HKIE Transactions, 2013

Vol. 20, No. 1, 34–51, http://dx.doi.org/10.1080/1023697X.2013.785084

Some explanatory/commentary notes on reinforced concrete detailing in accordance with the

Code of Practice for Structural Use of Concrete 2004
Law Chi Wai∗ and J.Y.W. Mak
Housing Department, HKSAR Government, Hong Kong, People’s Republic of China
(Received 18 November 2011; final version received 29 December 2011 )

A general review on the requirements of reinforced concrete detailing of a number of items by the Code of Practice
for Structural Use of Concrete 2004 has been conducted, with discussions on the underlying principles together with
comparisons with other reputable national and international codes of practice. These items include (1) omission of bends
and hooks in compression anchorages; (2) limitations of laps of bars in a single section; (3) transverse reinforcements
in lap zones; (4) transverse reinforcement for torsion; (5) use of 135◦ hooks; (6) maximum tension reinforcements in
beams; (7) maximum diameter of column reinforcing bars passing through beams when columns are designed to develop
plastic hinges in the end regions; (8) the “middle quarter rule” for main reinforcement lapping in columns; and (9)
Downloaded by [Law Chi Wai] at 18:34 01 July 2014

densified transverse reinforcements in columns at end regions. The adequacy and necessity of the detailing requirements
and difficulties in fixing are discussed, together with proposal of alternatives as appropriate.
Keywords: ductility; bond; transverse reinforcements; bends and hooks; plastic hinge

Introduction Omission of bends and hooks in compression

Certain aspects of new reinforced concrete (RC) detail- anchorages
ing requirements in the Code of Practice for Structural The requirement is stated in 8.4.6 of the Code read-
Use of Concrete 2004 [1] (hereafter referred to as “the ing “Bends and hooks do not contribute to compression
Code”) have some impacts on the culture of the local anchorages” which is identical to that stipulated in EC2
construction practice which was largely inherited from 8.4.1(3). In addition, similar requirements are also found in
the British Standards. The Code, though mainly based on the current American and New Zealand Standards, i.e. ACI
the British Standard BS 8110 [2], has incorporated some 318-08 (12.1.1) [5] and NZS 3101:2006 (8.6.10.4) [4].
practices from the Eurocode EC2 [3] and ductility require- Both codes state that hooks shall not be considered effec-
ments from the New Zealand Standard NZS 3101:2006 [4] tive in developing bars in compression. The requirement is
which are unfamiliar to Hong Kong and some of them not found in the British Codes and is not a usual practice
are difficult to implement by the industry. Though these in the local industry. Adoption of the new requirement
new requirements in detailing do improve the integrity, will affect the use of comparatively large reinforcement
ductility and durability of the structures and some of bar under compression which has to be bent and extended
them should be upheld, others, especially the ductility for some length beyond the bend when the bar has to be
requirements, may not be absolutely necessary as quite a anchored into structural members such as thick plates or
number of them are imposed in regions of strong seismic- beams.
ity whereas Hong Kong is in a region of low to moderate Explanations of the required omission are not explic-
seismicity. So alternatives achieving acceptable standards itly stated in the codes and their explanatory notes except
may be explored and employed instead. In the following NZS 3101:2006 has some elaborations in its commen-
sections of this paper, various aspects of these detailing tary part under C8.6.10.3 with the heading “Development
practice are reviewed with reference to other reputable length of standard hooks in tension” which reads “When
national/international codes of practice and publications. the same bar is required to develop yield strength in
The underlying principles are discussed and alternatives compression, the bent portion of the anchorage must
are proposed as appropriate. be disregarded in satisfying the requirements of 8.6.5.1

∗ Corresponding author. Email: [email protected]

© 2013 The Hong Kong Institution of Engineers

 HKIE Transactions 35

Anchorage Bar Anchorage Bar

in Tension in Compression

Bearing stress Fbt

Bearing stress (outside bend)
(inside bend) on Fbt / rf on
concrete of relative Fbt
concrete of relative
thick mass
thin mass that can
Fbt / rf
easily spall off

Concrete that
r may spall off
r

Figure 1. Bearing stress created by anchorage hook on concrete.

(for calculation of anchorage length). However, when the cross-sectional area of the bar, the designer has to
bars are anchored in column cores, as described above, rely on the provision of a cross bar with a diameter not
Downloaded by [Law Chi Wai] at 18:34 01 July 2014

the confinement may be considered to be equivalent to less than that of the main bar inside the bend for dissipa-
that implied in 8.6.5.3 (for calculation of compression tion of stress so as to be exempted from the checking of
anchorage length). The development of bars in compres- bearing stress in accordance with the same clause of the
sion will commence closer to the inner face of exterior Code.
columns.” The determination of bearing stress due to a bar in
The restriction in the use of bend and hook (and prob- compression can in fact be similar, with an even greater
ably also including the straight portion beyond the bend) projected length which is the external radius of the bend.
in compression anchorage of bar is due to the generally However, the allowable bearing stress should normally
lack of adequate concrete cover to resist the high bearing be below 2fcu /(1 + 2φ/ab ) for tension bar unless there
of the hook on concrete in compression anchorage as per is a comparable concrete cover or very good confine-
the usual practice of anchorage of bar at the far face as ment as described in the New Zealand Code such as the
illustrated in Figure 1. From the figure, bend in tension interior of the “column core” with a strong reinforce-
bar usually has a thick mass of concrete to resist the high ment cage formed by vertical bars and closed stirrups.
bearing created by the bend on the concrete while the con- Alternatively, cross bars as similar to that for tension bar
crete cover on the compression bar is relatively thin which can be added on the “outside” of the bend to effectively
can easily spall off. dissipate the bearing stress or “hold back” the hook. How-
The Code has provided a formula in its Equation (8.1) ever, the cross bars have to be well anchored into the
for checking of bearing stress on the inside of the bend of a concrete structure (say in form of U-bars) as the rela-
bar under tension. The checking is to ensure that the bear- tively thin concrete cover outside may spall under high
ing pressure (which is the force transmitted to the bend pressures. For better understanding of the phenomenon
divided by the “projected area of the bar” in the direction and quantification of the parameters such as cover thick-
of the force) will not exceed twice the cube strength of ness and anchorage details for general use, laboratory
concrete with some discount to adjacent bar having sim- tests should be conducted. So there should be alterna-
ilar bearing on the concrete. The equation is actually not tives allowing hooks and bends to be used in compression
easy to satisfy in practice. For example, a T40 bar with anchorages with provisions to ensure no concrete bearing
100 mm centre to centre spacing with other adjacent T40 failures instead of strictly not allowing bends to con-
bars will exert a bearing pressure of Fbt /rφ = 78.6 MPa tribute to compression anchorages. The effectiveness of
under the minimum internal bend radius (r = 4φ) accord- the straight portion of the bar beyond the bend should also
ing to Table 8.2 of the Code if its axial stress at the start of be studied.
the bend is 0.87 fy . The bearing stress is then well in excess
of the allowable bearing pressure of 2fcu /(1 + 2φ/ab ) =
37 MPa calculated by Equation (8.1) if the concrete grade Limitation of laps of bars in a single section
is 35 and cover is 50 mm. So unless there is a consid- The requirement of staggered laps is stated in 8.7.2 of the
erable length of bond before the start of the bend which Code which is identical to EC2 of the same clause no. By
can exhaust a large portion of the axial load in the bar the clause, laps should normally be staggered and comply
such that Fbt is substantially below 0.87 fy where As is with Figure 8.4 of the Code as extracted in Figure 2.
 36 C.W. Law and J.Y.W. Mak

Figure 2. Adjacent laps (Figure 8.4 of the Code).

