bs) 12. Rogowsky, D. M.; MacGregor, J. G.; and Ong, S. Y., “Tests of
ρh = horizontal shear reinforcement ratio (= Avh /bs2) Reinforced Concrete Deep Beams,” ACI JOURNAL, Proceedings V. 83,
xe = clear shear span measured from face of support to face of No. 4, July-Aug. 1986, pp. 614-623.
loading point 13. Kong, F. K., and Robins, P. J., “Web Reinforcement Effects on Light-
yr = the depth at which a typical web bar intersects a critical weight Concrete Deep Beams,” ACI JOURNAL, Proceedings V. 68, No. 7,
diagonal crack July 1971, pp. 514-520.
14. Kong, F. K.; Robins, P. J.; and Cole, D. F., “Web Reinforcement
REFERENCES Effects on Deep Beams,” ACI JOURNAL, V. 67, No. 12, Dec. 1970, pp.
1. British Standards Institution, “Structural Use of Concrete,” (BS 8110: 1010-1017.
Part 1. Code of Practice for Design and Construction), BSI, London, 1985. 15. Smith, K. N., and Vantsiotis, A. S., “Shear Strength of Deep Beams,”
2. Ove Arup & Partners, “The Design of Deep Beams in Reinforced ACI JOURNAL, Proceedings V. 79, No. 3, May-June 1982, pp. 201-213.
Concrete (CIRIA Guide-2),” Construction Industry Research and Information 16. Nilson, A. H., and Winter, G., Design of Concrete Structures, 11th
Association, London, Jan. 1977 (Reprinted 1984), 131 pp. edition, Chapter 4, McGraw-Hill, Inc., 1991.

ACI Structural Journal/September-October 1997 581