“Staggered laps” are favoured because of the phe- sizes in a particular layer to 40% of section breadth may
nomenon that “curtailment of a substantial amount of govern so that alternate lapping will still be required.
reinforcement at any section could lead to the development
of large longitudinal and transverse cracks” as pointed
Transverse reinforcements in lap zones
out by Kwan [6]. Kwan [6] further adds “It is there-
fore advisable to stagger the curtailment points in heavily The requirements are stipulated in 8.7.4 of the Code and
reinforced members.” In addition, the practice also helps again they are identical to EC2 of the same clause no.
ensure proper consolidation of concrete as the sum of By the clause, transverse reinforcements located between
Downloaded by [Law Chi Wai] at 18:34 01 July 2014

reinforcement sizes in a particular layer can be reduced. the bars and the surface of the concrete are required to be
Though a good practice and not too difficult to adopt, provided at the lap zones of main reinforcements of diam-
the staggered lapping can create some inconvenience to eter not less than 20 mm. The provision serves to reduce
drafting and construction. In fact, the practice is origi- cracking or “splitting” of the concrete cover and therefore
nated from EC2. However, EC2 does allow percentage of increases bond strength of the main bars as splitting of the
lapped bars at a section to be over 50% which implies concrete cover is a cause of bond failure. In fact, the pro-
inconsistency with Figure 2. By 8.7.3 and Table 8.3 of vision also improves ductility as failure of splices without
EC2, the coefficient α6 to account for the effect of dif- transverse reinforcement is sudden and complete while
ferent percentages of laps in a section on lap those with transverse reinforcement tend to exhibit a less
√ length over brittle failure and also possess residual strength beyond
the anchorage length is quantified as α6 = ρ1 /25 where
1.0 ≤ α6 ≤ 1.5 and ρ1 is the percentage of reinforcement the maximum load. [7] The following may be helpful in
lapped within 0.650 from the centre of lap length. So understanding the background of the requirement:
the coefficient is increased from 1.4 (same as the nor-
mally required coefficient in accordance with 8.7.3.2 of the (i) It is well established that a deformed bar bonded
Code) to 1.5 when staggered lap (ρ1 = 50) is changed to in concrete under tension or compression will cre-
100% lapping in one section, implying the requirement of ate “bond force” which “radiates” out into the
an increase of lap length by 7%. The lengthened lap length surrounding concrete from the bonding surface.
helps reduce crack formations at lap ends to satisfactory This force is originated from the component of
levels. the bearing force on the concrete by the ribs of
Similar provision is also found in ACI 318-08. 12.5.2 the deformed bar in the radiating direction per-
of ACI 318-08 classifies lap with (i) one-half or less pendicular to the bar. The bond force will in
of the total reinforcement within the lap length (includ- turn generate tensile stress in the circumferen-
ing staggered lapping as illustrated in Figure 2); and tial direction as shown in Figure 3 which may
(ii) area of reinforcement provided at least twice of that split the concrete. Tepfers [8] and Kemp [9] had
required over the entire lap length as Class A where the lap put forward equations for calculation of the ten-
length is 1.0 times the anchorage length. The anchorage sile stress by regarding the radial bond stress as
lengths of others (including 100% lapping in a section) hydraulic pressure on the inner surface of a thick-
are Class B with lap lengths 1.3 times the anchorage walled concrete cylinder surrounding the steel bar
lengths. in accordance with the work of Timoshenko [10].
So it is recommended that the Code should follow EC2 The equation for the elastic stage is shown as
by incorporating a provision such that 100% lapping in follows:
one section is allowed by an increase of lap length of (d/2)2 τ tan α
7% (comparing the factor 1.5 and 1.4), or a bit more, σt =
(Cy + d/2)2 − (d/2)2
say 10% generally. For very heavily reinforced sections  
where staggered lap is favoured regardless of this provi- (Cy + d/2)2
× 1+ , (1)
sion, other rules such as limiting sum of reinforcement r2
 HKIE Transactions 37

Circumferential
tension

r
Cy

2 Cy + d 22 Tension cracks induced in

d 2 t tan a
st = 1+ concrete due to shear by the
2 2 2
Cy + d 2 - d 2 r reinforcement bar on concrete

where s t is the concrete tensile stress in the

circumferential direction and t is the shear stress
by the reinforcement bar on concrete

Figure 3. Tensile stress induced on concrete due to bonds by reinforcement bar.

Downloaded by [Law Chi Wai] at 18:34 01 July 2014

Figure 4. Cracks (potential cracks) created in lapped bars intercepted by transverse reinforcements.

where the symbols are defined in Figure 3 except α concrete is

which relates the shear of the bar to the bond force
e = 0.486(Cy + d/2). (3)
which is taken as 45◦ when cover cracking begins.
By Equation (1), thinner concrete cover and higher The expression was later modified by Esfahani and
shear stress will incur higher tensile stress that Rangan [11,12] to account for the plastic defor-
may crack the concrete. In addition, if the tensile mation of the concrete by regression analysis on
stress induced on the concrete exceeds the tensile experimental data as
strength of concrete, tension cracks will take place
C/db + 0.5
in the weaker locations such as the concrete cover uc = 4.9 fct . (4)
and the space between two bars also illustrated in C/db + 3.6
Figure 3. In Equation (4), the symbols fcbc has been replaced
Tepfers [8] had further considered the “partly by uc as the bond strength of concrete; and Cy and
cracked elastic” stage by which the maximum d have been replaced by C and db , respectively.
tensile stress of concrete had been exceeded and (ii) As concrete is weak in tension, transverse rein-
concrete cracks within an inner core surrounding forcements can be placed to “intercept” the tension
the reinforcement bar while the outer core remains cracks and thus act as some provisions against
in the elastic stage. He considers that the bond “bond failure” due to concrete splitting (while
force from the reinforcement bar is transferred another failure mode by bond is the pulling out
through the “concrete teeth” in the cracked inner of the bar). So alternatively saying, bond strength
core to the outer elastic core and put forward a of concrete may be increased. In locations of
theoretical expression relating the maximum bond laps where the splitting forces from the bars are
stress fcbc that can be sustained by the bar to the additive, cracking will take place more easily.
tensile strength of concrete as Examples of provisions of the transverse rein-
forcements in critical locations of laps to intercept
Cy /d + 0.5
fcbc = fct , (2) the cracking (or potential cracking) are shown in
1.664 Figure 4.
where fct is the tensile strength of concrete (iii) 8.7.4 of the Code requires transverse reinforce-
and the radius of the inner core with cracked ments to be placed at the end thirds of the lap
 38 C.W. Law and J.Y.W. Mak

Area of concrete = Ac

s s0
Cy
s s0
Cy

l/2 l/2 Area of

steel = As Cx dx dx
2
r = As / Ac

Figure 5. Illustration of the modulus of displacement theory.

Shear Stress distribution of a T40 lapping with another Identical Bar

Elastic Plastic (no bond strength enhancement) Plastic (bond strength enhanced by transverse rebar)

0
–1000 –900 –800 –700 –600 –500 –400 –300 –200 –100 0 100 200 300 400 500 600 700 800 900 1000
–2

–4
Shear stress (MPa)

–6
Downloaded by [Law Chi Wai] at 18:34 01 July 2014

–8

–10

–12

–14

–16
Distance from centre of bar (mm)

Figure 6. Typical plots of variation of shear stress along lapped bars.

because higher splitting stresses are found at where τ1 and τ2 are the shear stress of the two bars
the ends due to the higher bond stresses in the in the lap at location x; x is zero in the middle
locations. By the work of Tepfers [13] and Giuri- of the lap; l/2 at the right end, −l/2 at the left
ani [14] and others, the shear stress (the bond end and l is the length of the lap; σs0 is the axial
stress) is not uniform along the lap. In Tepfers’ stress of each of the bar at its end; As is the cross-
“modulus of displacement theory” in which a con- sectional area of each of the reinforcement bar;
stant named the K-modulus (the proportionality u is the perimeter of each of the reinforcement
relating the shear stress of the bar to the “slip” bar; κ2 = AuKs Es
with K as the K-modulus measured
between the bar and concrete) was introduced and by Tepfers [13] as the secant modulus ranging
by considering equilibrium and compatibility and from 50 × 103 MN/m3 to 100 × 103 MN/m3 for
with the assumption of no through cracks within 0.2 mm slip depending on concrete grade; Es is
the lap zone, Tepfers [13] has derived the formulae Young’s modulus of steel; n is ratio of Young’s
of shear stress distribution of the two equal bars modulus of steel to concrete; and ρ is the steel
in a lap under constant bending moment as shown ratio as illustrated in Figure 5.
in Figure 5 as A typical plot for the shear stress of a T40 bar
lapping with another T40 bar for a lap length of 2 m
σs0 As κ2 cosh κ2 x
τ1 = − by Tepfer’s approach is presented in Figure 6. The
2u sinh(κ2 l/2) bars are stressed to 0.87 times of steel yield stress
σs0 As κ1 sinh κ2 x at the ends of the lap, i.e. σs0 = 400.2 N/mm2
+ , (5a)
2u(1 + 2nρ) cosh(κ2 l/2) and embedded in grade 35 concrete (long term
σs0 As κ2 cosh κ2 x Young’s modulus taken as 11.85 kN/mm2 ) and K
τ2 = is taken as 50 × 103 MN/m3 . It can be seen that
2u sinh(κ2 l/2)
the shear stresses are high at the ends of the laps
σs0 As κ1 sinh κ1 x and drop rapidly to low values in the middle for the
+ , (5b)
2u(1 + 2nρ) cosh(κ1 l/2) elastic stage. If the bond stresses of concrete at the
 HKIE Transactions 39

ends are capped (i) to the ultimate value imposed the minimum cross-sectional area of the transverse
by EC2 as 4.65 MPa and (ii) after enhancement by reinforcement according to the Code is provided
38% to 6.42 MPa due to the addition of transverse which is assumed to take up all the tension force in
reinforcement as required by 8.7.4 of the Code, the cover, it can easily be proven that the factor of
the stress profiles are re-analysed as the “plastic safety defined as the load carrying capacity of the
stage” and added in Figure 6. The enhancement transverse bar (also stressed to 0.87 fy ) divided by
of 38% is arrived at by the approach of Esfahani Ft will be π = 3.14. So the provision by the Code
and Kianoush [15] which will be described later. is very adequate for the ultimate limit state.
Bars of other sizes exhibit similar behaviour. (vi) It should be re-emphasised that transverse rein-
So placing transverse reinforcements at the outer forcements can only increase bond strength by
zones of the lap is an effective means against con- strengthening concrete against splitting whereas
crete splitting due to high bond stresses and thus bond failure can still take place by “pulling
enhances bond strength. out”. So the provision of transverse reinforce-
(iv) The Code has provided simple rules in its 8.7.4 for ment can only increase the bond carrying capacity
provision of transverse reinforcements at laps to to a certain extent. In addition, Esfahani and
reduce cracking of concrete. By the rules, no trans- Kianoush [15] pointed out that the stress in the
verse reinforcement is required when the diameter transverse reinforcement depends on the state
of the lapped bar is 16 mm or less, or the lapped of cracking of concrete cover. For non-cracked
bars in any section are less than 20%. Other- cover, the stress of transverse reinforcement is
Downloaded by [Law Chi Wai] at 18:34 01 July 2014

wise transverse reinforcements of total areas not negligible; thus the presence of transverse rein-
less than the cross-sectional area of the spliced forcement may not increase the bond strength sig-
bar should be provided in the space between the nificantly. After concrete cracks, the force transfer
spliced bar and the concrete surface. It is because will be larger especially at high Cd /d ratio.
for the smaller bars, the minimum reinforcement
is considered adequate to cope with the tensile Esfahani and Kianoush [15] have by experiments estab-
stresses generated at the laps in accordance with lished empirical formulae to account for the contribution
Beeby and Narayanan [7]. In addition, the trans- of the transverse reinforcements to the bond strength. They
verse reinforcements are to be positioned at the concluded that the bond strength can be increased by a fac-
outer thirds of the lap. The reason for such posi- tor of 1 + 0.015fR ACS
t Ab
if transverse reinforcement of area
tioning has been discussed in (iii) above. The 2
At in mm with spacing S in mm is provided. In the factor,
Code also requires extra transverse bars outside Ab is the area of the single spliced bar in mm2 and C (in
the lap for compression lap as some end-bearing mm) is the minimum of (1) the side cover to the spliced
by the main bars on the concrete may create burst- bar; (2) bottom (or top) cover to the spliced bar; and (3)
ing force outside the lap zone as explained by half of the clear spacing with the adjacent bar plus half
Kwan [6]. of diameter of the spliced bar in mm; fR is the factor to
(v) It may be worthwhile to check the structural account for the “rib effects” of the reinforcing bar which
adequacy of the Code requirement of transverse is 1 for bar with relative rib area ratios R < 0.11 and 1.6
bars by the work of Tepfers [13]. It can eas- for R ≥ 0.11. So for the T40 bars with transverse rein-
ily be deduced from Figure 7 that the tensile forcements of total area equal to that of the T40 added in
force required to maintain equilibrium by the the lap according to the Code, the bond strength can be
cover is Ft = dτ/2 where τ is the shear per cir- increased by 38%.
cumferential length of the bar over the entire Similarly other codes including EC2, ACI 318-08 and
lap length. If the longitudinal bar is stressed to NZS3101-2006 have included the effects of the transverse
0.87 fy so that π(d 2 /4)0.87fy = Ft = dτ/2 and reinforcements together with concrete cover thickness in

Reaction by
tension = dτ/2

Total horizontal “bond

force”=σ rd = dτ
Transverse
Bar to take up
Reaction by
the tension
tension = dτ/2

Figure 7. Tensile stress induced on concrete due to bonds by reinforcement bar.

 40 C.W. Law and J.Y.W. Mak

Comparison of Tension Lap Lengths of Deformed Bar among the

Code, EC2 and ACI318-08, NZS3101-2006 for Grade 35
CoPConc2004 EC2 ACI 318-08 NZS3101-2006
2000

1600

Lap Length (m)

1200

800

400

0
10 15 20 25 30 35 40
Bar Diameter (mm)

Figure 8. Comparison of lap lengths among the Code, EC2 and ACI 318 in Grade 35 concrete.

the determination of the bond and lap lengths. However, should be closed and be anchored by means of laps or
the Code has adopted a simple approach by BS 8110 in its hooked ends, see Figure 9.3, and should form an angle of
8.4.4 by relating bond strength to concrete grade and bar 90◦ with the axis of the structural element.” These words
type (plain or deformed bar) only. A comparison of the are identical to EC2 of the same clause no. However, there
required tension lap lengths for the 4 codes for T12 to T40 are differences and omissions in the illustrative diagrams
Downloaded by [Law Chi Wai] at 18:34 01 July 2014

bars in concrete grade 35 are plotted in Figure 8. The con- of the Code as compared with that of EC2 which are sum-
crete covers assumed are 25 mm for T12 to T20; 40 mm marised in the paragraphs that follow. The diagrams of
for T25 and 50 mm for T32 and T40. Transverse rein- the Code and EC2 are reproduced as Figures 9(a) and 9(b)
forcements required by the Code are provided where bond of this paper for ease of reference. In Figure 9(b), an
strength enhancements applicable to EC2, ACI 318-08 and additional arrangement suggested by the Concrete Code
NZS 3101-2006 are adopted. The lap lengths required by Handbook by Kwan [6] has been appended.
the Code are generally the longest whilst that of NZS 3101
and ACI 318 are the shortest as they allow the enhance- (i) In (a1) of EC2, the closed links with anchorage by
ment which can be up to 2.25 and 2.5 times the basic 135◦ hooks at the beam corner require only short
bond strength respectively by the aggregate factors due anchorage lengths beyond the hooks, presumably
to cover and transverse reinforcements. In addition, NZS 5 times bar diameter (Figure 8.1(c) of EC2) whilst
3101 allows lap length to be identical to anchorage lengths that in the Code requires full tension anchorage
while other codes require additional factors in the order length.
of 1.3–1.4 generally. The sudden increase of lap length (ii) In (a2) of EC2, one leg of the closed link can end
for T40 bar by EC2 is due to the reduction of ultimate in straight portion instead of a 90◦ bend in one of
bond strength for bar size in excess of T32 by EC2 (from the two alternatives whilst the Code has omitted
2.9–2.67 MPa). this alternative.
So it is advisable to take cover and transverse rein- (iii) (a3) in EC2 has been omitted in the Code. By
forcements into account in the determination of lap and (a3), requirements for the transverse reinforce-
anchorage lengths in the Code which is more realistic. ments are obviously less stringent when there are
In addition, shorter lap and anchorage lengths can be connections to the beam by adjacent slabs.
achieved. As the provision of transverse reinforcement
in the Code follows that of EC2, reference should there- The requirements by the ACI 318-08 are even simpler.
fore be made to EC2 8.4.4 for the determination of bond By its 11.5.4.2 and R11.5.4.2, 135◦ standard hook or seis-
strength enhancement due to cover and transverse rein- mic hook (hook with bends not less than 135◦ and each
forcements and subsequently the reduction of lap and with extension of 6 times bar diameters beyond the bend
anchorage lengths. The reduction can be effected by the into the interior of the beam) are required. In addition, 90◦
adoption of coefficients of α2 and α3 as indicated in hook can be used if there are adjacent slabs or flanges.
Table 8.2 of EC2. By the required provision of transverse Before commenting on the adequacy and necessity of
reinforcements in the Code, 10% reduction of lap length the details, it may be worthwhile to conduct a review
or anchorage length can generally be achieved. on the behaviour of torsion and the requirements of the
reinforcements in accordance with the work of Lam-
pert [16] and Mitchell and Collins [17]. Figure 10 shows
Transverse reinforcement for torsion how a cracked reinforced (or prestressed) concrete beam
The requirements of the transverse reinforcements for tor- which is in an ultimate limit state resisting torsion. In
sion are stated in 9.2.3 of the Code reading “Torsion links the figure, the torsion will first create shear stress in the
 HKIE Transactions 41

Torsion link arrangements in the Code Additional torsion link arrangement

in the Handbook by Kwan [6]
Downloaded by [Law Chi Wai] at 18:34 01 July 2014

Figure 9. (a). Torsion reinforcement arrangements in EC2. (b). Torsion reinforcement arrangements in the Code and Concrete Code
Handbook.
Note: The second alternative for (a2) (lower sketch) should have a full lap length along the top.

Corner can
easily spall off
transverse
outward thrust at beam rebar
corner by concrete
compression and held
back by transverse
tension
rebar crack

complementary
shear stress

Applied Torsion
shear stress

longitudinal diagonal diagonal

rebar compression compression diagonal
tension

Figure 10. The diagonal compression stress field model for torsion.

beam section. A cubical element as shown is enlarged “spall” off as shown. In fact, the compression will cre-
for detailed explanation. In the element, complementary ate outward thrusts on the whole beam. The longitudinal
shear stress, diagonal tension and compression are created and transverse reinforcement bars are therefore required to
according to the law of mechanics. In the ultimate limit hold the beam together in the two directions which prevent
state, the diagonal tensions will crack the concrete and the the beam from “splitting”. Adequate anchorages of both
diagonal compressions which “spiral” around the beam the longitudinal and transverse reinforcements are there-
will take place to achieve equilibrium. fore essential. So if the transverse reinforcement is in the
From Figure 10, it is obvious that the diagonal com- form of a stirrup lapped at a beam corner with 90◦ bends
pression will create outward thrust at the corner and the as shown in Figure 11(a), the stirrup may result in com-
concrete outside the transverse reinforcement may easily plete loss of anchorage upon spalling of the cover at high
 42 C.W. Law and J.Y.W. Mak

(a) (b)

Beam with Complete loss of Beam with Complete loss of

stirrups anchored anchorage of stirrups each anchorage of
by 90o bends at stirrup after formed by 2 stirrup after
corner. spalling of cover. lapped U-bars spalling of cover.

Figure 11. (a) Loss of Anchorage of Stirrups formed by 90◦ bend end. (b) Loss of Anchorage of Stirrups formed by lapped U-bars.

(a) (b) (c)

Downloaded by [Law Chi Wai] at 18:34 01 July 2014

Figure 12. Effective transverse reinforcements for torsion.

torsion, likewise the splice of the 2 U-bars if the stirrup is In the alternative (c) where there is a restraint by an
formed by lapping of 2 U-bars as shown in Figure 11(b). adjacent slab by which opening up of the stirrup leg can be
In such cases, the stirrup will become ineffective as they avoided, 90◦ hook can be used as allowed by ACI 318-08
“open up” and fail to hold the beam to prevent it from fur- R11.5.4.2 and Figure 11.5.4.2(b). There are other options
ther splitting. The use of a pair of U-bars as in Figure 11(b) that can be found in the Annex of the letter addressing to
is also explicitly forbidden in ACI 308-08 by its R11.5.4.1. all authorised persons, registered structural and geotechni-
However, the anchorage by 135◦ bend as shown in cal engineers and contractors by the Buildings Department
(a1) of Figure 9.6 of EC2 (reproduced in Figure 9(a) of dated 29 April 2011. [19]
this paper) will still be effective in holding the inner core
of the beam even the cover spalls off as the straight por-
tions beyond the 135◦ bends are well anchored in the inner Use of 135◦ link
core of the beam even without full anchorage length mea- The use of 135◦ link has been extensively specified
sured from the start of the bend. Tests on a hollow beam in many codes, especially the seismic codes includ-
at the University of Toronto [18] with 105◦ bends have ing ACI 318, NZS 3101 and the China Seismic Code
justified the effectiveness. Whilst fixing transverse rein- GB50011-2010 [20]. The Code requires 135◦ link under
forcements in accordance with the requirements of the the following circumstances:
Code are considered difficult by the local industry in gen-
eral, alternatives with bend ends well anchored into the (i) 9.5.3.2 requires links in rectangular or polygo-
interior of the beam to avoid “opening up of the links” nal columns be adequately anchored by means of
upon spalling of covers should be acceptable. Details sim- hooks bent though an angle of not less than 135◦
ilar to those shown in Figure 12 are considered effective and lapping circular links in circular column be
alternatives. In these alternatives, though the transverse anchored by terminating each end of the link with
reinforcements are made of two or more pieces of hooks, at least a 135◦ hook bent around a longitudinal bar
there are adequate anchorages into the interior of the beam. and overlapping the other end of the link.

The end of the 135o

90o bend can bend is anchored
easily open up and into the interior of
spall the concrete the concrete member
cover under and thus risk of
tensile load. “opening up” is
reduced.

Figure 13. Comparison of anchorage between 90◦ and 135◦ bends.

 HKIE Transactions 43

Figure 14. Demonstration of effects of transverse reinforcements on flexural strengths of columns by Sheikh and Khoury [21].

(ii) 9.9.1.2 and 9.9.2.2 requires links in beams and Similar works have been done by others including
columns contributing in the lateral load resisting Wehbe et al. [22] who had tested four RC columns with
Downloaded by [Law Chi Wai] at 18:34 01 July 2014

system should be adequately anchored by means transverse reinforcements comprising 135◦ hook periph-
of 135◦ or 180◦ hooks. ery ties and crossties. Subsequently, the failures were all
initiated by the opening up of the 90◦ crossties. Neverthe-
less, ACI 318-08 R21.6.4.2 allows consecutive crossties
As illustrated in Figure 13, the use of 135◦ (or a greater engaging the same longitudinal bar have their 90◦ hooks
bent angle) link is to ensure better anchorage of the link on the opposite sides of column.
in concrete by anchoring its end into the “inner core” of Though the 135◦ hook is unarguably the better option,
the concrete member so as to reduce the risk of “open- the 90◦ hook is more popular as it has the relative ease of
ing up” of the link under heavy and/or cyclic load (in an placement. The 135◦ hook is much more difficult to place
earthquake) which may subsequently lead to premature especially when the cast-in bars are misaligned. How-
buckling of the main bars. ever, if there are other physical restraints such as adjoining
The integrity of the link also serves to confine the beams or slabs preventing the opening up of the 90◦ hook
inner concrete core of the member even the surface spalls as similar to that illustrated in Figure 12(c) for torsion
off. The risk of total failure can therefore be reduced which is the main shortcoming of the hook, the use of it
and in addition, ductility is also increased. Sheikh and should be acceptable. More examples of these options can
Khoury [21] had performed load tests with cyclic flexure be found in the Annex of the letter addressing to all autho-
and shear on six RC columns subjected to constant axial rised persons and registered structural and geotechnical
load. Of the three specimens FS-9, ES-13, and AS-17 with engineers and contractors by the Buildings Department
the same longitudinal reinforcements and same quantities dated 29 April 2011.[17]
of transverse reinforcements (though bar sizes, spacing
and arrangement are different), the specimen FS-9 having
crossties (a crosstie is a link with a 90◦ bend at one end and
a 135◦ or greater bend at the other end) showed a sudden Maximum tension reinforcements in beams
drop of flexural strength at the ultimate curvature as shown 9.9.1.1 of the Code has limited the maximum area of
in Figure 14 at which the 90◦ bends opened up. Clearly, longitudinal tension reinforcement for beams contribut-
the performance of FS-9 was poorer to that of AS-17 with ing in the lateral load resisting system to 2.5% of the
135◦ bends links which possessed both higher strength and cross-sectional area of concrete which is also the abso-
higher ductility. Nevertheless, ES-13 showed the poorest lute maximum imposed by NZS3101-2006 9.4.3.3 for
performance as the side longitudinal bars without effective beams containing “ductile plastic regions” as an “addi-
tie buckled and pushed the tie outward, resulting in a loss tional design requirement for ductility in earthquakes”.
of confinement and a rapid deterioration of the specimen. The 2.5% is a reduction as compared with the previous
The test demonstrated the importance of the transverse version of the Code [23] which has allowed a higher
reinforcements in the confinement of the concrete core percentage of 4%. This reduction requires larger struc-
and prevention of buckling of the longitudinal bars which tural sizes to be used in some cases which may lead
lies not only on the amount of reinforcements added but to undesirable problems in structural and architectural
also in the arrangement and effectiveness of anchorages. planning.
 44 C.W. Law and J.Y.W. Mak

As the requirement is under the heading “ductility”, However, NZS3101-2006 has imposed similar restric-
the purpose of the reduction is to ensure certain ductil- tion of bar size to beams in its 9.3.8.4 which is not found
ity to be achieved in the beam because ductility decreases in the Code. The purpose of the restriction, as explained
with increase of tension reinforcement. [24,25] However, by the Commentary part of NZS3101-2006 in its C9.3.8.4
Kwan et al. [25] and Pam et al. [24] also demonstrate and C9.4.3.5 is to prevent premature bond slipping of the
that ductility increases with increase of compression bar under critical load combinations for flexure includ-
reinforcement. As the Code also requires compression ing earthquake actions. The phenomenon is in fact more
reinforcement percentage to be at least 50% of ten- critical in beams. C9.4.3.5 of the NZS3101-2006 says,
sion reinforcement in critical zones, a threshold ductility
level is thus “implied” by the Code at 2.5% tension At interior beam column joints, extremely high bond
stresses can develop when a frame sustains large inelastic
reinforcement percentage and 1.25% compression rein- deformations due to seismic actions. Beam bars may be
forcement percentage. Law [26] worked out that the forced to yield in tension at one column face and subject
implied threshold ductility can be met under the following to high compressive stress at the opposite face. Also, yield
conditions by which the maximum tension reinforcement penetration along beam bar from either face of an interior
percentage can be increased to the maximum of 4%: column may considerably reduce the effective anchorage
length of the bar.

(1) structural depth is not less than 400 mm; The phenomenon is illustrated by Figure 15 in which
(2) concrete grade ranges from 30 to 50; and the adequacy of the equation is also examined by using
(3) compression steel percentage is not less than 75% Equation (8.3) of the Code for the derivation of the bond
Downloaded by [Law Chi Wai] at 18:34 01 July 2014

of that of the tension steel percentage. strength. In the calculation, the partial strength factors of
concrete and steel are removed. Comparing the coefficient
Law’s [26] recommendation should therefore be adopted of 1.57 arrived in the calculation and 3.2 in the Equation
to allow more flexibility in the determination of structural (9.7), it seems that the diameter of the maximum bar has
sizes of beams. to be further reduced by more than half if this extreme
ultimate state has to be satisfied unless the bond strength
can be enhanced significantly by the confinement in the
Maximum diameter of column reinforcement bars beam-column joint. So there may be some bond slippage
passing through the beam when columns are designed at this extreme state when checked against bond strength
to develop plastic hinges in the end regions provisions in the Code.
Equation (9.7) in 9.9.2.1(a) of the Code restricts column Nevertheless, the provision is adequate when checked
bar size to a ratio of the depth of the adjoining beam which against provisions of the NZS3101 as NZS3101 C9.3.8.4 
is reproduced as follows: allows a maximum average bond strength of 1.5αf fc
 for beam bar passing through a beam-column joint by its
3.2h 0.8fcu C9.3.8.4 where αf = 0.85 for two way frame and αf = 1
φ≤ (6)
fy for one way frame. Column bar follows the same princi-
ple. This provision almost doubles the bond strength. The
where φ is the diameter of the column reinforcing bar and increase in bond strength can be justified by the compara-
h is the depth of the beam. tively high aggregate enhancement factor of 1.5 × 1.5 =
The requirement is identical to that of the New Zealand 2.25 allowed in NZS3101 due to cover and transverse
Code NZS3101-2006 10.4.6.6 in which fc instead of fcu is reinforcements confinement that is often provided in the
specified and the Code adopts fc = 0.8fcu . beam-column joint.

Tension
= f yπφ 2 / 4

Bond Strength (ultimate at bond slippage)

Reinforcing Bar = 0.5 × γ m f cu (Eqn 8.3 of the Code) γ m = 1.4
h ( )
∴ 2 f yπφ 2 / 4 ≤ 0.7 f cu πφ h
Beam-Column
Joint 1.4h f cu 1.57 h 0.8 f cu
⇒φ ≤ =
fy fy

Compression
= f yπφ 2 / 4

Figure 15. Checking adequacy of Equation (9.7) of the Code.

 HKIE Transactions 45

As explained by NZS3101, the requirement is in fact Department has relaxed the requirement by allowing cen-
more critical to beam. A column bar normally will not tre of laps be in the middle half (instead of middle quarter)
be having tension and compression simultaneously in of the floor height.
locations above and below the beam-column joint unless
the lateral load (probably due to very strong seismic
load) is so large that the compression load is entirely Densified transverse reinforcements in columns at
nullified. Besides, the requirement does place stringent end regions
restriction on column bar sizes which may lead to conges- 9.9.2.2 of the Code requires provision of densified trans-
tion. For example, bar size in column through a 400 mm verse reinforcements within “critical regions” of “limited
deep beam of grade 40 concrete is limited to below ductile high strength” RC columns. The critical regions are
16 mm. So, based on the rationale as discussed above, certain lengths of the column extending from the points of
the requirement should be exempted if the designer can maximum moments for some finite lengths defined in the
demonstrate that the column bar will not undergo ten- code as
sion and compression simultaneously above and below
the beam-column joint under the worst combination of (i) For 0 < N /(Ag fcu ) ≤ 0.1, the extent of critical
applied loads. region is taken as 1.0 times the greater dimen-
sion of the cross-section or where the moment
exceeds 0.85 of the maximum moment, whichever
The “Middle quarter rule” for the main is larger.
Downloaded by [Law Chi Wai] at 18:34 01 July 2014

reinforcement lapping in columns (ii) For 0.1 < N /(Ag fcu ) ≤ 0.3, the extent of critical
9.9.2.1(d) of the Code requires the centre of laps of main region is taken as 1.5 times the greater dimen-
reinforcements of columns to be in the middle quarter of sion of the cross-section or where the moment
the storey height unless it can be shown that plastic hinges exceeds 0.75 of the maximum moment, whichever
cannot develop in the column beneath the beam-column is larger.
joint. The requirement serves to shift the laps which are (iii) For 0.3 < N /(Ag fcu ) ≤ 0.6, the extent of critical
potential weak sections away from the ends of the columns region is taken as 2.0 times the greater dimen-
normally having high moments and thus are potential loca- sion of the cross-section or where the moment
tions for formation of plastic hinges. The requirement is exceeds 0.65 of the maximum moment, whichever
identical to that of NZS3101-2006 10.4.6.8.2(a) for design is larger.
for ductility in earthquakes. Similarly ACI 318-08 requires
lap splices be permitted only within the centre half of the Where N is the ultimate axial load and Ag is the cross-area
member (column) length for regions of high seismic risk. of column section.
To show that plastic hinges cannot develop in the The Code then requires “denser” transverse reinforce-
column so as to be exempted from the requirement, the ments within the critical regions as compared with the
“strong column – weak beam” design approach has to be regions outside. The determination of the critical regions
adopted. This can generally be achieved when the sum- is adopted from Ho [27] who uses the term “plastic hinge
mation of flexural strengths of the ends of the columns length” instead as the regions are where plastic hinges will
joining to a beam-column joint exceeds the summation of be formed. The definition of “limited ductile” is also based
the flexural strengths of the ends of the beams also joining on the thesis. By the thesis, a structural member is defined
to the beam-column joint. As a design approach adopted in to be limited ductile if its curvature ductility factor is close
many seismic codes, a factor of 1.2 is usually added to pro- to 10. The curvature ductility factor is defined in Figure 16.
vide some margins. The flexural strengths of the columns The imposition of limited ductile requirements instead
have to be based on axial loads being most unfavourable of fully ductile is to avoid the use of very congested rein-
to their flexural strengths. In equation form, it is forcement which would be unnecessary and uneconomical
  in Hong Kong as an area of low to moderate seismicity.
Mc ≥ 1.2 Mb , (7) The high-strength concrete which is the target for study in
Ho [27] refers to concrete of grades above 60. However,
where Mc and Mb are the flexural strengths of the columns the clause should also be applicable to normal strength
and beams, respectively. RC as similar provisions are also necessary for normal
The rule has incurred some difficulties for the local strength RC columns so as to enhance their strengths and
contractors who used to lap the main reinforcements of ductility under lateral load.
the column at floor levels. As the requirement which is For the determination of the length of the “critical
originated from NSZ3101-2006 and similar one from ACI regions” or the “plastic hinge length”, Ho [27] pointed
318-08 are for regions of strong seismicity, the require- out that, by previous research by Bayrak and Sheikh [28],
ment can be relaxed somehow for Hong Kong. Buildings Paultre et al. [29], the extent of plastic hinge region of RC
 46 C.W. Law and J.Y.W. Mak

Mmax
Moment of
Resistance

0.8Mmax

0.75Mmax
φy φu

Curvature

Figure 16. Definition of curvature ductility factor of a column. Notes: The graph is a typical one showing the variation of the flexural
strength of a column with its curvature (inverse of radius of curvature) under a constant axial load. The factor μ = φu /φy is known as
the “curvature ductility factor” which measures the ductility of the section. Obviously the higher the value of μ, the higher will be the
ductility or the ability to tolerate deformation.
Downloaded by [Law Chi Wai] at 18:34 01 July 2014

Figure 17. Observed “plastic hinge” formation on 325 mm × 325 mm columns under lateral load by Ho [27].

members subjected to flexure and compression axial load the region increases with axial compressive load level
relied purely on physical observation of the column region and concrete compressive strength, but decreases with the
that suffered extensive inelastic damage. Their observed volumetric ratio of transverse reinforcements. A formula
column plastic hinge length, including the region affected obtained by non-linear regression of experimental data is
by “stub effect” (to be defined in the coming paragraphs), listed as follows:
ranged from 1.0–2.0 h, where h is the cross-section dimen-  0.5  1.5  0.5
sion of the column. Determination of plastic hinge length p P fcu ρ
= 1.7 + 0.6, (8)
by observation can be demonstrated by test samples in h Ag fcu fy ρs
Ho [27] as extracted in Figure 17 for columns of sectional
size 325 mm × 325 mm. The columns were subjected to where p is the plastic hinge length, P is the applied load on
cyclic loads until failures. It can be seen that the columns the column, Ag is the cross-sectional area of the concrete
failed at the ends for length 1.0–2.0 h and the failure modes section, ρ is the steel ratio of the longitudinal steel, and ρs
are characterised by concrete spalling, buckling of the is the volumetric ratio of the transverse reinforcement.
main reinforcements and opening up of the transverse However, as Equation (8) is not user friendly because
reinforcements. it involves so many parameters, Ho proposed simpli-
Nevertheless, in addition to physical observation, fied and conservative guidelines which are adopted in
Ho [27] devised criteria in defining the plastic hinge 9.9.2.2 of the Code as listed above. It can be shown
length. He suggested that region with curvatures greater that the simplified guidelines together with the enhanced
than the yield curvatures (φy as defined in Figure 16) is transverse reinforcements are more conservative than
the plastic hinge region and by experiment the length of Equation (8).
 HKIE Transactions 47

The densified transverse reinforcements serve to buckling; and (iii) enhancing ductility against flexure. The
strengthen the column sections in these regions by (i) Code also mentions that regions with “stub effects” should
increasing the concrete strength within the transverse also be included in the “critical regions”. The “stub effect”
hoops by confinement; (ii) offering better restraints to is the phenomenon by which the start of the plastic hinge
the longitudinal reinforcements to prevent premature region has shifted slightly away from the beam-column
Downloaded by [Law Chi Wai] at 18:34 01 July 2014

Figure 18. (a) Typical arrangement of section of column for study of effect of confinement by transverse reinforcement. (b) Variation
of flexural strength against curvature for transverse reinforcement T8 @ 384 c/c. (c) Variation of flexural strength against curvature
for transverse reinforcement T8 @ 150 c/c. (d) P–M diagram of the column section under increasing transverse reinforcements. (e)
Variation of ductility factors with average axial stress and transverse reinforcements.
 48 C.W. Law and J.Y.W. Mak
Downloaded by [Law Chi Wai] at 18:34 01 July 2014

Figure 18. Continued.

interface, which is influenced by the region having appar- transverse reinforcements on flexural strength and duc-
ent extra confinement provided by the rigid beam, causing tility of column. Figure 18(a) shows the layout of main
the flexural strength and ductility within a small region and transverse reinforcements of the column in which the
near the beam-column interface to increase. transverse reinforcements can only vary in bar size and
Mander et al. [30] has developed a theoretical approach vertical spacing.
(verified by experimental data) to determine the confining In Figure 18(b) and 18(c), the variations of flexural
stresses by transverse reinforcements on the inner core of strengths of the column are worked out under increasing
the RC beams and columns and subsequently the enhanced curvature. Figure 18(b) is for the column with minimum
compressive strength of concrete. Based on Mander’s transverse reinforcements required by the Code outside
approach and the stress strain relationship of concrete by the “critical region” and Figure 18(c) for that inside the
Attard and Setunge [31] and Attard and Stewart [32], the “critical region”. Figure 18(d) shows the P–M interac-
variations of the moment of resistance (flexural strengths) tion diagram (axial strength verses flexural strength) for
of a grade 45 600 mm × 600 mm column with longitu- the column with varying transverse reinforcements and
dinal reinforcement of 12T32 under various transverse Figure 18(e) shows the variation of factors of ductility
reinforcements are determined, together with the “P–M defined in Figure 15. The fco used in defining axial stress
diagrams”. The plots are for illustration of effects of is taken as 0.9 × 0.8 = 0.72fcu which is the peak concrete
 HKIE Transactions 49

stress assumed in Attard and Setunge’s concrete stress assumed at 0.01 or ultimate strain steel at 0.12 as
strain model. It should be noted that except the first curve strains of these materials cannot be unlimited.
which is the design P–M curve to the Code in which
the partial strength factors have been applied, other P–M As a conclusion, use of densified transverse reinforcement
curves are the very ultimate ones without partial strength can improve both strength and ductility. This is also con-
factors. The first curve is included for comparison purpose. firmed by experiment by Ho [27]. The requirements in the
The foregoing analysis aims to demonstrate how Code should be adopted as they are not too excessive over
transverse reinforcements affect strengths and ductilities our usual requirements according to the British Standards
of a column which justify the use of densified trans- and the benefits are quite significant. As a way forward,
verse reinforcements at “critical sections”. The following confinement leading to increase of concrete strength may
observations and conclusions can be drawn: be taken into account in our structural design which can
achieve some economy.

(i) Figure 18(b) and 18(c) shows that flatter plateaus

are formed at low axial stress levels, indicat- Conclusions
ing that ductility of the column decreases with
The following concluding remarks can be drawn:
axial stress level. This is further justified by
Figure 18(e).
(ii) The flexural strength of the column section (the (1) The British Standards for RC works which were
Downloaded by [Law Chi Wai] at 18:34 01 July 2014

peak flexural strength in each flexural strength the bases of the previous versions of the code
versus curvature curve) increases first with axial of practice of Hong Kong are rather crude in
stress level and decreases after a threshold certain aspects such as (i) provision of ductility;
axial stress which is typical as exhibited by (ii) bonding and anchorage of reinforcement with
Figure 18(d). concrete; and (iii) confinement of concrete by rein-
(iii) Figure 18(d) shows that the quantity of trans- forcements. The provisions for these are either
verse reinforcement increases both the axial inadequate or over-simplified. There is thus room
compression and flexural strengths of the col- for improvement for the Code by incorporating
umn section. But the increases seem not too relevant practices from other codes into the current
significant as there are only about 8.8% and version. This has been done generally by which
4.4% increase in the maximum axial compres- the Code has made substantial advancements in
sion and flexural strengths respectively when the design and detailing of the RC work.
minimum densified transverse reinforcement in (2) However, difficulties have arisen as the current
critical region (T8@150c/c) is added as compared local construction practice and workmanship can-
with that in non-critical region (T8@384c/c). It not fully gear with the new requirements. In
is until with transverse reinforcement as heavy addition, some of the ductility requirements are
as T16@100c/c that both the axial and flexural considered too harsh and not really necessary
strength increases by 38% and 21%, respectively as they are originated from countries of strong
as the concrete strength in the inner core increases seismicity. The paper has reviewed a number
by 70% by confinement. Nevertheless, at the same of these aspects with respect to the underlying
axial stress level, say at high stress level of 0.8 principles, explored the necessity of the require-
fco (approximately 0.58 fcu ), the moment capac- ments and proposed alternatives where appropri-
ity confinement by T8@384c/c is only 799 kNm ate. It is hoped that the materials of the paper
while that by T8@150c/c is 982 kNm which is a can somewhat pave the way forward for future
23% increase. This shows that the transverse rein- improvement.
forcement is very effective in enhancing flexural (3) In addition, care should be exercised to ensure
strengths at high stress level. The flexural strength consistency and safety in the provisions of the
becomes 1604 kNm at 0.8 fco (a 100% increase) if Code when different requirements from differ-
the heavy transverse reinforcement T16@100c/c ent sources (codes) are incorporated. Moreover,
is used. as the deterministic parameters such as bond
(iv) Figure 18(e) shows that the quantity of transverse strengths, confinements by transverse reinforce-
reinforcement also increases ductility of the col- ments are greatly affected by the local materials,
umn section. However, it should be noted that the local workmanship and construction methods,
author has placed additional criteria for determi- tests should be conducted as appropriate to deter-
nation of the ultimate curvature φu which are the mine these parameters for realistic design and
attainment of either the ultimate concrete strain detailing.
 50 C.W. Law and J.Y.W. Mak

Notes on contributors [4] New Zealand Standard. NZS 3101:Part 1 & 2:2006 con-
crete structures standard; 2006.
Ir Law Chi Wai received [5] American Concrete Institute. ACI 318-08 and ACI
his B.Sc. and M.Sc. degrees, 318R-08 building code requirements for structural con-
respectively from the Univer- crete (ACI 318-08) and Commentary (ACI 318R-08),
sity of Hong Kong in 1978 and 2008.
1995. He obtained his profes- [6] Kwan AKH. Concrete code handbook. An explanatory
handbook to the code of practice for structural use of con-
sional qualifications in 1982. crete 2004. Hong Kong: The Hong Kong Institution of
He is currently serving as a Engineers Structural Division; 2006.
Senior Structural Engineer in [7] Beeby AW, Narayanan RS. Eurocode design handbooks
the Housing Department of designers’ handbook to Eurocode 2 Part 1.1: design of
the HKSAR. Before he joined the Housing Department, concrete structures. London: Thomas Telford; 1995.
[8] Tepfers R. Cracking of concrete cover along anchored
he has also worked as a Senior Engineer with Hyder Con- deformed reinforcing bar. Mag Concr Res. 1979;31(106):
sultants. Throughout his career, he has been engaged in 3–12.
design and administration of various projects involving [9] Kemp EL. Bond in reinforced concrete: behaviour and
tall buildings, deep basements, transfer structures, and design criteria. ACI J. 1986;83(1):50–57.
studies of various aspects of theories and methods in rein- [10] Timoshenko, S. Strength of materials. Part II: advanced
theory and problems. Princeton (NJ): D. van Nostrand
forced concrete structures analysis and design including Company Inc.; 1956. p. 205–210.
the finite element method, concrete shrinkage and creep, [11] Esfahani MR, Rangan BV. Bond between normal strength
ductility, etc. He has published 15 papers in both HKIE
Downloaded by [Law Chi Wai] at 18:34 01 July 2014

and high-strength (HSC) and reinforcing bars in splices in

Transactions and international journals. beams. ACI Struct J. 1998;95(3):272–280.
[12] Esfahani MR, Rangan BV. Local bond strength of rein-
Ir Prof J.Y.W. Mak is the forcing bars in normal strength and high-strength concrete
(HSC). ACI Struct J. 1998;95(2):96–106.
Chief Structural Engineer in [13] Tepfers R. Bond stress along lapped reinforcing bars. Mag
the Housing Department and Concr Res. 1980;32(112):135–142.
is looking after the stan- [14] Giurani E. On the effective axial stiffness of a bar in cracked
dardised designs, the Housing concrete. In: Bartos P, editor. Bond in concrete. London:
laboratory, all the appointed Applied Science Publishers; 1982. p. 107–126.
[15] Esfahani MR, Kianoush MR. Development/splice length
Direct Testing Laboratories, of reinforcing bars. ACI Struct J. 2005;102(1):
structural specification and 22–30.
standards and the off-territory [16] Lampert P. Torsion and bending in reinforced and
manufacture of building materials and components in the prestressed concrete. Proc Inst Civil Eng. 1972;50(4):
Development and Construction Division. In the past years, 487–505.
[17] Mitchell D, Collins MP. Detailing for torsion. ACI J.
he has been specialised in standard block designs and 1976;73(9):506–511.
prefabrication which have been widely used in Hous- [18] Mitchell D, Collins MP. The behaviour of structural con-
ing Authority projects. He is also very much devoted in crete beams in pure torsion. Civil Engineering Publication
the training of civil and structural engineering students No. 74-06, Department of Civil Engineering, University of
and graduates with over 15 years of experience in this Toronto, March 1974.
[19] Buildings Department HKSAR. Letter to all AP/RSE/
aspect. He is now the Head of Office for the Department RGE/RGBC/RSC informing acceptable alternatives of
structural engineering graduate training in the Hong Kong reinforcement fixing to code of practice for structural use
Institution of Engineers (HKIE). He is also currently the of concrete 2004; 2011.
Adjunct Professor of the Hong Kong Polytechnic Univer- [20]
sity. He has been actively involved in the HKIE since the Code for Seismic Design of Buildings GB50011
mid-1990s. He was the Chairman of the Structural Divi- – 2010; 2010.
[21] Sheikh SA, Khoury S. Confined concrete columns with
sion in 2004/2005 and a Council Member of the Institution stubs. ACI Struct J. 1993;90(4):414–431.
from 2007 to 2011. [22] Wehbe NI, Saiidi MS, Sanders DH. Seismic performance
of rectangular bridge columns with moderate confinement.
ACI Struct J. 1999;96(2):248–258.
[23] Buildings Authority Hong Kong. Code of practice for
References structural use of concrete 1987; 1987.
[1] Buildings Department HKSAR. Code of practice for struc- [24] Pam HJ, Kwan AKH, Ho JCM. Post-peak behaviour and
tural use of concrete 2004; 2004. flexural ductility of doubly reinforced normal- and high-
[2] British Standard. BS 8110-1:1997 structural use of con- strength concrete beams. Struct Eng Mech. 2001;12(5):
crete; 1997. 459–474.
[3] European Standard. Eurocode 2: design of concrete struc- [25] Kwan AKH, Ho JCM, Pam HJ. Flexural strength and duc-
tures – part 1-1: general rules and rules for buildings; tility of reinforced concrete beams. Proc Inst Civil Eng
2004. Struct Build. 2002;152(4):361–369.
 HKIE Transactions 51

[26] Law CW. Discussions on the maximum longitudinal rein- [29] Paultre P, Legeron F, Mongeau D. Influence of concrete
forcement ratio as a ductility requirement for reinforced strength and transverse reinforcement yield strength on
concrete beams by the code of practice for the structural behaviour of high-strength concrete columns. ACI Struct
use of concrete 2004 and recommendation of alternatives. J. 2001;98(4):490–501.
HKIE Tran. 2011;18(3):14–21 [30] Mander JB, Priestley MJN, Park R. Theoretical stress–
[27] Ho JCM. Inelastic design of reinforced concrete beams and strain model for confined concrete. J Struct Eng ASCE.
limited ductile high-strength concrete column [PhD the- 1988;114(8):1804–1825.
sis]. Pokfulam, Hong Kong: The University of Hong Kong; [31] Attard MM, Setunge S. The stress strain relationship of con-
2003. fined and unconfined concrete. ACI Mater J. 1996;93(5):
[28] Bayrak O, Sheikh SA. Confinement reinforcement design 432–442.
consideration for ductile HSC columns. J Struct Eng ASCE. [32] Attard MM, Stewart MG. A two parameter stress block for
1998;124(2):999–1010. high-strength concrete. ACI Struct J. 1998;95(3):305–317.
Downloaded by [Law Chi Wai] at 18:34 01 July 2014